Exploring Entrustment: Housestaff Autonomy and Patient Readmission
2014; Elsevier BV; Volume: 127; Issue: 8 Linguagem: Inglês
10.1016/j.amjmed.2014.04.013
ISSN1555-7162
AutoresShannon K. Martin, Jeanne M. Farnan, Andrea Flores, Lianne M. Kurina, David O. Meltzer, Vineet M. Arora,
Tópico(s)Healthcare Quality and Management
ResumoPerspectives Viewpoints•Attending physicians must decide when trainees can be entrusted to independently perform professional activities, although clinical implications of entrustment are unknown.•In this study of general medicine patients over an 8-year period, entrustment of autonomy was associated with decreased odds of 30-day hospital readmission and increased odds of transfer to the intensive care unit.•Entrustment appears to be linked to patient outcomes, although these associations were seen only for experienced senior residents.The model of progressive independence in graduate medical education balances oversight by clinical supervisors with progressive autonomy and independence for residents; it promotes development of clinical expertise while protecting patient safety.1Kennedy T.J. Regehr G. Baker G.R. Lingard L.A. Progressive independence in clinical training: a tradition worth defending?.Acad Med. 2005; 80: S106-S111Crossref PubMed Scopus (150) Google Scholar, 2Kilminster S.M. Jolly B.C. Effective supervision in clinical practice settings: a literature review.Med Educ. 2000; 34: 827-840Crossref PubMed Scopus (499) Google Scholar, 3Ericsson K.A. Deliberate practice and acquisition of expert performance: a general overview.Acad Emerg Med. 2008; 15: 988-994Crossref PubMed Scopus (921) Google Scholar Attending physicians serving as clinical supervisors must determine when residents can be trusted to independently perform a professional activity or what level of supervision they require.4Ten Cate O. Entrustability of professional activities and competency-based training.Med Educ. 2005; 39: 1176-1177Crossref PubMed Scopus (524) Google ScholarThe Accreditation Council for Graduate Medical Education Next Accreditation System has introduced entrustable professional activities (EPAs), specialty-specific patient care activities that learners must be "entrusted" to perform at increasing levels of autonomy and decreasing levels of supervision.4Ten Cate O. Entrustability of professional activities and competency-based training.Med Educ. 2005; 39: 1176-1177Crossref PubMed Scopus (524) Google Scholar, 5Nasca T.J. Philibert I. Brigham T. et al.The next GME accreditation system — rationale and benefits.N Engl J Med. 2012; 366: 1051-1056Crossref PubMed Scopus (1035) Google Scholar EPAs are benchmarks of readiness to progress through training and require attending physicians to make deliberate decisions about when to entrust residents to perform them.6Ten Cate O. Scheele F. Competency-based postgraduate training: can we bridge the gap between theory and clinical practice?.Acad Med. 2007; 82: 542-547Crossref PubMed Scopus (667) Google Scholar, 7Sterkenburg A. Barach P. Kalkman C. et al.When do supervising physicians decide to entrust residents with unsupervised tasks?.Acad Med. 2010; 85: 1399-1400Crossref PubMed Scopus (181) Google Scholar Entrustment has been found to be a complex and often implicit act that varies for different activities and is influenced by multiple factors, including attending physician characteristics, academic season, and resident characteristics.7Sterkenburg A. Barach P. Kalkman C. et al.When do supervising physicians decide to entrust residents with unsupervised tasks?.Acad Med. 2010; 85: 1399-1400Crossref PubMed Scopus (181) Google Scholar, 8Martin S.K. Farnan J.M. Mayo A. et al.How do attendings perceive housestaff autonomy? Attending experience, hospitalists and trends over time.J Hosp Med. 2013; 8: 292-297Crossref PubMed Scopus (10) Google Scholar, 9Kennedy T.J. Reghr G. Baker G.R. Lingard L. Point-of-care assessment of medical trainee competence for independent clinical work.Acad Med. 2008; 83: S89-S92Crossref PubMed Google Scholar However, to our knowledge, no studies to date have examined whether the decision to entrust autonomy to house staff may be linked with patient outcomes.One outcome of interest is readmission to the hospital or emergency department (ED) within 30 days, a costly and potentially dangerous patient outcome.10Jencks S.F. Williams M.V. Coleman E.A. Rehospitalizations among patients in the Medicare fee-for-service program.N Engl J Med. 2009; 360: 1418-1428Crossref PubMed Scopus (3826) Google Scholar, 11Forster A.J. Murff H.J. Peterson J.F. et al.The incidence and severity of adverse events affecting patients after discharge from the hospital.Ann Intern Med. 2003; 138: 161-167Crossref PubMed Scopus (1318) Google Scholar, 12Kripalani S. LeFevre F. Phillips C.O. et al.Deficits in communication and information transfer between hospital-based and primary care physicians: implications for patient safety and continuity of care.JAMA. 2007; 297: 831-841Crossref PubMed Scopus (1411) Google Scholar One could hypothesize that patients whose supervising attending physicians entrust autonomy to house staff may have lower rates of readmission, reflecting a higher quality of care. Alternatively, entrustment could diminish attending involvement and patients may actually fare worse. In this study, we tested whether attending physician perceptions of entrustment of autonomy to house staff during inpatient rotations were associated with odds of 30-day readmission for patients under that team's care, and specifically, whether it differed by resident experience. Two additional patient outcomes, length of stay and transfer to the intensive care unit (ICU), were examined as secondary outcomes.MethodsStudy DesignWe conducted a retrospective analysis of data obtained from the University of Chicago Hospitalist Project, a study of the care of hospitalized patients admitted to the general medicine service at the University of Chicago.13Meltzer D. Manning W.G. Morrison J. et al.Effects of physician experience on costs and outcomes on an academic general medicine service: results of a trial of hospitalists.Ann Intern Med. 2002; 137: 866-874Crossref PubMed Scopus (272) Google Scholar The study also surveys general medicine attending physicians about their experiences as ward attendings. This study was approved by the University of Chicago Institutional Review Board.Data CollectionAttendings and patients (or their proxies) consented to participate. Trained research assistants collected detailed patient sociodemographic and health information during a 30-minute intake inpatient interview, as well as from administrative records and chart review. Patients or proxies were also interviewed by telephone 30 days after discharge. Attending physicians received a 40-item, paper-based survey at the end of each general medicine rotation.Survey month and year were recorded along with information about house staff team members (sex and postgraduate year). Teams consist of one attending physician, one team resident, and one or 2 postgraduate year 1 (PGY-1) residents. Team residents may be in their second, third, or fourth postgraduate year; team management and supervisory responsibilities are irrespective of the PGY year. PGY-2 residents are in their first year of serving as a senior team resident, whereas PGY-3 and PGY-4 residents have had prior experience.Patient SampleWe retrospectively selected a subset of enrolled patients for the analysis (Figure). Patients admitted from June 1, 2001 to June 30, 2009 were considered for inclusion. The final sample included 11,351 patients; however, 4069 patients did not complete the 30-day follow-up interview and 170 were excluded to reduce bias from outlier effects. Missing values for predictor or outcome variables were not imputed, and because of missing data, only 9524 patients (61.1% of those enrolled) were included in the regression analyses.Outcome VariablesThe primary outcome was readmission to the hospital or ED visit within 30 days of discharge, obtained via self-report from the telephone follow-up interview. Secondary outcomes were patient length of stay and transfer to the ICU. Length of stay was defined as number of inpatient days and was determined from chart review and administrative data. ICU transfer was defined as transfer to the ICU during hospitalization and was abstracted from chart review. Readmission was the primary outcome because it is considered a negative patient outcome.11Forster A.J. Murff H.J. Peterson J.F. et al.The incidence and severity of adverse events affecting patients after discharge from the hospital.Ann Intern Med. 2003; 138: 161-167Crossref PubMed Scopus (1318) Google ScholarPredictor VariablesTo measure entrustment of autonomy, 2 statements were selected from the monthly attending survey: "My resident felt that s/he had sufficient autonomy this month" and "The intern(s) were truly involved in decision-making about their patients." These items have been used in previous work studying attending-resident dynamics and autonomy.8Martin S.K. Farnan J.M. Mayo A. et al.How do attendings perceive housestaff autonomy? Attending experience, hospitalists and trends over time.J Hosp Med. 2013; 8: 292-297Crossref PubMed Scopus (10) Google Scholar, 14Chung P. Morrison J. Jin L. et al.Resident satisfaction on an academic hospitalist service: time to teach.Am J Med. 2002; 112: 597-601Abstract Full Text Full Text PDF PubMed Scopus (43) Google Scholar, 15Arora V.M. Georgitis E. Siddique J. et al.Association of workload of on-call interns with on-call sleep duration, shift duration, and participation in educational activities.JAMA. 2008; 300: 1146-1153Crossref PubMed Scopus (83) Google Scholar These questions had Likert-type response options. Because the distributions of responses were skewed toward strong agreement, scores were collapsed into 2 categories: strongly agree versus do not strongly agree (ie, all other responses).8Martin S.K. Farnan J.M. Mayo A. et al.How do attendings perceive housestaff autonomy? Attending experience, hospitalists and trends over time.J Hosp Med. 2013; 8: 292-297Crossref PubMed Scopus (10) Google Scholar, 14Chung P. Morrison J. Jin L. et al.Resident satisfaction on an academic hospitalist service: time to teach.Am J Med. 2002; 112: 597-601Abstract Full Text Full Text PDF PubMed Scopus (43) Google Scholar Entrustment of resident autonomy was defined as a response of "strongly agree."Statistical AnalysisDescriptive statistics were used to examine baseline patient demographic and health characteristics as well as characteristics of attending and resident physicians. Pearson's chi-squared was used to compare proportions, and t tests were used to compare continuous variables.To explore the relationship between entrustment of autonomy and readmission, we used conditional logistic regression models that accounted for within-group variation by clustering on attending physicians. We also controlled for patient sociodemographic and health data, attending and resident factors, and secular trends (academic season and implementation of resident duty-hours restrictions in 2003). Separate models were constructed for perception of team resident autonomy and perception of PGY-1 resident involvement in decision-making. Finally, analyses were stratified by resident PGY levels as we hypothesized that entrustment decisions might vary based on resident experience.7Sterkenburg A. Barach P. Kalkman C. et al.When do supervising physicians decide to entrust residents with unsupervised tasks?.Acad Med. 2010; 85: 1399-1400Crossref PubMed Scopus (181) Google ScholarFor secondary outcomes, we used multivariate logistic regression for ICU transfer; a lack of within-attending variation precluded the use of conditional logistic regression. Length of stay was transformed on the logarithmic scale to achieve a more normal distribution, and clustered linear regression, clustering on attending physicians, was used for analysis. Analyses were done using Stata, version 11.2 (StataCorp, College Station, Tex).ResultsCharacteristics of Patients and Physicians Included in AnalysisFrom 2001 to 2009, 15,590 patients were enrolled in the Hospitalist Project and 9524 patients (61.1%) were included in the analysis. For attendings, 1103 monthly surveys were distributed to 120 unique physicians and 67% (n = 734) were returned. There were 274 unique team residents.Table 1 describes patients included in the analysis. Most patients were African-American (79.8%) and female (63.7%), with a mean age of 57.6 years (SD, 19.3). Nearly 90% of patients were discharged to home and 10.8% were discharged to another facility. Table 2 describes characteristics of attending and resident physicians in the study. Nearly 40% of attendings were female and 22.6% were hospitalists. Over 50% of resident physicians were PGY-2s, 42.0% PGY-3s, and 3.1% PGY-4s.Table 1Characteristics of Patients Participating in 1-Month Follow-Up Survey (n = 9524)∗Because of missing data, numbers may not correspond to exact percentages.CharacteristicValueAge (years) Mean (SD)57.6 (19.3) Median [IQR]59 (43-73) Female, n (%)6067 (63.7)Race, n (%) African-American7597 (79.8) White1601 (16.8) Asian/Pacific Islander76 (0.8) Hispanic167 (1.75) Other/unknown83 (0.85)Education, n (%) Less than high school graduate1930 (24.1) High school graduate2419 (30.2) Greater than high school graduate3653 (45.7)Insurance, n (%) Medicare4811 (51.4) Medicaid2156 (23.0) Private/grants2080 (22.2) No payor313 (3.34)Marital status, n (%) Married/living as if married2447 (29.7) Divorced/separated1444 (17.5) Widowed1764 (21.4) Single2593 (31.4)Charlson Comorbidity Index score Mean (SD)3.05 (2.24) Median [IQR]3 (1-5)Hospitalized in previous 12 months (%)4209 (51.3)Has a primary care provider (%)6733 (80.6)Discharge status Home8362 (88.5) Other facility (eg, skilled nursing facility, acute care hospital, rehabilitation, inpatient hospice unit)1023 (10.8) Left against medical advice55 (0.58) Died8 (0.08)IQR = interquartile range.∗ Because of missing data, numbers may not correspond to exact percentages. Open table in a new tab Table 2Characteristics of Physicians Caring for Study Participants∗Because of missing data, numbers may not correspond to exact percentages.CharacteristicValueAttending physicians (n = 120 unique physicians, 9524 physician-patient encounters) Female (%)39.9 Hospitalist (%)†Ascertained by survey question, "Do you consider yourself to be a hospitalist?"22.6 Experience (years since residency)Mean (SD)9.8 (8.1)Median [IQR]7 (3-15)0-4 years (%)33.85-11 years (%)32.7>11 years (%)33.6Resident physicians (n = 274 unique physicians, 9524 physician-patient encounters) Female (%)52.3 Postgraduate year (%)PGY-254.9PGY-342.0PGY-43.1 Months of experience as senior residentMean (SD)12.1 (7.2)Median [IQR]12 (6-18)IQR = interquartile range; PGY = postgraduate year.∗ Because of missing data, numbers may not correspond to exact percentages.† Ascertained by survey question, "Do you consider yourself to be a hospitalist?" Open table in a new tab Hospital readmission or ED visit within 30 days was reported by 34.7% of patients in the analysis. Mean length of stay was 4.36 days (SD 4.1), and 2.3% of patients had an ICU transfer.Compared with the patients included in the analysis, the 4069 patients who did not complete the follow-up interview were more likely to be male (40.5% vs 36.8%, P <.001), younger (mean age 55 vs 58.5 years, P <.001), and African-American (83.9% vs 79.6%, P <.001). They were also more likely to have been hospitalized in the last year (56.8% vs 52.4%, P <.001) and less likely to identify a primary care provider (72.8% vs 80.6%, P <.001). These patients were more often discharged to a facility other than home (14.8% vs 12.3%, P <.001). Patients not completing the follow-up interview had no difference in mean length of stay or ICU transfer.Attending Physician Ratings of EntrustmentAmong attending physicians, 50.2% entrusted autonomy to their residents and 43.9% perceived adequate PGY-1 involvement in decision-making. Team resident experience was associated with differences in entrustment, although not in a linear fashion, as 50.4% of attendings entrusted autonomy to their PGY-2 team residents, compared with 49.2% and 75.1% of attendings with PGY-3 and PGY-4 residents, respectively (χ2 = 79.8, P <.01). Similarly, 42% of attendings with a PGY-2 team resident perceived strong intern involvement in decision-making, compared with 45.1% and 56.9% of attendings with PGY-3 and PGY-4 residents, respectively (χ2 = 79.8, P <.01).Association of Entrustment of Autonomy with Readmission EventsIn the univariate analysis, patients cared for by attendings entrusting autonomy to residents had no significant difference in 30-day readmission compared with patients of attendings not entrusting autonomy (34.0% vs 35.3%, χ2 =1.93, P = .165). Patients of attendings perceiving high involvement of PGY-1s in decision-making also had rates of readmission similar to patients whose attendings did not perceive high involvement (34.4% vs 34.8, χ2 =1.93, P = .696).In multivariate analysis including only main effects, there remained no significant association between odds of readmission and entrustment of autonomy (Table 3). However, when stratifying on resident experience, there was a differential effect of entrustment by team resident seniority. In adjusted analysis, patients on teams with PGY-3 or PGY-4 residents had 21% lower odds of readmission when the attending entrusted autonomy to the team resident (odds ratio [OR] 0.79; 95% confidence interval [CI], 0.65-0.96, P = .018). Patients on teams with PGY-2 team residents had no difference in odds of readmission with entrustment (OR 0.99; 95% CI, 0.82-1.18, P = .873).Table 3Associations Between Odds of 30-day Patient Readmission Event and Attending Entrustment of Autonomy∗Multivariate conditional logistic regression models clustered on attending physicians and adjusted for possible confounding factors. Two separate models were constructed for each measure of entrustment. "Strongly agree" with entrustment measures was defined as a response of "Strongly Agree" or 5 on the Likert scale to each statement. Thirty-day readmission events were defined as a patient-reported rehospitalization or visit to the emergency department within 30 days of discharge. The models were adjusted for the following factors: patient factors: age, sex, race, education, marital status, insurance payor, Charlson Comorbidity Index, hospitalization within 12 months, has a primary care provider, discharge status; attending factors: hospitalist status, years of experience since residency, comfort with inpatient medicine; resident factors: sex, months of experience; and secular factors: 2003 duty hours limits, academic season.Entrustment MeasureOR of Readmission95% CIP ValueMy resident had sufficient autonomyUnstratifiedStrongly agree0.890.77-1.04.141Stratified on resident experience†P-value for interaction term = .052.PGY2Strongly agree0.990.82-1.18.873PGY ¾Strongly agree0.790.65-0.96.018The interns were truly involved in decision-making about their patientsUnstratifiedStrongly agree0.960.82-1.13.650Stratified on resident experience‡P-value for interaction term = .093.PGY2Strongly agree1.050.87-1.28.601PGY 3/4Strongly agree0.870.71-1.06.174CI = confidence interval; OR = odds ratio; PGY = postgraduate year.∗ Multivariate conditional logistic regression models clustered on attending physicians and adjusted for possible confounding factors. Two separate models were constructed for each measure of entrustment. "Strongly agree" with entrustment measures was defined as a response of "Strongly Agree" or 5 on the Likert scale to each statement. Thirty-day readmission events were defined as a patient-reported rehospitalization or visit to the emergency department within 30 days of discharge. The models were adjusted for the following factors: patient factors: age, sex, race, education, marital status, insurance payor, Charlson Comorbidity Index, hospitalization within 12 months, has a primary care provider, discharge status; attending factors: hospitalist status, years of experience since residency, comfort with inpatient medicine; resident factors: sex, months of experience; and secular factors: 2003 duty hours limits, academic season.† P-value for interaction term = .052.‡ P-value for interaction term = .093. Open table in a new tab Secondary OutcomesThere was no significant association between odds of ICU transfer and entrustment in the univariate or multivariate analyses (Table 4). However, when stratifying by resident experience, patients on teams with senior residents had higher odds of ICU transfer when the attending entrusted autonomy to the team resident (OR 1.68; 95% CI, 1.02-2.79, P = .043). For patients on junior resident teams, there was no difference in odds of ICU transfer with entrustment (OR 0.85; 95% CI, 0.56-1.30, P = .460).Table 4Associations Between Odds of ICU Transfer and Attending Entrustment of Autonomy∗Multivariate conditional logistic regression models clustered on attending physicians and adjusted for possible confounding factors. Two separate models were constructed for each measure of entrustment. "Strongly agree" with entrustment measures was defined as a responsive of "Strongly Agree" or 5 on the Likert scale to each statement. ICU transfer was defined as transfer to any ICU during hospitalization. The models were adjusted for the following factors: patient factors: age, sex, race, education, marital status, insurance payor, Charlson Comorbidity Index, hospitalization within 12 months, has a primary care provider, discharge status; attending factors: hospitalist status, years of experience since residency, comfort with inpatient medicine; resident factors: sex, months of experience; and secular factors: 2003 duty hours limits, academic season.Entrustment MeasureOR of ICU Transfer95% CIP ValueMy resident had sufficient autonomyUnstratifiedStrongly agree1.130.81-1.57.462Stratified on resident experience†P-value for interaction term = .038.PGY2Strongly agree0.850.56-1.30.460PGY 3/4Strongly agree1.681.02-2.79.043The interns were truly involved in decision-making about their patientsUnstratifiedStrongly agree0.810.58-1.14.233Stratified on resident experience‡P-value for interaction term = .554.PGY2Strongly agree0.750.48-1.16.195PGY 3/4Strongly agree0.910.55-1.49.699CI = confidence interval; ICU = intensive care unit; OR = odds ratio; PGY = postgraduate year.∗ Multivariate conditional logistic regression models clustered on attending physicians and adjusted for possible confounding factors. Two separate models were constructed for each measure of entrustment. "Strongly agree" with entrustment measures was defined as a responsive of "Strongly Agree" or 5 on the Likert scale to each statement. ICU transfer was defined as transfer to any ICU during hospitalization. The models were adjusted for the following factors: patient factors: age, sex, race, education, marital status, insurance payor, Charlson Comorbidity Index, hospitalization within 12 months, has a primary care provider, discharge status; attending factors: hospitalist status, years of experience since residency, comfort with inpatient medicine; resident factors: sex, months of experience; and secular factors: 2003 duty hours limits, academic season.† P-value for interaction term = .038.‡ P-value for interaction term = .554. Open table in a new tab No association between entrustment and length of stay was observed in either the univariate or multivariate analyses, even when stratifying by resident experience.DiscussionIn exploring the relationship between attending entrustment of autonomy to house staff and patient outcomes, we found---only for teams including senior residents---a significant inverse association between entrustment and patient 30-day readmission. Interestingly, for senior residents, an association was also observed between entrustment of autonomy and higher rate of transfer to the ICU. These findings suggest that entrustment of autonomy may be related to patient care outcomes for patients on general medicine teaching services.Readmission, the primary patient outcome of interest, is a complex event. Many successful efforts to reduce readmission have focused on care coordination.16Rennke S. Nguyen O.K. Shoeb M.H. Hospital-initiated transitional care interventions as a patient safety strategy: a systematic review.Ann Intern Med. 2013; 158: 433-440Crossref PubMed Scopus (170) Google Scholar, 17Shepperd S. Lannin N.A. Clemson L.M. et al.Discharge planning from hospital to home.Cochrane Database Syst Rev. 2013; : CD000313PubMed Google Scholar It is interesting that the association between entrustment of autonomy and readmission appears to vary by team composition and resident experience, which may represent an element of care coordination. Attending physicians may individualize entrustment more thoughtfully for more senior residents, resulting in smoother execution of patient care and better outcomes, whereas junior residents may receive a more "blanket" level of supervision.7Sterkenburg A. Barach P. Kalkman C. et al.When do supervising physicians decide to entrust residents with unsupervised tasks?.Acad Med. 2010; 85: 1399-1400Crossref PubMed Scopus (181) Google Scholar Additionally, the senior residents for whom attendings are entrusting autonomy are likely to be high-functioning and very competent residents, for whom entrustment is warranted. It is also possible that attendings who obtain better patient outcomes are generally skilled and capable attendings that also excel in appropriately prescribing autonomy and supervision.We also noted the unexpected finding that entrustment of autonomy conferred higher odds of ICU transfer for patients on teams with more senior residents. ICU transfer differs from readmission in a number of ways. Transfer to the ICU is rare (2.3% in this sample), and, interestingly, may not always represent a poor outcome. Although need for ICU-level care may indicate a worsening of the patient's condition, a prompt and controlled transfer may be the result of an organized team effort, a subtlety that may not be captured by chart review. Observing a senior team resident facilitating an ICU transfer may lead to more faith in the resident's clinical competence and thus, a higher level of entrustment by the attending. Junior residents who lack experience may require more attending involvement in transferring a tenuous patient to the ICU and thus, the attending may not perceive entrustment in the same manner.In contrast to readmission and ICU transfer, we did not observe a significant association between entrustment of autonomy and length of stay. Delays in discharge leading to prolonged hospitalizations commonly result from nonmedical rather than medical reasons; one study ascribed over 60% of delays to nonmedical reasons such as awaiting placement at another facility.18Carey M.R. Sheth H. Braithwaite R.S. A prospective study of reasons for prolonged hospitalizations on a general medicine teaching service.J Gen Intern Med. 2005; 20: 108-115Crossref PubMed Scopus (67) Google Scholar These types of delays may be outside a physician's realm of influence and relatively unaffected by entrustment of autonomy. It would be interesting to explore how entrustment may affect length of stay by focusing on delays due to medical reasons (such as decision-making or diagnostic test interpretation) for which physician influence and entrustment of autonomy may be more germane.There are several limitations to these findings. Generalizability is limited because it is a single-institution study restricted to one service. Although data were collected over 8 years, the most recent trends of residency training are not reflected. Readmission was measured by patient self-report, which may be affected by recall or nonresponse bias, particularly with the loss of over 4000 participants from the follow-up survey whose baseline characteristics differ in some ways from patients included in the analysis. Notably, however, these patients were also more likely to be discharged to another facility, such as a rehabilitation facility, where they may not have been able to be contacted for the telephone follow-up survey.It is important to note that predictor variables, the attending survey questions of interest, have been previously used to study autonomy but could be subject to variability in interpretation. The findings may not precisely reflect the relationship between entrustment and patient outcomes as the questions may be measuring a different construct, such as an attending's perception of a high-quality resident. It is also possible that an attending's knowledge of patient outcomes influenced answers about entrusting autonomy as the survey was administered weeks to months following rotations. However, even given these limitations, this work represents, to our knowledge, the first study to attempt to address clinical implications of entrustment.This study has significant implications for using entrustment in assessment of postgraduate training under the Next Accreditation System. EPAs and milestones (competency-based developmental outcome expectations specific to a specialty) are the framework for clinical assessment under the Next Accreditation System, although application and interpretation remain variable. A 2012 survey by the Association of Program Directors in Internal Medicine reported relatively low rates of incorporation of milestones into observation tools (28.8%) and criteria for promotion (30.8%).19Alliance for Academic Internal Medicine. APDIM survey data. Available at: http://www.im.org/toolbox/surveys/APDIMSurveyData/Pages/default.aspx. Accessed November 20, 2013.Google Scholar Future studies of attending entrustment should work toward an objective measure of entrustment, which could be validated and used for direct observation. Finally, attendings m
Referência(s)