Internal medicine interns in programs with flexible duty-hour policies were less satisfied with their educational experience than interns in programs with standard policies, according to findings from the Individualized Comparative Effectiveness of Models Optimizing Patient Safety and Resident Education (iCOMPARE) trial.
Researchers compared two duty-hour configurations, the standard policies adopted in 2011 by the Accreditation Council for Graduate Medical Education (ACGME) and more flexible policies that did not limit shift length or specify mandatory time off between shifts. They randomized 63 U.S. internal medicine residency programs to implement either the ACGME policies (n=31) or flexible policies (n=32).
From July 1, 2015, to June 30, 2016, researchers compared the groups on a variety of measures. Primary outcomes were interns' direct patient care and educational activities (assessed by observation), trainee and faculty satisfaction with education (assessed by survey), and interns' medical knowledge (assessed by score on the ACP In-Training Examination). Results on patient mortality outcomes, assessed using 2015 and 2016 Medicare data, are not yet available.
Results were published online on March 20 by the New England Journal of Medicine.
There were no significant differences between groups in the mean percentages of time that interns spent on direct patient care and education. In addition, there were no significant differences in trainees' reported perceptions of an appropriate balance between clinical work and education (response rate, 91%) or faculty members' assessments of whether trainees' workload exceeded their capacity (response rate, 90%).
Another survey (response rate, 49%) found that interns in flexible programs were more likely than those in standard programs to report dissatisfaction with various training aspects (odds ratio, 1.67; 95% CI, 1.02 to 2.73) and overall well-being (odds ratio, 2.47; 95% CI, 1.67 to 3.65). However, in a faculty survey (response rate, 98%), directors of flexible programs were less likely than those of standard programs to report dissatisfaction with various educational processes (odds ratio, 0.13; 95% CI, 0.03 to 0.49). Average test scores were 68.9% in flexible programs and 69.4% in standard programs, a difference which did not meet the noninferiority margin of two percentage points (difference, −0.43; 95% CI, −2.38 to 1.52; P=0.06 for noninferiority).
An accompanying editorial noted limitations to the study, such as limited generalizability, desirability bias in residents' survey responses, low response rates to some elements, substantial variations in outcomes between programs, and difficulties characterizing the actual differences in hours worked between groups. In addition, flexible programs applied that flexibility to a small minority of trainees' rotations, the editorialist noted.
According to the editorial, the iCOMPARE results represent a “sentinel plea of residents to reform our clinical learning environments to prioritize people,” since more than two-thirds of all respondents reported high or moderate levels of burnout (i.e., emotional exhaustion, depersonalization, and low perceptions of personal accomplishment). “The response of our profession to these clear warning signs should become the durable legacy of this trial,” the editorialist wrote.