learning objectives for this filepeople.musc.edu/~decristc/adv patho/unit 1 cell gene inflamm...

Adv Pathophysiology Unit 1: Cell, Gene, Inflamm, Immune of 23

File: advpatho_unit1_9disrisk.pdf Source: C. DeCristofaro, MD

Learning Objectives for this file: 1. Analyzing disease risk – RR, predictive value, P value, reliability, accuracy, validity 2. Normal & abnormal – the normal curve for biologic variables 3. Epidemiology – preventive health, screening, RCT



BASIC STATISTICAL ANALYSIS FOR CLINICIANS & INTERPRETING LAB RESULTS: Some statistical concepts: we are always COMPARING to the “general population” • prevalence: number of current cases per population at one point in time. • incidence: number of newly diagnosed cases in a given population over a period of time • if a subpopulation (pediatric, geriatric, men, women, etc.) must state in the analysis Validity: “pertinence” & “biologic plausibility” • how pertinent is this data or fact to the stated problem or disease? • are the results biologically plausible (compatibility with known biology) – do your numbers

match reality? o if my potassium is 9.0, I’d be dead – but here I am typing this file!! Invalid data!!

Reliability: reproducibility of results (data) • diagnostic test is repeated by either the same observer (intra-observer reliability) or different

observers (inter-observer reliability). • poor reproducibility indicates this data should not be used in decision-making Confounding factors: • separate factors that affect the data – may be related to the problem studied, or unrelated • example: basing decisions on data obtained from returned surveys – perhaps the type of

person who returns surveys is NOT similar to the general population!! What is a rare disease?

• Sometimes referred to as orphan diseases • Classified in the USA as diseases or conditions that affect fewer than 200,000 individuals • There are more than 6,800 rare diseases, affecting 25 to 30 million people in the USA

o For any one disorder, to be a rare disease it must affect <200,000 people • Pharmacological correlate: Orphan Drug Act of 1983 • See (Kesselheim, 2010, National Academy of Science):

http://www.ncbi.nlm.nih.gov/books/NBK56187/ • NIH website for rare diseases: https://rarediseases.info.nih.gov/

http://www.ncbi.nlm.nih.gov/books/NBK56187/

https://rarediseases.info.nih.gov/



Accuracy:

• Includes concepts of sensitivity, specificity, & predictive value (more below) • Sensitivity: proportion of persons with the condition who test positive (remember:

positive-positive). o Good sensitivity: correctly identify disease in those with the disease. o Poor sensitivity: produces a large proportion of "false negatives" will miss cases

& true cases will be told mistakenly that they are free of disease. • Specificity: proportion of persons without the condition who test negative (remember:

negative-negative). o Good specificity: correctly identify healthy persons as free of disease. o Poor specificity: large proportion of "false positives" consequence is that healthy

persons will be mistakenly told they actually have disease • leads to clinical correlates and terms such as "false positive" & "false negative":

o statisticians use terms such as Type I and Type II errors o “false positive” means that the test is positive, but the condition doesn’t actually

exist (Type I error) – a person will be told that they have disease when they really don’t have it – may result in expensive and even dangerous treatments or psychological damage (or self-fulfilling prophecies and development of conditions due to “labeling” of the individual)

o “false negative” means that the test is negative, but the condition does actually exist (Type II error) – a person will be told they are free of disease, when they really do have the disease – may result in delayed treatment and behaviors that transmit disease



What is the “p” value & the “power” of a study? • for any study using diagnostic data, there is a possibility that the results are due to pure

CHANCE and that there is no real relationship between the data • if the “p” value is less than 5% (0.05), then the results are more than 95% NOT due to

CHANCE and that the relationship is “true” • for most scientific studies, we accept a p < 0.05 as the MINIMUM requirement to accept the

results of the data • the lower the “p” value, the better!! (e.g. a p < 0.001 is excellent) • How to get a really good statistical significance:

o this usually requires LOTS of participants in a study (a “higher powered” study has MORE participants, e.g. tens of thousands)

• Look for the “p” value whenever reading the results of scientific studies • NOTE – in most studies, ALL data is reported – and then the statistical significance results are

also given o You may want to see the data even if it is not statistically significant – it may indicate a

trend o Most readers will discount any data that is not statistically significant – remember, this

means that the results may be due to chance more so than we are willing to accept What is the “Confidence Interval” (CI)? • Another statistical test to determine if we will accept the relationship of the data • An estimate of the range within which the true results lie

o For example, the 95% CI is the range of values within with we are 95% certain that the true value lies

o If the CI for the difference between two tested treatments includes zero, then you cannot exclude the possibility that there is really no difference between the two treatments (e.g. placebo and treatment are giving the same effect)

o If the range of values is very large, the final result may not be clinically relevant



Predictive value: • looking at a proportion of test results compared to population with or without the disease. • PPV is also affected by the prevalence rate of the condition in the tested population. If

condition is very rare (low prevalence rate) the PPV generates more false positives (rarity of finding true case).

• Screening tests: PPV must be high for a screening test to be of value. o Positive predictive value (PPV): proportion of positive tests results that are "true

positive." o Negative predictive value: proportion of persons with negative test who do not have

the condition ("true negative"). Example using PPV from researching a new way to do mammography: • Conventional screening mammography is less accurate in younger women (age < 40 yo)

because their breast tissue is denser than postmenopausal women, one reason we do mammograms less frequently in younger women .

• A new technology, “electrical impedance system” (EIS) uses electrical signals from breast tissue to enhance the mammographic image.

o patient holds a cylinder that serves as a ground & another instrument is held sequentially over the nine segments of the breast, including the nipple

o the electrical circuit created with the cylinder generates data that is imaged by computer (computer software categorizes findings as "normal" or "suspicious”)

• In younger women, age < 40 yo, the specificity was 96.2% (“negative-negative”) – which means that of 601 women with no risk of breast cancer, 578 had normal findings. Thus, this EIS screening test would NOT result in too many errors of “false positive”

• In women with actual breast abnormalities: o at least one benign lesion on follow-up imaging sensitivity of EIS was 6% o a low-risk benign lesion confirmed by biopsy the sensitivity of EIS was 33.3% o a confirmed carcinoma the EIS sensitivity was 66.7% (P< .001)

• Thus, a woman with a positive EIS result is seven times more likely to have breast cancer than a woman in the general population

Medical Calculators & Clinical Values – some weblinks:

• This link gives a listing of various statistical analysis examples – for example, how to calculate radiation risk

http://www.changbioscience.com/bio/cal0.htm • This link is a set of medical calculators, from Cornell Medical School:

http://www-users.med.cornell.edu/~spon/picu/calc/medcalc.htm • Many medical calculators, including drug dosing: http://www.globalrph.com/medcalcs.htm • Other calculators look for drug interactions: http://www.rxlist.com/drug-interaction-

checker.htm and http://umm.edu/health/medical/drug-interaction-tool A nice discussion of these topics at: http://www.merckmanuals.com/professional/special-subjects/clinical-decision-making/understanding-medical-tests-and-test-results?qt=&sc=&alt

http://www.changbioscience.com/bio/cal0.htm

http://www-users.med.cornell.edu/%7Espon/picu/calc/medcalc.htm

http://www.globalrph.com/medcalcs.htm

http://www.rxlist.com/drug-interaction-checker.htm

http://www.rxlist.com/drug-interaction-checker.htm

http://umm.edu/health/medical/drug-interaction-tool

http://www.merckmanuals.com/professional/special-subjects/clinical-decision-making/understanding-medical-tests-and-test-results?qt=&sc=&alt

http://www.merckmanuals.com/professional/special-subjects/clinical-decision-making/understanding-medical-tests-and-test-results?qt=&sc=&alt



LABORATORY MEDICINE – OVERVIEW – INTERPRETATION & ORDERING: "Normal Laboratory Values": what is “normal”??? • this is ONLY a statistical CONCEPT that we have created • first, look at the population – and actually test them to determine the empirical results

(empirical is what you get when you actually test, not just theory!) • then, find out the distribution of these results (see below) • after determining the distribution, you then find out how many people in the population are

very far away from the average value (or median value) and this determines your standard deviation (SD)

• by definition, normal values in a healthy population are those that fall within 2 SD of the mean/median (see bell-shaped NORMAL curve below)

Distribution of results for biologic data – the NORMAL curve: • for most biologic values, you see the data distribute in a “bell-shaped” curve, called a “normal

distribution” curve (see picture) • this tells us several things:

o the mean (average) = the median (the value at which 50% fall below, and 50% fall above)

o most values “cluster” around the mean, so that very few people are actually very far away from the mean – called a “measure of central tendency”

o thus, we are all very much alike! and, we can apply similar diagnostics and treatments to MOST people who wander into our clinical practice

• remember, “normal” means that the value falls within 2 SD of the mean/median (approx 96% of the population) – in a normal distribution curve

o you capture 68% of the population within 1 SD above & below the median/mean o you capture 95-96% of the population within 2 SD above & below the median/mean

• BUT, some folks fall outside the “normal” values and these are called “outliers” o there are 2% of people falling below and above the 2 SD spread that we have defined

as “normal” o clinical consequences:

may need a lot MORE medicine, or a lot LESS medicine, to get the same effect as on most people

their lab values may be “abnormal” even though they are perfectly healthy think of other examples (reporting pain more intensely, having more side effects

from medication at “normal” doses, etc.)

We are usually measuring what are called continous biologic variables that occur within a particular range. These usually will plot out as a “normal” bell-shaped curve. (Example: serum cholesterol, blood pressure, temperature)



Probability of abnormal results in the healthy population: • the statistical consequence is that 4-5% of normal persons will have "abnormal" lab

values for a specific test simply because they fall outside of the normal SD spread and are outliers

• every time you obtain a test, EACH test value mean is derived independently, so that the different tests do not depend on each other for normal range determination

• therefore, the probability of any one individual having multiple abnormal results is small (multiply 5% x 5% that many times for the number of lab tests in question).

• Clinical: o if you run multiple tests, you would expect 50% of patients to have at least one

abnormal value in a panel containing 10 tests, and 66% would have at least one abnormal value on a 20-test panel.

o The more tests, the higher likelihood of a single abnormal value occurring. • the idea of CLINICAL SIGNIFICANCE derives from this understanding – is your abnormal

value of significance in terms of health/disease, or not? do you followup? do you retest? do you simply chart “not clinically significant”?

o it is tempting to take the last option – BUT we are NOT allowed to simply “discard” lab values that don’t agree without working diagnosis

o before calling something “not clinically significant” you would at least retest once, and then usually bolster your evaluation with OTHER data (exam, history, additional lab or imaging diagnostics) that reinforces this decision

o explain why a value is not yielding the expected results for this condition o Example: in anemia we look at multiple supporting lab values (RDW, ferritin, RBC

indices, Coomb's test, reticulocyte count, Hb electrophoresis, vitamin levels) as well as physical findings (e.g. splenomegaly) to determine what kind of anemia this is (e.g. hemolytic anemia from thalassemia)

• in addition to all these statistical issues in interpretation of lab results, we must add in the fact that there will be other confounding factors such as

o lab error o clinical contributions to lab result deviations (subpopulations, morbidity, medication,

etc.) o we have to interpret lab data based on all these factors

Lab values & the Clinical Workup: • due to inherent limitations on utilizing lab data, you must bolster your diagnosis with multiple

values that support the diagnosis. • This reduces the chance that a diagnosis was made solely on the basis of a lab value that

may be in error. • If the expected combination of lab tests & physical findings do not agree internally with the

working diagnosis, a revision of the diagnosis or further testing is indicated.



What to order: Only order lab tests necessary for case management (not just out of “curiosity” !!) Order if you need to • Make (or confirm) the diagnosis

o lab results support the diagnosis o pathognomonic (“names the disease”) means that if you have this finding, you

DEFINITELY have a particular condition (often used in connection with physical exam findings, but can also be used for lab/imaging; example: “rat-tail deformity” on aortic imaging via transesophageal

echocardiography indicates diagnosis of aortoarteritis, this is pathognomic for aortoarteritis (see image below)

o sine qua non (“without which you don’t have”) means that if you DON’T have this finding, you DON’T have the condition example: acute leukemia requires > 20% clonal blasts in the bone marrow

biopsy to make the diagnosis. If you DON’T have this finding, you DON’T have acute leukemia. (see image below)

• Plan the therapeutic intervention o example: baseline renal function studies done before starting ACE inhibitor therapy in

heart failure; other drugs may have to be used if there is compromised renal function, and further studies (imaging of kidney vasculature) may also have to be done

• Monitor disease progression & resolution o example: observation of changes in EKG and cardiac enzymes over time for

progression of myocardial infarction • Determine therapeutic effectiveness & safety of interventions

o example: begin diet and drug therapy for hyperlipidemia and then measure cholesterol panel every 6 weeks until achieve therapeutic goal; liver function testing also done to monitor drug safety

Pathognomonic “names the disease” Sine qua non “without which it doesn’t exist”

Rat-tail deformity. Krishnamoorthy KM and Patle A. Imaging in aortoarteritis. Heart 2002;88:59.

Proliferation of small lymphoblasts in ALL (acute lymphoid leukemia) Scientific Data Center for the Atomic Bomb Disaster, Nagasaki University.



LAB ERROR: • there are expected lab errors (related to comorbidity, artefactual, etc.) • there are unexpected lab errors (related to technology – reagent, human error, etc.) Technology related: • automated methods increase errors

o example: traces of sample & reagent remaining in chamber o controls & calibrations less accessible to lab personnel.

• A lab should be glad to be alerted of clearly unexpected or incompatible values • Examples:

o Calcium results on panels: Common that Calcium result on automated Chemistry panels is unreliable Repeat with new sample, or request hand-done reagent test from the lab.

o Reagent interactions: “point-of-care” testing (testing the patient and getting the lab results immediately)

include bedside blood glucose testing with automated glucose meters – the same ones patients use at home to test their blood glucose

Maltose (sometimes included in IV products) will falsely elevate the blood glucose result, so the patient is treated as though they have hyperglycemia and this treatment can put them into hypoglycemic shock!!

Patient related: • circadian cycles (e.g. cortisol value taken at certain times of the day) • artefactual:

o such as sampling technique o example:

• pseudodilution:moving from standing to supine raises the plasma volume & therefore dilutes the values of some substances

• pseudohyperkalemia: prolonged tourniquet time, blood left standing at room temperature, drawing blood through very small needle (lyses RBC and causes elevated potassium value)

• confounding errors o due to other abnormal lab results in the same patient, OR patient comorbidities o examples:

• pseudohyponatremia: seen with elevated triglycerides • pseudohypoglycemia: seen with leukocytosis • Falsely elevated Cr in DKA: the acetoacetate organic acid in the DKA (diabetic

keto-acidosis) blood interferes with the automated reagent



LAB VALUES THAT CHANGE – ACCEPTABLE VARIABILITY OF VALUES: OVERVIEW:

• Many things can change a lab value – complete knowledge of the patient’s exam, history, clinical situation is therefore NECESSARY to interpret lab values

o lifespan o gender o comorbidity (associated conditions) o concurrent drug therapies o pregnancy

• NOTE that even NORMAL appearing values require this background information for interpretation (e.g. in a woman taking oral contraceptives or estrogen hormones, a NORMAL thyroid panel actually indicates that she has an UNDERactive thyroid)

• Normal values for age & sex are usually listed on the report • Your resource for this information can be texts, lab manuals, internet; and you can also call

the lab and ask to speak to the medical technologist PEDIATRIC POPULATION: Reference & Interpretation: • use a reference such as the Harriet Lane Handbook for normal values for age & sex • for more unusual conditions or tests, standard pediatric texts (e.g. Nelson's Pediatrics) • you may need to know the child’s age, height, weight and/or surface area in order to interpret

the values



The body surface area (BSA) Nomograms: • height-weight nomograms are used to determine these values • Children:

o the West body surface area (BSA) nomogram is a graphic display of linked data o connect a line from the weight to the height and where it crosses the middle line, that is

the Surface Area (SA) o See:

https://www.uncp.edu/sites/default/files/Images_Docs/Academics/Colleges_Schools_and_Departments/Departments/Nursing/Learning_Enhancement_Center/Nomogram.pdf

• Adult: o Calculations for adults: http://www.medcalc.com/body.html o BSA nomogram for adults:

http://onlinelibrary.wiley.com/doi/10.1002/9781444318364.app14/pdf



http://www.medcalc.com/body.html

http://onlinelibrary.wiley.com/doi/10.1002/9781444318364.app14/pdf



"Normally Abnormal": • lab values are expected to change over the child's life (premature neonate full-term infant toddler adolescent young adult)

• Common lab values that change: WBC differential count, RBC/Hb/Hct, haptoglobins, ferritin, serum iron, aldolase, base excess, bicarbonate, bilirubin, cholesterol, Creatinine, GGTP, iron, LDH, lipids, oxygen saturation, acid phosphatase, alkaline phosphatase, phosphorus, SGOT, SGPT, immunoglobulins, most hormones, Calcium.

• Examples: o due to bone growth, the alkaline phosphatase will almost always be elevated on the

chemistry panel (& more so during the pubertal growth spurt). o Thyroid & other endocrine values change radically throughout infancy childhood

adolescence adulthood senescence o LDH levels are usually higher than in adult. o Creatinine values require use of pediatric SA nomogram to interpret normalcy.

?? MALE OR FEMALE: • normal results change across many categories. • Examples: RBC/Hb/Hct, plasma volume, sed rate, cholesterol, uric acid, porphyrins, most

hormones, renal function tests. ASSOCIATED CONDITIONS (COMORBIDITIES) AND DRUG EFFECTS: Confounding variables: • changes in some lab values cause abnormals in others • Example: pseudohyponatremia of hyperlipidemia Drug therapies: • affect drug levels of other drugs (e.g. theophylline interaction with quinolone antibiotics). • Can also change other basic lab test values • Example: thyroid panel interaction with estrogen hormones

o thyroid test results will change in your patients on oral contraceptives (or any form of estrogen)

o The thyroid profile will be very abnormal looking, but in an expected way, due to a change in the amount of thyroid binding globulin (TBG).

o Normally, the T4, T3, TBG, & RAIU will be increased; the FT4, TSH and TRH test normal; and the T3U decreased.

o If the patient profile matches this, then all is normal. But, if it doesn't match, then there may be disease.

o If your patient has a normal T4 on estrogen or OC, she may actually be hypothyroid !!



GERIATRIC: Increased: serum copper, serum ferritin, BUN, serum creatinine (although may be normal even in the face of CRF until very late in disease process -- see below), serum alkaline phosphatase, serum immunoglobulin M, serum immunoreactive parathormone, serum cholesterol, serum uric acid, monoclonal gammopathy, serum fibrinogen, serum norepinephrine, serum triglycerides, serum glucose, PSA (normal value increases with age). Decreased: serum calcium, serum iron, serum phosphorus, serum thiamine, serum zinc, serum 1,25-di-OH-vitamin D, serum vitamin B6, serum vitamin B12, plasma vitamin C, serum selenium, plasma tocopherol (vitamin E), T3 (> age 75), serum testosterone, serum albumin, creatinine clearance. Unchanged (although many practitioners think these change due to age):

• hemoglobin, RBC count, WBC count, albumin, serum vitamin A, serum pantothenate, serum riboflavin, serum carotene, ESR (sed rate).

• Unclear at this time is effect of aging on thyroid profiles & other vitamin levels. • With the aged, unless you are sure that aging results in a significant change in a lab value,

assume a reported abnormal value really IS really abnormal.



EVIDENCE-BASED MEDICINE (EBM) & STATISTICAL TERMS: • Sensitivity (Sn): Percentage of patients with disease who have a positive test for the disease

in question • Specificity (Sp): Percentage of patients without disease who have a negative test for the

disease in question • Predictive value (positive and negative) (PV+ / PV- ): Percentage of patients with a positive

or negative test for a disease who do or do not have the disease in question • Pretest probability: Probability of disease before a test is performed • Post-test probability: Probability of disease after a test is performed • Likelihood ratio (LR): LR >1 indicates an increased likelihood of disease, LR <1 indicates a

decreased likelihood of disease. The most helpful tests generally have a ratio of less than 0.2 or greater than 5.

• Relative risk reduction (RRR): The percentage difference in risk or outcomes between treatment and control groups. Example: if mortality is 30 percent in controls and 20 percent with treatment, RRR is (30-20)/30 = 33 percent.

• Absolute risk reduction (ARR): The arithmetic difference in risk or outcomes between treatment and control groups. Example: if mortality is 30 percent in controls and 20 percent with treatment, ARR is 30-20=10 percent.

• Number needed to treat (NNT): The number of patients who need to receive an intervention instead of the alternative in order for one additional patient to benefit. The NNT is calculated as: 1/ARR. Example: if the ARR is 4 percent, the NNT = 1/4 percent = 1/0.04 = 25.

• Number needed to harm (NNH): The number of patients who need to receive an intervention instead of the alternative in order for one additional patient to experience an adverse event.

• 95 percent confidence interval (95% CI): An estimate of certainty. It is 95% certain that the true value lies within the given range. A narrow CI is good. A CI that spans 1.0 calls into question the validity of the result.

• Systematic review: A type of review article that uses explicit methods to comprehensively analyze and qualitatively synthesize information from multiple studies

• Meta-analysis: A type of systematic review that uses rigorous statistical methods to quantitatively synthesize the results of multiple similar studies

FROM: http://www.aafp.org/journals/afp/authors/ebm-toolkit/glossary.html The 5-step process in EBM:

• Step 1: Formulating a well-built question • Step 2: Identifying articles and other evidence-based resources that answer the question • Step 3: Critically appraising the evidence to assess its validity • Step 4: Applying the evidence • Step 5: Re-evaluating the application of evidence and areas for improvement

Tutorials on EBM (EBP):

• See (excellent): http://guides.mclibrary.duke.edu/ebmtutorial (excellent!) • See: http://medlib.bu.edu/tutorials/ebm/intro/ (quick) • See: http://library.downstate.edu/EBM2/contents.htm (also good)

http://www.aafp.org/journals/afp/authors/ebm-toolkit/glossary.html

http://guides.mclibrary.duke.edu/ebmtutorial

http://medlib.bu.edu/tutorials/ebm/intro/

http://library.downstate.edu/EBM2/contents.htm



POEMs VS. DOEs: Part of Evidence Based Practice:

• POEM: o Patient oriented evidence that matters o Outcomes studied that would change someone’s life (quality or length)

• DOE: o Disease oriented evidence o Outcomes studied that are physiologic or “surrogate” markers of health, but do not

directly look at morbidity, mortality, quality of life, symptoms More on POEMs and DOEs (EBM information here also): https://wilkes.libguides.com/c.php?g=191942&p=1266516 EBP Manual for family practice: http://web.squ.edu.om/med-Lib/MED_CD/E_CDs/Evidence-Based%20Family%20Medicine/family.pdf (all 192 pages of an excellent manual for family medicine)

https://wilkes.libguides.com/c.php?g=191942&p=1266516

http://web.squ.edu.om/med-Lib/MED_CD/E_CDs/Evidence-Based%20Family%20Medicine/family.pdf

http://web.squ.edu.om/med-Lib/MED_CD/E_CDs/Evidence-Based%20Family%20Medicine/family.pdf



GRADING THE MEDICAL LITERATURE – STRENGTH OF RECOMMENDATION TAXONOMY (SORT): FROM: http://www.aafp.org/afp/2004/0201/p548.html and http://www.aafp.org/journals/afp/authors/ebm-toolkit/strength.html (Rating the Strength of Evidence)

http://www.aafp.org/afp/2004/0201/p548.html

http://www.aafp.org/journals/afp/authors/ebm-toolkit/strength.html



EPIDEMIOLOGY: • Incidence rate is the number of new cases detected during a given period of time (e.g. one

year). • Mortality rate is the number of people dying form this disease during a given period of time.

Using these two rates yields the --> • Prevalence rate is the number of persons currently living with the disease, both old

accumulated cases and new cases. Expressed as number per persons in the population at large. Determined by incidence rate & mortality rate.

• Risk factors: o may be causal (directly causing the disease, e.g. hyperlipidemia & ASHD) o may be noncausal – are simply associated – do not directly cause the illness, but

factor is associated with the illness example is smoking & development of Grave's disease what does “associated” mean – well, firefighters show up at fires – did they

CAUSE the fire ?? They are certainly associated with fires !! • Relative risk (RR):

o the risk of developing the disease if a causal factor is present, compared to someone without the causal factor; expressed as a ratio.

o Relative Risk: RR > 1 means your risk is MORE than the general population RR = 1 means your risk is EQUAL to the general population RR < 1 means your risk is LESS than the general population

o Example: relative risk of 13.3 for smoking and lung cancer means a smoker is 13.3 times more likely to develop the disease compared to a matched nonsmoker

o Example: relative risk of developing hypertension increases above RR=1 at BP level of 130/80 (note the white dashed line representing RR of 1)

Relative Risk Increases with BP > 130/80



Other factors in evidence-based science: • Confounding factors can affect statistical analysis, so must carefully match studied

populations. • Prospective studies are considered to have less problems with confounding factors (subjects

can be excluded as confounding factors are discovered). • Retrospective studies more common, but more subject to confounding factors. • RCT (Randomized Clinical Trial) (also called Randomized Controlled Trial):

o Properly constructed and implemented clinical trial to test an intervention (drug, lifestyle, surgery, etc.) and determine how outcome is matched to the intervention

o Also could look at risk factors and development of disease o Requires matched cohorts of patients – matched as closely as possible by age, sex,

clinical condition, lifestyle, possibly race/ethnicity, etc. o Requires the use of placebo to test on one cohort and active drug (or other

intervention) on the other cohort o Usually it is “double-blind” (neither the investigator nor the subject knows if they are

receiving a placebo or an active drug) o Usually it is “double-dummy” (midway through the experiment, the placebo and active

cohorts are switched) o Clinicians expect that results obtained by properly powered (enough subjects) RCTs

are more statistically valid and useful



LABORATORY TESTING IN PERIODIC HEALTH SCREENING – OVERVIEW: Screening guidelines: population-based recommendations • Voluntary/professional associations such as the American Cancer Society (ACS), American

Academy of Pediatrics (AAP), American Academy of Family Practice (AAFP), US Public Service Task Force (USPSTF), etc. often disagree regarding the need, utility, or benefit of performing the different screening tests (e.g. mammography for breast cancer)

• As a provider, make a determination for each individual – requires clinical experience; knowledge of local epidemiological factors & case population (ethnicity, lifestyle), scientific knowledge, other factors (financial, third party insurance).

o Example: FNP might follow AAFP guidelines, or a PNP might follow AAP guidelines, or an endocrinologist might follow AACE guidelines). Many controversial areas (e.g. prostate cancer screening – digital rectal & PSA lab test)

• Periodic Health Exams: recommended by age/sex for general "health maintenance" & preventive medical care from AAFP http://www.aafp.org/patient-care/browse/all-recommendations-type.html and from AHRQ http://www.ahrq.gov/professionals/clinicians-providers/guidelines-recommendations/guide/index.html

Basis for recommendation – the clinical EVIDENCE base: • should be their utility in preventing illness (or diagnosing illness at an early enough stage to

provide lead time for adequate intervention to prevent significant morbidity & mortality) • should be derived from an evidence base that indicates the VALUE of such screening to

accomplish the first goal (see below) o REMEMBER THAT EVIDENCE-BASED PRACTICE (EBP)(Medicine)(EBM) is now

understood to be the “gold standard” of practice o not what we “think is correct” or “wish is correct” but what is ACTUALLY correct based

on examination of the facts (epidemiologic population-based analysis) o GREAT tutorial on EBM: http://guides.mclibrary.duke.edu/ebmtutorial o The strength of recommendations taxonomy (SORT) is based on the types of

available studies (good article http://www.aafp.org/afp/20040201/548.html ) • Example of applying these recommendations: annual Chest Xray (CXR) for smokers is no

longer recommended – by the time a pulmonary lesion is found this way, most will have such advanced disease that they will not benefit from therapy.

Graded Recommendations: based on levels of evidence in literature

• Level 1 – systematic reviews with homogeneous results, randomized controlled trials (RCTs), or all-or-none case series

• Level 2 – cohort studies and outcomes research • Level 3 – case-congtrol studies and nonconsecutive results • Level 4 – evidence indicates case series and poor-quality cohort studies • Level 5 – expert opinion

Strength of Recommendations Taxonomy (SORT)(SOR):

• SOR A – consistent Level 1 studies • SOR B – consistent Level 2 or 3 studies or extrapolations from Level 1 studies • SOR C – Level 4 studies or extrapolations from Level 2 or 3 studies • SOR D – Level 5 evidence or troublingly inconsistent or inconclusive studies of any level

http://www.aafp.org/patient-care/browse/all-recommendations-type.html

http://www.aafp.org/patient-care/browse/all-recommendations-type.html

http://www.ahrq.gov/professionals/clinicians-providers/guidelines-recommendations/guide/index.html

http://www.ahrq.gov/professionals/clinicians-providers/guidelines-recommendations/guide/index.html

http://guides.mclibrary.duke.edu/ebmtutorial

http://www.aafp.org/afp/20040201/548.html



Sterotyped testing:

• guidelines for diagnostics determined by analysis of risk factors based on ethnicity, demographics, age, sex, comorbidities, history (personal & family), lifestyle

• we are “profiling” our patients based on demographics – determination of risk factors to evaluate their level of risk and then applying screening protocols

Pre-employment & Industrial:

• you are employed by the prospective employer, not by the patient. • Therefore, your duty is to the employer, who may require specific tests (e.g. LS spine films

to rule out spondylolisthesis or spondylolysis). • There have been cases of non-disclosure of potentially life-threatening diagnostic findings,

with the courts holding that the practitioner had no obligation to inform the patient without specific permission of the sponsoring employer! (what about the ethics?)

• These may also include ongoing exams for employed individuals exposed to specific risk-behavior (hearing, toxins) that require special periodic testing (audiometry, PFT, LFT).

Pre-school, school-age, & sports:

• most follow AAP guidelines. • include vision, audiometry, periodic glucose, Hematocrit, lead; screen for scoliosis, inguinal

hernia, undescended testes; universal immunizations, PPD Signing up for listservs: • A listserv is a broadcast email sent to persons interested in that topic • Sign up for:

• NIH https://list.nih.gov/ (instructions at: https://list.nih.gov/LISTSERV_WEB/USERSGDE/appendix_c.htm ) all the NIH listservs are here

• Consider signing up for FDA Drug Info, FDA News Digest, FDA Newswire • NIMH https://public.govdelivery.com/accounts/USNIMH/subscriber/new (Mental

Health) • NLM http://www.nlm.nih.gov/listserv/emaillists.html

US Preventive Services Task Force (USPSTF) recommendations using app: Some recommendations are at a higher level of certainty

• In clinical practice, using level “A” and “B” recommendations are a best practice • See downloadable app for smartphone/tablet • Electronic Preventive Services Selector (ePSS):

o http://epss.ahrq.gov/PDA/index.jsp o Enter the individual’s demographic information (age, sex, etc.) and look for

recommendations by level

https://list.nih.gov/

https://list.nih.gov/LISTSERV_WEB/USERSGDE/appendix_c.htm

https://public.govdelivery.com/accounts/USNIMH/subscriber/new

http://www.nlm.nih.gov/listserv/emaillists.html

http://epss.ahrq.gov/PDA/index.jsp



TYPES OF PREVENTIVE MEASURES: PRIMARY PREVENTIVE MEASURES: • provided to individuals to prevent the onset of a targeted condition. • Examples: routine immunization of healthy children, using aspirin to prevent the first heart

attack. SECONDARY PREVENTIVE MEASURES: • identify and treat asymptomatic persons who have already developed risk factors or preclinical

disease but in whom the condition has not become clinically apparent. • Examples: PAP smear to detect cervical dysplasia before the development of cancer,

screening for high blood pressure. TERTIARY PREVENTIVE MEASURES: • preventive measures that are part of the treatment and management of persons with clinical

illnesses. • Examples: cholesterol reduction in persons with ASHD, allergen avoidance in persons with

asthma. NOTE: • in casual conversation, people often use different terminology to describe clinical trials • Example:

o a “primary prevention” trial means that the individuals studied are healthy with no risk factors for the disease (combining the terms “primary” and “secondary” above)

o a “secondary prevention” trial means we are studying individuals who already have had the illness (e.g. a heart attack) – so the word “secondary” really is referring to “tertiary”



CHARTING FOR PREVENTIVE HEALTH

“HEALTH MAINTENANCE” (HM): this term is used for ADULTS • usually the LAST entry in a SOAP note • indicates your recommendations for ongoing diagnostics & interventions in preventive health

o example: for visit with 49 yo woman, here today to check her blood pressure, you also advise that starting next year (at age 50 yo) she will get her mammograms and CBE (clinical breast exams) annually

o example: for visit with 18 yo man, here today for college physical, you recommend that he receive Menactra (meningitis) vaccine and Td booster prior to college entry, sometime in the coming months before going to live in the dormitory at college; update and discuss closing any gaps in vaccine-preventable illnesses (has he had 2nd MMR? has he had full 3 series of Hepatitis vaccines? has he had Tuberculin test?); and you teach GSE (genital self examination)

“ANTICIPATORY GUIDANCE”: this term is used in PEDIATRICS • usually the LAST entry in a SOAP note • indicates your recommendations for ongoing diagnostics, interventions in preventive health,

AND education of expected developmental issues between NOW and the NEXT VISIT • example: for visit with 15-month old, vaccine preventable illnesses are discussed and first

MMR vaccine is administered; caregiver/parent is counseled on house-proofing for toddler, seat belt, smoking/firearms in the home, and any periodic health examinations needed at this age. IMPORTANTLY you ALSO discuss what you EXPECT to occur in terms of development and indicate that if the normal timeline is not met, child should return sooner than recommended; THEN you decide WHEN to bring the child back for re-examination, intervention.

• What guideline to follow? Multiple choices – American Academy of Pediatrics (Bright Futures), American Academy of Family Physicians, US Public Services Task Force ?

• For instance, AAP 2 yo visit includes the following checklist: ANTICIPATORY GUIDANCE FOR A 2 YEAR OLD: � Forward facing safety seat in the back of vehicle � Smoke free environment � Ensure water safety; empty tub, buckets � Remove or lock up poisons/toxic substances □ If guns in home, keep unloaded and locked up � Helmet use and supervise play � Oral care brush twice daily w/ fluoride � Begin toilet training if child is ready � Read, and play together � Help child express emotions � Offer variety of healthy foods/ avoid struggles � Establish simple rules/ reinforce limits � Praise good behavior Note: the diagnostic recommendation for this visit is hematocrit and PPD



FORMS FOR PEDIATRIC VISIT PREVENTIVE SERVICES BY AGE ARE AT:

• http://brightfutures.aap.org/index.html • and PRINTABLE table: https://www.aap.org/en-

us/Documents/periodicity_schedule_oral_health.pdf (if it asks for a login, just wait and the website will respond)

EPSDT – FORMS FOR PEDIATRIC VISITS: Early and Periodic Screening, Diagnostic, and Treatment (EPSDT) Services in Medicaid: • Childhood developmental and health/oral screenings are part of mandated Medicaid

services • In addition to Bright Futures, many states and other agencies have created their own

guidelines to support EPSDT providers: o Example – some of this stuff is hard to find so here is an example: Virginia (drop-down

box choose EPSDT manuals): https://www.virginiamedicaid.dmas.virginia.gov/wps/portal/ProviderManual

http://brightfutures.aap.org/index.html

https://www.aap.org/en-us/Documents/periodicity_schedule_oral_health.pdf

https://www.aap.org/en-us/Documents/periodicity_schedule_oral_health.pdf

https://www.virginiamedicaid.dmas.virginia.gov/wps/portal/ProviderManual