Journal of Critical Care (2008) 23, 227–235

Improving communication in a pediatric intensive careunit using daily patient goal sheetsSwati Agarwal MDa,⁎, Lorry Frankel MD, MBAa, Susan Tourner MDa,Alex McMillan PhDb, Paul J. Sharek MD, MPHc

aDivision of Pediatric Critical Care, Department of Pediatrics, Stanford University School of Medicine, Palo Alto, CA, USAbHealth Research and Policy, Biostatistics, Stanford University School of Medicine, Palo Alto, CA, USAcDivision of General Pediatrics, Department of Pediatrics, Stanford University School of Medicine, Palo Alto, CA, USA

C

0d

Keywords:Communication;Patient goal sheet;Patient safety;Patient care;Pediatric intensivecare units

AbstractPurpose: The aim of the study was to determine if a pediatric intensive care unit (PICU) daily patientgoal sheet would improve communication between health care providers and decrease length ofstay (LOS).Materials and Methods: We evaluated a daily patient goal sheet's impact on questionnaire-basedmeasures of effectiveness of communication, nurses' knowledge of physicians in charge, and on LOS inthe PICU.Results: Four hundred nineteen questionnaires were completed by nurses and physicians before goalsheet implementation and 387 after implementation. Nurses and physicians perceived an improvedunderstanding of patient care goals (P b .001), reported increased comfort in explaining patient caregoals to parents (P b .001), and listed a higher number of patient care goals after goal sheetimplementation (P b .01). Nurses identified the patient's attending physician and fellow withincreased accuracy after goal sheet implementation (P b .001). Median PICU LOS was unchanged;however, mean LOS trended toward a reduction after goal sheet implementation (4.1 vs 3.7 days,P = .36). Seventy-six percent of respondents found the goal sheets helpful.Conclusions: Using a PICU daily patient goal sheet can improve communication between health careproviders, help nurses identify the in-charge physicians, and be helpful for patient care. By explicitlydocumenting patient care goals, there is enhanced clarity of patient care plans between healthcare providers.© 2008 Elsevier Inc. All rights reserved.

1. Introduction sentinel event database of the Joint Commission on

Effective communication among health care providers isimperative when taking care of critically ill patients. The

⁎ Corresponding author. INOVA Fair fax Hospital for Children, Fallshurch, VA 22042. Tel.: +1 703 776 6558; fax: +1 703 776 3503.E-mail address: swati.agarwal@inova.org (S. Agarwal).

883-9441/$ – see front matter © 2008 Elsevier Inc. All rights reserved.oi:10.1016/j.jcrc.2007.07.001

Accreditation of Healthcare Organizations identifies com-munication as the number one root cause of reported sentinelevents [1]. Two recent studies in adult intensive care unit(ICU) populations, which studied the impact of a novel dailypatient goal sheet in adult ICU patients, documented betterunderstanding of patient care goals among nurses andresident physicians as well as decreased adult ICU length

228 S. Agarwal et al.

of stay (LOS) [2,3]. The importance of these results ishighlighted in the findings from an earlier study in an adultICU that showed improved communication resulted inimprovements in mortality rate, ICU LOS, and ICU costs [4].

To date, there are no published studies translating theadult-based ICU patient goal sheet to the pediatric intensivecare unit (PICU) setting. The PICU is a unique environmentdifferent from an adult ICU with a patient population that ismore diverse in diagnosis and age. We undertook this studyto determine whether a similar patient goal sheet in the PICUwould improve communication and render more efficientpatient care.

We hypothesized that implementing a PICU-specific dailypatient goal sheet would improve communication betweenhealth care providers and decrease PICU LOS by facilitatinggoal-directed patient care.

2. Materials and methods

2.1. Study design

Using a longitudinal pre-post study design, we aimed todetermine if a daily goal sheet (1) improved communicationin the PICU; (2) was perceived as helpful or useful by PICUresident physicians and nurses; and (3) reduced patient LOS.This study was approved by the Stanford University Schoolof Medicine (Palo Alto, Calif) Institutional Review Board;and written, informed consent was obtained from all nursesand resident physicians involved before the initiation ofthe study.

2.2. Location

Lucile Packard Children's Hospital is a 254-bed quatern-ary care children's hospital adjacent to the StanfordUniversity School of Medicine and Stanford Hospital. LucilePackard Children's Hospital is the pediatric and obstetricteaching hospital for the Stanford University School ofMedicine, with 68 PICU nurses, 59 pediatric residents, 5pediatric critical care fellows, and 7 PICU attendingphysicians. The PICU is a 12-bed unit that houses a diversemix of postsurgical patients and medical patients includingtransplant patients, trauma patients, and patients withmultiorgan system failure from a variety of causes. ThePICU attending physician is the medical care coordinator forall PICU patients working jointly with subspecialty services,where appropriate, to deliver optimal patient care. Patientsand staff from the separate 12-bed cardiovascular ICU werenot included in this study.

2.3. Intervention

During the 2-week preintervention study period(August 1-14, 2005), nurses and resident physicians in the

12-bed PICU completed a brief, anonymous questionnaire(Fig. 1) for each patient under their care daily at 10:00 AM

and 10:00 PM. This questionnaire aimed to assess theirunderstanding of the patient care goals of the day. A dailypatient goal sheet (intervention) was then implementedbeginning on August 15, 2005, for all patients in the PICU(Fig. 2). This goal sheet was co-created by 2 PICU fellowsand the PICUmedical director using the goal sheet developedby Pronovost et al [2] as a model. This goal sheet was thenmodified based on feedback from the physicians, nurses, andother members of the interdisciplinary team accumulatedfrom a brief pilot test in the PICU completed before thepreintervention period. The patient goal sheet was completedby the PICU resident, fellow, or attending physician duringmorning walk rounds each day. All efforts were taken tocreate a goal sheet that was easy and quick to complete. Allmembers of the interdisciplinary team caring for the patientswere encouraged to modify/add goals to the daily goal sheetas they saw fit. This goal sheet was not a part of the permanentmedical record and was discarded each morning after a newdaily patient goal sheet was completed. The goal sheet waskept on the nurse's worktable at each patient's bedside andwas available for viewing by parents and members of otherinterdisciplinary teams. A 16-week implementation windowwas initiated and consisted of in-services and face-to-facereminders to ensure goal sheets were completed for everyPICU patient on a daily basis. In the postintervention 2-weekstudy period (December 5-18, 2005), the surveying methodswere repeated. This postintervention questionnaire includedthe additional question: “Do you find the Daily Patient GoalSheets helpful for patient care? Yes or No.”

2.4. Study population

All nurses and resident physicians caring for patients inthe PICU between August 1 and 14, 2005 (preintervention),and December 5 and 18, 2005 (postintervention), wereeligible and participated. Median and mean PICU LOS usingmidnight census data were calculated for all patientsadmitted for a 4-month period before goal sheet implementa-tion (December 1, 2004, to March 31, 2005) and for theanalogous 4-month period (December 1, 2005, to March 31,2006) after goal sheet implementation. All patients admittedduring the preintervention period were discharged from thePICU before goal sheet implementation. All patientsadmitted during the postintervention period were subjectedto the intervention. The daily patient goal sheet has beenused from implementation (August 15, 2005) through thepresent day.

2.5. Study outcomes

The primary outcome measures were (1) understanding ofpatient care goals for the day (“How well do you understandthe goals of care/what work needs to be accomplished for this

Fig. 1 Preintervention questionnaire.

229Improving communication in a PICU

patient today?” using a 5-point Likert scale) and (2) PICULOS. Length of stay was defined as the total number of daysthat a patient was in the PICU at midnight. Secondaryoutcomes included (1) the comfort in explaining patient caregoals by nurses and resident physicians (“How comfortablewould you be explaining the goals of care for this patient for

today to the parents?” using a 5-point Likert scale); (2) thenumber of goals subjects were able to list for each patientunder their care (“Please list up to 3 goals of patient care toaccomplish today”); (3) nurses' knowledge of the attendingphysician and fellow responsible for the patient during theirshift (“Do you know who the primary attending for this

Fig. 2 Daily patient goal sheet (intervention).

230 S. Agarwal et al.

patient is right now?” and “Do you knowwho the fellow is forthis patient right now?” using the options “yes” or “no”); and(4) the helpfulness of the goal sheets to nurses and residentphysicians (“Do you find the Daily Patient Goal Sheet helpfulfor patient care?” using the options “yes” or “no”).

2.6. Statistical analysis

Statistical analysis was conducted using Stata 7.0 (StataCorp, College Station, Tex). The Wilcoxon 2-sample ranksum test (2-tailed) was used to compare data found to be

Table 1 Subject demographics before and after goal sheetintervention

Subject characteristic Preinterventionn (%)

Postinterventionn (%)

Total 419 387OccupationNurse 229 (55) 174 (45)Physician (PGY2 or 3resident)

187 (45) 213 (55)

No occupation listed 3 (b1) 0 (0)Time working in PICUNurseb1 y 46 (20) 46 (20)1-3 y 98 (43) 98 (43)3-5 y 33 (14) 33 (14)N5y 49 (21) 49 (21)No answer listed 3 (1) 3 (1)Physicianb 1 y 187 (100) 213 (100)

Morning rounds with physicians a

Morning nurse total 95 97Present on rounds 76 (80) 71 (73)Not present on rounds 19 (20) 26 (27)

PGY indicates postgraduate year.a Reported for nurses who indicated a response for 3 questions

regarding “occupation,” “time of day,” and “present on rounds” onquestionnaire.

231Improving communication in a PICU

not normally distributed: Likert scale responses, thenumber of goals respondents listed, and nurses' years ofPICU experience. Fisher exact test was used to comparenurses' presence on morning rounds. Length of stay wascompared using the log-rank test and distribution of LOSas a time to discharge is presented using Kaplan-Meiercurves. Mean LOS was calculated and evaluated by meansof a permutation test (LOS data are skewed making thet test inapplicable). Before extracting data, it was calcu-lated that a sampling period of approximately 4 monthswould provide 80% power to detect a 20% to 25%reduction in median LOS. This calculation, however, couldnot take into account the fact that LOS was recorded in

Table 2 Mean scores (±SD) for nurses' responses to questionnaire

Question

Understand goals of care/what work needs to be accomplished today?(answers based on 5-point Likert scale: 1 = do not understand at all5 = understand completely)

Comfort in explaining goals of care for today? (answers based on 5-poLikert scale: 1 = not at all comfortable to 5 = completely comfortabl

No. of goals listed (maximum = 3)Knows who attending is for patientKnows who fellow is for patient

complete days. A P value of less than .05 was determinedto be statistically significant.

3. Results

General demographic data for the nurse and residentphysician participants are provided in Table 1. Nurseexperience, as ascertained by number of years working inthe PICU, and nurse presence during morning rounds weresimilar during the pre- and postintervention study periods(P = .13 and P = .31, respectively) and their survey answersdid not differ significantly based on years of PICUexperience. All resident physicians were second- and third-year pediatric residents or third-year emergency departmentresidents during the pre- and postintervention study periods.During the 2-week preintervention survey period, a total of419 questionnaires (overall return rate of 68%) werereturned: 229 from PICU nurses, 187 from residents, and 3with no occupation listed. During the 2-week postinterven-tion survey period, a total of 387 questionnaires (overallreturn rate of 65%) were returned: 174 from nurses and 213from residents. Nurses and residents both exhibitedimproved understanding of patient care goals, more comfortin their ability to explain patient care goals to parents, and theknowledge to list more patient care goals after the goal sheetswere implemented (see Tables 2 and 3). In addition, nurseswere significantly better able to identify the PICU attendingphysician and fellow responsible for the patient during theirshift after implementation of the goal sheet (Table 2).Seventy-six percent of all subjects, 87% of nurses, and 67%of resident physicians reported finding the goal sheets usefulfor patient care. The median patient LOS in the PICU was2.0 days both before (n = 299 patients) and after (n =342 patients) goal sheet implementation. The mean LOS was4.1 days preintervention and 3.7 days postintervention. Acomparison of the means based on 10000 random permuta-tions of the LOS data between the 2 periods is notstatistically significant (P = .36). The distribution of LOSas a time to discharge is shown in the Kaplan-Meier plot,which shows the fraction of patients still in the PICU at each

Preintervention(n = 229)

Postintervention(n = 174)

P

to4.2 ± 0.8 4.5 ± 0.6 b.001

inte)

4.1 ± 0.9 4.3 ± 0.8 .001

2.6 ± 0.7 2.8 ± 0.6 b.0175% ± 0.4% 92% ± 0.3% b.00179% ± 0.4% 93% ± 0.3% b.001

Table 3 Mean scores (±SD) for resident physicians' responses to questionnaire

Question Preintervention(n = 187)

Postintervention(n = 213)

P

Understand goals of care/what work needs to be accomplished today? (answers basedon 5-point Likert scale: 1 = do not understand at all to 5 = understand completely)

4.0 ± 0.6 4.7 ± 0.5 b.001

Comfort in explaining goals of care for today? (answers based on 5-point Likert scale:1 = not at all comfortable to 5 = completely comfortable)

3.9 ± 0.7 4.7 ± 0.5 b.001

No. of goals listed (maximum = 3) 2.0 ± 0.8 2.8 ± 0.5 b.001

232 S. Agarwal et al.

day during the 2 study periods (P = .71) (Fig. 3). This showsvery close correspondence between the 2 curves, with somepossible shortening in LOS between days 10 and 20 in thepostintervention group.

4. Discussion

This is the first pediatric study to describe the effect of adaily patient goal sheet on communication and LOS in aPICU. Literature in the professional and lay press highlightsthe frequency and seriousness of medical error, which occurswhen caring for children [5,6] and in critical care units [7,8].For example, in the Institute of Medicine report “To Err isHuman,” it was estimated that between 44000 and 98000

Fig. 3 Kaplan-Meier curves of the distribution of length

people die in US hospitals as a direct result of medical error[5]. As a result, patient safety has become a major area offocus for quality improvement for hospitals, regulators,insurance companies, and patients. The role of communica-tion in these deaths is not completely understood at present;however, it is clear that poor communication is a risk factoras revealed by a 2005 Joint Commission on the Accreditationof Healthcare Organizations report in which almost 70% ofsentinel events identified communication as a root cause [1].Accordingly, one of the 2006 National Patient Safety Goalsfrom the Joint Commission on the Accreditation ofHealthcare Organizations aimed to “improve the effective-ness of communication among caregivers [9].” Although therecommendation to enhance patient safety by improvingeffective communication is logical, this regulatory mandate

of stay in the pre- and postintervention time frames.

233Improving communication in a PICU

has very few evidence based strategies associated with it.Two previous adult-based studies [2,3], and now thispediatric-focused study, provide an effective strategy toenhance patient relevant communication.

In this study, we found that implementing a daily patientgoal sheet in the PICU improves communication betweenhealth care providers. Before goal sheet implementation,scores for understanding goals at the Lucile PackardChildren's Hospital were 4.2 for nurses and 4.0 for residentphysicians. These data are similar to the findings of theNarasimhan study (3.9 for nurses and 4.6 for physicians) [3]but higher than the baseline understanding of goals in thePronovost study, which described less than 10% of nursesand resident physicians scoring a 4 or 5 for understandinggoals [2]. After goal sheet implementation, our scoresincreased to 4.5 for nurses and 4.7 for resident physicians,again similar to Narasimhan's findings (4.8 for nurses and4.9 for physicians) [3], whereas Pronovost describes morethan 95% of nurses and resident physicians scoring a 4 or 5[2]. Each of these 3 studies identified improvement incommunication between health care providers with the use ofdaily patient goal sheets [2,3].

Interestingly, our LOS data did not show the samereduction in LOS found by Narasimhan et al [3] andPronovost et al [2] who described a decreased mean LOS of2 days and 1 day, respectively. Our mean LOS showed anonsignificant trend toward a reduction in LOS (4.1 days[pre] vs 3.7 days [post]), whereas our median LOS was2.0 days both before and after goal sheet implementation.Multiple explanations could account for this difference,with a few of the more likely described here. First, wecaptured the LOS data as whole days using the midnightcensus. This data capture strategy, used by our specificbilling software, may not have been granular enough toidentify significant shifts in LOS given the relatively shortlengths of stay in our PICU. Second, we did not evaluatethe “severity of illness” in the pre- and poststudy timeframes. It is conceivable that the postintervention patientshad a higher severity of illness that could have biased theresults inappropriately toward the mean. Finally, PICU LOSis related to both medical and nonmedical factors, and it ispossible that several nonmedical factors, including nursingavailability and bed availability on the medical-surgicalwards, affected the LOS data. Each of these explanationscould bias the findings in either direction and were not ableto be factored into the final LOS analysis.

The improved communication revealed after patient goalsheet implementation is reflected in improved understandingof patient care goals, increased comfort in explaining thegoals of care to families, and an increased listing of absolutenumber of goals. Responses of “mostly understand” or“mostly comfortable” improved to “understand completely”or “completely comfortable” after patient goal sheet imple-mentation. We believe that this increased understanding andhigher sense of comfort for caretakers optimize patient careand thus should be the standard to which all PICUs strive.

This conclusion is consistent with the finding that both nursesand residents were able to list more goals of patient care in thepostintervention period. Each of the improvements incommunication identified in the adult ICU literature werereplicated in this study of the PICU population; furthermore,we have expanded the literature to include evidence ofimproved “comfort with explanation of goals of care tofamilies” and increased number of goals listed to furthersolidify the importance of this communication tool.

Clinically, these surrogates of communication couldtranslate directly to enhanced patient safety and quality ofcare as numerous studies have highlighted the importanceof communication in health care and how it positivelyimpacts patient care outcomes [4,10-13]. The care of acritically ill child often requires numerous diagnostic tests,procedures, medications, and other interventions. Often,tasks are ordered simultaneously and priority must be givento which aspects of patient care need to be attended to first.Nurses and residents, who are the primary caregivers topatients, improve their understanding of patient care goalsusing a daily goal sheet and likely deliver more effective,goal-directed care. In fact, a recent commentary entitled“Safety in the pediatric ICU: the key to quality outcomes”and the Institute for Healthcare Improvement both recom-mend goal sheets as a way to foster team cohesiveness,articulate patient care goals, improve communication, andthereby promote safety in the ICU setting [11,14].

One unique and important aspect of this study, notpreviously evaluated in the adult studies, was determinationof nurses' ability to identify the attending physician andfellow for their shift in the PICU. In an academic institutionwith multiple care providers, including medical students,pediatric residents, emergency medicine residents, PICUfellows, PICU attending physicians, and numerous consult-ing services, it is important for nurses to be able to identifywho the fellow and attending physician are if questions arise.Accountability and communication are paramount in the careof critically ill children as evidenced by the news report ofthe 2003 Children's Hospital Boston case where a 5-year-oldboy with epilepsy died after apparent uncertainty over who(and what service) was in charge of the patient [15]. In ourstudy, nurses were better able to identify both the in-chargefellow and attending physician after implementation of thegoal sheets.

We ascertained the subjective usefulness and practicalityof a daily patient goal sheet after its implementation byasking respondents, “Do you find the Daily Patient GoalSheet helpful for patient care?” Nurses overwhelminglyfound the goal sheets helpful (87%). Numerous commented“If sheet completely filled out, it is very helpful—especiallyto know who the physicians/residents are for the patient, andif I am not able to be present for rounds then I still know theplan” and “helpful especially because we frequently getcalled away (during rounds).” In addition, 67% of residentsfound the goal sheets helpful. This was interesting as theywere often the ones burdened to fill out the goal sheet on

234 S. Agarwal et al.

rounds (ie, doing the “extra” work). Whereas we did notdetermine the mean length of time to fill out the goal sheet inour study, Narasimhan et al [3] report a mean length of timeof 1 minute to fill out their goal sheet. Realizing that healthcare providers are often overburdened with documentation[16], one of the aims in designing the PICU goal sheet was tomake it easy to complete.

There are several limitations in this study. First of all,given the nonindependence of data (duplication of respon-dents), there was a risk for overstating the precision of theresults. Our P values depend on the assumption ofindependence, which we cannot verify because the anonym-ity (anonymous questionnaire collection) prevented us frommatching questionnaires to individuals. This resulted in aninability to control for clustering. Second, the possibility ofvolunteer bias existed. Subjects, for example, who did (or didnot) find the goal sheet useful, may not have returned thequestionnaire. This effect would bias the results in anunknown direction. Third, it is possible that some patients inthe postintervention time frame did not have the goal sheetscompleted every day, which would bias the results in anunknown direction. We attempted to minimize the risks ofincomplete implementation by providing a 16-week learningphase before the postintervention data collection. Fourth, thisstudy was undertaken in a high-acuity academic teachinghospital PICU, and thus, these results may have limitedapplication in a community hospital or lower-acuity PICUswhere nurse-physician relationship and PICU roundingroutine are different. Fifth, the primary outcome measurein this study reflects self-reported data about nurses' andresidents' perceived understanding of patient care goals. Thisallows for a potential difference in perceived and actualknowledge, which is a common limitation of qualitativeresearch. Finally, this study suffers from the traditional biasesof a pre-post study design, specifically that we were unableto control for all potential confounding influences on theoutcomes. Despite these possible biases, we believe thatimplementation of patient-specific goal sheets indeedstandardizes and facilitates communication between healthcare providers in a high-risk, high-acuity PICU environment.The fact that our data are similar to 2 prior studies in differentunits with different health care professionals and patientpopulations is consistent with this conclusion. Although thisstudy did not set out to determine if this communication toolimproved patient outcomes, we suspect that standardized andenhanced communication strategies, such as a daily patientgoal sheet and Situation-Background-Assessment-Recom-mendation technique, has this effect [17-20].

5. Conclusion

We have shown that the use of a daily patient goal sheetin an academic PICU was associated with improvedcommunication between health care providers and was

found to be useful by both nurses and resident physicians. Inaddition to confirming the findings of previous adult goalsheet studies, we have expanded the literature to includeevidence of improved “comfort with explanation of goals ofcare to families,” increased number of goals listed, andimportantly, that nurses were better able to identify the in-charge fellows and attending physicians after goal sheetimplementation. A trend toward decreased mean LOS wasalso found after goal sheet implementation. Although goalsheets improve perceived communication of patient caregoals in this study, PICUs should use caution as they are inno way a substitution for verbal communication. Patient careteams should make efforts to have all caregivers present forrounds and address the questions/issues for each disciplineof care together in a multidisciplinary collaboration [21]. Inaddition, we encourage PICUs (and other pediatric hospitalsettings) to design goal sheets individualized to the care ofthe patient population in their units to standardize commu-nication, help all members of the health care team improvetheir understanding of patient care plans, and help allmembers of the health care team deliver more goal-directedcare for these high-risk patients. Future research shouldevaluate the effect of patient goal sheets on clinicaloutcomes including adverse drug events, and on patientand parental satisfaction.

Acknowledgments

We thank Linda Hargreaves, a registered nurse who iscertified in PICU by the American Association of Critical-Care Nurses and who is the PICU Research Representativefor Nursing Shared Leadership at the Lucile PackardChildren's Hospital, for her contributions and support duringthis research study.

References

[1] Joint Commission on Accreditation of Healthcare Organizations.Sentinel event statistics March 31, 2006. http://www.jointcommission.org/SentinelEvents/Statistics/ [last accessed August 24, 2006].

[2] Pronovost P, Berenholtz S, Dorman T, et al. Improving communicationin the ICU using daily goals. J Crit Care 2003;18:71-5.

[3] Narasimhan M, Eisen LA, Mahoney CD, et al. Improving nurse-physician communication and satisfaction in the intensive care unitwith a daily goals worksheet. Am J Crit Care 2006;15:217-22.

[4] Zimmerman JE, Shortell SM, Rousseau DM, et al. Improving intensivecare: observations based on organizational case studies in nineintensive care units: a prospective, multicenter study. Crit Care Med1993;21:1443-51.

[5] Kohn L, Corrigan J, Donaldson M, editors. To err is human: building asafer health system. Washington (DC): National Academy Press; 2000.

[6] Miller MR, Zhan C. Pediatric patient safety in hospitals: a nationalpicture in 2000. Pediatrics 2004;113:1741-6.

[7] Stockwell DC, Slonim AD. Quality and safety in the intensive careunit. J Intensive Care Med 2006;21:199-210.

[8] Sharek PJ, Horbar JD, Mason W, et al. Adverse events in the neonatalintensive care unit: development, testing, and findings of an NICU-

235Improving communication in a PICU

focused trigger tool to identify harm in North American NICUs.Pediatrics 2006;118:1332-40.

[9] Joint Commission on Accreditation of Healthcare Organizations. 2006critical access hospital and hospital national patient safety goals. http://www.jointcommission.org/PatientSafety/NationalPatientSafetyGoals/06_npsg_cah.htm [last accessed August 22, 2007].

[10] Baggs JG, Schmitt MH, Mushlin AI, et al. Association between nurse-physician collaboration and patient outcomes in three intensive careunits. Crit Care Med 1999;27:1991-8.

[11] Rice BA, Nelson C. Safety in the pediatric ICU: the key to qualityoutcomes. Crit Care Nurs Clin North Am 17 (2005) 431-40, xi

[12] Paying lip service to patient safety: better communication, improvedwork environment seen as lowering medical errors. Qual Lett HealthcLead 17 (2005) 2-10, 1

[13] Schmitt MH. Collaboration improves the quality of care: methodolo-gical challenges and evidence from US health care research. J InterprofCare 2001;15:47-66.

[14] Institute for Healthcare Improvement. ICU daily goals work-sheet. http://www.ihi.org/IHI/Topics/CriticalCare/IntensiveCare/

Tools/ICUDailyGoalsWorksheet.htm [last accessed March 4,2007.].

[15] Barnard A. Doctors were unsure of roles as boy died at Children's. TheBoston Globe, September 19, 2003.

[16] Excessive paperwork detracts from patient care, professional mentor-ing, and research. Oncology (Williston Park) 2001;15:834.

[17] Haig KM, Sutton S, Whittington J. SBAR: a shared mental model forimproving communication between clinicians. Jt Comm J Qual PatientSaf 2006;32:167-75.

[18] Hohenhaus S, Powell S, Hohenhaus JT. Enhancing patient safetyduring hand-offs: standardized communication and teamwork usingthe ‘SBAR’ method. Am J Nurs 2006;106:72A-B.

[19] SBAR initiative to improve staff communication. Healthcare Bench-marks Qual Improv 2005;12:40-1.

[20] JCAHO. The SBAR technique: improves communication, enhancespatient safety. Jt Comm Perspect Patient Saf 2005:5.

[21] Uhlig PN, Brown J, Nason AK, et al. John M. Eisenberg patient safetyawards. system innovation: concord hospital. Jt Comm J Qual Improv2002;28:666-72.