Louisiana State UniversityLSU Digital Commons

LSU Historical Dissertations and Theses Graduate School

1986

The Assessment of Health Status in HemodialysisPatients.Barbara Kathryn BruceLouisiana State University and Agricultural & Mechanical College

Follow this and additional works at: https://digitalcommons.lsu.edu/gradschool_disstheses

This Dissertation is brought to you for free and open access by the Graduate School at LSU Digital Commons. It has been accepted for inclusion inLSU Historical Dissertations and Theses by an authorized administrator of LSU Digital Commons. For more information, please contactgradetd@lsu.edu.

Recommended CitationBruce, Barbara Kathryn, "The Assessment of Health Status in Hemodialysis Patients." (1986). LSU Historical Dissertations and Theses.4223.https://digitalcommons.lsu.edu/gradschool_disstheses/4223

INFORMATION TO USERS

This reproduction was made from a copy of a manuscript sent to us for publication and microfilming. While the most advanced technology has been used to pho­tograph and reproduce this manuscript, the quality of the reproduction is heavily dependent upon the quality of the material submitted. Pages in any manuscript may have indistinct print. In all cases the best available copy has been filmed.

The following explanation of techniques is provided to help clarify notations which may appear on this reproduction.

1. Manuscripts may not always be complete. When it is not possible to obtain missing pages, a note appears to indicate this.

2. When copyrighted materials are removed from the manuscript, a note ap­pears to indicate this.

3. Oversize materials (maps, drawings, and charts) are photographed by sec­tioning the original, beginning at the upper left hand comer and continu­ing from left to right in equal sections with small overlaps. Each oversize page is also filmed as one exposure and is available, for an additional charge, as a standard 35mm slide or in black and white paper format.*

4. Most photographs reproduce acceptably on positive microfilm or micro­fiche bu t lack clarity on xerographic copies made from the microfilm. For an additional charge, all photographs are available in black and white standard 35mm slide format.*

*For more information about black and white slides or enlarged paper reproductions, please contact the Dissertations Customer Services Department.

T TA/f.I Dissertation U I V l l Information ServiceUniversity Microfilms InternationalA Bell & Howell Information C om pany300 N. Z eeb Road, Ann Arbor, M ichigan 48106

8629157

Bruce, Barbara Kathryn

THE A SSESSM EN T O F HEALTH STA TU S IN HEMODIALYSIS PATIENTS

The Louisiana S tate University a n d A gricu ltural an d M ech an ica l Col. Ph.D .

University Microfilms

International 300 N. Zeeb Road, Ann Arbor, Ml 48106

1986

The Assessment of Health Status in

Hemodialysis Patients

A DissertationSubmitted to the Graduate Faculty of the

Louisiana State University and Agricultural and Mechanical College

in partial fulfillment of the requirements for the degree of

Doctor of Philosophyin

The Department of Psychology

byBarbara Kathryn Bruce

B.A., Louisiana State University, 1976 M.S., University of New Orleans, 1981

May, 1986

Acknowledgements

I would like to express my warmest thanks to Dr. Phillip Brantley for his invaluable guidance and assistance in the direction of this dissertation. The instruction, vision, and encouragement he provided as my major professor made a tremendous impact upon my academic and professional goals. Special appreciation is extended to Dr. Thomas Cocke and Dr. G. T. McKnight and the staff and patients at Bio-Medical Applications of Baton Rouge, Inc. Without their enthusiastic assistance, this research project would not have been possible. Appreciation is also extended to Dr. Frank Gresham, Dr. Arthur Riopelle, Dr. William Waters, and Dr. Ray Buss for serving as committee members.

I would also like to thank my husband, parents, family, and friends for their understanding and support so generously provided throughout my graduate studies.

Table of Contents

Acknowledgements ....................................... iiTable of C o n t e n t s ........................................ii-iList of T a b l e s ......................................... ivList of Appendices...................................... viiiA b s t r a c t................................................ ixChapter I: Introduction ............................ 1Chapter II: M e t h o d ................................... 71Chapter III: R e s u l t s ................................. 79Chapter IV: Discussion ............................... 98References.............................................. 112T a b l e s ................................................... 134Appendices.............................................. 181Curriculum Vitae ....................................... 191Approval Sheets ....................................... 193

iii

List of Tables

Table 1. Survival Rates for Patients ReceivingDialysis or Undergoing Transplantation . . . . 135

Table 2. Summary of Sample Characteristics ............ 136Table 3» Means and Standard Deviations of

Health Status Measures ......................... 137Table 4. Correlations Among Measures of Health Status

and £ Values (N = 131 ) ........................... 138

Table 5. Means, Standard Deviations, Frequencies, andPercentages of Demographic/Medical Variables . 139

Table 6. Correlations Among Measures of Health Statusand Demographic/Medical Variablesand £ Values (N = 13 1 ) ............................ 140

Table 7. Means, Standard Deviations, Frequency, andPercentage of Hospitalizations, Hospital Days, and M o r t a l i t y .............................. 141

Table 8. Correlations Among Measures of Health Status,Demographic/Medical Variables andCriterion Measures .............................. 142

Table 9- Multiple Regression Summary Table forMeasures of Health Status and Number of Hospitalizations ................................ 143

Table 10. Stepwise Multiple Regression for the Measures of Health Status and Number ofHospitalizations ................................ 144

Table 114 Multiple Regression Summary Table for theDemographic/Medical Variables and Number of Hospitalizations............................. . 145

Table 12. Stepwise Multiple Regression for theDemographic/Medical Variables and the Number of Hospitalizations ............................ 146

iv

Table

Table

Table

Table

Table

Table

Table

Table

Table

Table

13* Multiple Regression Summary Table for the Combined Measures of Health Status and Demographic/Medical Variables and the Number of Hospitalizations ............................

14. Stepwise Multiple Regression for the Combined Measures of Health Status and Demographic/Medical Variables and the Number of Hospitalizations ............................

15. Multiple Regression Summary Table for Measures of Health Status and Number of Hospital Days (Outliers Included) ............................

16. Multiple Regression Summary Table for the Demographic/Medical Variables and Number of Hospital Days (Outliers Included) ............

17- Multiple Regression Summary Table for the Combined Measures of Health Status and Demographic/Medical Variables and Number of Hospital Days (Outliers Included) ...........

18. Stepwise Multiple Regression for the Combined Measures of Health Status and Demographic/Medical Variables and the Number of Hospital Days (Outliers Included) .........

19. Multiple Regression Summary Table for Measures of Health Status and Number of Hospital Days (Outliers Exluded) ..............................

20. Multiple Regression Summary Table for the Demographic/Medical Variables and Number of Hospital Days (Outliers Exluded) ..............

21. Stepwise Multiple Regression for the Dembgraphic/Medical Variables and the Number of Hospital Days (Outliers Excluded) .........

22. Multiple Regression Summary Table for the Combined Measures of Health Status and Demographic/Medical Variables and the Number of Hospital Days (Outliers Excluded) .........

147

148

149

150

151

152

153

154

155

156

v

Table 23. Stepwise Multiple Regression for the Combined Measures of Health Status and Demographic/Medical Variables and the Number of Hospital Days (Outliers Excluded) ........... 157

Table 24. Standardized Linear Discriminant Function (LDF)Coefficients for the Health Status Measures . 158

Table 25. Discriminant Analysis Classification for the Calibration Sample Based on the Health Status M e a s u r e s ........................................... 159

Table 26. Discriminant Analysis Classification for the Holdout Sample Based on the Health Status M e a s u r e s ........................................... 160

Table 27. Standardized Linear Discriminant Function (LDF) Coefficients for the Demographic/Medical V a r i a b l e s ....................................... 161

Table 28. Discriminant Analysis Classification for the Calibration Sample Based on the Demographic/ Medical Variables .............................. 162

Table 29. Discriminant Analysis Classification for the Holdout Sample Based on the Demographic/Medical Variables .............................. 163

Table 30. Discriminant Analysis Classification for the Calibration Sample Based on the Combined Measures of Health Status and Demographic/Medical Variables .............................. 164

Table 31. Discriminant Analysis Classification for the Holdout Sample Based on the Combined Measures of Health Status and Demographic/Medical V a r i a b l e s ......................................... 165

Table 32. Means of Dependent Measures for Survivor andDeceased Groups ................................ 166

Table 33. Analysis of Variance Summary Table for BloodUrea N i t r o g e n .................................... 167

Table 34. Analysis of Variance Summary Table forP o t a s s i u m ......................................... 168

vi

Table 35. Analysis of Variance Summary Table forP h o s p h o r o u s .......................................169

Table 36. Analysis of Variance Summary Table forHematocrit......................................... 170

Table 37. Analysis of Variance Summary Table for StaffRatings of Functional Status.................... 171

Table 38* Analysis of Variance Summary Table forPhysician Ratings of Physical Status ......... 172

Table 39- Analysis of Variance Summary Table for WahlerPhysical Symptoms Inventory.......................173

Table 40. Analysis of Variance Summary Table forSubjective Ratings of Functional Status . . . 174

Table 41. Analysis of Variance Summary Table forSubjective Ratings of Health.................... 175

Table 42. Analysis of Variance Summary Table for Age . . 176Table 43. Analysis of Variance Summary Table for Number

of Months on D i a l y s i s ........................... 177Table 44. Analysis of Variance Summary Table for Total

Number of Concurrent Diagnoses ................ 178Table 45. Analysis of Variance Summary Table for

Diabetes Mellitus .............................. 179Table 46. Analysis of Variance Summary Table for

Primary Hypertension ......................... 180

List of Appendices

Appendix A: Physician Ratings ........................ 182Appendix B: Wahler Physical Symptoms Inventory . . 183Appendix C: Functional Classification ............... 185Appendix D: Subjective Functional Rating .......... 186Appendix E: Subjective Rating of Health ............ 187Appendix F: Informed Consent ........................ 188Appendix G: Demographic/Medical Data .............. 190

viii

Abstract

The assessment of health status has become ofparamount importance as efforts expand to explore theimpact of psychological factors and lifestyle variables on disease. The need to clearly assess the efficacy ofinterventions designed to improve health and prevent health problems has provided an impetus for the development of assessment methodology in this area. There remains a need not only for vigorous research directed at the development of reliable and valid measures of health status, but also a critical need for the empirical validation of existing instruments. The complexity of variables influencing the health status of patients maintained on hemodialysis makes the assessment or quantification of health in this population exceedingly difficult. Although a variety of assessment procedures have been utilized clinically and in research, few investigations have directly assessed the psychometric properties of these measures. Of particular concern is the lack of demonstrated relations among various measures used to index health and important health outcomes for patients maintained on hemodialysis. The present study was designed to examine the concurrent and predictive validity of a number of measures of health status in a sample of hemodialysis patients. Subjects consisted of 131

adult outpatients with chronic renal failure who were receiving maintenance hemodialysis, and were followed over the course of one year. The measures of health status examined in the present investigation included biochemical analyses of blood urea nitrogen, potassium, phosphorous, and hematocrit; physician ratings of physical status; Wahler Physical Symptoms Inventory; staff ratings of functional status; subjective ratings of functional status; and subjective ratings of health. The relations among these measures of health status and other demographic and medical variables thought to impact health in this population were examined. Of primary importance were the relations among indices of health and subsequent medical utilization and survival. Results of the present investigation offer evidence for the concurrent validity of physician ratings of physical status, ratings of functional status, and the Wahler Physical Symptoms Inventory as measures of the construct of health status, and confirm previous findings which suggest that physician ratings are the most valid measure of health status.

x

Chapter I Introduction

The decade of the 1970's witnessed phenomenal growth in behavioral science applications to medicine, broadly focused upon the etiology, treatment, management, and prevention of illness (Agras, 1982). Much of the credit for psychology's recent expansion in medicine has been attributed to the successful application of behavioral techniques with medical patients (Blanchard, 1982). Health-related research and clinical activities have rapidly expanded in recent years, in large part, due to the finding that individual habits or "lifestyle" variables may account for over half of the annual deaths in the United States (U.S. Department of Health and Human Services, 1980). The escalating cost of health care has also provided a powerful economic impetus for the expansion of efforts directed at reducing medical utilization and altering lifestyle variables thought to contribute to health risk. For these reasons Cummings (1984) projected the area of health psychology to grow at a faster rate than any other aspect of clinical psychology. The rapidly developing fields of behavioral medicine and behavioral

2health share an emphasis on the development of technology directed at understanding health parameters as illustrated in their definitions. The Yale conference of the National Academy of Science adopted the following definition of behavioral medicine:

Behavioral medicine is theinterdisciplinary field concerned with the development and integration of the behavioral and biomedical science knowledge and techniques relevant to health and illness and the application of this knowledge and these techniques to prevention, diagnosis, treatment and rehabilitation (Schwartz & Weiss, 1977, p. 250).

Matarazzo (1980; 1982) further proposed a subcategory ofbehavioral medicine that he felt more closely addressed the areas of health maintenance and prevention of illness which he labeled behavioral health. He defined behavioral health as:

an interdisciplinary field dedicated to promoting a philosophy of health that stresses individual responsibility in the application of behavioral and biomedical science knowledge and techniques to the maintenance of health and the prevention of illness and dysfunction by a variety of self-initiated individual or shared activities (Matarazzo, 1982, p.8).

In the area of health, much of the early work has focused upon intervention techniques rather than thedevelopment of assessment procedures. Similarly, a general deficiency in assessment methodology has been noted to

characterize the behavioral medicine literature (Russo, Bird, & Masek, 1980). The assessment of health status has become of paramount importance as efforts expand to explore the impact of psychological factors and lifestyle variables on disease. Also, the need to clearly assess the efficacy of interventions designed to reduce disease incidence and affect the course of health problems has provided an impetus for the development of assessment methodology in this area. There remains a need not only for vigorous research directed at the development of reliable and valid measures of health status, but also a critical need for the empirical validation of existing instruments.

The assessment of health status will be reviewed, followed by an overview of chronic renal failure and dialysis procedures. This review is intended to provide a rationale for the proposed research objectives and methodology which will focus upon the assessment of health status in hemodialysis patients.

Assessment of Health Status

The available assessment methodology targeting health parameters continues to reflect the view that physical health and physical impairment are mirror images of one

another, that is, the presence of one is assumed to reflect the absence of the other. Although contemporarydefinitions of health have generally rejected the notion that health is merely the absence of illness, nosatisfactory measure of positive health is yet available. Consequently, current measures of health status focus primarily upon the assessment of illness, i.e., the nature and degree of organic pathology which is present; or illness behavior, i.e., those behaviors thought to be associated with the presence of illness or disease such as symptom-reporting, functional limitations, and medicalutilization.

The medical examination, physician ratings of health, and physician diagnosed medical disorders are considered to be more objective indices of health status than the more frequently employed self-report measures of illness behavior such as subjective ratings of health, self-monitoring of health-related events, interviews, and questionnaires. Similarly, morbidity data are viewed as more objective if obtained through hospital records, as opposed to the collection of identical information from self-report measures. The medical examination often serves as the standard against which other assessment strategies are compared, because physicians remain the ultimate health status definers due to their training and expertise (Twaddle, 1974). In fact, researchers strongly caution, against the use of self-report indices of health as

5measures of true morbidity without verification of organic pathology through the use of the medical examination,laboratory data, x-rays, etc. (Cleary, 1980; Marshall, Gregorio, & Walsh, 1982; Minter and Kimball, 1978). Sole reliance upon measures of illness behavior in the absence of medical validation is problematic, because a one-to-one correspondence is often falsely assumed between levels of illness behavior and the nature and degree of organic pathology present. Illness behavior can occur out of proportion to or even in the absence of an objectivedisease state. Illness behavior per se may be merely reflecting characteristic ways of responding to symptoms or a reporting tendency which may affect both the report of symptoms and the response to symptoms rather than organic illness.

The actual use of a medical examination or physician ratings in the assessment of health status for research purposes is frequently severely limited due to practicalconsiderations such as cost. Additionally, the use of these "objective" measures as the only criteria for the validation of health status measures presents difficulties due to their heavy reliance upon clinical judgment and the often questionable reliability associated with many diagnostic instruments (Jones & Vischi, 1979; Twaddle,1974).

6Subjective Ratings of Health

Subjective ratings of health are used extensively as an index of health status, and such measures are thought to assess a global, overall sense of well-being or tap the extent to which symptoms or disease limit functioning. Most commonly, subjects have been asked to rate their current health in terms of the following adjectives: excellent, good, fair, poor (e.g., Maddox, 1962; Otto,1979) or to estimate their overall health on a 10-point scale ranging from very poor to excellent (e.g., Garrity, Marx, & Somes, 1978).

Despite their widespread use, particularly in epidemiological and gerontological research, self-ratings of health are generally considered to be a somewhat questionable substitute for objective health status measures or are viewed merely as an indicator of general well-being. In fact, self-ratings of general health have been found to be highly correlated with life satisfaction (Palmore & Luikart, 1972). However, the relation between self-assessment of health status and objective health status as measured by a comprehensive medical examination has been examined, and results have revealed a high degree of congruence between the two methods of assessment (Maddox, 1962; Maddox & Douglass, 1973). Otto (1979) similarly found that individuals reporting a high number of symptoms were significantly more likely to rate their

health as "fair" or "poor” as compared to those who reported few symptoms. Other health parameters such as frequency of illness episodes and number of health problems have been found to be significantly related to subjective ratings of health in an elderly community sample, although no consistent relation was found for an elderly institutionalized sample (Fillenbaum, 1979). Additionally, recent studies provide strong evidence for an association between self-rated health and mortality (e.g., Singer, Garfinkel, & Cohen, 1976). Mossey and Shapiro (1982) found a single measure of self-rated health to be a powerful predictor of subsequent mortality, independent of objective health status, in a large representative sample of elderly adults. A review of 39 studies of self-perceptions of health revealed such ratings to be both reliable and reproducible (Ware, Davies-Avery, & Donald, 1978).

Self-monitoring of Health-related Events

Self-monitoring of health-related behaviors has been utilized in the assessment of health status, both as a primary source of data and also as a memory aid to reduce recall error in subsequent health interviews. In a recent review of nineteen studies that employed health diaries, it was found that only seven used the diary as a primary source of health data with the remainder using it either as a reliability check on other measures of health or as a

memory device (Verbrugge, 1980).Self-monitoring data has been obtained most often in a

diary format, but journals and wall calendars have also been employed (e.g., Marcus, 1982; Norman, McFarlane, Streiner, & Neale, 1982; Ridgeway & Mathews, 1982). The variables frequently recorded in such monitoring devices include frequency and description of physical symptoms, subjective ratings of severity of symptoms, and action taken in response to symptoms such as medication usage or medical utilization. Self-monitoring data is usually collected on a set schedule such as in daily, weekly or monthly intervals, although novel sampling techniques have been adopted to increase compliance over long periods of data recording. For example, Norman et al. (1982) had subjects complete health diaries on three randomly selected days in each two-week period over the course of one year.

The fact that self-monitoring procedures rely less extensively upon retrospective reporting of health-related events makes them quite attractive as health status measures. Data from health diaries have been found to contain two to three times higher rates of reported symptoms than retrospective health interviews (Verbrugge,1980). It is for this reason that self-monitoring formats have been widely utilized as memory aids to improve the reporting skills of respondents and reduce errors when answering questions in a subsequent interview. Additionally, self-monitoring provides a format for the

collection of data regarding mild or acute episodes ofillness, information that would not be reported to a physician nor accurately recalled in retrospective interviews. Thus, these procedures are thought to be a sensitive measure of a dimension of health that cannot be readily assessed through other health status measures.

The most noticeable limitation in the use of self-monitoring data is the finding that there is a decline in the completeness with which data is recorded over time. A significant drop of 5 to 25$ has been noted in thereporting of symptoms in health diaries over a three-month period, and has been attributed to respondent fatigue (Mooney, 1962; Sudman & Lannom, 1978; Sudman, Wilson, &Ferber, 1974). For this reason, self-monitoring of healthevents has only been used infrequently in long-term longitudinal studies. Further research is needed aimed at improving long-term compliance.

Interview Formats

Structured interview formats have frequently been used to collect a wide variety of health-related information, and serve as a common method of data collection in health surveys (e.g., Garrity, Somes, & Marx, 1977; Murphy & Brown, 1980; Parker, You-long, Long-gen, De-yu, & Hinman, 1982). Targets of assessment in either personal or telephone health interviews correspond to those of other

10measurement procedures and include frequency and description of physical symptoms, illness episodes, prevalence of diagnosed chronic disease, perceived severity of reported health problems, and resulting impairment in normal activities (e.g., days lost from work/school, time spent in bed).

The major limitation associated with the interview as a measure of health status is that the information obtained is of questionable accuracy. Health-related events have been found to be consistently underreported in such retrospective accounts (e.g., Madow, 1973; Meltzer, & Hockstim, 1970; Mooney, 1962; Moore, 1975). Bias in reporting has been linked primarily to faulty recall and has been found to be related to length of the time interval between the event and the interview. In addition, the perceived severity of a health event has been shown to affect recall, with minor symptoms or illnesses being recalled more poorly (Cannell, Marquis, & Laurent, 1977). It is for this reason that self-monitoring procedures such as the health diary have been used to improve the accuracy of information provided in health interviews.

Symptom Inventories

A number of inventories are currently available that focus upon physical symptoms, illness or disease. Symptom checklists are self-report inventories that provide

11

information related to physical health. Despite the limitations inherent in self-report methods of assessment, symptom inventories provide a large amount of data rapidly, and therefore can serve effectively both as a screening device and as a primary source of information regarding health status. Procedurally, such instruments require little professional time expenditure due to their ease of administration and simplicity of interpretation.

The Cornell Medical Index (Brodman, Erdman, Lorge, & Wolff, 1949; 1952; Brodman, Erdman, & Wolff, 1949) is oneof the most widely utilized symptom checklists. It is often employed by physicians as an adjunct to the medical examination and has been used in research as an index of symptomatology. The Cornell Medical Index is a 195 item true-false inventory that provides information regarding physical symptoms involving various organ systems, as well as vague physical complaints. Additional information is gained regarding past medical problems, family history ofdisease, and affective complaints. In fact, over 25# ofitems address psychological symptoms such as moods,feelings, and behavior patterns. Although data supporting the validity of this index is sparse, the Cornell Medical Index has been compared to an unstructured medical history interview and was found to be more inclusive in terms of symptom coverage (Brodman et al., 1949).

The Wahler Physical Symptoms Inventory (Wahler, 1968) is a 42 item symptom checklist designed to assess the

12

degree and specificity of somatic complaints. Each physical symptom is rated on a 6-point scale with regard to frequency ranging from 0 (almost never) to 5 (nearly every day). This inventory samples exclusively somatic symptoms or complaints and includes no items related to emotional and mood symptoms such as worry, anxiety or depression. Items contained in the scale were derived from the Cornell Medical Index (Brodman et al.,1949), MMPI Hypochondriasis Scale (Hathaway & McKinley, 1942), and clinician generated items. Data are available regarding the internalconsistency and test-retest reliability of this instrument. Internal consistency estimates based on the KR-20, an index of internal consistency based on scores from a single administration of an instrument, were found to range from .85 to .94 (M = .91) over several samples includinguniversity students, rehabilitation patients, psychiatric patients, and a sample of patients seeking disability compensation. Test-retest reliability has been assessed over varying time intervals in both normal and clinical samples. Coefficients of stability ranged from .69 to .94 (M = .81) over intervals of one day to one week, anddeclined to .45 to .84 (M = .67) over intervals from one to thirteen weeks. The Wahler Physical Symptoms Inventory was found to discriminate significantly between groups of subjects with different levels of expected symptomatology or different incentives for reporting and emphasizing somatic complaints such as healthy college students,

13

disability claimants, rehabilitation patients, andpsychiatric patients (Wahler, 1983). The primaryadvantages related to the use of this inventory as a measure of health status include the provision of a frequency rating of complaints and an exclusive focus on physical (somatic) symptoms.

Another self-report checklist, The Seriousness of Illness Rating Scale (Wyler, Masuda, & Holmes, 1968; 1970) has been used primarily for research purposes. It contains 126 commonly recognized health problems (e.g., peptic ulcer, fainting, leukemia, chest pain, dandruff, irregular heart beat, etc.), all of which have a general severity weight that reflects "seriousness". Weights weredetermined by having 117 physicians and 141 non-physicians rate each health problem with regard to variables such as threat to life, prognosis, degree of disability, and discomfort. The Seriousness of Illness Rating Scale hasbeen found to be significantly related to other measures ofseverity or seriousness of illness including subjective ratings of seriousness of health problems, number of days each health problem was experienced, number of days on which each health problem prevented the pursuit of normal activities, and ratings of overall health status (Garrity et al., 1978). This scale has been frequently employed in studies addressing the relation between stress and illness (e.g., Cooley, Miller, Keesey, Levenspiel, & Sisson, 1979; Garrity et al., 1977; 1978).

A symptom checklist that has been employed as a measure of health status primarily in health surveys is the Health Insurance Study Index (HIS; United States National Center for Health Statistics, 1957). The HIS is a checklist of chronic conditions including dental problems, stomach trouble, nerve trouble, skin problems, ear trouble, eye trouble not relieved by glasses, palsy, stroke, arthritis or rheumatism, and heart and circulation problems. Inconsistent findings have been reportedregarding the reliability of this scale, yet it has been used extensively to assess the health status of elderly populations residing in the community and in institutions (Chappell, 1981). A number of similar, but abbreviated symptom checklists have been constructed primarily for research purposes (e.g., Billings & Moos, 1981; Miller, Ingram, & Davidson, 1976; Murphy & Brown, 1980; Otto, 1979). Little data is available regarding the reliability and validity of these indices making comparisons across studies problematic.

Functional Status Measures

The assessment of functional status involves various attempts to quantify the extent of functional impairment resulting from health problems. An individual's ability to perform the activities required of daily living are generally targeted for assessment and include such

15

categories of behavior as: mobility (e.g., Are you able to drive a car?), self-care (e.g., Are you able to bathe without assistance?), physical activities (e.g., Do you have trouble getting up and down stairs?), and role activities (e.g., Are you able to perform lighthousework?). Although the majority of the generalpopulation will score perfectly on a specific measure of functional status despite the presence of various symptoms, the assessment of funtional ability becomes critical in attempts to measure health status in both chronic illness populations and the elderly. For these groups, a definition of health that merely encompasses the absence of symptoms or illness is meaningless. Therefore, a more important aspect of health is the absence of "debilitating" illness (Maddox & Douglas, 1973).

One of the most frequently used self-report measures of functional status is the Activities of Daily Living(Katz, Akpom, Papsidero, & Weiss, 1973). This scale primarily measures activities associated with self-care, that is, dressing, feeding, bathing, transfer, toileting, and continence. It is easily administered and scored, with the overall score reflecting level of dependency upon others in carrying out self-care activities. Similarinstruments that sample a variety of daily living tasks (e.g., financial matters, getting in and out of bed ) include the Index of Living Skills (Shanas, Townsend, Wedderburn, Friis, Milhoj, & Stehouwer, 1968); The Sickness

Impact Profile (Gilson, Gilson, Bergner, Bobbit, Kressel, Pollard, & Vesselago, 1975); and Shanas* Index of Disability (Shanas et al., 1968). The Functional Limitations Battery (Patrick, Bush, & Chen, 1973) assesses various categories of specific functioning, but additionally the status of each limitation is broken down into chronic limitations, those present for more than three months, and acute limitations, those present for three months or less. Another functional status measure, the Physical Capacities Battery (Hulka & Cassel, 1973) has incorporated a 21-item checklist of health problems for purposes of identifying the reason for each limitation endorsed.

Several health status questionnaires have combined the assessment of functional ability with measures of symptoms in an effort to provide a more sensitive measure of health status. It is thought that despite a given level of functional ability, the presence of symptoms can further impact upon health status. Adding a measure of symptoms also allows one to discriminate between individuals who are without functional limitations, but who are in obviously different states of health. The Health StatusQuestionnaire elicits information on a wide variety of chronic conditions and specific symptoms, and also assesses degree of disability in working, eating, dressing, and mobility. It has been found to be reliable and valid in comparison with medical records (Meltzer & Hockstim, 1970).

17Similarly, the Index of Well-Being (Kaplan, Bush, & Berry, 1976) assesses both the ability to perform activities related to mobility, social activity, self-care, and physical activity, and takes into account the presence of symptoms.

Although functional status has been primarily determined through the use of self-report inventories which target specific behaviors, global ratings of functional limitations have also been employed in research. For example, Naliboff, Cohen, and Yellen (1982) used a 4-point subjective rating scale to assess perceived functional limitations in a migraine headache sample. Theself-ratings were found to correlate .53 with spouse ratings of limitations, and moderate correlations were also found between self-ratings and physical therapist rated indices of physical performance (e.g., strength, range of motion). Ratings of functional status have also been completed by various staff members in a hemodialysis unit such as physicians, nurses, and social workers in an effort to obtain a more objective measure of physical limitations (e.g., Steidl, Finkelstein, Wexler, Feigenbaum, Kitse, Kliger, & Quinlan, 1980). Both self-report inventories and subjective ratings of functional status assess only perceived limitations, not actual performance. Therefore, caution should be used in the interpretation of data resulting from subjective measures of functional status when employed without validation through observation and/or

18ratings by others.

Physician Ratings

Physician ratings of various health parameters are frequently employed to quantify findings resulting from a medical examination. Although global overall ratings of health are most often used (e.g., Maddox & Douglas, 1973)> highly structured formats that provide a broad range of information are available. The Karnofsky Performance Status Scale (Karnofsky, 1949) is a physician-rated functional status measure that has been widely used in its original form since its introduction in 1949. It is an 11-point rating scale that assesses level of functioning in three important areas: self-care, level of activity, andability to work. Despite the long history of its use, reliability and validity data have only become available in the past few years. Recent studies examining thereliability and validity of this instrument reveal discrepant findings. Hutchinson, Boyd, and Feinstein(1980) found the Karnofsky Performance Status Scale to possess low interrater reliability when employed with emergency room patients and chronic hemodialysis patients. Additionally, low correlations were noted between physician ratings and patient self-ratings of functional status. However, a large sample of cancer patients were evaluated with this instrument, and it was found to be both reliable

19

and valid when employed by trained physicians using behaviorally based guidelines (Schag, Heinrich, & Gonz, 1984).

A highly structured "objective" health status indexwas developed by Mossey and Shapiro (1982) and used in astudy of health in an elderly population. The indexcontains a physician designated ICDA-8 three digit diagnostic code for each outpatient visit andhospitalization. The ICDA diagnoses were also classified by the physician according to the following healthparameters: expected duration of illness defined as acuteor chronic; degree of seriousness of illness episode ratedon a 5-point scale; and the presence of an associated risk to recovery. Extensive reliability, predictive validity, and construct validity data were presented to support the use of this index.

Medical Utilization

The rate of medical utilization is an aspect ofillness behavior frequently employed as an index of health status. The rate of health care utilization is generally defined as the frequency of outpatient physician contacts, number of hospitalizations, number of hospital days, and number of surgeries. Data suggests that only the more severe health problems are brought to the attention of a physician, indicating that medical utilization may be a

20relatively insensitive measure of overall health status in the general population (Norman et al., 1982). However, in more severely ill populations such as various chronic illness groups, the frequency of medical utilization is viewed as reflecting important changes in disease status (Wilson & Drury, 1984).

Although such data are frequently ascertained through various self-report formats, the information is considered to be most objective if obtained from medical/hospital records (e.g., Pope, 1982), insurance claim forms (e.g., Mossey & Shapiro, 1982) or a brief encounter form filled out by the physician at the time of an outpatient visit (e.g., McFarlane, Norman, Streiner, Roy, & Scott, 1980; Norman et al. , 1982). These sources yield not only datarelated to the frequency of utilization of health care services, but also provide information regarding presenting complaints, medical diagnosis, and prognostic variables. If information is obtained through self-report formats only, medical utilization as an indicator of health status may be confounded with "care-seeking" behavior. A large percentage of patients in the general population have been found to present to physicians with no demonstrable organic illness, and many instead evidence primarily psychological problems (Kaufman, 1959; Lipowski, 1967; Nigro, 1970). Further, those individuals in whom physicians can find no organic illness frequently demonstrate a high rate of medical utilization (Harrington, 1978). Therefore, when

21

medical utilization is employed as an index of true morbidity in a population, verification of organic pathology through physician diagnosis, laboratory tests and the like is essential.

Mortality

In addition to morbidity reports, mortality has frequently been used as an index of health. Information related to occurrence of death, date of death, and cause of death have been obtained through hospital insurance claim forms, hospital discharge summaries for those persons who die in the hospital, and death certificates. The widespread use of mortality statistics in determining the health of populations is due primarily to the availability of such information, and because such data is inexpensive to access (Chambers, 1982). However, limitations exist in the use of such routinely reported data due to its often questionable accuracy (Hunt, McKenna, McEwen, Backett, Williams, & Papp, 1980). Of particular concern is the lack of thoroughness frequently used in completing death certificates. Similarly, difficulties are encountered in establishing the correct cause of death or diagnosis in hospitalized patients. The problems associated with routinely reported mortality statistics together with a growing interest in collecting data relevant to other health parameters have contributed greatly to the expansion

22of assessment methodology.

Summary

There are problems associated with all currently available measures of health status, and no single measure appears to be superior (Brantley & Bruce, 1986). Therefore, an approach that employs the assessment of various parameters of health in determining physical status appears warranted at present. An example of such an approach is the Arthritis Impact Measurement Scales (Meenan, Gertman, & Mason, 1980). This instrument is composed of a patient self-administered scale and a physician supplement rating scale. The self-administered scale contains a detailed assessment of a number of areas including functional status, pain, medication usage, and psychological functioning. A subjective rating of health status and arthritis status are also included. The physician supplement is composed of a rating of disease activity, and a physician assessed functional status rating scale. A substantial amount of data regarding the reliability and validity of this instrument are available (Kazis, Meenan, & Anderson, 1983; Meenan, Gertman, Mason, & Dunaif, 1982).

23Dialysis

Chronic, intermittent or maintenance dialysis refers to a treatment modality that is performed on a regular, continuing basis to maintain life in individuals with irreversible renal failure or end stage renal disease. Approximately 56,000 Americans suffering from chronic renal failure were being kept alive in 1981 by either extracorporeal hemodialysis or peritoneal dialysis. Annually this treatment costs in excess of 1.2 billion dollars in governmental expenditures (Friedman, 1983).

Chronic renal failure and the resulting changes in physiological functioning will be discussed. Dialysis procedures will also be reviewed, as well as the associated medical complications and data related to survival.

Chronic Renal Failure

In healthy individuals, the kidneys sustain life by performing a number of critical functions which include the filtration of the body’s waste materials; the regulation of the volume and chemical composition of the blood; the production of erythropoietin which regulates the circulating hemoglobin through stimulation of the bone marrow; and assistance in the regulation of blood pressure and extracellular fluid volume through the secretion of the

24hormone, renin.

Chronic renal failure is a functional term that refers to a progressive and generally irreversible decline in kidney function. The etiology is varied, and a large number of diseases can result in chronic renal failure. Kokko (1983) reviewed data related to the causes of chronic renal failure in North American patients on maintenance dialysis. Glomerulonephritis, a heterogenous group of glomerular disorders, was noted to be the causal factor for chronic renal failure in 41.6$ of the patients. Cardiovascular disease and hypertension were the etiological factors in 13.5$; urinary tract disease 10.5$; unknown 8.4$; congenital abnormalities 7.6$; diabetes 7.2$; kidney infections 6.1$; and other known factors 5.1$.

Three stages of chronic renal failure have been described and include diminished renal reserve, renal insufficiency, and uremia (Hekelman & Ostendarp, 1979). Diminished renal reserve reflects reduced renal function, but the excretory and regulatory functions of the kidney are still preserved. Renal insufficiency refers to the accumulation of metabolic wastes in the serum and concomitant elevations in blood urea nitrogen (BUN) and serum creatinine that occurs when at least half of normal kidney function is lost. Uremia is a syndrome resulting from the inability of the kidneys to maintain homeostasis and generally occurs when overall renal function is less than 20 to 25$ of normal. Most organ systems continue to

25function normally during the early phases of advancing renal failure, but the uremic syndrome is characterized by derangement of function of many systems of the body. Treatment alternatives include conservative medical management until renal function decreases to 10 to 15$ of normal. At that stage, maintenance dialysis or renal transplantation are required to sustain life.

Uremic Syndrome

The myriad of clinical symptoms and physiological changes that result from advanced chronic renal failure are referred to as the uremic syndrome. Toxic substances in the body fluid which are normally excreted in the urine have been implicated as causal factors of the uremic syndrome. The consistent finding that azotemic patients improve symptomatically after the initiation of hemodialysis has supported a search for "uremic toxins" (Bergstrom & Furst, 1983). Initially, researchers thought that uremic symptoms were produced by high concentrations of urea, a by-product of protein metabolism that is eliminated primarily by the kidneys. But strong evidence indicates that all of the clinical features of chronic renal failure are not due solely to the accumulation of urea (Blachley & Knochel, 1981). A number of substances have been identified in elevated concentrations in the serum of uremic patients, but to date all of the clinical

26manifestations of the uremic syndrome cannot be explained by the accumulation of known compounds (Bergstrom & Furst, 1983).

The various symptoms of the uremic syndrome and the associated abnormalities in physiological functioning will be briefly reviewed.

Cardiovascular system. The most frequently occurring cardiovascular disorder associated with chronic renal failure is hypertension. Two causal factors have been identified in uremic patients and include the inability of the kidneys to excrete sodium and water resulting in hypervolemia, and derangement in the renin-angiotensin system resulting in the production of large quantities of renin. Approximately 70% of the hypertension in uremia due to fluid and sodium overload can be controlled with dietary reductions in sodium and fluid intake and/or fluid removal by dialysis (Delano, 1983). However, hypertensionresulting from increased renin production may require administration of antihypertensive medication or, in some cases, bilateral nephrectomy to remove the source of renin production (Onesti, Kim, Greco, del Guercio, Fernandes, & Swartz, 1975).

Hyperkalemia occurs in uremic patients due to the kidneys' diminished capacity to excrete potassium. In excess of 90% of the daily intake of potassium is excreted through the kidneys in healthy individuals. Patients with chronic renal failure are placed on a potassium restricted

27

diet, but failure to follow the prescribed dietary regimen can result in hyperkalemia. Other contributing factors include acidosis, ingestion of potassium in medications, blood transfusions, and bleeding. Hyperkalemia can produce ventricular tachycardia, fibrillation, and cardiac arrest, and has been termed the "silent killer" in renal disease, because such complications can occur with little warning.

Pericarditis, an inflammation of the pericardial sac occurs in approximately 25% of chronic renal failure patients (Comty, Wathen, & Shapiro, 1976). Although the etiology is unknown, uremic toxins are thought to inflame and irritate the pericardium. Treatment generally focuses upon frequent dialysis until the condition improves(Silverberg, Oreopoulos, Wise, Uden, Meindok, Jones, Rapaport, & de Veber, 1977). Complications include hemorrhagic effusion and cardiac tamponade.

Congestive heart failure is generally the first clinical manifestation of depressed cardiac function in chronic renal failure. Symptoms include dizziness, changes in mental status, and fatigue with minimum exertion.Although due to low cardiac output, such symptomatology is frequently attributed to depression, atherosclerotic disease, metabolic abnormalities or old age (Forst &O'Rourke, 1981). Contributing factors includehypertension, retention of sodium and water, and severe anemia.

28Hematological system. A number of hematological

abnormalities are associated with chronic renal failure, but the most universal is anemia. When the hematocrit falls below 20 - 25%, a number of symptoms generally occur such as pallor, dyspnea, fatigue or tachycardia. Theetiology of anemia secondary to chronic renal failure is multifactorial and includes inadequate production of erythropoietin; shortened red blood cell survival due to elevated uremic toxins; blood loss due to mucosalirritation; blood loss from iatrogenic blood testing; and loss of blood in hemodialysis tubing and coils (Delano, 1983). Hemodialysis will decrease the level of uremic toxins in the serum and increase the lifespan of red blood cells toward normal. However, treatment of severe anemia may also include blood transfusions. Androgen therapy has been used with some success to increase the hematocrit in these patients (Raga, Kramer, & Rosenbaum, 1977). Iron or folate supplements are administered if needed to correct deficiencies.

Platelet abnormalities are commonly observed inchronic renal failure and include a decrease in the numberof platelets and a defect in the quality of platelets(Kukko, 1983). Elevated uremic toxins are thought toincrease peripheral destruction of platelets and result inprolonged bleeding time, and a decrease in plateletadhesiveness and platelet aggregation. A hemorrhagic tendency manifested by excessive bleeding or bruising after

29

trauma is frequently observed. Dialysis can often correct these abnormalities.

Chronic renal failure is also associated with an increased susceptibility to infections which may be due to a number of factors (Montgomerie, Kalmanson, & Guze, 1968; Wilson, Kilpatrick, & Talmadi, 1965). Typically, there are abnormal responses to mitogenic stimulation. Thisabnormality may be due to uremic toxins, because lymphocyte function is found to return to nearly normal following vigorous dialysis (Kukko, 1983). The neutrophil count is generally normal in these patients, however, a transient decrease is found to occur following the onset of hemodialysis. Patients generally remain asymptomatic, and the count returns to normal within one to two hours. Additionally, the ability of leukocytes to phagocytize bacteria is decreased in uremic patients. Frequent exposure of dialysis patients to bacterial and viral agents contributes to the increased frequency of infections. Also, of concern is the high incidence of hepatitis in these patients thought to be due to increased exposure, as well as multiple blood transfusions (Galbreath, Portmann, Eddelston, Williams, & Gower, 1975).

Gastrointestinal system. Anorexia, nausea, vomiting, and hiccup are among the earliest clinical manifestations of uremia. Anorexia and vomiting are of particular concern, in that such symptoms can result in fluid and electrolyte imbalance and contribute to weight loss.

30Generally, these symptoms abate after the initiation of dialysis.

Other symptoms common in advanced uremia include the smell of urine and ammonia to the breath, gum ulceration, metallic taste, and stomatitis (Brundage, 1976). Such symptoms are thought to be due to the conversion of urea in the saliva to ammonia. Additional gastrointestinalproblems such as glossitis, gastritis, and enterocolitis result primarily from inflammation and ulceration of the gastrointestinal mucosa due to elevated uremic toxins.

Constipation is also a frequent complaint associated with chronic renal failure. This symptom results from the administration of aluminum hydroxide gels which are given as phosphate binders to treat osteodystrophy. Consequently, concomitant administration of a stool softener is frequently required (Stone, 1977).

Integument. In advanced chronic renal failure, the skin has a characteristic discoloration of a grayish-bronze color due to retained pigments which are normally excreted by the kidneys. As noted previously, pallor results from anemia. The skin is usually dry and scaly due to a decrease in the activity of oil glands and a decrease in the size of sweat glands. Pruritus (itching) is a frequent complaint in uremic patients and can be quite severe. Usually there are no dermatologic findings, and to date no single causative factor has been identified (Kukko, 1983)• Therefore, treatment is symptomatic and has included oral

antihistamine and sedative administration, dietary protein restriction, increased frequency of dialysis sessions, and exposure to ultraviolet light (Lazarus, 1981). Due to the platelet abnormalities observed in uremia and increased fragility of capillaries, purpura and ecchymoses frequently result from even minor blows to the skin.

Neurological abnormalities. The neurologicalmanifestations of uremia are numerous, and patients may evidence abnormalities in central and peripheral nervous system function (Kukko, 1983). Cerebral impairment may result from direct toxicity of the central nervous system producing symptoms of headache, lassitude, insomnia, drowsiness, stupor, coma, and convulsions. Cognitive functioning may be affected as evidenced by an inability to concentrate, shortened attention span, memory deficits, and confusion. Behavioral and affective changes may include irritability, withdrawal, apathy, lack of interest in the environment, decreased or absent libido, and depression, as well as psychotic behavior including delusional thinking.

Chronic renal failure is also commonly associated with a slowing of peripheral nerve conduction leading to peripheral neuropathy. The pathophysiology in mostpatients is not well understood. Peripheral neuropathy produces a number of clinical symptoms including painful paresthesias of the extremities, loss of deep tendon reflexes, muscular weakness, and occasionally sensory deficits. The most common motor nerve abnormality observed

32

in these patients is bilateral food drop (Lancaster & Pierce, 1979). The "restless legs syndrome" is one of the earliest signs of peripheral neuropathy and includes painful parasthesias, cramping, and itching of the lower extremities (Callahan, 1966; Teschan & Ginn, 1976). The "burning feet syndrome" is frequently observed and is characterized by bilateral burning sensations of the feet, as well as painful paresthesias. The symptoms of neuropathy have been improved by the use of large surface membrane dialyzers and by extending the period of dialysis.

Metabolic abnormalities. Uremic patients evidence a number of metabolic abnormalities which include aderangement in protein metabolism, carbohydrate intolerance, and hyperlipidemia.

Protein is broken down into amino acids during digestion. Urea and nitrogen are waste products resulting from amino acid metabolism. In healthy individuals, urea and nitrogen are eliminated by the kidneys and excreted in the urine. However, in chronic renal failure, theseprotein byproducts accumulate in the blood and are reflected in a rise in BUN (blood urea nitrogen). As noted previously, elevated concentrations of urea areresponsible, in part, for the development of the uremic syndrome. By restricting the amount of protein ingested, the concentration of urea and nitrogen in the serum can be reduced. Therefore, uremic patients are frequently placed on low-protein diets.

The impaired glucose tolerance associated with chronic renal failure is referred to as "uremic pseudodiabetes mellitis" (DeFronzo, Tobin, Rowey, Sapir, Kramer, & Andres, 1974). Insulin is filtered and metabolized by the kidneys in healthy individuals. But in chronic renal failure,slowed insulin degradation by the kidneys leads to an increased half-life and elevated serum levels. Although increased levels of insulin are available, the insulin is not effectively utilized due to peripheral resistance. The decreased sensitivity of peripheral tissues to insulinfound in these patients is thought to be due to an enzyme abnormality associated with uremia (Lancaster & Pierce, 1979). There does not appear to be a defect in theproduction of insulin in response to glucose in uremic patients, but a delay in production has been found to occur. Maintenance dialysis appears to improve insulin and glucose metabolism, but does not return them to normal(DeFronzo et al., 1974).

•7

Abnormalities in lipid metabolism associated with chronic renal failure include elevated triglycerideconcentrations and near normal or slightly elevated plasma cholesterol levels. Etiologic factors in hyper­triglyceridemia are not well understood, but defectsspecific to triglyceride removal appear to be involved(e.g.,, Ibels, Reardon, & Nestel, 1976). The serum level of triglycerides does not decrease after the initiation of dialysis. In fact, chronic dialysis patients evidence a

34higher incidence of this metabolic abnormality than nondialyzed uremic patients (Lancaster & Pierce, 1979). Hyperlipidemia has been linked to atherosclerosis, and many dialysis patients develop this disease.

Skeletal System. The skeletal changes commonlyassociated with uremia are collectively referred to as renal osteodystrophy. They include osteitis fibrosa,osteomalacia, osteoporosis, and osteosclerosis which are manifested by such symptoms as bone pain, fractures, and periarticular calcifications. Some degree of bone disease is almost universal in renal failure (Kukko, 1983).Hyperparathyroidism, alterations in vitamin D metabolism, and chronic low grade metabolic acidosis are causal factors in renal osteodystrophy, and together result in thedemineralization of the skeleton. Prevention and treatment of renal osteodystrophy includes lowering of serum phosphate with the administration of phosphate binding mechanisms; administration of vitamin D supplements to increase calcium levels; and, in cases of severe bone disease, parathyroidectomy may be indicated.

Endocrine Abnormalities. Among the associatedendocrine problems in chronic renal failure are abnormalities in gonadal function. Advanced renal failure commonly results in infertility, amenorrhea, cessation of ovulation, decreased libido, and impotence (Bailey, 1977). Research regarding the etiology of sexual dysfunction in uremic patients has implicated psychological, as well as

35

pathophysiological factors (Bommer, Tschoppe, Ritz, & Andrassy, 1976). Maintenance dialysis is associated with the resumption of ovulation and menstruation, and some improvement in sex drive. But renal transplantation often leads to significant improvement in gonadal function (e.g., Holdsworth, Dekretser, & Atkins, 1978).

In summary, chronic renal failure affects the functioning of nearly all major systems of the body. Maintenance dialysis is able to improve or correct some of the abnormalities associated with the uremic syndrome, but other symptoms and derangements persist.

Dialysis Procedures

Dialysis involves the removal of accumulated metabolic waste products from the blood, and the restoration of water, electrolyte, and acid-base balance. This process can be achieved either through the use of synthetic membranes (extracorporeal hemodialysis) or the peritoneal membrane within the abdominal cavity (peritoneal dialysis). Both procedures rely on diffusion down a concentration gradient and across a semipermeable membrane which selectively permits the transfer of particles according to their molecular weight. In this way, adjustments in concentration and content of blood solutes are achieved.

Extracorporeal hemodialysis requires access to the circulatory system which involves either an internal or

36external device which shunts the blood between an artery and vein. The blood is passed through an "artificial kidney" which contains a semipermeable membrane and dialyzing fluid similar in composition to normal plasma. Fluid and wastes are removed from the serum and returned to the patient through a vein. Factors such as degree of residual renal function, body size, dietary intake, and clinical status determine the duration of dialysis, type and size of the dialyzer, and rate of blood flow. Stable patients are typically dialyzed for three to four hours per treatment, and treatment sessions are scheduled usually three times a week (Lazarus, 1981).

m r ■

Peritoneal dialysis requires access to the peritoneal space where a catheter is inserted, and dialysis fluid is introduced and then drained. It is a simpler procedure than hemodialysis and requires less specialized equipment. Peritoneal dialysis is occasionally employed in patients with chronic renal failure who develop complications in access availability and complications due to heparinization. Clearly, the major disadvantagesassociated with peritoneal dialysis are the time required for treatment and increased risk of infection.

Maintenance dialysis can be performed in an outpatient clinic setting staffed by trained nurses and technicians or at home. Even though self-dialysis at home is an efficient way to decrease the cost of long-term dialysis and is preferable to some patients, the majority of patients

receive treatment at outpatient centers (Blagg & Scribner, 1980; Rennie, 1978). In fact,. the 1970's witnessed a sustained decrease in the proportion of all hemodialysis patients treated at home from a reported 39.856 in 1972 to 12.9$ in 1977. The decline in number of patients on dialysis at home has been attributed to physician bias rather than a clinical failure of this mode of therapy (Friedman, Delano, & Butt, 1978). Other factors which mitigate against home dialysis include severe medical complications, the lack of an adequate dialysis partner or appropriate home setting, and current federal reimbursement practices (Wauters, Hunziker, & Brunner, 1983). In-center dialysis can be conducted by trained staff or by the patient. In self-care dialysis, the patient hasresponsibility for managing his own therapeutic regimen, and staff are available for assistance as needed.

Dialysis Treatment Regimen. In addition to regularly scheduled dialysis sessions, the treatment regimen for these patients includes replacement therapy and dietary restrictions. Ascorbic acid, folic acid, and B-complex vitamins are removed to a significant extent during dialysis due to their size and water solubility. In order to prevent depletion, multiple vitamins and folate are routinely prescribed. Aluminum hydroxide is administered to decrease serum phosphorous levels. Intake of vitamin D may be recommended to decrease hypocalcemia and possibly prevent or reverse secondary hyperparathyroidism. Iron

38supplements and blood transfusions are used in the treatment of anemia.

Dietary restrictions are extensive and of critical importance in the dialysis treatment regimen. Fluid control helps to decrease hypertension and/or edema. Phosphorous intake is restricted in order to augment the phosphorous-binding gels and decrease serum levels. Sodium contributes to hypertension, as well as hypotension, and excessive thirst, and intake is closely controlled. Ingestion of potassium through diet or medications is restricted to prevent hyperkalemia. An adequate level of protein is needed to maintain positive nitrogen balance, but intake is controlled in order to keep blood urea nitrogen (BUN) levels low.

Adequate Treatment. The adequate management ofpatients with chronic renal failure on maintenance dialysis remains a significant challenge. A well delineated and acceptable medical definition of adequate dialysis does not exist and guidelines are not available (Lazarus, 1981; Newman, 1983). No single clinical or biochemical measure such as BUN or creatinine or a combination of measures has proven to be acceptable in defining proper treatment.

However, expectations regarding patient functioning as a result of adequate treatment are quite similar. Delano (1983) defined adequate dialysis as, "that which permits the patient to be rehabilitated, eat a reasonable diet, make blood, maintain a normal blood pressure, and prevent

39

the development or progression of neuropathy." Newman(1983) noted the following description of adequate treatment, "in addition to acceptable blood chemistries, a stable and well-functioning mind and body during and between treatments." And Henderson (1983) proposed that adequate dialysis should "maintain a patient free of symptoms and signs of flagrant uremia (e.g., prevention of

peripheral neuropathy, pericarditis, pleuritis)".

Medical,Complications

The medical complications associated with maintenance dialysis include various mechanical and/or iatrogenic complications that can occur during the dialysis procedure; medical complications specific to dialysis treatment; and persistent uremic problems.

Despite the complexity of the dialysis procedure itself, hemodialysis is considered to be relatively safe. The risk of death resulting from technical or human error in the hemodialysis procedure has been estimated to be 1 in 76,138 (Friedman & Lundin, 1982).

Dialysis. Hypotension, dialysis disequilibriumsyndrome, muscle cramps, and complications associated with anticoagulation are problems frequently encountered as a result of the dialysis procedure (Lazarus, 1981).

Episodes of hypotension are common in hemodialysis, but rarely occur in peritoneal dialysis. Approximately 20

40to 30^ of hemodialysis sessions involve symptomatic hypotension which consists of a drop in blood pressure associated with symptoms of pallor, dizziness, malaise, unexplained anxiety, nausea, vomiting, and cramps (Blagg, 1983). These symptoms are thought to be related to the volume of excess fluid removed and the rate of removal during hemodialysis. Treatment includes administration of hypertonic mannitol, isotonic and hypertonic saline.

Dialysis disequilibrium syndrome is no longer a frequent complication of maintenance dialysis, although it still occurs in association with severe azotemia or when rapid dialysis takes place (Lazarus, 1981). The onset of symptoms generally occurs at the end of the treatment session and may include headache, restlessness, nausea, vomiting, blurring of vision, muscle twitching, muscle cramps, disorientation, tremors, seizures or coma (Delano, 1983). The exact etiology is unknown, but a consistent finding in these patients is elevated cerebral spinal fluid and cerebral edema (Blagg, 1983). It is thought that the disequilibrium syndrome may be due to a more rapid clearance of urea from the blood than from the cerebral spinal fluid. This complication may be prevented or minimized by utilizing shorter and more frequent dialysis sessions.

Painful muscle cramps are common in maintenance dialysis patients and occur both during and between treatment sessions. They are thought to be the result of

41hypoosmolality or rapid extracellular volume removal induced by sodium shifts out of the muscle cell (Battista, 1979). Symptomatic relief can usually be obtained with increased sodium intake or saline replacement.

Anticoagulants such as heparin are routinely used in hemodialysis to prevent clotting of the arteriovenous access. Complications arise from an increased incidence of spontaneous bleeding episodes which can include subdural hematoma, cerebral hemorrhage, gastrointestinal hemorrhage, hypermenorrhagia, hemorrhagic effusion in the pericardium, pleura, joints, and retroperitoneal space. In addition to the administration of anticoagulants, other factors which contribute to bleeding episodes include functional platelet abnormalities which prolong bleeding time and possible increased prostacyclin activity (Blagg, 1983).

Uremia. A number of the symptoms and complications associated with the uremic syndrome are not corrected by maintenance dialysis.

Although the platelet abnormalities and bleeding tendencies associated with chronic renal failure improve with maintenance dialysis, anemia persists. Hemodialysis may contribute to anemia through added blood loss resulting from the dialysis procedure.

Hypertension remains a significant problem in maintenance dialysis patients contributing to the high incidence of coronary heart disease. Hypertriglyceridemia which has been associated with atherosclerotic disease also

42persists despite adequate dialysis. Myocardial infarction and cerebrovascular accidents are the most common causes of death in chronic renal failure patients on dialysis (Lindner, Chara, Sherrard, & Scribner, 1974).

Infection is the second most common cause of death in maintenance dialysis patients (Henderson, 1983). Infections of the vascular access, diverticulosis, diverticulitis, and urinary tract infections are common. These patients are particularly at risk for viral hepatitis.

Secondary hyperparathryoidism is of particular concern in maintenance dialysis patients. If not adequately managed, this conditiion can result in severe bone disease. Despite adequate dialysis, many patients exhibitintermittent and mild chronic acidosis which may contribute to metabolic bone disease.

Maintenance dialysis will generally halt the progression of peripheral neuropathy, but preexisting neuropathy will not be significantly improved (Lazarus, 1981). However, diabetic patients with chronic renal disease may evidence progression of peripheral neuropathy despite adequate treatment efforts.

Additional uremic symptoms which persist despite adequate dialysis include insomnia, restlessness, andpruritus. These symptoms are recurrent and largelyunresponsive to treatment.

43Morbidity

Patients maintained on dialysis receive asignificantly high level of on-going medical supervision from physicians and specialized nursing staff as a function of their treatment regimen. Consequently, the rate of hospitalization, as an index of morbidity, is considered to be an extremely important indicator of health and is viewed as reflecting more serious alterations in status (Wilson & Drury, 1984).

The frequency and duration of hospitalizations over a one-year period in 1981 were examined in a largemulticenter regional study of 946 dialysis patients (Carlson, Duncan, Naessens, & Johnson, 1984). Data were included for patients maintained on hemodialysis, continuous ambulatory peritoneal dialysis, and intermittent peritoneal dialysis. Results revealed that the mean number of hospitalizations per patient during the year was 1.77, and the mean duration of hospitalizations was 15.38 days. Of the patients who were hospitalized, 15% experienced more than three admissions. A trend was noted for an increase in the frequency of hospitalizations with advancing age. Not surprisingly, both the number of hospitalizations andthe number of hospital days were found to be highest fordialysis patients who died during the study period. Of the 1256 of the sample that died during the year, 75% had been hospitalized and 57% died in the hospital. Factors

44associated with increased morbidity also included a diagnosis of diabetes and other concurrent medical disorders such as arteriosclerotic heart disease, cerebrovascular accident, malignant lesion, and chronic obstructive pulmonary disease. Patients maintained on peritoneal dialysis were found to experience a higher rate and longer duration of hospitalization than patients on hemodialysis. This finding confirmed a similar report by Bovberg, Diamond, Held, and Pauly (1983). No differences were found in the rate of hospitalization for patients maintained on home versus center dialysis programs.

The major causal factors associated with hospitalization in this sample, accounting for 50$ of the admissions, were reported to be dialysis access problems, gastrointestinal diseases, and cardiac disease (Carlson et al., 1984). The most common gastrointestinal problem found was hemorrhage, and the most frequently occurring cardiovascular problems associated with hospitalization were ischemic heart disease, cardiac arrhythmias, and pericarditis. Other studies of morbidity in dialysis patients have reported similar findings. Hirschman, Wolfson, Mosimann, Clark, Dante, and Wineman (1981) examined 1,049 patients on dialysis for a period of one year, and found that 4856 required hospitalization at least once . The reasons noted for hospitalization included vascular access problems (26%), cardiovascular complications (16$), and infections (11$).

Long-term hospitalization, defined as a duration of at least three months, was examined in European patients on hemodialysis and in patients who had received a first renal transplant (Robinson & Hawkins, 1980). Hospital admissions for routine dialysis, routine transplant evaluations or research were excluded from the analysis. During the first year of treatment, 9% of the patients in both groups were found to have been hospitalized for periods exceeding three months. Again, the data revealed more frequenthospitalizations with advancing age across treatment modalities. Although the findings reflected a lower rate of hospitalization for transplant patients across all age groups as compared to hemodialysis patients, transplantfailures were associated with a marked increase in morbidity.

Mortality

Survival data are available for patients with chronic renal failure across treatment modalities includingdialysis procedures and renal transplantation. Krakauer,Grauman, McMullan, and Creede (1983) presented survivalrates for the United States as a whole based on data compiled by the Medical Information System of the Health Care Financing Administration. All recipients of medicare benefits for end-stage renal disease are included in this data which represents 93% of the U.S. population with the

disease. Over 70,000 patients who had begun to receive dialysis or who had received a transplant between 1977 and 1980 were included in the analysis. Overall,transplantation success, i.e., graft retention, showed progressive and substantial increases over this time period, whereas survival rates for both dialysis and transplant recipients remained relatively stable. Dialysis patients were found to be subject to a nearly constant risk of death over the first four years of treatment with the risk remaining at approximately 19$ per year. Short-term survival, defined as one year, for dialysis patients under the age of 50 was found to be superior. The data alsosupported a finding previously noted by others (Reiman, 1982; Rostand, Kirk, Rutsky, & Pate, 1982) that theincidence of chronic renal failure is over three times higher among blacks than whites in the U.S. The increased incidence in blacks has been attributed to the elevated rate of severe hypertensive disease (Reiman, 1982). Blacks on dialysis were found to have a lower death rate than whites. Table 1 contains one-year and three-year survival rates for patients receiving dialysis and transplant patients categorized by sex, age, race, and primary disease. Comparable one-year survival rates were reported for European patients under the age of 55 as compiled by the European Dialysis and Transplant Association (1979;1980; 1981). European data did reflect increased survivalfor all dialysis patients at three years. Survival data

47were also analyzed in a multicenter regional study (End-Stage Renal Disease Network 7) of 2,725 patients who began chronic dialysis or received a renal transplant during a five and one-half year study period ending in 1983 (Hellerstedt, Johnson, Ascher, Kjellstrand, Knutson, Shapiro, & Sterioff, 1984). Overall, survival rates compared favorably with those of the nation as a whole as reported by Krakauer et al. (1983).

Long-term survival of patients with chronic renal failure appears to be quite variable across studies. A survival rate of 18$ was reported after 12 years in one sample of hemodialysis patients (Vollmer, Wahl, & Blagg, 1983), whereas 85$ of another sample receiving dialysis were alive after 10 years (Charra, Calemard, Cuche, & Laurent, 1983). In a review of six large scale studies which examined survival in chronic hemodialysis patients, Gabriel (1984) noted that the reasons for such a wide variation in long-term success is not clear at this time. A factor which may significantly influence survival data is the variation found in the criteria employed for the selection of patients with chronic renal failure offered treatment. Criteria have been found to differ with respect to factors such as age, diagnoses, and associated diseases. In 1973, .the Social Security Administration assumed responsibility for payment of 80$ of the costs of treating chronic renal failure by hemodialysis or renal

T

transplantation, thereby allowing selection criteria to be

48significantly relaxed in centers across the U.S. However, variable selection criteria are still found across European Centers (Charra et al., 1983).

Factors Affecting Survival. Although it is generally accepted that the survival of patients with chronic renal failure is clearly age and diagnosis dependent, recent studies suggest that other factors may also be important (Friedman, Delano, & Butt, 1978). Hellerstedt et al.,(1984) analyzed over 35 risk factors in a large regional study of survival in patients with chronic renal failure over five and one-half years. Demographic variables and various health parameters were examined in 2,725 patients. Results revealed that survival rates were appreciably lower for patients with diabetes, irrespective of treatment modality. The age of the patient at the time of initiation of treatment significantly affected survival with an increased risk of mortality found to be associated with advancing age. The lowest survival rates were associated with a primary diagnosis of malignant lesions including renal tumors or multiple myeloma, primary hypertensive disease, and diabetes. The highest five-year survival rates were found for the primary diagnoses of glomerulonephritis, systemic lupus erythematosus, and focal glomerulosclerosis. Also the presence of the following concomitant morbid conditions at the time of initiation of treatment negatively affected survival, irrespective of treatment modality: arteriosclerotic heart disease, chronic

pulmonary disease, peripheral vascular disease, and noncutaneous malignant lesions. The coexistence of multiple morbid conditions were also found to negatively effect survival. The outcome of 220 patients who received either their first dialysis treatment or first renal transplant between 1970 and 1975 was analyzed (Hutchinson, Thomas, & MacGibbon, 1982). A number of variables present at the time of initiation of treatment were examined to determine their effects on prognosis. Three variables, accounting for the majority of the variance, were found to be related to survival in both dialysis and renal transplant patients and included age, diagnosis of diabetes mellitus, and the presence of left-sided heart failure as evidenced by pulmonary edema. The presence of depression was also found to be significantly related to survival, although this factor was only present in a small percentage of the 220 patients studied.

The effects of psychological, demographic, and physiological variables on survival were examined in a study of 285 home dialysis patients followed over the course of 18 months (Wai, Richmond, Burton, & Lindsay, 1981). Psychological variables of anxiety, depression, and denial were assessed through the use of a structured interview. Physiological parameters assessed included frequency of excessive interdialysis weight gain, anemia, systolic blood pressure, serum creatinine, total protein, albumin, potassium, and phosphate levels. Only four

variables were found to significantly discriminate between survivors and nonsurvivors which included age, depression, level of self-reported stress associated with dialysis, and albumin level. No significant differences were found for variables thought to be highly related to survival such as hypertension or parameters usually indicative ofnoncompliance with the medical regimen such as excessive fluid gain between dialysis sessions or elevated levels of phosphorous. Similarly, Farmer, Bewick, Parsons, andSnowden (1979) and Foster, Cohn, and McKegney (1973) found no significant relation between biochemical measures of creatinine, blood urea nitrogen, and hematocrit andsurvival. However, the two latter studies examined a large number of variables in small samples limiting theconclusions that can be drawn from their data.

Psychological correlates of survival in hemodialysis patients were examined using the Minnesota Multiphasic Personality Inventory (MMPI; Ziarnik, Freeman, Sherrard, & Calsyn, 1977). Subjects were tested prior to the initiation of dialysis, and three groups were compared: patients who died during the first year of dialysis, patients who remained on dialysis three to seven years, and patients who survived seven to ten years. Significant differences were found between patients who only survived one year and the other two groups on the MMPI. The patients who died during the first year evidenced feelings of helplessness, high levels of anxiety and depression, and

a preoccupation with somatic difficulties. Patients surviving three years or more were characterized by dependency and mild levels of depression. Prior to drawing any conclusions from the data, the presence of associated medical problems such as heart disease, hypertension, angina, obesity, diabetes, and stroke were examined for all three groups. Significant differences were found, with patients who died within the first year of dialysisdemonstrating significantly more associated medical problems than the other two groups. It was proposed that this factor may have accounted for both the differences in survival, as well as the differences in MMPI results.Other researchers have suggested that severity of illness may account for the consistent elevations observed on scales 1(Hypochondriasis), 2(Depression), and 3(Hysteria) of the MMPI in hemodialysis patients (Osberg, Meares,McKee, & Burnett, 1982). However, De-Nour and Czaczkes(1976) reported data that suggests that severe psychiatric complications such as depression may have a significant impact on the survival of patients in the early stages of dialysis. A sample of 120 patients beginning dialysis were followed over the course of 36 months. Of the total sample, 12.5% died within the first six months of treatment. A large percentage of deaths were attributed toa refusal to continue treatment and suicide.

Recent studies have examined characteristics of long-term survivors on dialysis in an effort to delineate

factors associated with increased risk. Guttman (1983) compiled data from Sweden and the U.S. on 25 patients who had survived on dialysis for 14 years or more. A number of health parameters were assessed, and generally these patients at the time of the survey were characterized by a low prevalence of hypertension and smoking, low rate of appreciable cardiovascular morbidity, and compliance with the dialysis regimen as reflected by low interdialysis weight gain and satisfactory averages and ranges of selected chemistries which included serum calcium, phosphorous, and alkaline phosphate concentrations. The major medical complication experienced by these patients was musculoskeletal disease. As noted earlier, Charra et al. (1983) reported an excellent 10-year survival rate of85% in a sample of hemodialysis patients. The authors postulated that good control of blood pressure achieved through strict regimens of fluid control and low-salt diet contributed significantly to the high survival rate. The sample was described as completely free of clinical evidence of cerebrovascular or ischemic heart disease, major causes of death in dialysis patients.

Causes of Death. The most common causes of death in patients with chronic renal failure continue to be myocardial infarction and cerebrovascular accident. Robinson and Hawkins (1980) summarized data on European patients on both hemodialysis and peritoneal dialysis from the European Dialysis and Transplant Association. Between

the years 1970 and 1980, the five leading causes of death for patients on hemodialysis were listed as follows: cardiac arrest, cause unknown; cerebrovascular accident; other causes of cardiac failure; myocardial infarction; and septicemia. For those patients maintained on peritoneal dialysis, the leading causes of death were cardiac failure, cerebrovascular accident, septicemia, and cachexia. The proportion of deaths due to various causes were remarkably similar for hemodialysis patients and peritoneal dialysis patients.

Three additional causes of death in hemodialysis patients warrant review: patients in which a decision ismade to terminate treatment; patients who die from suicide; and patients who through their own apathy or neglect express an obvious disinterest in survival, i.e., the noncompliant patient. Rodin, Chmara, Ennis, Fenton, Locking, and Steinhouse (1981) reviewed a sample of hemodialysis patients treated between 1976 and 1979. In 26$ of the 80 deaths that occurred during this time, an active decision was made to terminate further treatment either initially proposed by the medical staff or by the patient, himself. In cases where the staff proposed termination, the majority of these patients were considered to be mentally incompetent due to severe organic brain syndromes usually associated with a grave medical condition.

Suicidal behavior has been reported to occur with increased frequency in hemodialysis patients. Abram, Moore, and Westervelt (1971) estimated the incidence of suicide in hemodialysis patients in the United States to be 100 times that of the general population (10 per 100,000). Their findings were based on a survey completed by 125 of 201 dialysis centers contacted across the United States. If patients who had abandoned treatment, i.e., requested withdrawal from dialysis programs or refused to accept further treatment, were included as suicides, the rate rose to 400 times that of the general population.

A more comprehensive study was conducted on the occurrence of suicide in hemodialysis patients and patients with a functioning renal transplant in Switzerland and Europe (Haenel, Brunner, & Battegay, 1980). Questionaires were completed by 25 of the 30 hemodialysis centers in Switzerland and six renal transplant units. The incidence of suicide in hemodialysis patients was found to be similar to that of renal transplant patients. The suicide rate among both groups was reported to be 10 times that of the general population. Again, if the death rate due to refusal of treatment is included with the actual suicides, the incidence of suicide rose to 25 times that of the general population. Data were also obtained from the European Dialysis and Transplant Association Registry (EDTA) to determine death rates due to suicide and refusal of treatment in Europe. For the year 1977, the death rate

attributed to suicide and refusal of treatment in European dialysis patients was reported to be 108 per 100,000. This finding is considerably higher than that of the general population. The World Health Statistics estimated thedeath rate due to suicide to be 4-5 per 100,000 in Mediterranean countries and 20-25 per 100,000 in Central Europe and Scandinavian countries. As Levy (1978) noted, suicide statistics are probably in error and exceedingly low due to the difficulties involved in accuratelyestablishing the number of deaths due to suicide.Nonetheless, the available data does indicate asignificantly higher rate of suicide in hemodialysis patients than in the general population. It is difficult to determine whether suicide rates for hemodialysis patients differ from other groups of chronic illness patients due to the absence of comparative studies.

The suicidal behavior found in hemodialysis patients has most commonly been associated wih impaired quality of life and the high prevalence of depression found in these patients (Foster, Cohn, & McKegney, 1973; Levy and Wynbrandt, 1975). The high incidence of suicide has also been attributed to the fact that hemodialysis patients have ready access to means for ending their lives which are not available to others (Haenel, Brunner, & Battegay, 1981). A method of suicide in these patients is exsanguination,i.e., disconnecting or cutting of the arteriovenous shunt. Also, noncompliance with the dialysis regimen such as

ingestion of large amounts of fluid or foods forbidden by the dialysis regimen can result in death. For example, two patients in one study died as a result of hyperkalemia in which excessive intake of foods rich in potassium caused death. Other suicides were reportedly caused by excessive salt and fluid intake (Haenel et al., 1980).Interestingly, Abrams et al., (1971) reported that the majority of deaths in hemodialysis patients, not necessarily suicides, were related to noncompliance with the dialysis regimen. An inability or refusal to adhere tothe numerous restrictions included in the dialysis regimenhas been reported to occur in one-third to one-half of patients at any one time (De-Nour, 1970). At what point this problematic behavior, i.e., noncompliance, is also considered to be suicidal behavior is a perplexing issue and may contribute to the high rates of suicide reported by various dialysis centers. Other contributing factors include the higher than average age of patients on maintenance dialysis, and the high incidence of additional medical problems found in these patients such as diabetes mellitus and neoplastic disease. Further research isclearly needed to examine the effects of psychological factors such as depression and noncompliance on the survival of dialysis patients. Nonetheless, noncompliance can adversely affect the health status of dialysispatients, and remains an area of particular concern to medical staff.

57

Summary

The prognosis for patients with chronic renal failure continues to improve as advances in technology become available, particularly in the area of renal transplantation (Knapp, 1982). The young patient with chronic renal failure, under the age of 40 who is normotensive and has no coexisting medical problems, has the best chances for survival at this time (Gabriel, 1984)-.

It is generally agreed that for patients with chronic renal failure a higher quality of life results from a well functioning renal transplant than the very best maintenance dialysis (Henderson, 1983). However, for many patients this alternative is not readily available. Patients can be maintained on dialysis for an indefinite period of time, but as reviewed, a number of complications are associated with this treatment modality and some abnormalities in physiological functioning persist despite adequate treatment. Additionally, patient noncompliance with the complex and demanding medical regimen is a common finding in dialysis patients and may contribute significantly to the occurrence of medical complications. No other chronic illness requires the degree of management by both medical personnel and the patient as chronic renal failure.

1

58

Assessment of Health Status in Dialysis Patients

Patients with chronic renal failure maintained on dialysis continue to exhibit a wide range of abnormalities in physiological functioning and associated physical complaints resulting from uremia and the dialysis regimen, itself. As in other chronic diseases, the course is highly variable with the variance in patient discomfort, level of functioning, response to treatment, and disease process controlled by a number of complex and interacting variables which include both the medical treatment and the influence of environmental factors (Cataldo, Russo, Bird, & Varni,1980). Prior to 1973, older patients and patients with significant coexisting morbid conditions were excluded from dialysis treatment programs. However, since that time the rigid selection criteria have relaxed considerably, and the population of patients with chronic renal failure currently receiving treatment are both older and more ill. The mean age of 93% of the patients with chronic renal failure in the U.S. was reported to be 53 (Hellerstedt et al., 1984). Also, a large percentage of these patients suffer from serious concomitant medical conditions that further impact upon health. Calsyn, Sherrard, Freeman, Hyerstay, and Curtis (1978) reported that approximately 1/3 of a sample of 107 hemodialysis patients suffered from one or more of

the following conditions, in addition to renal failure, at the time that dialysis was initiated: cancer, stroke,myocardial infarction, insulin-dependent diabetes, obstructive airways disease, alcoholism, drug addiction or psychosis. Charra et al., (1983) noted a shift in the age and morbidity of patients entering dialysis treatment programs in France in recent years. Patients age 54 and older were found to comprise 23/6 of the patients currently beginning dialysis treatment. Also, the percentage of patients considered to be at high risk due to diagnoses of nephrosclerosis, diabetes, and collagen diseases was found to have risen from 7% in 1971 to 25% in 1983.

The complexity of variables impacting upon the health status of patients maintained on dialysis makes the assessment or quantification of health in this population exceedingly difficult. Nonetheless, the assessment of health parameters remains of critical importance as physicians continue to monitor and manage the treatment of these patients. The problems inherent in the assessment of health in dialysis patients are reflected in the absence of clearly defined guidelines for physicians to use in assessing the adequacy of treatment (Newman, 1983). Also, researchers must have reliable and valid methods for assessing health status in order to examine the impact on health of various interventions, as well as control for the influence of health factors as a potential confound in research. A major criticism of psychological research in

this area has been the failure to include an appraisal of the physiological status of dialysis patients (Osberg et al., 1982). Yanagida and Streltzer (1979) noted that biological factors present in many dialysis patients such as anemia, hypertension, glucose intolerance, abnormal electrolyte levels, fluid overloading, and various medications can significantly influence psychological test data. A reason given for the omission of illness measures in research has been the lack of available assessment methodology (Stauch-Rahauser et al., 1977), as well as confusion on the part of nonmedical researchers regarding the selection of measures considered to be most appropriate.

The methods used to assess the health status of dialysis patients have included biochemical measures,physician ratings, symptom checklists, and measures of

functional status.

Biochemical Measures

Serum chemistries reflect a variety of physiological parameters of critical importance in patients with chronic renal failure who are maintained on hemodialysis. Laboratory values of creatinine, blood urea nitrogen,hematocrit, potassium, phosphorous, calcium, sodium, carbon dioxide, chloride, alkaline phosphatase, and others arecarefully observed and continuously monitored by the

medical staff to assess the health status of dialysis patients (Hekelman & Ostendarp, 1979). These data are used to assess the adequacy of the treatment regimen in an effort to identify potential medical problems. Theon-going evaluation of serum chemistries can impact uponthe dialysis treatment regimen suggesting a need for changes in dialysis parameters, initiation ordiscontinuation of drug therapy or modification of diet. Many of the laboratory values such as potassium and phosphorous are viewed as objective measures of patient compliance (Agashua, Lyle, Livesley, Slade, Winney, & Irwin, 1981). For example, Blackburn (1977) assessedcompliance with the dialysis regimen through laboratory values, and noted that only 25$ of patients in one sample were potassium compliant, and 60$ were phosphorous compliant.

In addition to the routine monitoring of chemistries by the medical staff, researchers have also employed laboratory values as an index of physical status in dialysis patients. Trieschmann and Sand (1971), in an investigation of psychological functioning in predialysis patients, used creatinine level, an index of the degree of impairment in renal function, to differentiate patients into "most sick” and "least sick" groups. Yanagida and Streltzer (1979) stressed the importance of including physiological measures in dialysis research to avoid misinterpreting differences among groups as due to

psychological factors when biological factors may be causitive. In a further study, these researchers examined two groups of dialysis patients for differences in physiological status through an assessment of serum levels of blood urea nitrogen, hematocrit, creatinine, calcium, potassium, phosphorous, sodium, chloride and carbon dioxide in order to ensure equality on a number of potentially confounding variables (Yanagida, Streltzer, & Siemsen,1981). The finding that significant correlations exist between both blood urea nitrogen and creatinine and performance on tasks that assess visuomotor speed and accuracy in dialysis patients led researchers to warn others about the need to control for, or at least assess, the potentially confounding effects of physiological status in future research (Hart, Pederson, Czerwinski, & Adams, 1983).

As noted previously, medically, the physiological parameters reflected in serum chemistries are viewed as critical to patient survival. However, longitudinal studies of factors related to survival in dialysis patients

f

generally do not consistently find any one biochemical measure to be highly predictive of longevity (e.g., Wai et al., 1981). It is likely that when such measures areentered into predictive equations with powerfuldemographic variables such as age, their impact are considerably lessened. Further, studies completed to date such as Foster et al., (1973) include methodological

63shortcomings such as grossly inadequate sample size which preclude the drawing of firm conclusions. Further research is clearly needed directed at the relation between biochemical measures and other measures of health status.

Physician Ratings

Physician ratings are frequently employed in an effort to quantify the physical status of dialysis patients. Dialysis patients receive a significantly high degree of routine medical supervision. In-center dialysis patients are generally seen by a physician during each dialysis session. As a result, staff physicians (nephrologists), who usually complete such ratings, have a thorough and on-going knowledge of each patient's medical status. Chambers (1982) pointed out the potential bias in physician's assessments which can result from prior expectations and past knowledge of the patient. De-Nour and Czaczkes (1974) examined the criteria employed by staff physicians in the assessment of their patients, and compared physician ratings with other indices of patient functioning. Overall, their data suggested that physician ratings were biased, because they tended to minimize patient problems and rate their patients as doing better than other indicators would suggest. Additionally,physician rating scales present difficulties due to the reliance on clinical judgment, making them hard to

64standardize and calibrate. However, these methodological concerns can be adequately controlled through the use of two independent ratings and demonstrating high levels of interrater reliability.

A variety of formats for physician rating scales have been used with dialysis patients. Most commonly, a global rating scale has been employed in which the medical status or overall health of patients are rated on a 4-point scale ranging from good to very bad (e.g., De-Nour, 1982). Devins, Binik, Hollomby, Barre, and Guttman (1981) had attending physicians rate each dialysis patient's general nonrenal health along a 5-point scale, in an effort to exclude patients with serious nonrenal pathology from further study. A modified Karnofsky Performance Status Scale was used to assess the physical activity level of 25 dialysis patients (Guttman, 1983). Staff physicians rated each patient on a 5-point scale ranging from very active to moribund. A 7-point rating scale was employed to measure overall medical condition in 23 dialysis patients (Steidl et al., 1980). Three staff physicians and the head nurse of the dialysis unit rated each patient, and satisfactory interrater reliability was reported. Staff ratings were also validated against the Cornell Medical Index, and significant correlations were obtained between the two measures.

65Physical Symptom Checklists

As previously reviewed, patients maintained on hemodialysis continue to experience a number of physical symptoms resulting from persistent physiological abnormalities secondary to chronic renal failure and the dialysis procedure, itself. As Kutner, Fair, and Kutner(1985) emphasized, dialysis patients are likely to experience a myriad of somatic symptoms such as nausea, dizziness, muscle cramps, and chronic fatigue, even in the absence of serious coexisting medical conditions. The occurrence and frequency of physical symptoms have been viewed as an important index of health status in dialysis patients, and such measures are routinely employed in the assessment of the adequacy of dialysis treatment (Newman, 1983).

Strauch-Rahauser, Schafheutle, Lipke, and Strauch(1977) developed an instrument specifically designed to assess the presence of physical symptoms in dialysis patients. The symptom checklist consists of 54 somatic items which cover a broad range of symptomatology. All items were judged to be of significance in the assessment of physical status in dialysis patients by 100 nephrologists in Germany. It is easily administered, and was found to produce a wide range of scores in a sample of dialysis patients. Adequate interrater reliability was found for this scale, but limited data is available related

to its validity. The KDS-2 form (Kupfer & Detre, 1972) was also used to assess physical symptomatology in 190 dialysis and transplant patients (Bonney, Finkelstein, Lytton, Schiff, & Steele, 1978). The KDS-2 is a symptom checklist containing 64 specific physical symptoms. Significant differences were found between dialysis patients and transplant patients with 3856 of the dialysis group endorsing more than 20 symptoms, whereas only 956 of the transplant group did so. These findings are contrasted with a group of healthy controls where none were found to reply positively to more than four symptoms on this scale. Other researchers have employed widely used instruments such as the Cornell Medical Index to assess symptomatology in dialysis patients (Steidl et al., 1980). Other more global ratings of overall physical symptom severity based upon interview data have also been used (Farmer, Bewick, Parsons, & Snowden, 1979).

Functional Status Measures

The functional status of patients maintained on dialysis is of primary concern in the rehabilitation process. Levy (1977) defined rehabilitation as efforts to "restore to a condition of health for useful and constructive activity". The goal for treatment of patients with chronic renal failure is not only to sustain life, but also reduce the limitations and restrictions in functioning

that are commonly observed (Bonney et al., 1978). Frequently, successful adjustment on dialysis is viewed as a return to the level of employment and physical and social activities that a patient experienced before becoming severely uremic. However, a sizable literature is available regarding the functioning of dialysis patients, and for many such a level of functional status is not easily achieved. For example, the proportion of patients on dialysis working full-time, an index of vocational rehabilitation, was been reported to range from Q% to 51 it (Disney & Row, 1974; Strauch, Huber, Rahauser, Werner, Walzer, & Hefner, 1971).

The assessment of functional status in dialysis patients generally targets the areas of mobility, occupation, activities of daily living, and general activity level (Chyatte, 1979). More specific areas of functioning such as strength, range of motion, sensation, timed mobility, and fatigue may also be assessed. Most often, researchers have employed the functional classification system recommended by the National Kidney Foundation. This 4-point rating scale primarily assesses general physical activity ranging from 1 (capable of performing all usual types of physical activity) to 4 (severe limitation of usual physical activity). Steidl et al., (1980) used the classification system to assess the overall medical condition of a sample of dialysis patients. The staff physician, nurse, and social worker completed the

scale for each patient, and satisfactory interraterreliability was reported. Other researchers utilized this rating scale together with a structured interview to elicit an overall assessment of functional status (Bonney et al., 1978). Of the 190 dialysis and transplant patients, 35?6 of the dialysis patients and 10£ of the transplant patients received a functional status rating of 3 or 4, suggesting a substantial interference in their ability to perform usual daily activities. A fairly comprehensive questionnaire, the Psychosocial Adjustment to Illness Scale (PAIS), was designed to assess the functioning level of dialysis patients across seven areas including health careorientation, vocational environment, domestic environment, sexual relationships, extended family relationships, social environment, as well as psychological distress (De-Nour,1982). It is a self-administered format that assessesperceived limitations in functioning. It contains 44 items which are rated on a 4-point scale ranging from 0 (no disturbance) to 3 (marked disturbance). Adequatepsychometric properties support( its use in obtaining information related to global adjustment and psychological functioning in specific areas.

69

The Present Research

Increasing emphasis is being placed upon the development of assessment methodology targeting health parameters. The availability of reliable and valid measures of health status for use with dialysis patients is of critical concern to both medical professionals and researchers. Although a variety of assessment procedures have been used clinically and in research, little data is available regarding the psychometric properties of these instruments. There remains a need to examine the interrelation among the various methods of assessing health in dialysis patients. Also, the relation between measures of health status and other variables previously found to impact the health of these patients needs to be examined. Of particular concern is the lack of a demonstrated relation between various measures used to index health and important health outcomes for patients maintained on dialysis. The relation between health status measures and subsequent medical utilization has not been investigated. There is also clearly a need to examine the utility of various health indices in discriminating between those patients on dialysis who will continue to survive and those patients at risk for dying.

The present study was designed to examine the concurrent and predictive validity of a number of health status measures including biochemical measures, physician

ratings of physical status, staff ratings of functional status, Wahler Physical Symptoms Inventory, subjective ratings of health, and subjective ratings of functional status in hemodialysis patients. Also, of interest are the demographic and medical variables of age, length of time on hemodialysis, diagnosis of diabetes mellitus, diagnosis of primary hypertension, and total number of concomitant morbid conditions which have been previously identified to be significantly related to health in hemodialysis patients, and their relation to measures of health status. If evidence for the concurrent and predictive validity of these measures of health status can be established, such data will be useful to physicians, as well as researchers in determining the health status of hemodialysis patients.

The following questions are of most importance:1. What is the relation among various measures of health

status in hemodialysis patients?2. What is the relation among measures of health status

and previously identified significant demographic and medical variables?

3. What is the relation of various indices of health to subsequent medical utilization and survival?

4. What measure of health status best predicts subsequent medical utilization and survival?

5. What is the relation among previously identified significant demographic and medical variables to both subsequent medical utilization and survival?

Chapter II Method

Subjects

Subjects consisted of adult outpatients with chronic renal failure who were receiving maintenance hemodialysis at Bio-Medical Applications of Baton Rouge, Inc. An effort was expended to recruit all routine hemodialysis patients in order to provide a full range of scores on the measures of health status. Of the 158 patients receivinghemodialysis at the two outpatient clinics in the community, 91% served as subjects in the initial assessment phase of the present investigation. Three patients declined participation, and 12 patients were excluded because of severe physical impairment resulting from stroke, blindness, mental retardation or critically ill medical status that significantly interferred with completion of assessment instruments. Eight subjects who initially participated in assessment received a renal transplantation during the one-year course of the investigation and were subsequently dropped from the study. Four additional subjects transferred to another dialysis

72

facility, and complete data for these patients were not available. Data from 131 subjects were available for analysis.

The age of participating hemodialysis patients ranged from 20 to 78 years, with a mean age of 52 years. The sample was predominantly black (71%), unemployed (8456), and of lower socioeconomic status as measured by the Hollingshead Two-Factor Index of Social Status. The length of time on hemodialysis, measured in months since initial onset of maintenance treatment, ranged from 1 to 152 months, and the mean length of time for the sample was 39 months at the time of the initial assessment phase of the investigation. These and other demographic data are presented in Table 2.

Measures of Health Status

Biochemical analyses of blood urea nitrogen, potassium, phosphorous, and hematocrit; physician ratings of physical status; Wahler Physical Symptoms Inventory; staff ratings of functional status; subjective ratings of functional status; and subjective ratings of health served as measures of health status.

Biochemical analyses. Plasma levels of blood urea nitrogen (normal values: 10-26 mg/dl); potassium (normalvalues: 3.5-5.0 mg/dl); phosphorous (normal values: 2.5-4.2 mg/dl); and hemocrit (normal values: 38-5256) were

determined for each subject by assays performed on a Technicon-SMAC-2 by Lifechem Laboratories, Rockleigh, New Jersey. This laboratory performs routine biochemical assays on a monthly basis for all patients and staff at Biomedical Applications of Baton Rouge, Inc., as well as numerous dialysis centers throughout the United States. The reliability of the assay procedures is continuously monitored, and the instrumentation is calibrated after every ten samples. Normal samples are included at random intervals as a further check on reliability. Additionally, if a sample is found to be significantly abnormal, the laboratory requests that another plasma sample be drawn and the analysis is repeated.

Physician ratings of physical status. An adaptation of a seven-point rating scale developed by Steidl et al. (1980) was used to obtain an assessment of each subject's current medical condition. A previous application of this scale with hemodialysis patients yielded satisfactory interrater reliability (.77 to .89), and physician ratings were found to be significantly correlated with both ratings of functional status (.75) and the Cornell Medical Index (.50). Two staff physicians at Bio-Medical Applications of Baton Rouge, Inc. independently rated each subject. Physicians were asked to define a score of 1 as the patient is doing well, in a global sense, in terms of medical adjustment; and a score of 7 as the patient is suffering from and is incapacitated by extremely severe medical

74complications of his/her illness. An anchored scale was provided to assist physicians in making this determination (see Appendix A). Interrater reliability was calculated, and mean scores were used to provide a single rating for

each subject.Wahler Physical Symptoms Inventory. The Wahler

Physical Symptoms Inventory is a brief self-report questionnaire that was administered to each subject. Reliability and validity data regarding this inventory have been previously reviewed. Forty-two physical symptoms are included, and respondents were asked to provide a frequency rating for each item on a scale of 0 (almost never) to 5 (nearly every day). Additional anchors are provided to assist subjects in providing frequency ratings (see Appendix B). The total score reflects the sum of all ratings provided after the number of items omitted or given more than one score are subtracted.

Staff ratings of functional status. Functionalstatus or degree of disability was assessed through ratingsof each subject provided by the staff social worker and

(

staff dietician at Bio-Medical Applications of Baton Rouge, Inc. Each subject was rated on a 4-point scale according to the functional classification system recommended by the National Kidney Foundation (see Appendix C). In this system, a score of 1 is defined as the patient is capable of performing all usual types of physical activity, and a score of 4 is defined as the patient is experiencing severe

75

limitations in usual activity level, may need assistancewith some aspects of self-care, may have impairedmentation, may be confined to bed. Additional anchors were provided to assist in determining ratings. Interrater reliability was calculated, and mean scores were obtained to provide a single rating per subject. Previous interrater reliability coefficients ranged from .71 to .88 (Steidl et al., 1980).

Subjective ratings of functional status. A 4-point rating scale that parallels the ratings of functional status completed by the staff social worker and dietician were completed by each subject. Subjects were asked to rate their current level of physical activity on a scale with a score of 1 defined as the patient is capable of performing all usual types of physical activity, and a score of 4 defined as the patient is experiencing severe limitations in usual activity level, may need assistancewith some aspects of self-care, may have impairedmentation, may be confined to bed (see Appendix D).

Subjective ratings of health. A 4-point rating scale was used for subjective ratings of current health. Subjects were asked to define a score of 1 as excellent, 2 as good, 3 as fair, and a score of 4 as poor (see AppendixE). As reviewed, self-assessed health status is often considered to be one of the best single indicators of health particularly due to its relation with subsequentmortality (Kaplan & Camacho, 1983).

76

Demographic/Medical Variables

A number of demographic and medical variables have been previously identified to be significantly related to survival in hemodialysis patients. The demographic andmedical variables examined in the present investigation included age, length of time on hemodialysis, diagnosis of diabetes mellitus, diagnosis of primary hypertension, and total number of coexisting morbid conditions present in addition to chronic renal failure. Demographic and medical information was obtained from the medical records atBio-Medical Applications of Baton Rouge, Inc. Theinformation contained in the Medical records was compared to available hospital discharge summaries in an effort to insure completeness of diagnostic data. The total number of coexisting morbid conditions was determined by the presence of the following diagnoses: hypertension,congestive heart failure, diabetes mellitus, arteriosclerotic cardiovascular disease, chronicobstructive pulmonary disease, malignancy, liver disease, and systemic lupus erythematosus.

Medical Utilization

The rate of medical utilization was defined as thefrequency of hospitalizations and number of hospital days for each subject during a one-year interval. This data was

77

obtained from the medical records at Bio-Medical Applications of Baton Rouge, Inc. Hospitalizations and hospital days associated with a renal transplant procedure or dialysis access problems were omitted from this data.

Survival

Survival data for the entire sample over a one-year interval were also obtained from the medical records at Bio-Medical Applications of Baton Rouge, Inc. Date and cause of death for deceased subjects were ascertained in order to determine length of survival for each subject.

Procedure

All subjects were required to complete one 20-minute assessment session which was scheduled during routine hemodialysis administration. A consent form (see AppendixF) was signed by both the subject and the experimenter at the onset of the assessment session. Subjects were told that the objective of the present study was to examine measures of health status in hemodialysis patients.

During the assessment session, subjects completed the Wahler Physical Symptoms Inventory, subjective ratings of functional status, and subjective ratings of health. These measures were administered orally by the experimenter, precluding the need for subjects to be able to write while

78receiving hemodialysis. This precaution was taken because the hemodialysis access site is frequently located in a patient's dominant arm.

Pertinent demographic and medical information including age, sex, length of time on dialysis, medical diagnoses, and medications were obtained from each subject's medical records at Bio-Medical Applications of Baton Rouge, Inc. (see Appendix G). Biochemical analyses of blood urea nitrogen (BUN), potassium, phosphorous, and hematocrit were also obtained from the medical records. Plasma samples are routinely drawn from each patient on a monthly basis and assayed by Lifechem Laboratories, and these data are recorded in each patient's medical chart.

Physician ratings of each subject were completed independently by two staff physicians, and ratings coincided with the assessment session and drawing of monthly plasma samples from each subject. Functional classification of each subject by the staff social worker and dietician were performed on the same day as physician ratings.

The rate of medical utilization, i.e., frequency of hospitalizations, number of hospital days, and survival data (date and cause of death) were obtained for each subject over a 12-month period following the initial assessment of health status. This data was contained in the medical records at Bio-Medical Applications of Baton Rouge, Inc.

Chapter III

Results

Concurrent Validity

The means and standard deviations for the measures of health status are presented in Table 3- Also, the range of normal values for each biochemical measure are included for comparison purposes. Physician ratings of physical status reflect the mean rating of two staff physicians; interrater reliability, r = .73. Staff ratings of functional status reflect the mean rating of the staff social worker and dietician; interrater reliability, r = .77.

The interrelations among the measures of health status were examined using Pearson product-moment correlations as presented in Table 4. The correlation matrix contains unadjusted alpha levels to facilitate comparison with other studies. However, the Bonferroni procedure (Dunn, 1961) for correlations was employed to control the experiment-wise error rate likely because of the large number of correlations. Utilizing this procedure, correlations in Table 4 are considered to be significant only if at the .001 level. Blood levels of blood urea nitrogen (BUN), potassium, and phosphorous were significantly intercorrelated in the present sample, but

79

showed little relation to the other measures of health status. Further, hematocrit failed to correlate significantly with any other measure of health status. Substantial correlations were found among physician ratings of physical status, staff ratings of functional status, the Wahler Physical Symptoms Inventory, subjective ratings of health, and subjective ratings of functional status. Only the correlation between subjective ratings of health and staff ratings of functional status failed to reach significance. Staff and physician ratings were highly correlated, r = .73, £. < .0001, suggesting a significantdegree of agreement in the dialysis staff's perceptions of patients' level of illness and disability. Of note, staff and patient ratings of functional status showed substantial agreement, r = .68, £ < .0001.

The relations among the measures of health status and a number of demographic and medical variables identified in previous research to be significantly related to health in hemodialysis patients was also examined in the presentinvestigation. Means, standard deviations, frequency, andpercentage of the demographic/medical variables arepresented in Table 5. The total number of concurrent diagnoses reflects the sum of the eight diagnostic categories contained in Table 5.

The intercorrelations among the measures of health status and the demographic/medical variables are presented in Table 6 along with £ values. Again, the Bonferroni

81

procedure (Dunn, 1961) for correlations was employed to control the experiment-wise error rate likely due to the large number of correlations. Utilizing this procedure, correlations in Table 6 are only considered statistically significant at the .0005 level. Inspection of Table 6

reveals that none of the biochemical measures weresignificantly correlated with the demographic/medicalvariables. Also, no significant relations were foundbetween subjective ratings of health and anydemographic/medical variable. Physician ratings of physical status were substantially correlated with adignosis of diabetes mellitus, r = .34, £ ̂ *0001, and thetotal number of concurrent diagnoses present in addition to chronic renal failure, £ = .44, £ < .0001. Functionalstatus, as reflected by both staff and subjective ratings, respectively, was significantly correlated with age, r = .49, .38; £ < .0001. Additionally, subjective ratings offunctional status showed a significant relation to adiagnosis of diabetes mellitus, r = .31, £ < .0003. Scoreson the Wahler Physical Symptoms Inventory were found to correlate significantly only with total number of concurrent diagnoses, r = .37, £ < .0001. Interestingly, a diagnosis of primary hypertension or nephrosclerosis showed no relation to the measures of health status, nor did thisvariable correlate significantly with any otherdemographic/medical variable. It is likely that this finding is due to the high prevalence of primary

82

hypertension in the present sample, having been diagnosed in 48.156 of the subjects.

In summary, the correlational data presented in Tables 4 and 6 suggest that the biochemical measures may be assessing a separate aspect of functioning in dialysis patients as compared with the other measures of healthstatus. Moreover, the pattern of correlations lendssupport to the concurrent validity of physician ratings ofphysical status, staff ratings of functional status, the Wahler Physical Symptoms Inventory, and subjective ratings of functional status as measures of the construct of health status.

Predictive Validity

Multiple regression and discriminant analysis procedures were used to examine the relations between indices of health status, demographic/medical variables and subsequent medical utilization and survival in the present sample of hemodialysis patients. The rate of medical utilization, defined as the frequency of hospitalizationsand number of hospital days, and survival for the sample over a one-year interval served as criterion variables. Table 7 contains the means, standard deviations, frequency, and percentage of the criterion variables. As presented in Table 7, 18 subjects (13.7/6 of the sample) died during the one-year study period. The number of hospitalizations and

83

number of hospital days were scaled in subsequent analyses in order to control for the limited number of available hospitalizations and/or hospital days for the subjects who died during the study interval. The scaling procedure utilized the proportion of hospitalizations and hospital days, respectively, to the number of available days for each subject over the one-year interval.

Table 8 contains the Pearson-product moment correlations between the measures of health status, demographic/medical variables, and the criterion variables. Although these correlations provide some information regarding the relation between the predictors and the criterion variables, the correlational procedure does not take into account the degree of intercorrelation among the predictors. To compensate for the moderate intercorrelations among the predictors, separate multiple regression analyses were utilized to predict number of hospitalizations and number of hospital days.

Number of hospitalizations. Separate multiple regression analyses were performed using number of hospitalizations as the criterion variable and the following sets of predictor variables: measures of health status, demographic/medical variables, and the combined measures of health status and demographic/medical variables.

The measures of health status were entered into a standard multiple regression analysis. As displayed in

84

Table 9, the analysis was significant, F(9, 115) = 2.24, £2

< .02, R = .15. Table 10 contains the subsequent stepwise multiple regression analysis for the measures of health status. As can be seen from Table 10, physician ratings of physical status were the best predictor of number of hospitalizations, F(1, 123) = 10.62, £ < .001, accountingfor 856 of the variance. The addition of phosphorous level accounted for an additional 4$ of the variance, yielding a significant two-variable model, F(2, 122) = 8.24, £ <.0004. Subjective ratings of health accounted for anadditional 256 of the variance, F(3, 121) = 6.37, £ < .0006. The addition of the Wahler Physical Symptoms Inventory, subjective ratings of functional status, potassium, and hematocrit, respectively, did not add significantly to the prediction.

The demographic/medical variables of age, length of time on dialysis, diabetes mellitus, primary hypertension, and total number of coexisting morbid conditions were entered into a standard multiple regression analysis. The analysis, summarized in Table 11, was significant, F(5,

2123) = 2.73, £ < .02, R = .10. Table 12 contains thesubsequent stepwise multiple regression analysis for the demographic/medical variables. Inspection of Table 12 reveals that the total number of coexisting morbid conditions best predicted the number of hospitalizations, F( 1, 127) = 5. 34, £ < .02, accounting for 456 of the

85

variance. The addition of the length of time on dialysis in a two-variable model accounted for an additional 4% of the variance, F(2 , 126) = 5.31, £ < .006. Thethree-variable model of total number of concurrent morbid conditions, length of time on dialysis, and age accounted for a total of 9% of the variance in number ofhospitalizations, F(3 , 125) = 3*96, £ < .009. The addition of the variables of diabetes mellitus and primaryhypertension did not add significantly to the prediction.

The combined measures of health status and thedemographic/medical variables were entered into a standardmultiple regression analysis. As displayed in Table 13,the analysis was significant, F(14, 108) = 2.02, £ < .02,2

R = .21. The subsequent stepwise multiple regressionanalysis for the combined variables is presented in Table 14. As displayed in Table 14, physician ratings of physical status best predicted number of hospitalizations, F(1, 121) = 9 *66, £ <.002, accounting for 7% of the variance. Theaddition of length of time on dialysis accounted for anadditional 6% of the variance, F(2, 120) = 8.70, £ < .0003. Phosphorous level accounted for an additional 3% of the variance, F(3, 119) = 7.56, £ < .0001. Only the addition of subjective ratings of health in a four-variable model added significantly to the prediction, F(4, 118) = 6.21, £ <.0001.

Number of hospital days. Two subjects evidenced

86

extreme scores on the criterion variable of number of hospital days, totally 93 and 142 hospital days, respectively. Because of the deleterious effects that extreme scores may have on multiple regression analyses, the data were analyzed intially with the outliers included and then with the two outliers excluded according to recommendations by Tabachnick and Fidell (1983, p. 92). Both analyses are presented for comparison.

Separate multiple regression analyses were performed for thd criterion variable of number of hospital days with the outliers included utilizing the following sets ofpredictor variables: health status measures,demographic/medical variables, and combined measures of health status and demographic/medical variables.

The measures of health status were entered into a standard multiple regression analysis utilizing thecriterion of number of hospital days with the outliersincluded. As displayed in Table 15, the analysis was not

2significant, F(9, 115) = 1.56, £ < .14, R = .11.Therefore, no stepwise multiple regression analysis was performed.

The demographic/medical variables were entered into a standard multiple regression analysis utilizing thecriterion variable of number of hospital days with theoutliers included. As displayed in Table 16, the analysis

2was not significant, F(5, 123) = 1.175, £ < .13, R = .07.

87

Therefore, no subsequent multiple regression analysis was performed.

The combined measures of health status and demographic/medical variables were entered into a standard multiple regression analysis utilizing the criterion variable of number of hospital days with the outliers included. As displayed in Table 17, the analysis was

2significant, F(14, 108) = 1.91, p. < .03, R = .20. Asubsequent stepwise multiple regression analysis was performed and is presented in Table 18. Physician ratings of physical status were found to be the best predictor of number of hospital days, accounting for 1% of the variance, F(1, 121) = 8.67, p < .004. The variable of length of timeon dialysis accounted for an additional 1% of the variance, F(2, 120) = 9•63, p < .0001. Only the addition of adiagnosis of diabetes mellitus in a three-variable model significantly improved the prediction, F(3, 119) = 8.25, p < .0 0 0 1.

Separate multiple regression analyses were performed with the outliers excluded from the criterion variable of number of hospital days utilizing the same sets of predictor variables which included the health status measures, demographic/medical variables, and the combined measures of health status and demographic/medical variables.

The measures of health status were entered into a

88

standard multiple regression analysis utilizing thecriterion variable of number of hospital days with theoutliers excluded. As displayed in Table 19, the multipleregression analysis was not significant, F(9, 113) = 1*74,

22 < .09, R = .12. Therefore, no subsequent stepwisemultiple regression analysis was performed.

The demographic/medical variables were entered into a standard multiple regression analysis to predict the criterion variable of number of hospital days with theoutliers excluded. As presented in Table 20, the standard multiple regression analysis was significant F(5, 121) =

24.40, £ < .001, R = .15. A subsequent stepwise multipleregression analysis was performed and is summarized in Table 21. As displayed, the length of time on dialysis was the best predictor of the number of hospital days, F(1, 125) = 14.18, 2 < .0003, accounting for 10^ of thevariance. The total number of concurrent diagnoses accounted for an additional 3% of the variance, F(2, 124) = 9.34, 2 < .0002. Only the addition of the variable ofdiabetes mellitus in a three-variable model added significantly to the prediction, F(3, 123) = 6.72, 2 <.0004.

The combined measures of health status and the demographic/medical variables were entered into a standard multiple regression analysis to predict the criterion of number of hospital days with the outliers excluded. As

89

presented in Table 22, the analysis was significant, F(14,2

106) = 2.39, £ <.006, R = .24. A subsequent stepwisemultiple regression analysis was performed and is presented in Table 23* As shown, the length of time on dialysis was the best predictor of hospital days, F(1, 119) = 15.69, £ < .0001, accounting for 12$ of the variance. Physician ratings accounted for an additional 4$ of the variance, F(2, 118) = 11.11, £ < .0001. Only the addition of thevariable blood urea nitrogen in a three-variable model added significantly to the prediction, F(3, 117) = 8 .98, £< .0 0 0 1.

In summary, the measure of health status that best predicted medical utilization in the present sample was physician ratings of physical status. Even when considered in combination with important demographic and medical variables, physician ratings were consistently the best predictor of subsequent medical utilization.

Survival. Separate direct linear discriminant function analyses were performed utilizing the measures of health status, the demographic/medical variables, and the combined health status measures and demographic/medical variables as predictors of survival. The discriminant analyses were used to classify subjects into groups (survivor versus deceased) on the basis of their scores on the independent variables.

For each set of predictor variables, the total sample

90

(N = 131) was randomly divided into a calibration sample and a holdout sample. The calibration sample was used to develop the discriminant function model, which was then applied to the holdout sample. In this way, the model was cross validated to enhance the external validity of the findings. Also, to minimize the effects of chance, thelargest possible calibration or modeling sample was utilized. Although a 75:25 split between the calibration and holdout samples is viewed as optimal, this was not feasible due to the small size of the deceased group.Consequently, a 67:33 split was employed for eachsubsequent discriminant function analysis.

A direct discriminant function analysis procedure, a method for computing the discriminant function which considers all independent variables concurrently, was performed using the measures of health status as predictors of survival. The calibration sample consisted of 69

survivors and 13 deceased subjects. The standardized linear discriminant function coefficients developed on the calibration sample with .84 prior probability of correct classification are presented in Table 24. The resultantclassification of subjects into survivor versus deceased groups is summarized in Table 25. In the calibration sample, 85$ (70 from a total of 82) of the subjects were correctly classified and 15$ (12 from a total of 82) were incorrectly classified into survivor and deceased groups. As indicated, 67 subjects or 97.10$ of the surviving group

91

were correctly classified as survivors, while 2 subjects or 2.90% of the surviving group were incorrectly classified as deceased. In the deceased group, 3 subjects or 23.08% were correctly classified as deceased, while 10 subjects or 76.92% from the deceased group were incorrectly classified as survivors.

The holdout sample consisted of 38 survivors and 5 deceased subjects, yielding a maximum chance criterion of .88. The model was applied to the holdout sample, and the resultant classification of subjects into survivor versus deceased groups is summarized in Table 26. As displayed, 91% (39 from a total of 43) of the subjects were correctly classified, and 9% (4 from a total of 43) were incorrectly classified into survivor and deceased groups. Thus, themodel surpassed the maximum chance criterion. As indicated in Table 26, 37 subjects or 97-37% from the surviving group were correctly classified as survivors, while 1 subject or 2.63% from the surviving group were incorrectly classified as deceased. However, in the deceased group, only 2 subjects or 40% of the deceased group were correctly classified as deceased, while 3 subjects or 60% from thedeceased group were incorrectly classified as surviving. Therefore, despite a significant difference in theproportion of subjects classified into the survivor and deceased groups as indicated by a Fisher Exact Probability Test = .03, the model misclassified the majority of the deceased group. It is important to note that the Fisher

92

Exact Probability Test, like the Chi-Square Test, analyzes deviation in any direction. In this case the deviation is significant, but it is in the wrong direction. In summary, the discriminant model based on the measures of health status failed to discriminate accurately between thesurvivor and the deceased groups.

The demographic/medical variables were entered into a direct discriminant function analysis. The calibration sample consisted of 71 survivors and 12 deceased subjects. The standardized linear discriminant function coefficients developed on the calibration sample with .86 priorprobability of correct classification are presented inTable 27. The resultant classification of subjects intosurvivor and deceased groups is summarized in Table 28. In the calibration sample, 86? (71 from a total of 83) of the subjects were correctly classified and 14? (12 from a total of 83) of the subjects were incorrectly classified into survivor and deceased groups. As indicated in Table 28, 68 subjects or 95.77? from the surviving sample were correctly classified as survivors, while 3 subjects or 4.23? from the surviving sample were incorrectly classified as deceased. In the deceased sample, 3 subjects or 25? were correctly classified as deceased, while 9 subjects or 75? from the deceased sample were incorrectly classified as surviving.

The holdout sample consisted of 41 survivors and 5 deceased subjects, yielding a maximum chance criterion of

t

.89. The model was applied to the holdout sample, and the

resultant classification of subjects into survivor versus deceased groups is summarized in Table 29. As displayed, 8956 (41 from a total of 46) of the subjects were correctlyclassified, and 1156 (5 from a total of 46) of the subjects were incorrectly classified into survivor and deceased groups. Thus, the model based on the demographic/medical variables equalled the maximum chance criterion. That is, the model predicted as well as a chance criterion that

i

involved the placement of all subjects into the larger (survivor) group. As indicated in Table 29, 41 subjects or 100/6 from the surviving group were correctly classified as survivors, while no subjects from the surviving sample were incorrectly classified as deceased. In the deceasedsample, no subjects were correctly classified as deceased, while 5 subjects or 100/6 of the deceased group wereincorrectly classified as surviving. Neither the Chi-Square test nor the Fisher Exact Probability Test could be used on this data to determine whether the two groups differed in the proportion falling into survivor anddeceased groups, because of the size of the cellfrequencies. Nonetheless, it can be readily observed that the model misclassified 100/6 of the deceased group. Insummary, the discriminant model based on thedemographic/medical variables also failed to discriminate accurately between the surviving and deceased groups.

A direct discriminant function analysis procedure was performed using the combined health status measures and the

demographic/medical variables as predictors of survival. The calibration sample contained 69 survivors and 12 deceased subjects, yielding a prior probability of correct classification of .85. The resultant classification of subjects into survivor and deceased groups is summarized in Table 30. In the calibration sample, 100% (81 of a total of 81) of the subjects were correctly classified and none (0 from a total of 81) were incorrectly classified into survivor and deceased groups. As indicated in Table 30, 69 subjects or 100%'of the surviving sample were correctly classified as survivors, while no subjects from the surviving sample were incorrectly classified as deceased. In the deceased group, 12 subjects or 100% of the deceased group were correctly classified as deceased, while no subjects from the deceased group were incorrectly classified as surviving.

The holdout sample consisted of 37 survivors and 5 deceased subjects, yielding a maximum chance criterion of .88. The model was applied to the holdout sample, and the resultant classification of subjects into survivor versus deceased groups is summarized in Table 31. As displayed, 88% (37 from a total of 42) of the subjects were correctly classified and 12% (5 from a total of 42) of the subjectswere incorrectly classified into survivor and deceased groups. Thus, the model based on the combined health status measures and demographic/medical variables only equalled the maximum chance criterion. As indicated in

95

Table 31, 37 subjects or 100$ of the survivors werecorrectly classified as survivors, while no subjects from the surviving sample were incorrectly classified as deceased. In the deceased group, no subjects were correctly classified as deceased, while 5 subjects or 100^ of the deceased group were incorrectly classified as surviving. Again, neither the Chi-Square Test nor the Fisher Exact Probability Test was appropriate for use with this data due to the size of the cell frequencies. But, as can be seen from Table 31, the model misclassified 10056 of the deceased group. Therefore, the model does not accurately predict survival from the combined measures of health status and the demographic/medical variables.

In summary, each of the three sets of predictor variables which included the measures of health status, the demographic/medical variables, and the combined health status measures and demographic/medical variables failed to discriminate accurately between the survivor and deceased groups.

Despite the failure of the measures of health status and the demographic/medical variables to discriminate significantly between survivor and deceased groups, group differences on these variables were assessed by means of a multivariate analysis of variance (MANOVA) and subsequent one-way analyses of variance (ANOVAs).

The means for the surviving and deceased groups, respectively, on the dependent measures of blood urea

96

nitrogen, potassium, phosphorous, hematocrit, staff ratings of functional status, physician ratings of physical status, the Wahler Physical Symptoms Inventory, subjective ratings of functional status, subjective ratings of health, age, length of time on dialysis, and total number of concurrent diagnoses are presented in Table 32. Subsequent analyses also included the dummy coded variables of the diagnoses of diabetes mellitus and primary hypertension.

Multivariate analysis of variance revealed a highly significant effect, JP (14, 108) = 2.92, £ < .0009.Subsequent one-way analyses of variance were performed, and the results are presented in Tables 33 - 46. Significantgroup differences were in the anticipated direction; that is, surviving subjects appeared to be healthier than subjects who subsequently died during the study interval. Surviving subjects were initially rated as functioning at a higher level than the subjects who subsequently died during the study interval as indicated by both staff ratings of functional status, F(1, 121) = 26.64, £ < .0001, andsubjective ratings of functional status, F(1, 121) = 8.06, £ < .005. Further, physicians also viewed the surviving subjects as "healthier" or experiencing less severe medical complications than the subjects who subsequently died, F(1, 121) = 7.29, £ < .008. Similarly, the groups differed onscores on the Wahler Physical Symptoms Inventory. The subjects who subsequently died during the study interval endorsed a greater frequency of physical symptoms, F(1,

121) = 10.00, £ < .002. The deceased group was also foundto have been maintained on chronic hemodialysis for a significantly longer time, F(1, 121) =5.67, £ < .02, and to be suffering from a greater number of concurrent diagnoses in addition to chronic renal failure, F(1, 121) = 4.84, £ < .03, than the survivor group. Despite previous findings linking subjective ratings of health to mortality, no significant differences were found between the two groups on this measure. Additionally, no significant differences were found between the surviving and deceased groups on age or a diagnosis of diabetes mellitus or primary hypertension. Congruent with previous findings in this study, no significant differences were found between the two groups on any biochemical measure.

Chapter IV

Discussion

The present study was designed to assess the concurrent and predictive validity of a number of health status measures. As such, the interrelations among the measures of health status were examined, as well as the relation between the measures of health status and subsequent medical utilization and survival in a sample of hemodialysis patients. Overall, the present investigation offers evidence for the concurrent validity of physician ratings of physical status, ratings of functional status, and the Wahler Physical Symptoms Inventory as measures of the construct of health status. The data also confirm previous findings which suggest that physician ratings are the most valid measure of health status (Twaddle, 1974).

Significant intercorrelations were found among physician ratings of physical status, ratings of functional status, and the Wahler Physical Symptoms Inventory. Additional support for the validity of these measures of health status was provided by the significant relations found between these measures and other important health variables such as age, length of time on dialysis, diagnosis of diabetes mellitus, and total number of concurrent diagnoses present in addition to chronic renal

98

99

failure.Interestingly, no significant relations were found

between the biochemical measures of blood urea nitrogen, potassium, phosphorous, and hematocrit, and the remaining measures of health status. Further, the biochemicalmeasures showed little relation to other variables found in previous research to have a significant impact upon thehealth of hemodialysis patients such as a diagnosis of diabetes mellitus (Hellerstedt, et al., 1984; Krakauer, et al., 1983). The data suggest that the biochemical measures used may have little relation to the construct of health status.

In the present sample, the biochemical measures of blood urea nitrogen, potassium, and phosphorous were moderately intercorrelated, however, hematocrit failed to correlate significantly with any other measure of healthstatus or demographic/medical variable. As noted previously, the biochemical measures of blood ureanitrogen, potassium, and phosphorous are among the most commonly used measures of compliance in the hemodialysis

f

literature. It is not surprising to find significantintercorrelations among various measures of compliance. Previous studies have suggested that compliance with one aspect of the dialysis medical regimen is predictive of compliance with the other components of the regimen as well (Cummings, et al., 1981).

Dietary restrictions and medication intake are viewed

as important aspects of the dialysis treatment regimen, and noncompliance with this regimen and the resultant changes in blood chemistries are assumed to have a direct impactupon health status (Cummings, Becker, Kirscht, & Levin,1981; Czaczkes & De-Nour, 1978). Seime (1980) pointed out the paucity of data related to compliance with the medical regimen and the daily functioning of dialysis patients, particularly with regard to feelings of well-being and physical symptoms. He proposed that noncompliance may have little relation to physical symptoms, yet have direlong-term consequences reflected in increased morbidity and/or mortality. The results of the present study are in contrast to the notion that compliance is a critical variable influencing both physical functioning and survival. Specifically, the present research did not support Seime's observations. Not only did the biochemical measures fail to correlate significantly with the other measures of physical symptomatology and functional status, but the measures of compliance showed little relation to medical utilization as. well. Further, no significantdifferences were found between the surviving and deceased groups on any biochemical measure.

Despite the commonly held view that the biochemical measures of blood urea nitrogen, potassium, and phosphorous, as well as hematocrit, reflect parameters critical to health status in dialysis patients, previous research has also failed to consistently link any one or

combination of these measures to longevity or survival (Armstrong & Woods, 1983; Farmer, et al., 1979; Foster & McKegney, 1978; Foster, et al., 1973; Richmond, Lindsay, Burton, Conley, & Wai, 1982; Wai, et al., 1981). In fact, previous research designed to examine this relation has produced inconsistent and conflicting results. For example, Wai and colleagues (1981) studied the influence of a number of physiological and psychological variables on the survival of 285 home dialysis patients. Included in the investigation were measures of blood urea nitrogen, potassium, phosphorous, as well as hematocrit. Hematocrit and total protein levels were found to significantly discriminate between survivors and nonsurvivors. However, other investigations have reported that the physiological indices of blood urea nitrogen, potassium, phosphorous, and hematocrit show no relation to survival in samples of in-center dialysis patients (Foster & McKegney, 1978; Foster, et al., 1973) and home dialysis patients (Richmond, et al., 1982).

Subjective ratings of health showed an inconsistent pattern of correlations with the other measures of health status and the demographic/medical variables in this sample. Although significant correlations were found among subjective ratings of health and physician ratings, the Wahler Physical Symptoms Inventory, and subjective ratings of functional status, no significant relation was found between this measure of health status and staff ratings of

102

functional status or any biochemical measure. Additionally, subjective ratings of health failed to significantly correlate with any demographic/medical variable assessed in this study. Further, no significant differences were found between survivor and deceased groups on this measure. These findings are in contrast to recent longitudinal studies which suggest that subjective ratings of health may be one of the best single indicators of health status. This measure has been linked to mortality in extensive longitudinal studies conducted on an elderly sample (Mossey & Shapiro, 1982) and a representative community sample (Kaplan & Camacho, 1983).

Denial has been repeatedly implicated as a primary coping strategy employed by hemodialysis patients in their adjustment to chronic maintenance dialysis (Short & Wilson, 1969; Yanagida, Streltzer, & Siemsen, 19 81 ). It is not surprising to find dialysis patients having difficulty in accurately assessing and rating their own health status, because of their strong tendency to deny problems and present themselves in a socially desirable manner (Strauch-Rahauser, Schafheutle, Lipke, & Strauch, 1977; Yanagida & Streltzer, 1979). The rating scale used in the present study may have been particularly problematic,because dialysis patients were asked to rate their healthin terras of the following adjectives: excellent, good,fair, and poor. The other subjective rating scalecompleted by subjects in this study consisted of a rating

103

of functional status. This rating scale did not appear to present difficulties in accurate completion, probably because the scale offers extensive behavioral descriptors of each scale interval. The current data would support the suggestions of other researchers in the cautious use of subjective assessment devices with hemodialysis patients (Yanagida & Streltzer, 1979). Further, the present data offers little evidence for the validity of subjective ratings of health when used as an index of health status in hemodialysis patients.

The purpose of the present study was also to examine the predictive validity of the measures of health status. Physician ratings of physical status were found to be the best predictor of medical utilization in this sample.Physicians have been termed the ultimate health statusdefiners due to their training and expertise (Twaddle, 1974). For this reason, physician ratings are frequentlythe criterion against which other measures of health status are compared (De-Nour & Czaczkes, 1974; Steidl, et al., 1980). Nonetheless, previous investigations have generally failed to validate physician ratings against appropriate criteria such as medical utilization.

These results are important in terms of the application of the measures of health status both clinically and in research. Evidence for the validity of physician ratings allows rational choices to be made among the available measures of health status. Physician ratings

104

appear to be the best single measure of health status in hemodialysis patients. As such, this measure can be useful in assessing the efficacy of interventions designed to impact upon the health status of these patients, as well as in examining the relation between psychological factors and lifestyle variables on chronic renal disease.

Despite strong recommendations in the past for the use of physician ratings in the assessment of health in research, physician ratings or the medical examination have been used only infrequently due to practical considerations such as cost (Cleary, 1980). Hemodialysis centers generally have staff physicians or nephrologists who are familiar with each patient's medical status as a result of a high degree of on-going medical supervision. Consequently, physician ratings of dialysis patients can be completed with a minimum of time expenditure, making the routine use of this measure of health status feasible.

Research is needed to further explore the psychometric properties of this measure. At the time of the present investigation, data were not available regarding the test-retest reliability of the physician rating scale used in this study. Also, the level of agreement achieved (.73) between the two physicians in this study on the rating scale was only moderate. Research is continuing that was designed to assess the test-retest reliability of this measure. Additionally, the utility of physician training in the use of explicitly defined behavioral anchors for the

105

rating scale is being examined in an effort to improve interrater reliability. Additional research is also needed to examine the sensitivity of the physician rating scale to changes in health status. A line of investigation that warrants further study involves the delineation of the components of physical status in dialysis patients that comprise physician global ratings. No single clinical or biochemical measure has proven to be acceptable in defining proper or adequate dialysis treatment. However, it is likely that physicians evaluate and differentially weight a vast amount of information that includes a number of health indices such as biochemical measures or physical symptoms in arriving at a single global rating of physical status. Information related to the components involved in physician ratings of physical status would have implications relevant to the development of more sensitive measures of health status in dialysis patients.

Although physician ratings of physical status were found to be the best predictor of medical utilization, the predictive validity of this measure and the other indices

t

of health status examined in the present investigation are of concern. Despite statistical significance, the measures of health status accounted for a surprisingly small amount of the variance in medical utilization. Further, even though statistically significant differences were found between survivor and deceased groups on the measures of physician ratings, ratings of functional status, and the

106

Wahler Physical Symptoms Inventory, not enough of the variance was accounted for by these measures for them to be predictive of survival. These findings may be due to limitations inherent in the measures of health status examined in the present study and/or limitations in the methodology employed.

Because reliability has strong implications for the validity of a measure, the reliability of the measures used in this investigation warrant further study. Specifically, a factor which may limit the conclusions that may be drawn from the data involves the questionable reliability or stability of the biochemical measures of blood urea nitrogen, potassium, phosphorous, and hematocrit. Although previous research would suggest stability in the biochemical measures (De-Nour & Czaczkes, 1976; Seime, 1980), no study to date has directly assessed the reliability of these measures in dialysis patients.

Previous investigations have utilized various sampling procedures ranging from one (Wai, et al., 1981 ) to multiple (Cummings, et al. , 1984; Foster, et al., 1973)observations, where the mean of the observations is used for analyses. Only a single observation of each biochemical measure was used in the present investigation. It is possible that a different pattern of correlations might have emerged had multiple measures been obtained and the mean of these measures used for analyses. However, the inconsistency in sampling procedures found across previous

107

studies may also account for the disparate findings reported regarding the relation of biochemical measures to morbidity and mortality.

Additional research is needed to assess the reliability or stability of the biochemical measures of blood urea nitrogen, potassium, phosphorous, and hematocrit in dialysis patients. In this way, the degree of variability in these measures will be assessed, and a uniform sampling procedure can be proposed which will have implications for the compliance literature as a whole. A uniform measurement technique will then allow comparisons to be made across studies. Also, before concluding that compliance with the treatment regimen has no relation to health status in hemodialysis patients, further research is needed to examine the relation of compliance to other indices of physical functioning such as physical symptoms and functional status, as well as its relation to important health outcomes such as morbidity or mortality.

Also as mentioned previously, data were not available regarding the test-retest reliability of physician ratings of physical status or ratings of functional status. Further research is needed to address the stability of these measures.

Given the chronically ill status of the present sample, the possibility of a restricted range of responses on the measures of health status was examined. However, the measures of health status revealed considerable

108

variability among subjects, suggesting that the lack of association was not a restriction of range problem.

A potential limitation of the study involves the length of the study interval. One year may be too short to provide meaningful information regarding the criterion variables of medical utilization and survival in this sample. Researchers involved in the assessment of factors thought to impact upon medical utilization such as drug, alcohol or mental health interventions generally stress that one year is the absolute minimum needed to assess meaningful trends in the use of health care services (Jones & Vischi, 1979; Longobardi, 19 81). The time interval frequently recommended to discern trends in health care utilization is up to five years. The problem of the length of the study interval is also reflected in the small size of the deceased group at the end of only one year. The small size of this group presented difficulties in statistical analyses. Consequently, the results of the discriminant analyses are viewed as less stable. It may be that more meaningful data related to the predictive validity of these measures may be gained by extending the length of the investigation and continuing data collection. For this reason, it is recommended that these data be reanalyzed at the end of two years.

The predictive validity of the demographic/medical variables also examined in this study warrant further comment. The demographic/medical variables included age,

109

length of time on dialysis, diagnosis of diabetes mellitus, diagnosis of primary hypertension, and total number of concurrent diagnoses present in addition to chronic renal failure. Similar to the measures of health status, these variables significantly predicted medical utilization. However, although statistical significance was reached, only a small amount of the total variance was accounted for by these variables. More importantly, thedemographic/medical variables failed to accurately predict survival in this sample, despite statistically significant differences that were found between survivor and deceased groups on length of time on dialysis and total number of concurrent diagnoses. These results appear to be in contrast to previous large-scale research which consistently link such demographic/medical variables to survival in hemodialysis patients (Hellerstedt, et al., 1984; Krakauer, et al., 1983). Two additional factors may account for such disparate findings, and include the design of the present study and the nature of the sample.

The current study utilized a cross-sectional design in which all available hemodialysis patients served as subjects. The participating subjects were then followed over the course of one year, during which time medical utilization and survival for the sample were assessed. In the present study, demographic and medical information was obtained for each subject at the time of the initial assessment phase of the investigation. Previous studies

which have examined similar variables as predictors ofsurvival in hemodialysis patients are longitudinal in design. In these studies, demographic/medical variables are assessed at the time of initiation of dialysis treatment. For example, age in the present study meant age of the subject at the time of the initial assessment phase of the investigation. However, in the data reviewed by Krakauer and colleagues (1983) on the U. S. dialysis population, the variable of age represented age when the patient initially began dialysis treatment. It is possible that a diagnosis of diabetes mellitus at the time ofinitiation of dialysis versus at a later point in hemodialysis treatment may have differing levels of impact upon survival. Also the length of the study interval inthe present investigation is considerably shorter than previous large-scale longitudinal studies.

As noted previously, the present sample waspredominantly black (71/6), unemployed (84/6), and of lower socioeconomic status. Additionally, 55/6 of the sample didnot complete high school. In this respect, the sample was not representative of the hemodialysis population in theUnited States. For example, Krakauer, et al. (1983)reported 17,194 blacks as compared with 43,990 whites nation-wide receiving maintenance dialysis. It is possible that other demographic/medical variables not assessed inthis study may be more critical to survival in this particular sample.

Ill

In summary, this study is viewed as an initial attempt to examine the validity of a number of health status measures in a chronically ill population. The results of this investigation provide empirical data related to the strengths and weaknesses of the available measures of health status with hemodialysis patients. As such, the results provide for potential avenues of future research.

Evidence is presented for the concurrent validity of physician ratings of physical status, ratings of functional status, and the Wahler Physical Symptoms Inventory as measures of the construct of health status. Data also revealed that physician ratings were the best predictor of subsequent medical utilization in this sample. Although results suggest that physician ratings of physical status are the most valid measure of health status for use with hemodialysis patients, clinicians and researchers should be cautious in the application of these findings. While a potentially useful measure, further research establishing the reliability and validity of the physician rating scale is needed. Future research should also investigate the psychometric properties of the other measures of health status, and expand this work with other chronic illness groups to determine the generalizability of these findings.

References

Abram, H. S., Moore, G. L., & Westervelt, F. B. (1971). Suicidal behavior in chronic dialysis patients.American Journal of Psychiatry, 127, 119-124.

Agashua, P. A., Lyle, R. C., Livesley, W. J., Slade, P. D., Winney, R. J., & Irwin, M. (1981). Predicting dietary non-compliance of patients on intermittent haemodialysis. Journal of Psychosomatic Research, 25, 289-301.

Agras, W. S. (1982). Behavioral medicine in the 1980’s:Nonrandom connections. Journal of Consulting and Clinical Psychology, 50, 797-803.

Armstrong, S., & Woods, A. (1983). Patient self-reported adjustment and health beliefs in compliant versus noncompliant hemodialysis patients. In N. B. Levy (Ed.), Psychonephrology: Psychological Problems in Kidney Failure And Their Treatment (pp. 79-91). New York: Plenum Medical Book Company.

Baily, G. L. (1977). The sick kidney and sex. NewEngland Journal of Medicine, 296, 1288-1291.

Battista, L. (1979). Hemodialysis. In F. P. Hekelman &C. A. Ostendarp (Eds.), Nephrology Nursing: Persectives Of Care (pp. 120-202). New York: McGraw-Hill Book

113

Company.Bergstrom, J., & Furst, P. (1978). Uraemic toxins. In W.

Dukker, F. M. Parsons, & J. F. Maher (Eds.),Replacement of Renal Function by Dialysis (pp. 354-390). Boston: Martinus Nijhoff Publishers.

Billings, A. G., & Moos, R. H. (1981). The role of coping responses and social resources in attenuating the stress of life events. Journal of Behavioral Medicine, 4̂, 139-157.

Blachley, J. D., & Knochel, J. P. (1981). In B. M. ■Brenner & J. H. Stein (Eds.), Chronic Renal Failure (pp. 28-45). New York: Churchill Livingston.

Blackburn, S. L. (1977). Dietary compliance of chronic haemodialysis patients. Journal of the American Dietetics Association, 70, 31-37.

Blagg, C. R. (1983). Acute complications associated with hemodialysis. In W. Dukker, F. M. Parsons, & J. F.Maher (Eds.) Replacement of Renal Function by Dialysis (pp. 611-629). Boston: Martinus Nijhoff Publishers.

Blagg, C. R., & Scribner, B. H. (1980). Long-term dialysis: Current problems and future prospects.American Journal of Medicine, 68, 633-635.

Blanchard, E. B. (1982). Behavioral medicine: Past,present, and future. Journal of Consulting and Clinical Psychology, 50, 795-796.

Bommer, J., Tschoppe, W., Ritz, E., & Andrassy, K. (1976). Sexual behavior of hemodialyzed patients. Clinical

Nephrology, (3, 315-319.

Bonney, S., Finkelstein, F. 0., Lytton, B., Schiff, M., & Steele, T. E. (1978). Treatment of end-stage renal failure in a defined geographic area. Archives of Internal Medicine, 138, 1510-1513.

Bovberg, R. R., Diamond, L. H., Held, P. J., & Pauly, M. V. (1983). Continuous ambulatory peritoneal dialysis: Preliminary evidence in the debate over efficiency and cost. Health Affiliates, 2, 96-102.

Brantley, P. J., & Bruce, B. K. (in press). Assessment in -behavioral medicine. In A. R. Ciminero, K. S. Calhoun,& H. E. Adams (Eds.), Handbook of Behavioral Assessment. New York: Wiley.

Brodman, K., Erdman, A. J., Jr., Lorge, I., & Wolff, H. G. (19^9). The Cornell Medical Index: an adjunct to medical interview. Journal of the American Medical Association, 140, 530-534.

Brodman, K., Erdman, A. J., Jr., & Wolff, H. G. (1949). Cornell Medical Index Health Questionalre. New York: Cornell University Medical College.

Brodman, K., Erdman, A. J., Jr., Lorge, I., & Wolff, H. C. (1952). The Cornell Medical Index Health Questionaire: The relationship of patients' complaints to age, sex, race, and education. Journal of the American Medical Association, 147, 112-119.

Brundage, D. V. (1976). Nursing Management of Renal Problems. St. Louis: Mosby.

115

Callaghan, N. (1966). Restless legs syndrome in uremic neuropathy. Neurology, 16, 359-361.

Calsyn, D. A., Sherrard, D. J., Freeman, C. W., Hyerstay,B. J., & Curtis, F. K. (1978). Vocational adjustment, psychological assessment and survival on hemodialysis. Trans American Society for Artificial Organs, 24, 125-126.

Cannell, C. F., Marquis, K. H., & Laurent, A. (1977). A summary of studies of interviewing methodology (Vital and Health Statistics, Series 2, No. 69). Washington, .DC: National Center for Health Statistics.

Carlson, D. M., Duncan, D. A., Naessens, J. M., & Johnson, W. J. (1984). Hospitalization in dialysis patients. Mayo Clinic Proceedings, 59, 769-775.

Cataldo, M. F., Russo, D. C., Bird, B. L., & Varni, J. W. (1980). Assessment and management of chronic disorders. In A. Ferguson & C. Taylor (Eds.), The Comprehensive Handbook of Behavioral Medicine (Vol. 3, pp. 67-95). New York: Spectrum Publications.

Chambers, L. W. (1982). Health program review in Canada: Measurement of health status. Canadian Journal of Public Health, 73, 26-34.

Chappell, H. L. (1981). Measuring functional ability and chronic health conditions among the elderly: A researchnote on the adequacy of three instruments. Journal of Health and Social Behavior, 22, 90-102.

Charra, B., Calemard, E., Cuche, M., & Laurent, G. (1983). Control of hypertension and prolonged survival on maintenance hemodialysis. Nephron, 33, 96-99*

Chyatte, S. B. (1979). Rehabilitation medicine in chronic renal failure. In S. B. Chyatte (Ed.), Rehabilitation In Chronic Renal Failure. Baltimore: Williams &Wilkins.

Cleary, P. J. (1980). A checklist for life events research. Journal of Psychosomatic Research, 24, 199-207.

Combined report on regular dialysis and transplantation in Europe. (1979-1981). Proceedings of Europe Dialysis and Transplant Association , _1_6, 2-73; J_7> 2-86; _1_8, 2-58.

Comty, C.M., Wathen, R. L., & Shapiro, F. L. (1976).Uremic pericarditis. In D. H. Spodick & A. N. Brest (Eds.), Cardiovascular Clinics (pp. 219-236). Philadelphia: F. A. Davis.

Cooley, E. J., Miller, A. W., Keesey, J. C., Levenspiel, M. J., & Sisson, C. F. (1979). Self-report assessment of life change and disorders. Psychological Reports, 44, 1079-1086.

Cummings, K. M., Becker, M. H., Kirscht, J. P., & Levin, N. W. (1981). Intervention strategies to improve compliance with medical regimens by ambulatory hemodialysis patients. Journal of Behavioral Medicine,

4, 111-127.

117

Cummings, N. A. (1984). The future of clinical psychology in the United States. The Clinical Psychologist, 37, 19-2 0 .

Czaczkes, J. W., & De-Nour, K. (1978). Chronic Hemodialysis As A Way of Life. New York:Brunner/Mazel.

DeFronzo, R. A., Andres, R., Edgar, P., & Walker, W. G. (1973). Carbohydrate metabolism in uremia : A review. Medicine, 52, 469-480.

Delano, B. G. (1983). Regular dialysis treatment. In W..Dukker, F. M. Parsons, & J. F. Maher (Eds.), Replacement of Renal Function By Dialysis(pp. 391-409). Boston: Martinus Nijhoff Publishers.

De-Nour, K. (1970). Psychotherapy with patients onchronic hemodialysis. British Journal of Psychiatry, 116. 207-215.

De-Nour, K. (1982). Psychosocial Adjustment To Illness Scale (PAIS): A study of chronic hemodialysis patients. Journal of Psychosomatic Research, 26, 11-22.

De-Nour, K ., & Czaczkes, J. W. (1974). Bias inassessment of patients on chronic dialysis. Journal of Psychosomatic Research, 18, 217-221.

De-Nour, K ., & Czaczkes, J. W. (1976). The influence of patient's personality on adjustment to chronic dialysis. The Journal of Nervous and Mental Disease, 162, 323-333.

Devins, G. M., Binik, Y. M., Hollomby, D. J., Barre, P. E., & Guttman, R. D. (1981). Helplessness and depression

118

in end-stage renal disease. Journal of Abnormal

Psychology, 90, 531-545.Disney, A. P., & Row, P. G. (1974). Australian

maintenance dialysis survey. Medical Journal of Australia, 2, 651-656.

Dunn, 0. J. (1961). Multiple comparisons among means. Journal of the American Statistical Association, 56, 52-64.

Farmer, C. J., Bewick, M., Parsons, V., & Snowden, S. A.(1979). Survival on home haemodialysis: Its

.relationship with physical symptomatology, psychosocial background and psychiatric morbidity. Psychological Medicine, 9_, 515-523.

Fillenbaum, G. G. (1979). Social context andself-assessments of health among the elderly. Journal of Health and Social Behavior, 20, 45-51.

Forst, D. H., & O'Rourke, R. A. (1981). Cardiovascular complications of chronic renal failure. In B. M.Brenner & J. H. Stein (Eds.), Chronic Renal Failure (pp. 84-152). New York: Churchill Livingstone.

Foster, F. G., Cohn, G. L., & McKegney, F. P. (1973). Psychobiologic factors and individual survival on chronic renal hemodialysis - a two year follow-up : Part I. Psychosomatic Medicine, 35, 64-82.

Foster, F. G., & McKegney, F. P. (1978). Small group dynamics and survival on chronic haemodialysis. International Journal of Psychological Medicine, 8̂,

119

105-1 1 6.Friedman, E. A. (1983). Oral sorbents in uremia. In W.

Dukker, F. M. Parsons, & J. F. Maher (Eds.), Replaoement of Renal Function by Dialysis. (pp. 3*11-353). Boston: Martinus Nijhoff Publishers.

Friedman, E. A., Delano, B. G., & Butt, K. M. (1978).Pragmatic realities in uremia therapy. The New England Journal of Medicine, 298, 368-371.

Friedman, E. A., & Lundin, A. P. (1982). New England Journal of Medicine, 306, 167-169.

Gabriel, R. (1984). Morbidity and mortality of long-term haemodialysis : A review. Journal of the Royal Society of Medicine, 77, 595-601.

Galbreath, R. M., Portmann, D., Eddleston, A. L., Williams, R., & Gower, P. E. (1975). Chronic liver disease developing after an outbreak of HBs antigen negative hepatitis in a hemodialysis unit. Lancet, ii, 886-890.

Garrity, T. F., Marx, M. B., & Somes, G. W. (1977). Theinfluence of illness severity and time since life change on the size of the life change-health change relationship. Journal of Psychosomatic Research, 21, 377.

Garrity, T. F., Marx, M. B., & Somes, G. W. (1978). Therelationship of recent life change to seriousness oflater illness. Journal of Psychosomatic Research, 22, 7-1 2.

120

Garrity, T. F., Somes, G. W., & Marx, M. B. (1977). Personality factors in resistance to illness after recent life changes. Journal of Psychosomatic Research, 21, 23-32.

Gilson, B. S., Gilson, J. S., Bergner, M., Bobbitt, R. A., Kressel, S., Pollard, W. E., & Vesselago, M. (1975).The Sickness Impact Profile: Development of an outcome measure of health care. American Journal of Public Health, 65, 1304-1325.

Guttman, R. A. (1983). Characteristics of long-term (14 .years) survivors of maintenance dialysis. Nephron, 33, 111-115.

Haenel, T., Brunner, F., & Battegay, R. (1980). Renal dialysis and suicide: Occurrence in Switzerland and inEurope. Comprehensive Psychiatry, 21, 140-145.

Harrington, R. L. (1978). Mental health component of multiphasic health testing services. In M. F. Collen (Ed.), Multiphasic Health Testing Services. (pp. 413-436). New York: Wiley Press.

Hart, R. P., Pederson, J. A., Czerwinski, A. W., & Adams,R. L. (1983). Chronic renal failure, dialysis, and neuropsychological function. Journal of Clinical Neuropsychology, j>, 301-312.

Hathaway, S. R., & McKinley, J. C. (1942). The Minnesota Multiphasic Personality Inventory. Minneapolis: University of Minnesota Press.

121

Hekelman, F. P., & Ostendarp, C. A. (1979). Nephrology Nursing: Perspectives of Care. New York: McGraw-Hill.

Hellerstedt, W. L., Johnson, W. J., Ascher, N.,Kjellstrand, C. M., Knutson, R., Shapiro, F. L., & Sterioff, S. (1984). Survival rates of 2,728 patients with end-stage renal disease. Mayo Clinic Proceedings, 59, 776-783.

Hirschman, G. H., Wolfson, M., Mosimann, J. E., Clark, C. B., Dante, M. L., & Wineman, R. J. (1981).Complications of dialysis. Clinical Nephrology, 15,.66-74.

Holdsworth, S. R., Dekretser, D. M., & Atkins, R. C.(1978). A comparison of hemodialysis and transplantation in reversing the uremic disturbance of male reproductive function. Clinical Nephrology, 10, 146-152.

Hulka, B. S., & Cassel, J. C. (1973). The AAFP-UNC study of the organization, utilization, and assessment of primary medical care. American Journal of Public Health, 63, 494.

Hunt, S. M., McKenna, S. P., McEwen, J., Backett, E. M., Williams, N., & Papp, E. (1980). A quantitative approach to perceived health status: A validation study. Journal of Epidemiology and Community Health, 34, 281-286.

Hutchinson, T. A., Boyd, N. F., & Feinstein, A. R. (1980). Scientific problems in clinical scales, as demonstrated

122

in the Karnofsky Index of Performance Status. Journal of Chronic Disease, 32, 661-666.

Hutchinson, T. A., Thomas, D. C., & MacGibbon, B. (1982). Predicting survival in adults with end-stage renal disease: An Age Equivalence Index. Annals of Internal Medicine, £6, 417-423.

Ibels, L. S., Reardon, M. F., & Nestel, P. J. (1976).Plasma post-heparin lipolytic activity and triglyceride clearance in uremic and hemodialysis patients and renal allograft recipients. Journal of Laboratory and .Clinical Medicine, 87, 648-651.

Jones, K. R., & Vischi, T. R. (1979). Impact of alcohol, drug abuse, and mental health treatment on medical careutilization : A review of the research literature.Medical Care, 17, 1-23.

Kaplan, R. M., Bush, J. W., & Berry, C. C. (1976). Healthstatus: Types of validity for an index of well-being.Health Services Research, 11, 478-507.

Kaplan, G. A., & Camacho, T. (1983). Perceived health andmortality: A nine-year follow-up of the HumanPopulation Laboratory Cohort. American Journal of Epidemiology, 117, 292-304.

Karnofsky, D. (1949). Clinical evaluation ofchemotherapeutic agents in cancer. In C. N. MacLeod (Ed.), Evaluation of Chemotherapeutic Agents. New York: Columbia University Press.

Katz, S., Akpom, C. A., Papsidero, J. A., & Weiss, S. T.(1973). Measuring the health status of populations. In R. L. Berg (Ed.), Health Status Indexes. Chicago: Hospital Research and Educational Trust.

Kaufman, M. M. (1959). Psychiatric findings in admission to a medical service in a general hospital. Journal of Mt. Sinai Hospital, 26, 160.

Kazis, L. E., Meenan, R. F., & Anderson, J. J. (1983).Pain in the rheumatic diseases: Investigation of a keyhealth status component. Arthritis and Rheumatism, 26,• 1017-1022.

Knapp, M. S. (1982). Renal failure - dilemmas anddevelopments. British Medical Journal, 284, 847-850.

Krakauer, H., Grauman, J., McMullan, M., & Creede, C.(1983). The recent U.S. experience in the treatment of end-stage renal disease by dialysis and transplantation. The New England Journal of Medicine, 26, 1558-1563.

Kupfer, D. J., & Detre, T. P. (1972). The KDS- 1 scale for symptom discrimination. Psychological Reports, 30» 915-919.

f

Kutner, N. G., Fair, P. L., & Kutner, M. H. (1985). Assessing depression and anxiety in chronic dialysis patients. Journal of Psychosomatic Research, 29, 23-31.

Lancaster, L. E., & Pierce, P. (1979). Total bodymanifestations of end stage renal disease and related medical and nursing management. In L. E. Lancaster (Ed.), The Patient With End Stage Renal Disease

124

(pp. 1-60). New York: John Wiley & sons.Lazarus, J. M. (1981). Hemodialysis in chronic renal

failure. In B. M. Brenner & J. H. Stein, (Eds.),Chronic Renal Failure (pp. 153-192). New York:Churchill Livingston.

Levy, N. B. (1978). Psychological sequelae to hemodialysis. Psychosomatics, 19, 329-331.

Levy, N. B. (1983). Psychonephrology: Psychological Problems In Kidney Failure and Their Treatment. New York: Plenum Medical Book Company.

Levy, N. B., & Wynbrandt, G. D. (1975). The quality of life on maintenance hemodialysis. Lancet, _i, 1328-1330.

Linder, A., Chara, B., Sherrard, D., & Scribner, B.(1974). Accelerated atherosclerosis in prolonged maintenance hemodialysis. New England Journal of Medicine, 290, 697-704.

Lipowski, Z. J. (1967). Review of consultation psychiatry and psychosomatic medicine: Clinical aspects. Psychosomatic Medicine, 29, 201-224.

Longobardi, P. G. (1981). The impact of a briefpsychological intervention on medical care utilization in an Army health care setting. Medical Care, 19, 665-671.

Maddox, G. L. (1962). Some correlates of differences in self-assessment of health status among the elderly. Journal of Gerontology, 17, 180-185.

125

Maddox, G. L., & Douglas, E. (1973). Self-assessment of health: A longitudinal study of elderly subjects.Journal of Health and Social Behavior, 14, 87-93.

Madow, W. G. (1973). Net differences in interview data on chronic conditions compared with information derived from medical records (PHS Publication No. 1000, Series 2, No. 57). Washington, DC: U.S. Government Printing Office.

Marcus, A. C. (1982). Memory aids in longitudinal health surveys: Results from a field experiment. American .Journal of Public Health, 72, 567-573-

Marshall, J. R., Gregorio, D. I., & Walsh, D. (1982). Sex differences in illness behavior: Care seeking among cancer patients. Journal of Health and Social Behavior, 23, 197-204.

Matarazzo, J. D. (1980). Behavioral health and behavioral medicine: Frontiers for a new health psychology.American Psychologist, 35, 807-817.

Matarazzo, J. D. (1982). Behavioral health's challenge to academic, scientific, and professional psychology. American Psychologist, 37, 1-14.

McFarlane, A. H., Norman, G. R., Streiner, D. L., Roy, R.,& Scott, D. J. (1980). A longitudinal study of the influence of the psychosocial environment on health status: A preliminary report. Journal of Health and Social Behavior, 21, 124-133.

Meenan, R. F'., Gertman, P. M., & Mason, J. H. (1980). Measuring health status in arthritis: The ArthritisImpact Measurement Scales. Arthritis and Rheumatism,23, 146-152.

Meenan, R. F., Gertman, P. M., Mason, J. H., & Dunaif, R. (1982). The Arthritis Impact Measurement Scales:Further investigations of a health status measure. Arthritis & Rheumatism, 25, 1048-1053.

Meltzer, J., & Hochstim, J. (1970). Reliability and validity of survey data on physical health. Public Health Report, 85, 1075-1086.

Miller, P. M., Ingram, J. C., & Davidson, S. (1976). Life events, symptoms and social support. Journal of Psychosomatic Research, 20, 515-522.

Minter, R. E., & Kimball, C. P. (1978). Life events and illness: A review. Psychosomatics, 19, 334-342.

Montgomerie, J. Z., Kalmanson, G. M., & Guzc, L. B.(1968). Renal failure and infection. Medicine, 1-4.

Mooney, H. W. (1962). Methodology in two Californiahealth surveys (PHS Publication No. 942, Public Health Monograph No. 72). Washington, DC: U.S. Government Printing Office.

Moore, J. R. (1975). Accuracy of a health interviewsurvey in measuring chronic illness prevalence. Health Services Research, 10, 162-167.

Mossey, J. M., & Shapiro, E. (1982). Self-rated health: A predictor of mortality among the elderly. American

127Journal of Public Health, 72, 800-808.

Murphy, E., & Brown, G. W. (1980). Life events,psychiatric disturbance and physical illness. British Journal of Psychiatry, 36, 326-338.

Naliboff, B. D., Cohen, M. J., & Yellen, A. N. (1982).Does the MMPI differentiate chronic illness from chronic pain? Pain, 13, 333-341.

Nigro, S. A. (1970). A psychiatrist's experience ingeneral practice in a hospital emergency room. Journal of the American Medical Association, 214, 1657.

Norman, G. R., McFarlane, A. H., Streiner, D. L., & Neale, K. (1982). Health diaries: In compliance with relationto other measures. Medical Care, 20, 623-629.

Onesti, G., Kim, K. E., Greco, J. A., del Guercio, E. T., Fernandes, M., & Swartz, C. (1975). Blood pressure regulation in end stage renal disease and anephric man. Circulation Research, 36, 145-149.

Osberg, J. W., Meares, G. J., McKee, D. C., & Burnett, G.B. (1982). The MMPI as a measure of the emotional correlates of chronic hemodialysis: A review. Journal of Personality Assessment, 46, 268-278.

Otto, R. (1979). Negative and positive life experience among men and women in selected occupations, symptom awareness and visits to the doctor. Social Science and Medicine, 13, 151-164.

Palmore, E., & Luikart, C. (1972). Health and socialfactors related to life satisfaction. Journal of Health

128

and Social Behavior, 13, 68-80.Parker, R. L., You-long, G., Long-gen, S., De-yu, H., &

Hinman, A. R. (1982). The sample household health interview survey. American Journal of Public Health,72, 65-70.

Patrick, D. L., Bush, J. W., & Chen, M. M. (1973). Toward an operational definition of health. Journal of Health and Social Behavior, 14, 6.

Pope, C. R. (1982). Life-styles, health status andmedical care utilization. Medical Care, 20, 402-413.

Raga, R. M., Kramer, M. S., & Rosenbaum, J. L. (1977).Long term use of nandrolone decanoate and iron in dialysis patients. Dialysis and Transplantation, 6_,

36-39.Reiman, A. S. (1982). Race and end-stage renal disease.

The New England Journal of Medicine, 306, 1290-1291. Rennie, D. (1978). Home dialysis and the cost of uremia.

New England Journal of Medicine, 287, 339-400.Richmond, J. M., Lindsay, R. M., Burton, H. J., Conley, J.,

& Wai, L. (1982). Psychological and physiological factors predicting the outcome on home hemodialysis. Clinical Nephrology, 17, 109-113.

Ridgeway, V., & Mathews, A. (1982). Psychological preparation for surgery: A comparison of methods.British Journal of Clinical Psychology, 21, 271-280.

Robinson, B. H., & Hawkins, J.- B. (1980). Proceedings of the European Dialysis and Transplant Association.

129

London: Pitman Medical.Rodin, G. M., Chmara, J., Ennis, J., Fenton, S., Locking,

H., & Steinhouse, K. (1981). Stopping life-sustaining medical treatment: Psychiatric considerations in thetermination of renal dialysis. Canadian Journal of Psychiatry, 26, 540-544.

Rostand, S. G., Kirk, K. A., Rutsky, E. A., & Pate, B. A. (1982). Racial differences in the incidence of treatment for end-stage renal disease. The New England Journal of Medicine, 306, 1276-1279.

Russo, D. C.,. Bird, B. L., & Masek, B. J. (1980).Assessment issues in behavioral medicine. Behavioral Assessment, 2, 1-18.

Schag, C. C., Heinrich, R. L., & Ganz, P. A. (1984). Karnofsky Performance Status revisited: Reliability, validity, and guidelines. Journal of Clinical Oncology,

2, 187-193.Schwartz, G. E., & Weiss, S. M. (1977). Proceedings of

the Yale conference on behavioral medicine (NIH Publication No. 78-1424). Washington, DC: U.S. Government Printing Office.

Seime, R. J. (1980). The relationship of compliance,self-reported feelings and symptoms, and staff attention in dialysis: A pilot study. In D. J. Withersty, J. M.Stevenson, and R. H. Waldman (Eds.), Communication and Compliance In A Hospital Setting (pp. 65-90). Springfield, Illinois: Charles C. Thomas.

Shanas, E., Townsend, P., Wedderburn, d., Friis, H.,Milhoj, P., & Stehouwer, J. (1968). Old People In Three Industrial Sooieties. New York: Atherton Press.

Short, M. J., & Wilson, W. P. (1969). Roles of denial in chronic hemodialysis. Archives of General Psychiatry, 20, 433-437.

Silverberg, S., Oreopoulos, D. G., Wise, D. J., Uden, D.E., Meindok, H., Jones, M., Rapaport, A., & de Veber, G. A. (1977). Pericarditis in patients undergoinglong-term hemodialysis and peritoneal dialysis.American Journal of Medicine, 63, 874-891.

Singer, E., Garfinkel, R., & Cohen, S. M. (1976).Mortality and mental health: Evidence from the midtown Manhattan restudy. Social Science and Medicine, 10, 517-525.

Steidl, J. H., Finkelstein, F. 0., Wexler, J. P.,Feigenbaum, J., Kitsen, J., Kliger, A. S., & Quinlan, D. M. (1980). Medical condition, adherence to treatmentregimens, and family functioning. Archives of General Psychiatry, 37, 1025-1027.

Stone, W. J. (1977). Therapy of constipation in patients with chronic renal failure. Dialysis and Transplantation, 6̂, 7-13.

Strauch-Rahauser, G., Schafheutle, R., Lipke, R., &Strauch, M. (1977). Measurement problems in long-term dialysis patients. Journal of Psychosomatic Research,

V

Strauch, M.f Huber, W . f Rahauser, G., Werner, J., Walzer, P., & Hafner, H. (1971). Rehabilitation in patients undergoing maintenance haemodialysis: Results of a questionnaire in 15 dialysis centers. Proceedings of the European Dialysis and Transplantation Association,8, 28-32.

Sudman, S., & Lannom, L. A. (1978). A comparison of alternate panel procedures for obtaining health data. Urbana-Champaign, IL: University of Illinois, Survey Research Laboratory.

Sudman, S., Wilson, W., & Ferber, R. (1974). Thecost-effectiveness of using the diary as an instrument for collecting health data in household surveys (Report to the Bureau of Health Services Research and Evaluation). Urbana-Champaign, IL: University of Illinois, Survey Research Laboratory.

Tabachnick, B. G., & Fidell, L. S. (1983). Using Multivariate Statistics. New York: Harper & Row.

Teschan, P. E., & Ginn, H. E. (1976). The nervous system. In S. G. Masery & A. L. Sellers (Eds.), Clinical Aspects of Uremia and Dialysis (pp. 3-33). Springfield: CharlesC. Thomas.

Trieschmann, R. B., & Sand, P. L. (1971). WAIS and MMPI correlates of increasing renal failure in adult medical patients. Psychological Reports, 29, 1251-1262.

Twaddle, A. C. (1974). The concept of health status. Social Science and Medicine, j), 29-38.

132

United States Department of Health and Human Resources(1980). Ten leading causes of death in the U.S., 1977. Washington, DC: U.S. Government Printing Office.

United States National Center for Health Statistics (1957). Mortality from selected causes by age, race, and sex in the U.S., 1955. Vital Statistics Special Reports (National Survey), 4_6, 5.

Verbrugge, L. (1980). Health diaries. Medical Care, 18,73.

Vollmer, W. M., Wahl, P., & Blagg, C. R. (1983).Long-term survival on dialysis. Lancet, ii, 1198.

Wahler, H. J. (1968). The Physical Symptoms Inventory:Measuring levels of somatic complaining behavior.Journal of Clinical Psychology, 24, 207-211.

Wahler, H. J. (1983). The Wahler Physical SymptomsInventory Manual. Los Angeles: Western Psychological Services.

Wai, L., Richmond, J., Burton, H., & Lindsey, R. (1981). Influence of psychosocial factors on survival of home-dialysis patients. Lancet, ii, 115-1156.

Ware, J. E., Davies-Avery, A., & Donald, C. A. (1978). Conceptualization and measurement of health for adults in the health insurance study. General Health Perceptions (Vol V). Santa Monica: The Rand Corporation.

Wauters, J., Hunziker, A., & Brunner, H. (1983).Regionalized self-care hemodialysis: A solution to the

133increasing cost. Journal of the American Medical Association, 250, 59-62.

Wilson, R. W., & Drury, T. F. (1984). Interpreting trends in illness and disability: Health statistics and healthstatus. Annual Review of Public Health, jj, 83-106.

Wilson, W. E., Kilpatrick, C. H., & Talmade, D. W. (1965). Suppression of immunologic responsiveness in uremia. Annals of Internal Medicine, 62, 1-9.

Wyler, A. R., Masuda, M., & Holmes, T. H. (1968). Seriousness of Illness Rating Scale. Journal of Psychosomatic Research, 11, 363-375.

Wyler, A. R., Masuda, M., & Holmes, T. H. (1970).Seriousness of Illness Rating Scale: Reproducibility. Journal of Psychosomatic Research, 14, 59-64.

Yanagida, E. H., & Streltzer, J. (1979). Limitations of psychological tests in a dialysis population. Psychosomatic Medicine, 41, 557-567.

Yanagida, E. H., Streltzer, J., & Siemsen, A. (1981).Denial in dialysis patients: Relationship to complianceand other variables. Psychosomatic Medicine, 43, 271-280.

Ziarnik, J. P., Freeman, C. W., Sherrard, D. J., & Calsyn,D. A. (1977). Psychological correlates of survival onrenal dialysis. The Journal of Nervous and MentalDisease, 164, 210-213.

Tables

Table 1Survival Rates for Patients Receiving Dialysis or Undergoing Transplantation

DIALYSIS TRANSPLANTATIONUNRELATED DONOR RELATED DONOR

Survival Survival No. of Survival Survival No. of Survival Survival No. ofto 1 Year to 3 Years Cases to 1 Year to 3 Years Cases to 1 Year to 3 Years Cases

Total 81*0.2$ 56*0.3$ 65,270 86*0.4$ 78*0.5$ 7,595 95*0.4$ 91*0.6$ 3,491SexMale 81*0.3 55*0.4 36,600 86*0.6 77*1.0 4,818 95*0.5 90*1.0 2,C79Female 82*0.4 58*0.4 28,581 86*1.0 78*1.0 2,777 94*0.6 91*1.0 1,410Total 65,181 7,595 3,489

RaceBlack 85*0.3 62*0.6 17,194 86*1.0 78*1.0 1,625 93*1.0 87*2.0 364White 80*0.2 54*0.3 43,990 86*0.5 77*1.0 5,588 95*0.4 91*1.0 2,905Total 61,184 7,213 3,269

Age (yr)<10 90*2.0 82*3.0 601 89*3-0 80*3.0 155 93*2.0 89*3-0 150

11-20 95*0.5 88*2.0 2,620 92*1.0 87*1.0 844 97*1.0 95*1.0 66221-30 91*0.5 78*1.0 6,240 92*1.0 87*1.0 1,792 97*1.0 95*1.0 1,08031-40 89*0.5 71*1.0 7,883 85*1.0 78*1.0 2,016 93*1.0 89*1.0 79641-50 88*0.4 68*0.7 9,995 81*1.0 70*1.0 1,737 91*1.0 82*2.0 490>50 77*0.2 48*0.3 35,991 79*1.0 68*2.0 1,051 88*2.0 81*3.0 208

Total 63,250 7,595 3,488Primary diseasePrimary hypertension 82*0.5 56*1.0 7,049 85*1.0 77*2.0 593 93*2.0 86*3.0 159Glaaerulonephritis 91*0.5 69*1.0 6,538 87*1.0 81*1.0 1,760 97*1.0 93*1.0 811Diabetic nephropathy 75*1.0 39*1.0 5,856 79*2.0 65*2.0 518 89*2.0 82*3.0 268Polycystic kidneys 95*1-0 78*2.0 1,973 81*2.0 77*3-0 293 91*3-0 82*4.0 75

(Reprinted from Krakauer, Grauman, McMullan, and Creede, 1983, p. 1560)

136Table 2

Summary of Sample Characteristics

Characteristic Frequency (N =131) Percentage (%)

SexMale 73 56.0Female 58 44.0

RaceWhite 38 29.0Black , 93 71.0

Marital StatusMarried 72 55.0Single 21 16.0Separated/Divorced 21 16.0Widowed 17 13.0

Household CompositionLives alone 19 14.3Lives with others 111 85.0Institution 1 0.7

Education (years)0-6 25 19.07-9 25 19.010-11 22 17.012 32 25.013 16 12.016 8 6 . 017+ 3 2.0

137

Table 3Means and Standard Deviations of Health Status Measures

Measure M SD

Biochemical Analyses Normal ValuesBlood urea nitrogen (BUN) 76.73 20. 43 10.0-26.0 mg/dlPotassium 5. 11 • 78 3-5- 5.0 mg/dlPhosphorous 6.41 2.05 2.5- 4.2 mg/dlHematocrit 25.90 5.07 38.0-52.0 %

Physician Ratings 3-95 . 1.29Staff Ratings of Functional Status 2.00 • 93

Wahler Physical Symptoms Inventory 1.37 .68

Subjective Ratings of Health 2. 84 .86

Subjective Ratings of Functional Status 2.53 .96

Table 4Correlations Among Measures of Health Status and p Values (N =131)

BUN K P Hct PR SRFS WPSI SHS SFSBlood Urea Nitrogen (BUN)

.46*

.0001.47*.0001

-.03• 74

-. 12. 17

-. 17.06

-. 10.27

. 16

.08-. 18.04

Potassium(K)

• 39*.0001

.0009

.99.02.81

-.03.72

-.009.92

.06

.52.007.94

Phosphorous(P)

.03

.73. 12 . 16

-.006.94

.07

.44.09.30

-.11.20

Hematocrit (Hct)

-.07. 41

-.05.60

-.002.98

-. 16.08

.05

.55Physician Ratings (PR)

.73*

.0001.57*.0001

.33*

.0002.58*.0001

Staff Ratings of Functional

.47*

.0001.26.003

.68*

.0001Status (SRFS)Wahler Physical .41* .53*Symptoms Inventory .0001 .0001(WPSI)Subjective Health .43*Status (SHS) .0001Subjective Functional Status (SFS)

* £ <.001 u>00

139

Table 5

Means, Standard Deviations, Frequencies, and Percentages of Demographic/Medical Variables

VariableAgeLength of Time on Dialysis (in months)Total Number of Concurrent Diagnoses

Mean5239

1.7

Frequency (N=131)

Diabetes Mellitus 31Primary Hypertension 63Concurrent Diagnoses

Hypertension 107Congestive Heart Failure 31Diabetes Mellitus 31Arteriosclerotic 23Cardiovascular Disease

Chronic Obstructive 13Pulmonary Disease

Malignancy 7Liver Disease 1Systemic Lupus Erythematous 4

Standard Deviation 14. 20

33-90

1.04

Percentage

23.7 48.1

81.723.7 23-7 17-6

9.9

5.3 0.8 3-1

Table 6Correlations Among Measures of Health Status and Demographlc/Hedlcal Variables and p Values (N=131)

BUN

Blood Urea Nitrogen (BUN)

Potassium(K)Phosphorous(P)Hematocrit(Hct)Physician Ratings (PR)Staff Ratings of Functional Status (SRFS)

Wahler Physical Symptoms Inventory (WPSI)

Subjective Health Status (SHS)Subjective Functional Status (SFS)

Age

K P Hct PR

H6* . H7* -.03 -. 120001 .0001 . 7H . 17

• 39* .0009 .02.0001 .99 .81

.03 .12

.73 . 16

-.07.H1

SRFS WPSI SHS

-. 17 -. 10 . 16.06 .27 .08

-.03 -.009 .06.72 .92 .52

-.006 .07 .09. 9H . HH .30

-.05 -.002 -. 16.60 .98 .08

.73* .57* • 33*

.0001 .0001 .0002

. H7* .26

.0001 .003

.HI*

.0001

Months on Dialysis (MOD)Diabetes Mellitus (DM)

Primary Hyper­tension (PHT)Total Concurrent Diagnoses (TDX)

» £ < .0005

SFS Age MOD DM PHT TDX

. 18 -.20 . 15 -.06 -.03 -.09

.OH .02 . 10 .53 .75 • 31

.007 -. 10 .07 .03 -.08 -.00H

.9H .28 . H6 .73 .35 .96

.11 -. 19 .09 .OH -.06 .06

.20 .03 • 3H .69 . H8 .50

.05 . 18 .23 -.06 -.13 — 02

.55 .05 .01 . H9 .13 .80

.58* .29 -.09 . 3H* -. 11 .HH'*

.0001 .001 • 3H .0001 .21 .0001

.68* . H9* -.06 .27 -. OOH .39*

.0001 .0001 .52 .002 .96 .0001

.53* .09 .02 .21 -.06 .37*

.0001 • 30 .8H .02 .52 .0001

. H3* .02 -.02 . 10 -.07 .21

.0001 .81 .81 . 2H . H3 .02

.38* -. 18 .31* .005 .37*

.0001 .05 .0003 .95 .0001

.001 .07 .02 . 3H*• 99 .HO .86 .0001

-.23 -.09 .03.01 .29 .70

-. 10 . H1 •.23 .0001

-.06.H6

O

141

Table 7Means, Standard Deviations, Frequency, and Percentage of Hospitalizations, Hospital Days, and Mortality

Variable Mean Standard DeviationNumber of Hospitalizations 1.2 1.4Number of Hospital Days 12.4 19*6

Frequency PercentageMortality 18 13.7

142

Table 8Correlations Among Measures of Health Status, Demographic/Medical Variables and Criterion Measures

Number of Hospital­izations

Number of Hospital

DaysSurvival

Blood Urea Nitrogen .11 .05 - .01Potassium . 12 - .04 - .07Phosphorous .24 . 11 .05Hematocrit - .01 .0008 .04Physician Ratings .28 .25 .23Staff Ratings of Functional Status

. 16 .21 .40

Wahler Physical Symptoms .25 .22 C\1•

Subjective Health Status .23 .19 . 18Subjective Functional Status

• 13 . 14 .24

Age - .02 .08 . 14Months on Dialysis .20 .23 .21Diabetes Mellitus - .003 - .07 .09Primary Hypertension - .06 .03 - .03Total Concurrent Diagnoses

.20 .07 .20

Number of Hospitalizations

1.00 .73 .48

Number of Hospital Days .73 1.00 .49Survival .48 .49 1.00

143

Table 9Multiple Regression Summary Table for Measures of Health Status and Number of Hospitalizations

Source df Sura of Squares Mean Square

ModelErrorTotal

9115124

61.41 350.13 411.54

6.823-04

2.24 .02 .15

VariableBUNPotassiumPhosphorousHematocritWahlerStaff Functional RatingsPhysician RatingsSubjective Ratings of Functional Status

Beta Weight .0002 .1304 . 1346 .0090 .2873 .0122 . 33 02

-.2099

Subjective Ratings of Health . 3072

Table 10Stepwise Multiple Regression for the Measures of Health Status and Number of Hospitalizations

StepNumber

VariableEntered

Multiple2

Rdf F £ Beta

Weight

1 Physician Rating • o CD 1, 123 10.62 <.001 .30972 Phosphorous . 12 2, 122 8.24 <.0004 .17593 Subjective

Rating of Health. 14 3, 121 6.37 <.0006 .2943

145

Table 11Multiple Regression Summary Table for the Demographic/Medical Variables and Number of Hospitalizations

Source df Sura of Squares2

Mean Square F £ R

Model 5 39.64 7-93 2.73 .02 .10Error 123 357.55 2.91Total 128 397.18

Variable Beta WeightAge - .0132Months on Dialysis .0082Diabetes - .4893Hypertension - .2237Total Number of Concurrent Diagnoses

.4658

Table 12Stepwise Multiple Regression for the Demographic/Medical Variables and the Number of Hospitalizations

Step Variable MultipleO df F £ BetaNumber Entered R Weight

1 Total Number of Concurrent Diagnoses

.04 1, 127 5.34 <.02 .3854

2 Months on Dialysis .08 2, 126 on•LfN <.006 .0100

3 Age .09 3, 125 3.96 <.009 .0125

- OON

147

Table 13Multiple Regression Summary Table for the Combined Measures of Health Status and Demographics/Medical Variables and the Number of Hospitalizations

2Source df Sum of Squares Mean Square F £ R

Model 14 81.04 5.79 2.023 <-02 .21Error 108 309.07 2.86Total 122 390.11

Variable Beta WeightBUN -.0021Potassium .1255Phosphorous .1362Hematocrit -.0125Wahler .0251Staff Functional Ratings .0249Physician Ratings .3406Subjective Ratings of .0183Functional StatusSubjective Ratings of Health .1874Age -.0133Months on Dialysis .0097Diabetes -.7262Hypertension -.1036Total Number .2429of Concurrent Diagnoses

Table 14Stepwise Multiple Regression for the Combined Measures of Health Status and Demographic/Medical Variables and the Number of Hospitalizations

StepNumber

VariableEntered

Multiple2

Rdf F £ Beta

Weight

1 Physician Rating .07 1, 121 9.66 <.002 • 33152 Months on Dialysis . 13 2, 120 8.70 <.0003 .01133 Phosphorous . 16 3, 119 7.56 <.0001 . 15804 Subjective Health . 17 4, 118 6.21 <.0001 .2598

149

Table 15Multiple Regression Summary Table for Measures ofHealth Status and Number of Hospital Days (Outliers Included)

Source df Sum of Squares2

Mean Square F £ R

Model 9 11807.30 1311.92 1.56 .14 .11Error 115 96921.35 842.79Total 124 108729.00

Variable Beta WeightBUN . 1011Potassi um 3.0110Phosphorous • 9913Hematocrit .2360Wahler 3.4926Staff Functional Ratings 4.0182Physician Ratings 4.0085Subjective Ratings of Functional Status

-3.7148

Subjective Ratings of Health 3.7317

Table 16 Multiple Regression Summary Table for the Demographic/MedicalVariables and Number of Hospital Days (Outliers Included)

Source df Sum of Squares Mean Square F2

£ R

Model 5 6988.31 1397-66 1.75 o•onr—•

Error 123 98387.09 799.90Total 128 105375.00

Variable Beta WeightAge . 1033Months on Dialysis .1747Diabetes -6.3364Hypertension 1 .2931Total Number of Concurrent Diagnoses

2.2521

151

Table 17Multiple Regression Summary Table for the Combined Measuresof Health Status and Demographic/Medical Variables andNumber of Hospital Days (Outliers Included)

Source df Sum of Squares Mean Square F2

£ R

Model 14 20692.84 1478.06 1.91 .03 .20Error 108 83444.82 772.64Total 122 104138.00

Variable Beta WeightBUN .0464Potassi um 2.8364Phosphorous 1.4492Hematocrit . 2564Wahler • 9513Staff Functional Ratings 2.6556Physician Ratings 5.9270Subjective Ratings of Functional Status

.6712

Subjective Ratings of Health 2.2188Age .0233Months on Dialysis .2132Diabetes -12.4591Hypertension 2.4021Total Number of Concurrent Diagnoses

- 1.6181

Table 18Stepwise Multiple Regression for the Combined Measures of Health Status and Demographic/Medical Variables and the Number of Hospital Days (Outliers Included)

Step Variable Multiple2

Number Entered Rdf F £ Beta

Weight

1 Physician Rating .07 1, 121 8.67 <.004 8.04312 Months on Dialysis .14 2, 120 9.63 <.0001 .20113 Diabetes .17 3, 119 8.25 <.0001 13.9312

153

Table 19Multiple Regression Summary Table for Measures ofHealth Status and Number of Hospital Days (Outliers Excluded)

Source df Sum of Squares Mean Square F2

£ R

ModelError

9113

6090.02 676.67 1.74 44003.12 389.41

.09 .12

Total 122 50093-14

Variable Beta WeightBUN . 1633Potassium 1.3701Phosphorous • 3312Hematocrit .3375Wahler 3.2056Staff Functional Ratings 1.5549Physician Ratings 2.8234Subjective Ratings of Functional Status

-4.5506

Subjective Ratings of Health 4.0340

154

Table 20Multiple Regression Summary Table for the Demographic/Medical Variables and Number of Hospital Days (Outliers Excluded)

Source df Sum of Squares2

Mean Square F £ R

Model 5 7129.47 1425.89 4.40 .001 .15Error 121 39200.66 323.97Total 126 46330.13

Variable Beta WeightAge - .0905Months on Dialysis . 1536Diabetes 5.9283Hypertension 3.6623Total Number of Concurrent (Diagnoses

4.4660

Table 21

Stepwise Multiple Regression for the Demographic/Medical Variables and the Number of Hospital Days (Outliers Excluded)

StepNumber

Variable Multiple2

Entered Rdf F £ Beta

Weight

1 Months on Dialysis . 10 1, 125 14.18 .0003 . 16132 Total Number of

Concurrent Diagnoses. 13 2, 124 9.34 .0002 3-9829

3 Diabetes . 14 3, 123 6.72 .0004 5.0707

156

Table 22Multiple Regression Summary Table for the Combined Measures of Health Status and Demographic/Medical Variables and the Number of Hospital Days (Outliers Excluded)

2Source df Sum of Squares Mean Square F £ R

Model 14 10832.11 773-72 2.390 .006 .24Error 106 34368.37 324.23Total 120 45200.47

Variable Beta WeightBUN .1301Potassium -7935Phosphorous - .0392Hematocrit - .0717Wahler - .6272Staff Functional Ratings 1.5174Physician Ratings 2.6654Subjective Ratings of - .7869Functional StatusSubjective Ratings of Health 1.8459Age .1118Months on Dialysis .1626Diabetes - 9.2715Hypertension -3.0029Total Number 3.2949of Concurrent Diagnoses

Table 23Stepwise Multiple Regression for the Combined Measures of Health Status and Demographic/Medical Variables and the Number of Hospital Days (Outliers Excluded)

Step Variable Number Entered

Multiple2

Rdf F 2 Beta

Weight

1 Months on Dialysis . 12 1, 119 15.69 .0001 . 19252 Physician Rating . 16 2, 118 11.11 .0001 3.44033 BUN . 19 3, 117 8.98 .0001 . 1615

Ln

158

Table 24Standardized Linear Discriminant Function (LDF) Coefficients for the Health Status Measures

Variable LDF CoefficientsSurvivors Deceased

Blood Urea Nitrogen .0287 .0397Potassium 8.8153 8.3276Phosphorous .0329 . 1574Hematocrit 1.2664 1.3662Wahler Physical Symptoms Inventory .3868 .4686Staff Ratings of Functional Status 3.0363 4.6752Physicians Ratings 1.3063 .7578Subjective Health Status 4.8907 5.3164Subjective Functional Status -1.5661 -1 .8728

Table 25Discriminant Analysis Classification for theCalibration Sample Based on the Health Status Measures

Actuala

ClassificationPredicted Classification

SurvivorsPercent

Classified DeceasedPercent

Classified Total

Survivors 67 97.10 2 2.90 69Deceased 10 76.92 3 23.08 13Total 77 93.90 5 6.10 82aprior probability = .84

Table 26Discriminant Analysis Classification for theHoldout Sample Based on the Health Status Measures

Actuala

ClassificationPredicted Classification

SurvivorsPercent

Classified DeceasedPercent

Classified Total

Survivors 37 97.37 1 2.63 38Deceased 3 60.00 2 40.00 5Total 40 93.02 3 6.98 43aprior probability = .88

161

Table 27Standardized Linear Discriminant Function (LDF) Coefficients for the Demographic/Medical Variables

Variable LDF CoefficientsSurvivors Deceased

Age .2312 .2507Months on Dialysis .0416 .0669Diabetes 1.5097 2.2032Hypertension 2.7862 2.7429Total Number of Concurrent Diagnoses .3473 .8618

Table 28Discriminant Analysis Classification for theCalibration Sample Based on the Demographic/Medical Variables

Actuala

ClassificationPredicted Classification

SurvivorsPercent

Classified DeceasedPercent

Classified Total

Survivors 68 95.77 3 4.23 71Deceased 9 75.00 3 25.00 12Total 77 92.77 6 7.23 83aprior probability = .86

Table 29Discriminant Analysis Classification for theHoldout Sample Based on the Demographic/Medical Variables

Actuala

ClassificationPredicted Classification

SurvivorsPercent

Classified DeceasedPercent

Classified Total

Survivors 41 100.00 0 0.00 41Deceased 5 100.00 0 0.00 5Total 46 100.00 0 0.00 46aprior probability = .89

Table 30Discriminant Analysis Classification for theCalibration Sample Based on the Combined Measures of Health Status and Demographic/Medical Variables

Actuala

Classi ficationPredicted Classification

SurvivorsPercent

Classified DeceasedPercent

Classified Total

Survivors 69 100.00 0 0.00 69Deceased 0 0.00 12 100.00 12Total 69 85. 19 12 14.81 81aprior probability = .85

Table 31Discriminant Analysis Classification for theHoldout Sample Based on the Combined Measures of Health Status andDemographic/Medical Variables

Actuala

ClassificationPredicted Classification

SurvivorsPercent

Classified DeceasedPercent

Classified Total

Survivors 37 100.00 0 0.00 37Deceased 5 100.00 0 0.00 5Total 42 100.00 0 0.00 42aprior probability = .88

166

Table 32Means of Dependent Measures for Survivor and Deceased Groups

Survivors (n=106)

Deceased(n=17)

Blood Urea Nitrogen 76.67 75.59Potassium 5. 10 4.95Phosphorous 6. 32 6.66Hematocrit 25.75 26.35Staff Ratings of Functional Status 1.83 2.97Physicians Ratings 3.79 4.68Wahler Physical Symptoms Inventory 1.28 1.82Subjective Functional Status 2.42 3. 12Subjective Health Status 2.77 3.18Age 51.26 57.35Months on Dialysis 36.44 57.24Total Number of Concurrent Diagnoses 1.58 2. 18

I

167

Table 33Analysis of Variance Summary Table for Blood Urea Nitrogen (BUNT

Source df Sum of Squares Mean Square F E2

R

Model 1 17-14 17. 14 .04 .84 .0003Error 121 52057.56 430.23Total 122 52074.70

168

Table 34Analysis of Variance Summary Table for Potassium

Source df Sum of Squares Mean Square F £2

R

Model 1 .31 • 31 • 53 .47 .004Error 121 70.68 .58Total 122 70.99

169

Table 35Analysis of Variance Summary Table for Phosphorous

Source df Sum of Squares Mean Square F2

£ R

Model 1 1.70 1.70 .41 .52 .003Error 121 500.32 4. 14Total 122 502.63

170

Table 36

Analysis of Variance Summary Table for Hematocrit

Source df Sum of Squares Mean Square F £2

R

Model 1 5.24 5.24 .20 .65 .002Error 121 3143.50 25.98Total 122 3148.75

171

Table 37Analysis of Variance Summary Table for Staff Ratings of Functional Status

-

Source df Sum of Squares Mean Square F £ R

Model 1 19-21 19.21 26.64 .0001 .18Error 121 87.26 .72Total 122 106.47

172

Table 38

Analysis of Variance Summary Table for Physician Ratings of Physical Status

Source df Sum of Squares Mean Square F2

£ R

Model 1 11.45 11.45 7.29 .008 .06Error 121 189.90 1.57Total 122 201.35

173

Table 39Analysis of Variance Summary Table for Wahler Physical Symptoms Inventory

Source df Sum of Squares Mean Square F £2

R

Model 1 4.25 4.25 10.00 .002 00o

Error 121 51.37 .42Total 122 55.62

174

Table 40Analysis of Variance Summary Tablefor Sub.jective Ratings of Functional Status

2Source df Sum of Squares Mean Square £ £ R

Model 1 7.04 7.04 8.06 .005 .06Error 121 105.66 .87Total 122 112.70

175

Table 41Analysis of Variance Summary Table for Sub.jective Ratings of Health

; 2Source df Sum of Squares Mean Square F £ R

Model 1 2.38 2.38 3-38 .07 .03Error 121 85.04 .70Total 122 87.41

176

Table 42 Analysis of Variance Summary Tablefor Age

2Source df Sum of Squares Mean Square F £ R

Model 1 543. 14 543.14 2.74 . 10 .02Error 121 24024.49 198.55Total 122 24567.63

177

Table 43Analysis of Variance Summary Table for Number of Months on Dialysis

Source df Sum of Squares Mean Square F £2

R

Model 1 6333.41 6333.41 5.67 .02 .04Error 121 135261.22 1117.86Total 122 141594.63

178

Table 44Analysis of Variance Summary Table for Total Number of Concurrent Diagnoses

Source df Sum of Squares Mean Square F L2

R

Model 1 5.29 5.29 4.84 .03 .04Error 121 132.37 1.09Total 122 137.66

»

179

Table 45Analysis of Variance Summary Table for Diabetes Mellitus

Source df Sum of Squares Mean Square F £2

R

Model 1 • 09 .09 .49 .49 .004Error 121 21.54 . 18Total 122 21.63

180

Table 46Analysis of Variance Summary Table for Primary Hypertension

, _ Source df Sum of Squares Mean Square I! £ R

Model 1 .04 .04 .15 .70 .001Error 121 30.47 .25Total 122 30.50

Appendices

182Appendix A

PHYSICIAN RATINGS

Please estimate each patient's current medical condition on the attached sheet. Please indicate the number of the scale item contained below which best describes the patient’s overall physical status:

1. Patient is largely free of medical problems and isdoing well in a global sense in terms of medicaladjustment

2. Very mild3. Mild4. Moderate5. Severe6. Very severe7. Patient is suffering from and is incapacitated by

extremely severe medical complications of his/her illness

183

APPENDIX B WAHLER PHYSICAL SYMPTOMS INVENTORY

Name ___________________ Age Sex: M F Date________WHAT YOU ARE TO DO:

Below is a list of physical troubles. Please indicate how often each of these bothers you. Do this by circling the number to the right of each trouble which shows how often you are bothered by that trouble. Keep in mind that the LARGER the number the MORE OFTEN the trouble bothers you. Please DO NOT SKIP any troubles. You may take as much time as is necessary.

Nearly every day -------IIAbout twice a week ---- i

■ iAbout once a week I I

i i i

About once a month — ! ! !i i i ii i i i

About once a year — ! ! ! IAlmost never ------ii

1. Nausea (Feeling like throwing up). 0 2 3 4 5

2. Headaches. 0 2 3 4 53. Trouble with ears or hearing. 0 2 3 4 54. Neck aches or pains. 0 2 3 4 55. Feeling hot or cold regardless

of the weather. 0 2 3 4 56. Arm or leg aches or pains. 0 2 3 4 57. Shakiness. 0 2 3 4 58. Swelling of arms, hands,

legs, or feet. 0 2 3 4 59. Stuttering or stammering. 0 2 3 4 5

10. Difficulty sleeping. 0 2 3 4 511. Losing weight. 0 2 3 4 512. Backaches. 0 2 3 4 513. Intestinal or stomach trouble. 0 2 3 4 514. Difficulty with urination

(Passing water). 0 2 3 4 515. Heart trouble. 0 2 3 4 516. Trouble with teeth. 0 2 3 4 517. Numbness, or lack of feeling

in any part of the body. 0 2 3 4 518. Aches or pains in hands or feet. 0 2 3 4 5

184

Nearly every day About twice a week

About once a week ----About once a month ---II

About once a year --- Ii ii i

Almost never ------- ! i

19. Fainting spells. 0 1 2 3 4 520. Excessive perspiration. 0 1 2 3 4 521. Abnormal blood pressure. 0 1 2 3 4 522. Paralysis (Unable to move

parts of the body). 0 1 2 3 4 523. Trouble with eyes or vision. 0 1 2 3 4 524. Burning, tingling, crawling

feelings in the skin. 0 1 2 3 4 525. Skin trouble (Rashes, boils

or itching). 0 1 2 3 4 526. Feeling tired. 0 1 2 3 4 527. Muscular weakness. 0 1 2 3 4 528. Dizzy spells. 0 1 2 3 4 529. Muscular tensions. 0 1 2 3 4 530. Any trouble with the senses

of taste or smell. 0 1 2 3 4 531. Difficulty breathing (Short

of breath, asthma, etc.). 0 1 2 3 4 532. Twitching muscles. 0 1 2 3 4 533. Poor health in general. 0 1 2 3 4 534. Excessive gas. 0 1 2 3 4 535. Difficulty swallowing. 0 1 2 3 4 536. Seizures. 0 1 2 3 4 537. Gaining weight. 0 1 2 3 4 538. Difficulty with appetite. 0 1 2 3 4 539. Bowel trouble (Constipation

or loose bowels). 0 1 2 3 4 540. Vomiting. 0 1 2 3 4 541. Chest pains. 0 1 2 3 4 542. Hay fever or other allergies. 0 1 2 3 4 5Please write down any important physical symptoms notlisted above which trouble you:

185APPENDIX C

FUNCTIONAL CLASSIFICATION

Please estimate(patient's name) current general level of physical activity. Please circle the number of the scale item which best describes the patient's current functional status or degree of disability:

1. Capable of performing all usual types of physicalactivity

2. Unable to perform the most strenuous of usual physicalactivities (e.g. sports, fast walking, shoveling snow, lawn mowing)

3. Unable to perform usual daily activities on more than a part-time basis (e.g. housework, employment, driving a car, playing with children)

4. Severe limitation of usual physical activity, may needassistance with some aspects of self care, may haveimpaired mentation, may be confined to bed

186

APPENDIX D

SUBJECTIVE FUNCTIONAL RATING

Please estimate your current level of physical activity. Please circle the number of the scale item which best describes your current functional status or degree of disability:

1. Capable of performing all usual types of physical activity

2. Unable to perform the most strenuous of usual physical activities (e.g. sports, fast walking, shoveling snow, lawn mowing)

3. Unable to perform usual daily activities on more than a part-time basis (e.g. housework, employment, driving a car, playing with children)

4. Severe limitation of usual physical activity, may need assistance with some aspects of self care, may have impaired mentation, may be confined to bed

187

APPENDIX E

SUBJECTIVE RATING OF HEALTH

Please estimate, in your own opinion, your current level of health. Circle the number of the scale item which best describes your overall health ranging from excellent to poor.

1. Excellent

2. Good

3. Fair

4. Poor

APPENDIX F188

INFORMED CONSENT

I, ______________________________________, freely andwillingly consent to be a participant in a research project investigating the assessment of health status in hemodialysis patients. This project is being directed by P. J. Brantley, Ph.D., T. B. Cocke, M.D., and G. T. McKnight, M.D., of the LSU Medical School.

As a research subject, I agree to participate in one assessment interview lasting 20 minutes, in which I will be asked to answer several paper and pencil questionnaires. Additionally, I agree to allow the research team to review any of my pertinent medical records. I understand that there are no risks involved by participation in this project.

I understand that I may withdraw from participation in this study at any time with no adverse consequences. In addition, any information I provide during this project will be kept in strict confidence, and if this information is presented publicly (i.e. conferences, journal articles), no information will be identified with me personally.

189

I realize that I have a right to ask questions at any time and to have these questions answered to my satisfaction. I have read and thoroughly understand this consent form.

Participant Date

Witness Date

190

APPENDIX G

DEMOGRAPHIC/MEDICAL DATAPatient’s Name ________________________ ID# _________________Age _______ Sex Race/Ethnic ____________________Occupation/Vocation Current job status __________Marital Status ____ How long? ____ Age of partner ___Education: Gr Hs Coll_____ Grad S ____Source(s) of income ____________________________________________Living arrangements ____________________________________________Primary Medical Diagnoses: Date of initial diagnosis:(Renal)

Other Medical Diagnoses:

Current Medication:Drug Home/Unit Dose Administration Regimen

Current Dialysis:_______ hrs. per visit Date dialysis beganRelevant treatment history ________________________

Curriculum Vitae

PERSONAL DATAName: Barbara Kathryn Bruce

January 26, 1954 Baton Rouge, Louisiana

Date of Birth:Birthplace:Present Position: Instructor, Department of Medicine,

Louisiana State University School of Medicine

EDUCATIONBroadmoor High School Baton Rouge, Louisiana Graduation May, 1972.Louisiana State University Baton Rouge, Louisiana Major: PsychologyReceived B.A. May, 1976University of New Orleans New Orleans, Louisiana Major: Physiological PsychologyReceived M.S. May, 1981Louisiana State University Baton Rouge, Louisiana Major: Clinical PsychologySpecialty Area: Behavioral Medicine Minor: Behavioral NeurologyReceived Ph.D. May, 1986.

RESEARCHMaster’s thesis: The Effects of Adrenalectomy on

Hypothalamic Hyperphagia and Obesity.Intermediate Research Project: Convergent-Discriminant

Validity In The Assessment of Adolescent Social Skills.

192

Dissertation: The Assessment of Health Status InHemodialysis Patients.

PUBLICATIONS:Murphy, J. K., Bruce, B. K., & Williamson, D. A. ( in

press). Couples treatment of obesity; A four year follow-up. Behavior Therapy.

Brantley, P. J., & Bruce, B. K. (in press). Assessment in behavioral medicine. In A. R. Ciminero, K. S. Calhoun, & H. E. Adams (Eds.), Handbook of Behavioral Assessment (Vol.2). New York: John Wiley and sons.

Williamson, D. A., Moody, S. C., Granberry, S. W., Lethermon, V. R., Lemanek, K. L., & Bruce, B. K. (in press). Development of a standardized, norm-referenced role-play test of children’s social skills. Journal of Clinical Child Psychology.

Gresham, F. M., Bruce, B. K., Veitia, M. C., & Jones, G. (in press). Same-sex, cross-sex, and age differences in peer assessment of adolescent social skills. Journal of Early Adolescence.

King, B. M., Banta, A. R., Tharel, G. N., Bruce, B. K.,& Frohman, L. A. (1983). Hypothalamic hyperinsulinemia and obesity: Role of adrenal gluco-corticoids. American Journalof Physiology, 83, 194-199.

Bruce, B. K., King, B. M., Phelps, G. R., & Veitia, M.C. (1982). Effects of adrenalectomy and corticosterone administration on hypothalamic obesity in rats. American Journal of Physiology, 243, 152-157.

E X A M IN A T IO N A N D THESIS R EPO R T

Candidate: Barbara K. Bruce

Major Field: Psychology

T itle of Thesis: The Assessment of Health Status in Hemodialysis Patients

Approved:

lv)Dean of the Graduate School

EXAMINING COMMITTEE:

L o

1$ A i n U l i d — ________

Date of Examination:

Anril 25. 1986