Surgeons as Educators A Guide for Academic Development and Teaching Excellence

178

their duty hours on realms of patient safety, patient care, operative experience, and
education. Conversely, the flexible-policy group was more likely to perceive a nega-
tive effect of their duty hours on realms related to time outside of the hospital (i.e.,
time with family and friends, extracurricular activities, rest, health, etc.). While the
flexible-policy group perceived their duty hours to have less of a negative impact on
their sense of professionalism, there was no difference in terms of job satisfaction,
career choice decision, or morale. Fatigue as it relates to patient personal safety was
similar between both groups. The standard-policy group was significantly more
likely have left during an operation and missed an operation or handed off a patient
with active issues in the past month than their cohort in the flexible-policy group.
These results demonstrate that more flexible duty hours do not result in inferior
patient outcomes, and resident satisfaction was maintained. Other single-center ret-
rospective studies have shown that patient outcomes in other specialties are inferior
when resident exceed 80 hours per week [ 52 ]. Some limitations of the study exist
and are lucidly outlined by Billmoria et  al. in their companion article [ 46 ]. The

Table 9.3 (continued)

Outcome

Standard-

policy group

Flexible-

policy group

P valuea

Odds ratio for

flexible-policy

no./total no.(%) group (95% Cl)b P value

Occurrence during past month owing to duty-hour regulationsf

Left during an

operation

256/1944

(13.2)

128/1821

(7.0)

<0.001 0.46

(0.32−0.65)

<0.001

Missed an

operation

817/1944

(42.0)

544/1821

(29.9)

<0.001 0.56

(0.45−0.69)

<0.001

Handed off an

active patient issue

901/1944

(46.3)

583/1821

(32.0)

<0.001 0.53

(0.45−0.63)

<0.001

Denominators represent the number of respondents per survey item in the trial sample of residents.
Response rates varied across survey items, ranging from 84 to 87%. When the Bonferroni correc-
tion was applied to the 34 resident outcomes assessed, the level of significance was adjusted from
0.05 to 0.0015, and the differences between the study groups were no longer significant for three
outcomes: time for rest, quality and ease of handoffs and transitions in care, and professionalism
a Cluster-corrected P values were calculated by means of a chi-square test of association between
study-group assignment and dichotomized resident outcome
b Odds ratios and 95% confidence intervals (Cl) and two-tailed P values were calculated by means
of two-level hierarchical logistic regression with program-level random intercepts. Models
assessed the association between outcomes and study-group assignment, with adjustment for
program- level strata based on 30-day rates of postoperative death or serious complications in
2013 (stratifying variable for randomization). Significant odds ratios of less than 1.00 favor flex-
ible policies over standard policies. Significant odds ratios of more than 1.00 favor standard poli-
cies over flexible policies
c The numerator represents the number of residents who reported being “very dissatisfied” or “dis-
satisfied” versus “neutral,” “satisfied,” or “very satisfied”
d The numerator represents the number of residents who perceived a “negative effect” of 2014–
2015 institutional duty hours versus “no effect” or a “positive effect”
e The numerator represents the number of residents who reported that fatigue “always” or “often”
affects personal safety or patient safety versus “sometimes,” “rarely,” or “never”
f The numerator represents the number of residents who reported one or more occurrences in the
past month versus no occurrence

D.J. Rea and M. Smith