Initial Characterization of Internal Medicine Resident Resilience and Association with Stress and Burnout

Introduction. Burnout is prevalent in medical trainees. Little data exists on resident resilience.Methods. Anonymous surveys were provided to a convenience sample of internal medicine residents. Resilience was assessed using the Connor-Davidson resilience scale. Responses were categorized into low (<70), intermediate (70–79), and high (80–100) resilience. Results. 77 residents from six institutions completed surveys. 26% of residents had high resilience, 43% intermediate, and 31% low.Themean resilience score was 73.6 ± 9.6 and lower than the general population (mean 80.4 ± 12.5, p < 0.001). Trainees with high resilience were more likely to never have stress interfere with their relationships outside of work (high: 40%; low: 0%; p < 0.001). High resilience residents were more likely to have the skills to manage stress and burnout (high: 80%; low: 46%; p = 0.02) and less likely to feel inferior to peers (high: 20.0%; low: 70.8%; p < 0.001). There was a trend towards those with high resilience reporting less burnout (high: 40.0%; intermediate: 27%; low: 16.7%; p = 0.08). Only 60% report a program outlet to discuss burnout. Conclusions. There is a wide range of resilience among IM residents and scores were lower than the general population. Low resilience is associated with more stress interfering with relationships, feeling inferior to peers, and fewer skills to manage stress and burnout.


Introduction
Burnout among residents is widespread and is associated with depression, suicide, and poor clinical performance [1][2][3].While depression across the entire medical profession is at least as common as in the general population, studies looking at depression and depressive symptoms in medical and surgical residents place the prevalence higher, at nearly 28% [1,2].Further studies have shown that symptoms of depression in medical trainees are associated with perceived medical errors, increased work hours, and stressful life events [4].Despite increased recognition of burnout in residents, little data exists on how it relates to resilience and few interventions focus on resident resilience.
Resilience is the ability to adapt well or thrive in the face of adversity [5].It is important to cultivate resilience in trainees to enhance quality, promote wellness, and sustain the workforce [6,7].Resilience training in residents has been highlighted recently as an important area for graduate medical education (GME) training, although there have been few studies that have sought to define resilience levels in physicians [6,7].A 2008 multicenter study of medical and surgical interns found that baseline resilience scores in first year residents were significantly less than resilience scores in the general population and correlated with depression scores over a 6-month period [8].However, no studies have evaluated resilience further across the GME continuum and none have evaluated how resilience relates to burnout.Although many resident interventions have targeted burnout and wellness, little is known about the ability to augment resilience in medical trainees through education.One resilience program designed for a family medicine residency program demonstrated acceptance by trainees and increased self-care activities [9].Our objective was to complete a preliminary characterization of internal medicine (IM) residents' resilience and how resilience relates to stress and selfreported burnout in medical trainees through surveys of IM residents from multiple institutions in Northern Illinois.

Methods
Anonymous surveys were provided to IM residents from all years attending the 2014 Northern Illinois Regional American College of Physicians Associates meeting in October 2014.This is an educational conference held annually for residents to network and present research.Not all residents attending the conference are presenters.Participation was voluntary, but as an incentive, participants were entered into a raffle to receive a gift card after completing surveys.Baseline resilience was assessed using the Connor-Davidson resilience scale (CD-25), a validated 25-item self-report scale [5].
Responses were categorized into low (<70), intermediate (70-79), and high (80-100) resilience based upon previous literature.Residents were also surveyed on burnout, resilience training, and involvement in medical errors using Likert scale questions ranging from "strongly disagree" to "strongly agree" (Supplementary Survey in Supplementary Material available online at http://dx.doi.org/10.1155/2016/3508638).This portion of the survey was piloted with a group of firstyear residents at the University of Chicago during the spring of 2014.Descriptive statistics were examined using Stata version 13.0 software (StataCorp LT, College Station, TX).Chi-square tests, -tests, and trend analysis for proportions were used as appropriate to compare different resilience groups.This study was granted exemption status from the University of Chicago Institutional Review Board.

Results
In total, 254 residents attended part of the meeting and 77 residents (30%) completed surveys.Participants were from six residency programs; however, data was not complete for how many residents were from each institution.Residents surveyed were both categorical (residents who would complete internal medicine residency) (63%, 49) and preliminary residents (residents who would complete only one year of internal medicine prior to their chosen specialty) (13%, 10) with 23% (18) of residents not responding to this question.Other demographic questions including gender and age had high nonresponse rates.
The majority of residents believed burnout is an important issue during residency (92%) (Table 1).About half (56%) received training on burnout during residency and 66% received training on burnout during medical school.Only 60% of residents reported an outlet within their program to discuss stress and burnout and 70% reported the skills to manage stress and burnout (Table 1).When questioned about burnout, only 27% of residents reported low levels of burnout (never or rarely feeling burnout) (Table 1).Similarly, only 17% of residents reported that stress from work has never interfered with their patient care (Table 1).

Discussion
In this multi-institutional regional sample of IM residents, there was a wide range of resilience scores.Resilience was lower than the general population and similar to a previous sample of medical and surgical interns [5,8].Thus our findings are concordant with the one previous study examining resident resilience, demonstrating that residents have lower resilience than the general population.These findings further highlight the problem of resident burnout and the need to augment resilience training in residents.
In addition, a significant fraction of residents, approximately one-third, demonstrated low resilience.Low resilience was associated with more stress interfering with relationships, feeling inferior to peers, and fewer skills to manage stress and burnout.These residents with low resilience may be a group that requires extra attention and intervention.Many residents reported burnout but about half had received training on burnout during residency and 60% had an outlet to discuss stress and burnout in their program.Thus despite more attention being called to burnout in GME training, existing programs are likely still not sufficient.Although involvement in medical errors was prevalent, reporting of medical errors was rare and no association was seen between resilience and medical errors.However, since experience with medical errors was prevalent for all residents, helping residents learn how to cope after medical errors may still be an important part of resilience training and needs more investigation.
Our study has several limitations.Due to nonresponse to demographic questions, we cannot analyze associations between resilience and any demographic factors.We also cannot determine if certain institutions or community or academic programs were overrepresented which may have influenced our results and introduced bias.We did not ask residents about their year of training which may affect our results.We also cannot determine if there was any selection bias introduced by using residents attending a regional conference; for example, these residents may be more organized or engaged and thus have different resilience and burnout levels compared to those not attending the conference.Our response rate was likely low since many residents were only present for a portion of the conference and participation was voluntary.External validity may be limited as only one region of Illinois was included in this study.However, residents from both academic and community residency programs were surveyed.The questions regarding stress, burnout, and medical errors were not validated; thus residents may have interpreted the questions differently than intended.Burnout was assessed by resident self-report and not by a validated scale.Our questions regarding stress did not specify if we were asking about the intensity or frequency of stress and may have been interpreted differently.Although we asked about burnout training, we did not ask about specific information about what type of training they received.We also piloted our survey questions with first-year residents, a subgroup of the sample, which may have affected our results.Although the CD-25 has been validated in the general population, it has not been validated with physicians, which may impact our results.The medical error data was based on self-report, which may introduce bias.
Larger studies of resilience with a more representative sample of residents from different specialties, regions, and types of programs are needed to characterize resident resilience and determine how it relates to burnout.Additionally, correlation of resilience to burnout using a validated measure of burnout, such as the Maslach Burnout Inventory, is needed.Based on our preliminary study, however, resilience in residents is likely lower than the general population and there is an important population of residents with lower resilience who may benefit from interventions focused on enhancing resilience.This highlights that resilience, in addition to burnout, is an area that needs more attention and cultivation in residents.Although one resilience intervention in family medicine was positive, few educational innovations to enhance or teach resilience to residents have been identified and are an important area for future curriculum development.Determining if these interventions can impact resilience or burnout will also be important.

Conclusion
Our initial data investigating resident resilience demonstrates that there is variance in resilience among residents and a significant number of residents possess low levels of resilience.Resilience in residents is lower than the general population.Certain aspects of stress and burnout may be associated with lower resilience.More outlets to address stress, burnout, and resilience are needed in residency programs.

Table 1 :
Resident resilience and burnout survey measures.

Table 2 :
Association of survey measures with high and low resilience a .Resilience was measured using the Connor-Davidson resilience scale.Low resilience was defined as a score <70 and high resilience as a score of 80-100. a