Background and Objectives: Burnout is prevalent among clinicians and entails negative personal, professional, and organizational consequences. Assessments of burnout are typically anonymous to facilitate psychological safety. This limits the capacity of leadership to help struggling providers and reduces the level of demographic detail. Nonanonymous, confidential assessments may facilitate outreach to individuals or targeted interventions for at-risk populations.
Methods: We administered the Maslach Burnout Inventory to physician faculty and advanced practice providers in an academic department of family medicine. We identified a wellness officer within the department who served as an honest broker to keep nonanonymous survey responses confidential. Respondents had the option of taking the survey anonymously or confidentially. Anonymous respondents were allowed to withhold demographic information to ensure anonymity.
Results: Sixty-seven of 109 providers responded (61% response rate), with 46 (69%) doing so confidentially. Burnout rates were similar between groups: 48% among confidential respondents, and 43% among anonymous respondents (P=.71). Subscales of the MBI also showed no significant differences. Because a large proportion of anonymous respondents withheld demographic data, no demographic trends could be identified among them. Younger confidential respondents were more likely to exhibit depersonalization (P=.01).
Conclusions: Most participants chose to respond confidentially. There was no significant difference in the level of burnout between confidential and anonymous respondents. Our findings refute the conventional wisdom that clinicians require anonymity to respond to burnout surveys. This finding has the potential to open a new line of inquiry regarding burnout, its drivers and potential solutions.
Burnout is characterized by exhaustion, cynicism, and inefficacy.1 It has significant negative personal and professional consequences.2 Burnout is frequent among clinicians.3
Burnout remains stigmatized and thus hidden by providers,4 and so is typically measured anonymously.5 Anonymous assessments are used to create systematic interventions to address burnout.6 This approach has two important shortcomings. First, for surveys performed within an organization, true anonymity is threatened by detailed demographics.5 As such, organizations must choose between incomplete anonymity or incomplete demographics. Limiting demographic detail may reduce knowledge of at-risk groups for targeted interventions. Second, no individuals can be contacted to discuss their issues. This reduces the ability of local leadership to assess individual drivers of burnout, generate custom solutions, and reassess the impact of interventions. A nonanonymous burnout survey might meet those needs.
Nonanonymous surveys assessing burnout have been used previously.7-9 Research suggests survey participants prefer anonymity for sensitive information,10-11 yet other studies have demonstrated high response rates on nonanonymous surveys and willingness to share sensitive information.12-15 To our knowledge, no nonanonymous burnout surveys have been reported within an academic department.
We assessed the feasibility of conducting a nonanonymous burnout survey in a large academic department. We hypothesized: (1) most providers would respond anonymously, and (2) burnout rates would be higher among anonymous respondents.
The wellness officer was chosen on September 17, 2017, and announced to the department on January 5, 2018. Table 1 details wellness officer responsibilities. We administered the survey between January 5, 2018 and February 6, 2018. Participants were 109 faculty physicians and advanced practice providers (APPs) employed by the department of family and community medicine in a tertiary care university hospital system with 14 urban, suburban, and rural ambulatory clinical sites. Physicians provide inpatient coverage in two hospitals, and prenatal care and obstetrical care in one hospital. Faculty demographics at the time of the survey were 57 (52%) women, mean of 16.7 years since residency graduation (physicians), mean work effort 0.92 full-time equivalents, 34 (31%) associate professor or professor (APPs have no rank), 15 (14%) provide daytime in-patient care, 23 (21%) provide overnight in-patient call, and six (6%) provide obstetrical care.
We assessed burnout using a licensed copy of the Maslach Burnout Inventory Human Services Survey for Medical Personnel (MBI) online using REDCap. Overall burnout scores represent respondents in the top tercile of either the emotional exhaustion or depersonalization scales.16
The wellness officer informed participants that (1) nonanonymous responses would remain with the wellness officer and would only be made available to others with their consent, and were thus referred to as “confidential”; (2) confidential respondents may be contacted by the wellness officer to discuss their results; and (3) anonymous respondents would be able to withhold any demographics through a “prefer not to share” option, while confidential respondents would be expected to share all demographics except race/ethnicity and relationship status (for which they also could “prefer not to share”). Anonymous respondents needed to be able to withhold demographics due to the local nature of this survey, where a few key demographics might allow identification of an individual. Respondents had the option to respond anonymously or confidentially after completing the burnout questionnaire.
We conducted descriptive statistics using the R software version 3.5.1 (R Foundation for Statistical Computing, Vienna, Austria) to generate a reproducible statistical report. We analyzed differential assessments of burnout between the confidential and anonymous groups. The R package compareGroups17 (version 3.4.0) was used to calculate and display the statistical results incorporating appropriate tests including Kruskall-Wallis test for continuous data and Fisher exact test for categorical data.
Our institution’s review board determined this project to be exempt.
Sixty-seven of 109 (61%) participants responded. Responses were 46/67 (69%) confidential and 21/67 (31%) anonymous. The most complete demographic data among the respondents was for gender, with 26 (39%) male, 32 (48%) female, and no response 9 (13%), with no significant difference between respondents and faculty roster at time of survey (P=.72). No other comparisons on demographic features were possible.
Burnout rates were similar between groups: 22 of 46 (48%) among confidential respondents, and 9 of 21 (43%) among anonymous respondents (P=.71). Subscales of the MBI also showed no significant differences (Figure 1). No demographic trends could be discerned among anonymous respondents because many of them withheld demographic data (Figure 2). Among confidential respondents, the average age of providers with high levels of depersonalization was 39 years (SD=5.8 years) versus an average age of 47 years (SD=12.9) without it (P=.010).
Given the overwhelming use of anonymous survey strategy to assess burnout, we hypothesized that most respondents would be anonymous. Instead, most respondents completed the burnout survey confidentially. The response rate was 60th percentile for nonmanagerial employees.18 Overall burnout was present in 31 of 67 respondents (46%), a level consistent with other reports in the literature.3 Therefore, the high number of confidential respondents is not likely to be due to low burnout rates or poor response rate. Our hypothesis that burnout rates would be higher among anonymous respondents was also not supported, with both confidential and anonymous respondents having similar rates.
Younger confidential respondents were more likely to report depersonalization. Previous work has demonstrated varying results regarding the relationship between age and depersonalization.19-22 Potential reasons for this finding include older respondents having acclimated to the challenging environment, or depersonalization leading to attrition.23 Anonymous respondents withheld demographic data, reinforcing the limited ability of anonymous surveys to identify at-risk groups.
Limitations of this study are its single site, small sample size, and absence of comprehensive department-wide demographic data to assess representativeness of sample. Future studies might explore providers’ motivations for responding anonymously or confidentially, and the impact of one-on-one outreach to burned-out respondents.
The literature on burnout and its consequences is mostly confined to anonymous assessment. This represents a limitation on both the fidelity of the information gathered and the scope of interventions available. Our study demonstrates that confidential surveys of provider burnout are feasible in a large academic department. This may reveal new methods to assess and address the burnout epidemic by creating system-wide and individual-level interventions.
References
- Maslach C, Jackson S, Leiter M. Maslach Burnout Inventory Manual. 3rd ed. Palo Alto, CA: Consulting Psychologists Press; 1996.
- Shanafelt TD, Noseworthy JH. Executive Leadership and Physician Well-being: Nine Organizational Strategies to Promote Engagement and Reduce Burnout. Mayo Clin Proc. 2017;92(1):129-146. https://doi.org/10.1016/j.mayocp.2016.10.004
- Shanafelt TD, Hasan O, Dyrbye LN, et al. Changes in Burnout and Satisfaction With Work-Life Balance in Physicians and the General US Working Population Between 2011 and 2014. Mayo Clin Proc. 2015;90(12):1600-1613. https://doi.org/10.1016/j.mayocp.2015.08.023
- Wainwright E, Fox F, Breffni T, Taylor G, O’Connor M. Coming back from the edge: a qualitative study of a professional support unit for junior doctors. BMC Med Educ. 2017;17(1):142. https://doi.org/10.1186/s12909-017-0978-0
- Saari LM, Scherbaum CA. Identified employee surveys: potential promise, perils, and professional practice guidelines. Ind Organ Psychol. 2011;4(4):435-448. https://doi.org/10.1111/j.1754-9434.2011.01369.x
- Panagioti M, Panagopoulou E, Bower P, et al. Controlled interventions to reduce burnout in physicians: a systematic review and meta-analysis. JAMA Intern Med. 2017;177(2):195-205. https://doi.org/10.1001/jamainternmed.2016.7674
- Kluger MT, Bryant J. Job satisfaction, stress and burnout in anaesthetic technicians in New Zealand. Anaesth Intensive Care. 2008;36(2):214-221. https://doi.org/10.1177/0310057X0803600212
- Johns MM III, Ossoff RH. Burnout in academic chairs of otolaryngology: head and neck surgery. Laryngoscope. 2005;115(11):2056-2061. https://doi.org/10.1097/01.MLG.0000181492.36179.8B
- Aström S, Nilsson M, Norberg A, Winblad B. Empathy, experience of burnout and attitudes towards demented patients among nursing staff in geriatric care. J Adv Nurs. 1990;15(11):1236-1244. https://doi.org/10.1111/j.1365-2648.1990.tb01738.x
- Spiers J, Buszewicz M, Chew-Graham CA, et al. Barriers, facilitators, and survival strategies for GPs seeking treatment for distress: a qualitative study. Br J Gen Pract. 2017;67(663):e700-e708. https://doi.org/10.3399/bjgp17X692573 PMID:28893766
- King FW. Anonymous versus identifiable questionnaires in drug usage surveys. Am Psychol. 1970;25(10):982-985. https://doi.org/10.1037/h0029859
- O’Malley PM, Johnston LD, Bachman JG, Schulenberg J. A comparison of confidential versus anonymous survey procedures: effects on reporting of drug use and related attitudes and beliefs in a national study of students. J Drug Issues. 2000;30(1):35-54. https://doi.org/10.1177/002204260003000103
- Latimer WW, O’Brien MS, Vasquez MA, Medina-Mora ME, Rios-Bedoya CF, Floyd LJ. Adolescent substance abuse in Mexico, Puerto Rico and the United States: effect of anonymous versus confidential survey formats. J Child Adolesc Subst Abuse. 2006;16(1):69-89. https://doi.org/10.1300/J029v16n01_06
- van de Looij-Jansen PM, Goldschmeding JEJ, Van de Wilde E. Comparison of anonymous versus confidential survey procedures: effects on health indicators in dutch adolescents. J Youth Adolesc. 2006;35(4):659-665. https://doi.org/10.1007/s10964-005-9027-0
- Bjarnason T, Adalbjarnardottir S. Anonymity and confidentiality in school surveys on alcohol, tobacco, and cannabis use. J Drug Issues. 2000;30(2):335-343. https://doi.org/10.1177/002204260003000206
- Dyrbye LN, West CP, Satele D, et al. Burnout among U.S. medical students, residents, and early career physicians relative to the general U.S. population. Acad Med. 2014;89(3):443-451. https://doi.org/10.1097/ACM.0000000000000134
- Subirana I, Sanz H, Vila J. Building bivariate tables: the compareGroups Package for R. J Stat Softw. 2014;57(12). https://doi.org/10.18637/jss.v057.i12
- Anseel F, Lievens F, Schollaert E, Choragwicka B. Response rates in organizational science, 1995-2008: a meta-analytic review and guidelines for survey researchers. J Bus Psychol. 2010;25(3):335-349. https://doi.org/10.1007/s10869-010-9157-6
- Eckleberry-Hunt J, Kirkpatrick H, Taku K, Hunt R. Self-report study of predictors of physician wellness, burnout, and quality of patient care. South Med J. 2017;110(4):244-248. https://doi.org/10.14423/SMJ.0000000000000629
- Garcia HA, McGeary CA, Finley EP, McGeary DD, Ketchum NS, Peterson AL. The influence of trauma and patient characteristics on provider burnout in VA post-traumatic stress disorder specialty programmes. Psychol Psychother. 2016;89(1):66-81. https://doi.org/10.1111/papt.12057
- Garrosa E, Moreno-Jiménez B, Liang Y, González JL. The relationship between socio-demographic variables, job stressors, burnout, and hardy personality in nurses: an exploratory study. Int J Nurs Stud. 2008;45(3):418-427. https://doi.org/10.1016/j.ijnurstu.2006.09.003
- Ang SY, Dhaliwal SS, Ayre TC, et al. Demographics and Personality Factors Associated with Burnout among Nurses in a Singapore Tertiary Hospital. BioMed Res Int. 2016;2016:6960184. https://doi.org/10.1155/2016/6960184
- Brewer E, Shapard L. Employee burnout: a meta-analysis of the relationship between age or years of experience. Hum Resour Dev Rev. 2004;3(2):102-123. https://doi.org/10.1177/1534484304263335
There are no comments for this article.