Preliminary Findings From a Pilot Professional Coaching Program on the Components of Burnout in a Diverse Group of Physician Leaders

Abstract

Background

Burnout impacts physicians at alarming rates; physician leaders are not immune. While burnout of physician leaders is associated with decreased leadership efficacy, physician coaching has been associated with reduced burnout.

Objective

This study aimed to assess the effect of a physician-leader specific coaching program on the components of burnout.

Methods

A 10-week Wellness Coaching Program for physician medical directors was conducted and burnout was measured pre and post intervention using the Maslach Burnout Inventory-General Survey (MBI-GS).

Results

All subscales of the MBI-GS trended towards a decrease in the components of burnout with a statistically significant decrease in Cynicism.

Conclusion

Wellness coaching programs have the potential to positively impact the threat of burnout experienced by physician leaders in today’s health care workforce.

Background & Objective

Burnout impacts physicians at alarming rates, with 62.8% of physicians reporting at least 1 symptom of burnout in 2021. ¹ Physician leaders are not immune. Higher burnout scores in physician leaders are associated with decreased leadership efficacy, ² whereas effective leadership decreases burnout and increases physician wellbeing. ³

Physician coaching provides a results-oriented and stigma free method to address burnout ⁴ and has been associated with reduced burnout and improved well-being. ⁵ However, prior studies focus on practicing physicians and not specifically physician leaders. Previously studied coaching interventions are time intensive and difficult to scale. We posited that a 10-week virtual group coaching program specific to physician leaders would positively impact the components of burnout.

Methods

We piloted a 10-week virtual wellness coaching program for physician medical directors in a national medical group of approximately 5000 clinicians. All medical directors in the group received an email application. Three emails were sent to the medical director listserv over 6 weeks to solicit applicants.

To ensure a diverse cohort, the applications received were divided by gender and subdivided by region and specialty. Applications were then reviewed, and applicants from all regions and specialties, as well as an equal number of male and female applicants were intentionally selected to ensure participant diversity. As this was the first pilot program, participants were limited to a total of ten to ensure they each had sufficient opportunity for robust discussion and coaching.

Participants had 6 virtual group sessions and 4 virtual individual coaching sessions. Group sessions were one hour; 50% of the time was spent on didactics and 50% on group coaching. The group sessions were scheduled biweekly. The weekly didactics focused on cognitive awareness, work-life balance, procrastination, self-regulation and stress management, human-centered leadership, and personal resiliency. The group sessions were confidential and intentionally created without the traditional hierarchical structure, allowing for psychological safety and the ability for participants to build collegiality and compassion through reflections and discussing shared experiences.

The first individual coaching session was 45 min in length; all subsequent sessions were 30 min each. The first individual session was scheduled within the first 2 weeks of the program. The remaining sessions were divided throughout the 10 weeks and scheduled during weeks when there were no group sessions. Scheduling was based on the availability of the participants.

All sessions were led by a physician and former medical director of 8 years, who was also a Master Certified Coach (MCC) with 6 years of coaching experience, certified through the Life Coach School (LCS). The coaching theory and application used was similar to other recently published physician coaching studies, where the physician coaches were also certified through LCS.^6-8 The physician MCC also had a background in diversity, equity, and inclusion and served as the National Wellness Director for a national physician organization, meeting all 6 of the recommended core competencies for physician coaching. ⁹

Burnout and its components were measured using the Maslach Burnout Inventory-General Survey (MBI-GS). ¹⁰ The MBI-GS is defined by 3 subscales: (1) Emotional Exhaustion, (2) Cynicism, and (3) Professional Efficacy. Each item is a seven-point question on a Likert-type scale.

The MBI-GS was administered pre- and post-intervention, and participants completed the surveys independently of each other. Pre-assessment was completed within the 2 weeks prior to the intervention. Post-assessment was measured within 2 weeks after the intervention.

The study was provided an exemption from review by the Franciscan Health Institutional Review Board as defined in 45 CFR 46.101(b).

Results

A total of 269 medical directors were invited to apply and 36 completed the application, all of whom were considered eligible. Ten were selected, of which 50% identified as female and 50% identified as male. All participants were medical directors of hospital-based community medicine practices, representing geographically diverse practice locations across the United States and diverse specialties including Anesthesia, Critical Care, Emergency Medicine, Hospitalist Medicine, and Post-Acute Care. All 10 participants completed the pre-intervention survey, and 9 participants completed the post-intervention survey.

Analyses included only participants who had both a pre- and a post-intervention survey score (n = 9). Participants’ subscale scores were calculated as a total sum of the questions in each construct as recommended by the MBI fourth edition for ease of comparison between research reports. ¹⁰ A paired t test analysis was conducted using the Analyze Data function in Excel, version 2307, to assess for change in participants’ self-assessments. Due to the small sample size, a Shapiro-Wilk test was conducted on all pre- and post-intervention scores by construct with all results supporting the null hypothesis (P > .05) that the data are distributed normally. Study design and distribution of scores supported assumptions of paired t test, namely matched pair independent observations with a continuous dependent variable that is normally distributed. ¹¹

The pre-intervention baseline scores of the participants were abnormal on 2 subscales, Cynicism and Professional Efficacy. The baseline score of the third subscale, Emotional Exhaustion, was in the normal range. The following pre-intervention mean subscale scores of the cohort included: Cynicism, 12.9 (SD = 6.9); Professional Efficacy, 30.44 (SD = 4.3); and Emotional Exhaustion, 16.6 (SD = 7.4).

Of the 3 measured subscales of burnout, we found that the scores of all subscales trended towards a decrease in the symptoms of burnout (Figure 1) using a one-tailed measure of significance. Post-intervention mean subscale scores of the cohort included: Cynicism, 8.4 (SD = 5.3); Professional Efficacy, 32.9 (SD = 2.6); and Emotional Exhaustion, 15.1 (SD = 6.1). There was a statistically significant decrease in Cynicism subscale scores (t (8) = 3.0, P < .01), as well as an apparent trend towards increase in the Professional Efficacy subscale scores (t (8) = -1.8, P = .06).

Figure 1.

Maslach burnout inventory pre and post intervention subscale scores^a. Pre- and post-scores for cynicism showed a statistically significant decrease (P < .01); scores for professional efficacy showed a positive trend. ^aA decrease in cynicism and emotional exhaustion and an increase in professional efficacy denotes a positive effect of the intervention.

Discussion

Though there is no perfect measure for clinical burnout, the MBI has been the most robust and validated tool to evaluate burnout in physicians. ¹² Clinical burnout is complex and cannot be assessed as a dichotomous variable. Instead, it should be viewed as a spectrum with Cynicism, Emotional Exhaustion, and Professional Efficacy as symptoms of the syndrome.^10,13 While various definitions of burnout utilizing the subscales of the MBI have been proposed, Maslach has suggested that various profiles may exist between the experiences of burnout and engagement. By this definition, burnout is defined by an abnormal score on all 3 subscales; conversely, an individual with normal scores on all subscales would be considered engaged. Strongly abnormal scores on the Cynicism, Emotional Exhaustion, or Professional Efficacy subscales alone suggest an individual who is disengaged, overextended, or ineffective respectively.^14,15 Scores considered abnormal include Cynicism ≥10, Emotional Exhaustion ≥27, and Professional Efficacy ≤31 (this scale has an inverse relationship to burnout). ¹⁰ By this definition the pre-intervention scores of this cohort suggest a group that is disengaged rather than experiencing burnout.

In this pilot coaching program for physician leaders, participants who received 6 h of group coaching and 2 and a quarter hours of individual coaching had a statistically significant reduction in Cynicism. Participants also trended toward a reduction in Emotional Exhaustion and an increase in Professional Efficacy, though not statistically significant. Our findings are consistent with prior coaching studies that have shown a positive effect in some aspects of burnout, though not all. ⁵ The overall improvement on each subscale may suggest that this intervention could positively impact burnout in physician leaders.

Our wellness intervention was specific to physician leaders, a group with great demands and limited time. As such, we chose to measure an intervention requiring limited time commitment from the participants. The intervention was virtual and required 8 and a quarter hours over 10 weeks from the participants. As health systems have multiple budgetary constraints, we chose an intervention that required limited resources, including only 1 physician coach and an online virtual meeting platform rather than in-person meetings which may require budget for time, meeting space, and travel expenses. Given the flexibility of a virtual platform, the limited time required from the participants, and the minimal resources required for this intervention, the intervention could likely be reproduced and scaled if found effective in larger trials.

Study limitations included, first, a small sample size. Second, participants voluntarily applied. Some physician leaders who would benefit from this program may not choose to apply as they may already be experiencing burnout and prefer to not add to their workload. Third, we did not measure long-term effects of the intervention on physician-leader wellness. Larger, randomized trials with follow-up time-points are needed to further evaluate the effectiveness and sustainability of the likely positive impact of coaching on prevention or amelioration of burnout and its components.

Our study suggests that institutional investment in professional wellness coaching may be a useful approach to positively impact burnout and support career longevity in physician leaders.