Review Reports - Coaching the Caregivers: A Burnout Intervention for Military Physicians and Medical Students

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

In this highly commendable study, the authors tackle a critical public health gap by rigorously evaluating the impact of a novel, virtual group coaching intervention designed specifically for military obstetrician-gynecologists (OB/GYNs) and medical students.

By focusing on this uniquely high-stress cohort, the investigators provide invaluable, forward-thinking insights into mitigating burnout and elevating professional well-being within military medicine. However, a disconnect exists between the intervention design and the targeted outcomes. The data suggest that although individual coaching was provided, the most notable improvements occurred in systemic organizational perceptions, as reflected in AWS scores for workload, control, and fairness.

The discussion section should further examine how individual cognitive reframing or boundary-setting may translate into perceived changes in macro-level workplace fairness and workload. As an observational study without a randomized parallel control arm, it is methodologically challenging to distinguish the effects of the coaching intervention from natural temporal confounding variables or external environmental changes within the Military Health System (MHS) during the three-month period. This threat to internal validity should be explicitly acknowledged as a primary limitation.

As noted in the limitations, the cohort began with exceptionally high baseline scores in Personal Accomplishment (PA). Within military structures, professional duty and identity often buffer PA scores. Since traditional burnout definitions require a decrease in PA alongside high emotional exhaustion and depersonalization, this baseline asymmetry likely skewed or masked the aggregate therapeutic signal of the MBI.

On Page 2, line 53 states: "59 participants completed coaching and surveys from August–October 2024." 2024". On Page 5, line 189 states:"A total of 66 participants enrolled from April 8, 2024, to June 20, 2024...." There is a clear mismatch of data. Clarify the precise timeline distinguishing initial enrollment, the 12-week intervention period, and data completion.

In Table 3 (Page 9), the row for "Overall score?" under the PCQ framework is missing its corresponding baseline and post-intervention data cells. Similarly, the "Overall life satisfaction index?" row at the bottom of the FFWEL section is entirely blank. These entries should be completed or removed.

Ensure standard formatting for all p-values throughout the text and tables (e.g., avoid alternating between uppercase and lowercase).

Rather than relegating the session-adjusted data to Supplement 2, elevate this analysis to the main results section.

Elaborate on the Selection Choice for the 3-Profile LPA.

Add 1–2 sentences explicitly justifying why you overrode the VLMR signal of the 5-profile model (e.g., preventing clinical over-stratification).

Address the theoretical tension between individual-level coaching and macro-level workplace factors. Coaching cannot alter administrative burdens, lack of support staff, or military schedules.

Explain why coaching did significantly improve perceptions of structural workplace constraints (AWS scores). Frame this clearly in your discussion.

Create a short synthesis paragraph in the discussion defining the overlapping structural vulnerabilities of this cohort. Specifically, explain how the military's top-down command structure naturally suppresses "Control" and "Autonomy ".

Propose a Direct Implementation Pathway for the MHS, meaning translate your conclusion into practical implications.

Explicitly state in your conclusion the consequence of OB/GYNs exiting the service early due to burnout.

Ensure that references to Supplement 1- 4 in the body of the text exactly match the headings used in your appendix. In Table 2, ensure it is clear whether you are presenting standard deviation (SD) or interquartile range (IQR).

Use the same font in Tables as in Text

Comments on the Quality of English Language

No comment

Author Response

Reviewer #1

Thank you, we included the following in the manuscript, “Since coaches included boundary setting and cognitive re-framing in coaching groups, it is possible that these techniques provided new tools for participants to use in their professional lives, and that these approaches may have affected participant survey responses.” And in the limitations section, “Furthermore, we cannot determine whether coaching or other environmental changes in the Military Health System may have affected participant responses on surveys.”

Thank you for this comment, please see figure 1 for an overview of participant enrollment. We identified that one person had been double-counted (they signed up with 2 different email addresses, so we updated Figure 1 to reflect 69 people expressing interest, 4 declining to participate, and 65 completing pre-surveys, with 6 dropouts (not completing post-surveys), and a total of 59 participants (unchanged). 6 participants did not complete the post coaching survey and were excluded from the analysis. We clarified that only participants who completed both pre- and post- coaching surveys were included in the analysis, “Only participants who completed both pre- and post- coaching surveys were included in the data analysis.”

Thank you, we have added this data to the table.

Ensure standard formatting for all p-values throughout the text and tables (e.g., avoid alternating between uppercase and lowercase).

Thank you for this comment, we have verified that all are listed as p-values (lower case p)

Rather than relegating the session-adjusted data to Supplement 2, elevate this analysis to the main results section.

Thank you for this comment. This was changed to Table 4.

Elaborate on the Selection Choice for the 3-Profile LPA.

Add 1–2 sentences explicitly justifying why you overrode the VLMR signal of the 5-profile model (e.g., preventing clinical over-stratification).

Thank you for this comment, we added the following sentence, “While the 5-profile solution has a significant VLMR, indicating improvement from the 4-profile solution, the minimum class probabilities were quite low, which would lead to difficult interpretation and generalization. Therefore, the 3-profile model was preferred to prevent clinical over-stratification.”

Address the theoretical tension between individual-level coaching and macro-level workplace factors. Coaching cannot alter administrative burdens, lack of support staff, or military schedules.

Thank you for this comment. We agree, and have included the following in the discussion, “Involvement of healthcare systems and leadership is a critical step toward addressing macro-level workplace issues such as administrative burdens, support staff shortages, and scheduling issues. While individual coaching can address personal responses to challenges, and empower physicians to speak up, durable changes need support and investments from healthcare systems.”

Explain why coaching did significantly improve perceptions of structural workplace constraints (AWS scores). Frame this clearly in your discussion.

Thank you, we have clarified this in the discussion, “Since coaches included boundary setting and cognitive re-framing in coaching groups, it is possible that these techniques provided new tools for participants to use in their professional lives, and that these approaches may have affected participant survey responses.”

Thank you, we have included additional discussion on this topic, and believe that such hierarchical structures also exist outside the military in civilian healthcare systems. The following was added to the text, “The military (and civilian healthcare systems) includes a hierarchical structure that can suppress individual control and autonomy. Such hierarchical structures can make it more challenging for individual participants to speak up and share opportunities for improvements within healthcare systems. Thus, it is important for healthcare systems to provide formal and informal avenues for clinicians to share concerns and ideas for system improvements.”

Propose a Direct Implementation Pathway for the MHS, meaning translate your conclusion into practical implications.

Thank you for this suggestion, we have added this to the paper:

“A proposed direct implementation pathway for the military health system (or other healthcare system) interested improving physician autonomy, reducing administrative burden and increasing clinical efficiency includes convening a group of interested parties, assessing current administrative burdens, eliminating or re-designing low-value tasks and inefficiencies, increasing physician decision-making authority and involvement, optimizing use of technology, and monitoring and refining the system based on physician feedback and efficiency metrics. Using formal quality improvement techniques such as FOCUS-PDSA (plan-do-study-act)³⁸ and lean six sigma,³⁹can provide valuable structure when improving processes.”

Explicitly state in your conclusion the consequence of OB/GYNs exiting the service early due to burnout.

Thank you, we have added the following to the conclusions: “Military OB/GYN physicians cited lack of support, lower compensation and inefficiencies as reasons why physicians leave military service.”

Thank you we have verified and corrected references to supplements

In table 2, each row is labeled to define n(%), mean (SD), and median IQR.

Use the same font in Tables as in Text

Thank you, the fonts have been verified and are now consistent

Reviewer 2 Report

Comments and Suggestions for Authors

I commend the authors on excellent work with this study and I found the manuscript to be very well written and clear. I have several comment and suggestions by section below to consider in an effort to further strengthen this manuscript.

Title: Non-physician trainees (i.e., medical students) were also included as a fairly substantial portion of the study population and the title should reflect this

Highlights: excellent

Abstract:

- Line 41: suggest editing to "personal accomplishment" as it's referring to the sense of accomplishment and not accomplishments per se

Introduction:

While rates are still disproportinately high among physicians (compared to non-physicians) burnout has been decreasing since the pandemic, more recent good data demonstrating 45.2% prevalence among US Physicians (see reference: https://www.mayoclinicproceedings.org/article/S0025-6196(24)00668-2/fulltext)
Line 104: is there data that show that these unique factors experienced by military OB/GYNs contribute to increased risk of burnout? If not, suggest adding "may" contribute to a hightened risk...
Line 108: it's noted that "individualized" professional coaching may improve burnout... however, reference 17 is regarding an intervention that was group-based coaching. Would clarify that both individual and group-based coaching improves burnout. This has been already demonstrated in RCTs, including those you have referenced and additional, see the following which should be cited as RCTs demonstrating the impacts of this intervention (understanding not specifically among OB/GYN physicians)
- https://doi.org/10.1001/jamainternmed.2019.2425
- https://doi.org/10.1037/ocp0000180
- https://doi.org/10.1001/jamanetworkopen.2024.5645
- https://doi.org/10.1001/jamanetworkopen.2023.35541
- https://doi.org/10.1007/s11606-025-09653-w

Materials and Methods:

Would include description of the cost of the coaching intervention, this is especially critical as group coaching is often cheaper and scalable compared to individual coaching. If unable to include cost, this should be presented in the limitations
Would benefit from more clarity regarding the makeup of the groups, it's listed in line 129 that the groups had 5 to 13 participants, but it would be helpful to know more detail, especially in terms of standardization and the ability for others to replicate this work
Regarding the sample size calculation, it appears that the burnout rate used is based on attending physician data (from the national physician samples); however, this study is including students and trainees which tend to have lower burnout rates and could confound the analysis. this should be described in limitations
While the details of the coaching intervention may be variable based on the coach and coaching sessions often can divert to different topics, it would be helpful to include more details regarding the basic themes of coaching and if there were any particular themes that were reviewed for this intervention. For example, a table can be created that reviews the structure and topics reviewed during coaching. this will help with replication of the study. if there is not a way to do this, then I suggest including it in the limitations as having a standardized intervention is important to be able to assess its efficacy appropriately

Results

Would consider adding a subanalysis evaluating burnout as a dichotomous variable; i.e., participants with high EE or high DP are considered burned out and evaluate pre and post burnout prevalence. Note: PA is typically not associated with burnout among physicians and in most studies of burnout interventions among physicians, it is not used in calculations as physicians generally have higher PA
Line 191: would include the mean number of coaching sessions attended by participants (i.e., 3.5?)

Discussion

Line 313: would re-word female-"dominated"
Line 322-323: high levels of PA are unlikely to skew burnout rate as among physicians, PA is not typically used when evaluating burnout given baseline high levels. High EE and/or high DP are the primary markers used to indicate burnout
Is there more detail that can be shared regarding what the coaches led in their sessions? Was there any standardization of the intervention besides the number of sessions over a set period of time? These are important limitations to discuss if not. See an example of an coaching RCT for physicians that utilized a structured and themed-based approach to the coaching intervention (https://link.springer.com/article/10.1007/s11606-025-09653-w)
25% of the cohort were medical students and one may argue that they face a set of challenges and experiences that are very different than attending physicians. would further discuss this in the limitations; similarly a significant subset are resident physicians who also face different barriers
Pre/post studies can only demonstrate correlation and an RCT may be better at delineating if this intervention can truly make an impact, this is a limitation
Line 326: Any thoughts regarding why session attendance was low, would explain the "various reasons"
I believe it would be helpful to include more explanation regarding the utility of the LPA and discussion regarding why this may be important. why might it be helpful to identify distinct burnout profiles?

Tables/Figures: no comments

Author Response

Reviewer #2:

Title: Non-physician trainees (i.e., medical students) were also included as a fairly substantial portion of the study population and the title should reflect this

Thank you for this feedback, we have modified the title to include medical students:

Coaching the Caregivers: A Burnout Intervention for Military Physicians and Medical Students

Highlights: excellent

Abstract:

- Line 41: suggest editing to "personal accomplishment" as it's referring to the sense of accomplishment and not accomplishments per se

- Thank you for this feedback , this was edited to ‘personal accomplishment’

Introduction:

While rates are still disproportinately high among physicians (compared to non-physicians) burnout has been decreasing since the pandemic, more recent good data demonstrating 45.2% prevalence among US Physicians (see reference: https://www.mayoclinicproceedings.org/article/S0025-6196(24)00668-2/fulltext)

Thank you for this updated reference. We have updated reference #6 in the manuscript accordingly:

Shanafelt TD, West CP, Sinsky C, Wang H, Carlasare LE, Dyrbye LN. Changes in Burnout and Satisfaction With Work-Life Integration in Physicians and the General US working Population Between 2011 and 2023. May Clin Proc 2025;100:P1142-1158.

Line 104: is there data that show that these unique factors experienced by military OB/GYNs contribute to increased risk of burnout? If not, suggest adding "may" contribute to a hightened risk...

Thank you, we updated this to “may” as suggested

Line 108: it's noted that "individualized" professional coaching may improve burnout... however, reference 17 is regarding an intervention that was group-based coaching. Would clarify that both individual and group-based coaching improves burnout. This has been already demonstrated in RCTs, including those you have referenced and additional, see the following which should be cited as RCTs demonstrating the impacts of this intervention (understanding not specifically among OB/GYN physicians)

https://doi.org/10.1001/jamainternmed.2019.2425
https://doi.org/10.1037/ocp0000180
https://doi.org/10.1001/jamanetworkopen.2024.5645
https://doi.org/10.1001/jamanetworkopen.2023.35541
https://doi.org/10.1007/s11606-025-09653-w

Thank you we clarified that group coaching has also demonstrated a benefit in the paper, and included two additional relevant references from this list:

Khalili J, Miotto K, Wang T, Mafi JN, Kyababchyan E, Sanford J, et. al. Professional Coaching to Reduce Physician Burnout: A Randomized Clinical Trial. J Gen Intern Med. DOI: 10.1007/s11606-025-09653-w
Mann A, Shah AN, Thibodeau PS, Dyrbye L, Syed A, Woodward MA, et. al. Online Well-Being Group Coaching Program for Women Physician Trainees. JAMA Netw Open 2023;6:e2335541.

Materials and Methods:

Would include description of the cost of the coaching intervention, this is especially critical as group coaching is often cheaper and scalable compared to individual coaching. If unable to include cost, this should be presented in the limitations

Thank you, we added the cost of the professional coaching:

“Coaching fees for the professional coaches were negotiated, and the cost was $750 per one hour group session, with a total of six sessions for 8 groups (48 sessions), this was $36,000.”

Would benefit from more clarity regarding the makeup of the groups, it's listed in line 129 that the groups had 5 to 13 participants, but it would be helpful to know more detail, especially in terms of standardization and the ability for others to replicate this work.

Thank you, we clarified this in the methods:

“Groups self-selected based on the best time for participation based on their work and clinical schedules. The maximum number of participants in each group was limited to no more than 13 to allow for participation of all members.”

Regarding the sample size calculation, it appears that the burnout rate used is based on attending physician data (from the national physician samples); however, this study is including students and trainees which tend to have lower burnout rates and could confound the analysis. this should be described in limitations

Thank you for bringing this to our attention, we have added this to the limitations: “Additionally, the sample size calculation was based on the burnout rate for attending physicians using national physician samples, and students and trainees may have different rates of burnout that would affect the calculated sample size.”

While the details of the coaching intervention may be variable based on the coach and coaching sessions often can divert to different topics, it would be helpful to include more details regarding the basic themes of coaching and if there were any particular themes that were reviewed for this intervention. For example, a table can be created that reviews the structure and topics reviewed during coaching. this will help with replication of the study. if there is not a way to do this, then I suggest including it in the limitations as having a standardized intervention is important to be able to assess its efficacy appropriately

Thank you, for this comment, we added additional detail of the coaching topic discussed in Supplement 1. We also included an additional note in the limitations, “However, the variability of coaching topics and coaching techniques may have limited our assessment of coaching efficacy.”

Supplement 1 – Coaching topics discussed

Coaching group	Topics Discussed
Medical Students	Residency application/match process How to develop a competitive residency application AAMC tools for self-awareness and specialty selection Self-care Priorities in school, life, etc. Specialty disrespect Research opportunities How to address equity topics e.g. women in underrepresented fields Rotations
Resident Physicians	Team cohesion, coping Career options Work life balance
Junior Physicians (< 10 years from residency)	Cognitive Thought Model (CTFA) Circumstance thought leads to a feeling leads to an action (propelling emotions vs. emotions that keep us stuck) actions come from feelings and create results. Give yourself permission to feel your feelings, then you can decide if you want to continue to have those feelings. Accept thoughts, don't judge. Your thoughts and feelings matter. Accept it before you work on fixing it. The smarter you are and the more operational you are, you are more likely to drift into black and white thinking. Thoughts are not the truth. Shift your thought from negative to neutral. Re-wire your brain to set boundaries Saying yes - Think through - if I say yes, what am I saying no to! You don't want to be the person you disappointed at the end of your life Feelings behind procrastination Relaxation and meditation techniques
Senior Physicians (≥ 10 years from residency)	Wheel of life assessment sheet Complaining exercise Induced thought model Boundary setting: define, communicate, consequence Facts vs. circumstances Thought model (intentional and unintentional); things we can't control/change (e.g. big system undergoing lots of changes), if you can't change the system, you can adapt or leave. How do you convince people to do what you want? (negotiations) - Tactical empathy = trust and collaboration, influencing people by influencing their thoughts. Be clear about what you are thinking and feeling BEFORE you start talking. Use the thought model and use the part that includes you. Channel a thought - eg. certainty (all problems fall into 3 buckets: 1. we know what it is and we can fix it, 2. we rule out bad things and feel safe, 3. we don't know what it is but we keep working to find a solution).

Results

Would consider adding a subanalysis evaluating burnout as a dichotomous variable; i.e., participants with high EE or high DP are considered burned out and evaluate pre and post burnout prevalence. Note: PA is typically not associated with burnout among physicians and in most studies of burnout interventions among physicians, it is not used in calculations as physicians generally have higher PA.

Thank you for this suggestion, however, we used the Maslach Burnout Inventory scoring manual’s recommendation to calculate the individual MBI scale scores separately, and did not combine into a single “burnout score.” The standard deviations from these scale scores (EE, DP and PA) were used to calculate the values for the profiles in Supplemental Table 4. Since these z-scores were used to define the profile, we cannot use them for a dichotomous variable as viewing the three scales jointly offers the description of interest. No changes were made to the paper.

Line 191: would include the mean number of coaching sessions attended by participants (i.e., 3.5?) Thank you, we included this number in the text, “. On average, participants attended three to four of the six coaching sessions (mean 2.83-4.78).”

Discussion

Line 313: would re-word female-"dominated"

Thank you, we changed this to female-predominant

Line 322-323: high levels of PA are unlikely to skew burnout rate as among physicians, PA is not typically used when evaluating burnout given baseline high levels. High EE and/or high DP are the primary markers used to indicate burnout

Thank you for bringing this up, we have clarified this in the text: “Additionally, all participants began the study with very high levels of personal accomplishment (PA), which may have skewed the burnout results to appear less severe (with more individuals having overextended (high EE, low to moderate DP, and low to moderate PA) or disengaged profiles (low to moderate EE, high DP, and low to moderate PA), rather than burnout profiles (high EE, high DP and low PA)^27”

Reference 27. Leiter MP, Maslach C. Latent burnout profiles: A new approach to understanding the burnout experience. Burnout Research 2016;89-100.

Is there more detail that can be shared regarding what the coaches led in their sessions? Was there any standardization of the intervention besides the number of sessions over a set period of time? These are important limitations to discuss if not. See an example of an coaching RCT for physicians that utilized a structured and themed-based approach to the coaching intervention (https://link.springer.com/article/10.1007/s11606-025-09653-w)

Thank you, we agree, and with the different groups coaches were allowed to tailor their discussions to what the groups felt was most beneficial. We have included the topics of discussion in Supplement 1:

Supplement 1 – Coaching topics discussed

Coaching group	Topics Discussed
Medical Students	Residency application/match process How to develop a competitive residency application AAMC tools for self-awareness and specialty selection Self-care Priorities in school, life, etc. Specialty disrespect Research opportunities How to address equity topics e.g. women in underrepresented fields Rotations
Resident Physicians	Team cohesion, coping Career options Work life balance
Junior Physicians (< 10 years from residency)	Cognitive Thought Model (CTFA) Circumstance thought leads to a feeling leads to an action (propelling emotions vs. emotions that keep us stuck) actions come from feelings and create results. Give yourself permission to feel your feelings, then you can decide if you want to continue to have those feelings. Accept thoughts, don't judge. Your thoughts and feelings matter. Accept it before you work on fixing it. The smarter you are and the more operational you are, you are more likely to drift into black and white thinking. Thoughts are not the truth. Shift your thought from negative to neutral. Re-wire your brain to set boundaries Saying yes - Think through - if I say yes, what am I saying no to! You don't want to be the person you disappointed at the end of your life Feelings behind procrastination Relaxation and meditation techniques
Senior Physicians (≥ 10 years from residency)	Wheel of life assessment sheet Complaining exercise Induced thought model Boundary setting: define, communicate, consequence Facts vs. circumstances Thought model (intentional and unintentional); things we can't control/change (e.g. big system undergoing lots of changes), if you can't change the system, you can adapt or leave. How do you convince people to do what you want? (negotiations) - Tactical empathy = trust and collaboration, influencing people by influencing their thoughts. Be clear about what you are thinking and feeling BEFORE you start talking. Use the thought model and use the part that includes you. Channel a thought - eg. certainty (all problems fall into 3 buckets: 1. we know what it is and we can fix it, 2. we rule out bad things and feel safe, 3. we don't know what it is but we keep working to find a solution).

And have also included this in the limitations:

“Another limitation was different coaches, curriculum, and topics (Supplement 1) with potential variability in coaching styles which may have influenced results and differences between groups. Coaches independently planned and administered the sessions based on their expertise and what was most relevant to each group. While this may have introduced variability across groups, this practice best served the different career stages. However, the variability of coaching topics and coaching techniques may have limited our assessment of coaching efficacy.”

25% of the cohort were medical students and one may argue that they face a set of challenges and experiences that are very different than attending physicians. would further discuss this in the limitations; similarly a significant subset are resident physicians who also face different barriers.

Thank you for this comment, we agree, and have highlighted this in the discussion:

“The diversity of participants included in this study was another potential limitation. While including diverse groups allowed us to provide coaching to students and physicians at different stages in their careers, students and residents face different challenges and experiences than physicians, and some of these differences may have impacted participation, with resident physicians having the lowest participation of all groups.”

Pre/post studies can only demonstrate correlation and an RCT may be better at delineating if this intervention can truly make an impact, this is a limitation

Thank you, we agree, and have added the following as a limitation in the text, “As this study was an observational study, we cannot determine causality.”

Line 326: Any thoughts regarding why session attendance was low, would explain the "various reasons"

Thank you, we clarified this in the text, “For varying reasons (such as time constraints, work and other commitments), most participants attended only four of the six coaching sessions.”

I believe it would be helpful to include more explanation regarding the utility of the LPA and discussion regarding why this may be important. why might it be helpful to identify distinct burnout profiles?

Thank you, we have clarified this in the text, “Using LPA allows for the assessment of unique patterns in participants, allowing for an individualized assessment of the three dimensions for different people and identifying unique subpopulations.²⁷”

Tables/Figures: no comments

Reviewer 3 Report

Comments and Suggestions for Authors

I would like to thank the authors for their interesting work. Beyond the value of the data reported, I will refer to some critical issues below:

Introduction

The introduction is well-reasoned. It effectively establishes the gravity of the OB/GYN shortage, the rise of maternity care deserts, and the unique dual-professional stressors (clinical and military) experienced by active-duty obstetrician.

From an occupational health and preventive medicine perspective, burnout interventions must be categorized. While structural workplace modifications represent primary prevention, individual-focused interventions like coaching are secondary prevention (enhancing individual coping and cognitive appraisal). The authors should explicitly clarify in the introduction that while coaching is a highly valuable tool to mitigate the consequences of work-life incongruence, it serves as a secondary prevention strategy and does not replace the primary prevention need for systemic organizational reforms.
While the overall objective of evaluating the coaching program is clear, the introduction lacks explicit hypotheses regarding the secondary outcomes (AWS and PCQ dimensions). The authors should state whether they hypothesized improvements in specific subscales (control, efficacy) even in the absence of a global reduction in Maslach Burnout Inventory (MBI) scores.
To strengthen the international relevance of this work, the introduction should briefly frame how physician burnout and professional fulfillment have evolved globally in the post-COVID-19 era. Please integrate a brief discussion of recent international findings on post-pandemic physician burnout, such as the cross-sectional study by https://doi.org/10.3390/healthcare14040454 (Burnout Among Italian Medical Doctors: A Cross-Sectional Study on Emotional Exhaustion, Depersonalization, and Gratification Post-COVID-19). This will help illustrate that high levels of emotional exhaustion and depersonalization remain a persistent global threat post-COVID-19, and will set the stage for discussing how professional gratification and personal accomplishment function as vital coping resources.

Materials and Methods

The recruitment, cohort stratification, and survey administration procedures are thoroughly detailed. However, several methodological weaknesses limit the internal validity of the study and must be addressed.

This is the primary methodological limitation of the study. Because this is a single-arm, pre-post observational design without a non-intervention control group, it is impossible to definitively rule out confounding external factors (seasonal changes in patient volume, clinical schedule adjustments, or regression to the mean, particularly for those starting with extreme burnout scores). The authors must explicitly acknowledge this in the Methods section and adopt a more cautious tone when discussing causal relationships.
The study sample pools medical students, residents, junior faculty, and senior faculty. From an occupational health standpoint, the daily duties, levels of clinical autonomy, and systemic stressors of a medical student are fundamentally different from those of an attending physician. While stratifying the actual coaching groups by training level was an excellent decision, analyzing them as a single group in the pre-post analysis introduces substantial heterogeneity. The authors should justify why these distinct career stages were pooled for the overall quantitative analysis.
The manuscript states that "coaches determined the content of the coaching sessions tailored to their specific audiences and group interest". While beneficial for engagement, this introduces a high degree of inter-group variability and a lack of intervention standardization. The authors must discuss how this variability in curriculum and coaching styles might have influenced the outcomes.
The power analysis indicated that 62 participants were required to detect a significant decrease in burnout. Although the authors planned to enroll 80 individuals to account for attrition, they only enrolled 66, and only 59 completed both pre- and post-surveys. Technically, the study is slightly underpowered relative to its initial statistical plan. The authors must explicitly state this epidemiological limitation in the methods and limitations section.

Results

The results are presented clearly, and the use of tables and the LPA profile plot is highly informative.

Out of 66 enrolled participants, 7 (10.6%) failed to complete the post-intervention surveys. The authors must report whether a sensitivity analysis was conducted to compare the baseline demographics or baseline burnout scores of these 7 drop-outs against the 59 completers. This is crucial to rule out potential attrition bias.
The authors mention extracting common themes from open-ended questions. However, they do not describe the formal qualitative methodology used (thematic analysis according to Braun & Clarke, double-coding to assess inter-rater reliability, or the use of qualitative software). The authors must provide details on how these qualitative data were systematically analyzed to ensure scientific rigor.
A striking finding of the LPA is that all three profiles, including Profile 3 ("burnout"), demonstrated high levels of Personal Accomplishment. This is an atypical finding in general burnout literature. The authors should discuss whether this represents a self-selection bias (highly motivated individuals volunteering for coaching) or a unique sociocultural feature of military medical officers.

Discussion

The discussion does a commendable job of linking the findings to existing theories of job congruence and psychological capital.

The most intriguing finding is that while global MBI scores did not change, participants in the burnout profile experienced significant post-coaching improvements in their perceptions of workload, control, and fairness. Crucially, the objective workload, staffing levels, and administrative demands of the military treatment facilities did not change during these 12 weeks. This indicates that the coaching modified the clinicians' cognitive appraisal and coping strategies rather than the actual hazards. The authors must engage in a deeper, critical discussion of this distinction, emphasizing that secondary coping interventions must not be used by leadership to excuse or ignore hazardous systemic workloads.
In the qualitative responses, participants highlighted EHR inefficiencies and administrative burdens as major barriers to daily efficiency. To address these systemic hurdles, healthcare organizations are increasingly turning to Digital Health and Artificial Intelligence solutions to automate clinical documentation. However, introducing technology without proper training can increase cognitive load. Please discuss this using the study by https://doi.org/10.1093/eurpub/ckaf161.728 (Mapping AI and Digital Health training: are we preparing the public health workforce?) to argue that administrative burnout mitigation requires pairing well-being programs with structured digital literacy and technological training.

Limitations and Conclusions

The limitations identified by the authors (lack of generalizability, attendance rates, coaching variability) are highly appropriate.

The authors must explicitly add the lack of long-term follow-up as a limitation. Surveys were completed immediately post-intervention. In public health and occupational medicine, it is vital to know if the cognitive improvements in workload perception and self-efficacy persist 6 or 12 months after the coaching program ends.
Given the single-arm design, potential selection bias, and the fact that the final sample size fell slightly below the calculated power, the conclusions must be phrased more conservatively. The authors should tone down any causal assertions and frame their positive findings as preliminary trends that warrant confirmation in future multi-center randomized controlled trials.

Author Response

Reviewer #3

I would like to thank the authors for their interesting work. Beyond the value of the data reported, I will refer to some critical issues below:

Introduction

1. From an occupational health and preventive medicine perspective, burnout interventions must be categorized. While structural workplace modifications represent primary prevention, individual-focused interventions like coaching are secondary prevention (enhancing individual coping and cognitive appraisal). The authors should explicitly clarify in the introduction that while coaching is a highly valuable tool to mitigate the consequences of work-life incongruence, it serves as a secondary prevention strategy and does not replace the primary prevention need for systemic organizational reforms.

Thank you for this insight. We agree, and have modified the text to explain this difference, “Interventions to improve burnout include primary prevention using structural workplace modifications such as reductions in workload¹¹ and secondary prevention such as mindfulness-based interventions,¹² and enhancing individual coping and cognitive approaches through coaching.”

2. While the overall objective of evaluating the coaching program is clear, the introduction lacks explicit hypotheses regarding the secondary outcomes (AWS and PCQ dimensions). The authors should state whether they hypothesized improvements in specific subscales (control, efficacy) even in the absence of a global reduction in Maslach Burnout Inventory (MBI) scores.

Thank you for this comment, we clarified that we hypothesized an improvement on the surveys administered in the study, “We hypothesized that a coaching intervention would improve participant scores on the MBI²¹ and AWS²² with improvements in the PCQ,²³ PSS,²⁴ FFWEL.²⁵ Table 1”

3. To strengthen the international relevance of this work, the introduction should briefly frame how physician burnout and professional fulfillment have evolved globally in the post-COVID-19 era. Please integrate a brief discussion of recent international findings on post-pandemic physician burnout, such as the cross-sectional study by https://doi.org/10.3390/healthcare14040454 (Burnout Among Italian Medical Doctors: A Cross-Sectional Study on Emotional Exhaustion, Depersonalization, and Gratification Post-COVID-19). This will help illustrate that high levels of emotional exhaustion and depersonalization remain a persistent global threat post-COVID-19, and will set the stage for discussing how professional gratification and personal accomplishment function as vital coping resources.

Thank you for this comment, we have added the following to the discussion, “Furthermore, the effects of the COVID-19 pandemic may have heightened burnout in clinicians,¹¹ particularly related to high levels of emotional exhaustion and depersonalization.¹¹ Although some studies note a decrease in burnout post-pandemic, levels remained higher than prepandemic levels and certain specialties such as primary care reporting the highest levels of burnout.¹²”

Materials and Methods

1. This is the primary methodological limitation of the study. Because this is a single-arm, pre-post observational design without a non-intervention control group, it is impossible to definitively rule out confounding external factors (seasonal changes in patient volume, clinical schedule adjustments, or regression to the mean, particularly for those starting with extreme burnout scores). The authors must explicitly acknowledge this in the Methods section and adopt a more cautious tone when discussing causal relationships.

Thank you, we have clarified this as a limitation, “As this study was an observational study (with a pre- post- analysis of a single group of physicians and students), we cannot determine causality. Furthermore, we cannot determine whether coaching or other environmental changes (or other confounding external factors) in the Military Health System may have affected participant responses on surveys.”

2. The study sample pools medical students, residents, junior faculty, and senior faculty. From an occupational health standpoint, the daily duties, levels of clinical autonomy, and systemic stressors of a medical student are fundamentally different from those of an attending physician. While stratifying the actual coaching groups by training level was an excellent decision, analyzing them as a single group in the pre-post analysis introduces substantial heterogeneity. The authors should justify why these distinct career stages were pooled for the overall quantitative analysis.

We agree, and while we designed the study to include trainees (students and residents) and physicians (junior and senior physicians), we did complete a subgroup analysis of these 2 groups as noted in the results section, “Participant MBI-HSS(MP) mean scores were aggregated based on level of training into two groups, divided into trainees (students and residents) and faculty (junior and senior faculty) to assess differences between groups, and compared pre-coaching and post-coaching as shown in Supplement 4. Trainees had higher rates of engagement than faculty at baseline, but faculty engagement increased after coaching, Supplement 4.”

We also added the following to the discussion, “

When comparing trainees (students and residents) to practicing physicians (junior and senior faculty), we identified that although trainees had higher levels of engagement with the initial surveys, the faculty engagement increased following coaching sessions. While we cannot determine if this may be due to increased faculty participation in coaching or other confounding factors, this difference is interesting particularly considering the differences in duties and responsibilities in the healthcare system and suggests that interventions including coaching should be tailored to healthcare professionals’ needs.

3. The manuscript states that "coaches determined the content of the coaching sessions tailored to their specific audiences and group interest". While beneficial for engagement, this introduces a high degree of inter-group variability and a lack of intervention standardization. The authors must discuss how this variability in curriculum and coaching styles might have influenced the outcomes.

Thank you, we added the following to the paper, “However, the variability of coaching topics and coaching techniques may have limited our assessment of coaching efficacy. The diversity of participants included in this study was another potential limitation related to intergroup variation and lack of interventions standardization. While including diverse groups allowed us to provide coaching to students and physicians at different stages in their careers, students and residents face different challenges and experiences than physicians, and some of these differences may have impacted participation, with resident physicians having the lowest participation of all groups.

4. The power analysis indicated that 62 participants were required to detect a significant decrease in burnout. Although the authors planned to enroll 80 individuals to account for attrition, they only enrolled 66, and only 59 completed both pre- and post-surveys. Technically, the study is slightly underpowered relative to its initial statistical plan. The authors must explicitly state this epidemiological limitation in the methods and limitations section.

Thank you we have added the following to the limitations section, “. “Due to attrition, with some participants choosing not to complete post-coaching surveys, we were underpowered, with only 59 participants compared to the calculated 62 participants.”

Results

The results are presented clearly, and the use of tables and the LPA profile plot is highly informative.

1. Out of 66 enrolled participants, 7 (10.6%) failed to complete the post-intervention surveys. The authors must report whether a sensitivity analysis was conducted to compare the baseline demographics or baseline burnout scores of these 7 drop-outs against the 59 completers. This is crucial to rule out potential attrition bias.

Thank you for this suggestion. We identified that one person had been double-counted (they signed up with 2 different email addresses, so we updated Figure 1 to reflect 69 people expressing interest, 4 declining to participate, and 65 completing pre-surveys, with 6 dropouts (not completing post-surveys), and a total of 59 participants (unchanged). For these 6 dropouts (9.2%), we completed a sensitivity analysis comparing demographic information between the dropouts and participants who completed the study and found no significant differences. We added the following to the results:

“Six participants did not complete post-coaching surveys, and were not included in the analysis. A sensitivity analysis was conducted comparing the baseline demographics of participants who dropped out of the study to those who completed the study, and there were no significant differences at the 0.05 level.”

2. The authors mention extracting common themes from open-ended questions. However, they do not describe the formal qualitative methodology used (thematic analysis according to Braun & Clarke, double-coding to assess inter-rater reliability, or the use of qualitative software). The authors must provide details on how these qualitative data were systematically analyzed to ensure scientific rigor.

Thank you for this comment, we clarified the following in the methods section, “. Themes were coded by LMF, with review and agreement among all authors. Analysis was conducted using thematic analysis,33 themes were coded by hand and organized using a Microsoft Excel spreadsheet.”

3. A striking finding of the LPA is that all three profiles, including Profile 3 ("burnout"), demonstrated high levels of Personal Accomplishment. This is an atypical finding in general burnout literature. The authors should discuss whether this represents a self-selection bias (highly motivated individuals volunteering for coaching) or a unique sociocultural feature of military medical officers.

Thank you for this comment, we agree, and have added the following to the discussion, “It was also interesting that we found high levels of personal accomplishment on the LPA for all 3 profiles, including the profile with high levels of EE and DP (Figure 2). Though we cannot determine the exact cause, we postulate that this finding may be related to a unique sociocultural aspect of serving in military medicine, or perhaps selection bias with highly motivated individuals self-selecting to participate in a coaching program.”

Discussion

The discussion does a commendable job of linking the findings to existing theories of job congruence and psychological capital.

1. The most intriguing finding is that while global MBI scores did not change, participants in the burnout profile experienced significant post-coaching improvements in their perceptions of workload, control, and fairness. Crucially, the objective workload, staffing levels, and administrative demands of the military treatment facilities did not change during these 12 weeks. This indicates that the coaching modified the clinicians' cognitive appraisal and coping strategies rather than the actual hazards. The authors must engage in a deeper, critical discussion of this distinction, emphasizing that secondary coping interventions must not be used by leadership to excuse or ignore hazardous systemic workloads.

Thank you for this comment and insight. We agree, and have added the following to the paper, “Since coaches included boundary setting and cognitive re-framing in coaching groups, it is possible that these techniques provided new tools for participants to use in their professional lives, and that these approaches may have affected participant survey responses. Such potential improvements in cognitive framing and coping strategies may be important areas that individuals can use when faced with challenges. However, it is also critical that healthcare organizations recognize the importance of improving and correcting systemic issues to mitigate negative effects on clinicians and improve system functioning.

2. In the qualitative responses, participants highlighted EHR inefficiencies and administrative burdens as major barriers to daily efficiency. To address these systemic hurdles, healthcare organizations are increasingly turning to Digital Health and Artificial Intelligence solutions to automate clinical documentation. However, introducing technology without proper training can increase cognitive load. Please discuss this using the study by https://doi.org/10.1093/eurpub/ckaf161.728 (Mapping AI and Digital Health training: are we preparing the public health workforce?) to argue that administrative burnout mitigation requires pairing well-being programs with structured digital literacy and technological training.

Thank you for this suggestion. We have modified the paper to include the following:

“In the open-ended responses, participants frequently cited administrative burdens and electronic health record (EHR) inefficiencies as sources of inefficiency and frustration. Burdensome EHR systems are frequently cited contributors to burnout, including excessive data entry, high volumes of messaging, and suboptimal useability resulting in additional workload beyond clinical hours.⁴³ Focused opportunities for improvements in the quality of data and alerts, and improvements in efficiency can reduce the burden on clinicians and improve EHR efficiency. Health systems are often initiating digital health and artificial intelligence solutions to automate documentation, however, this requires comprehensive training for effective integration, and structured technological training and digital literacy are often lacking.⁴⁴ Thus organizations must work to ensure wellbeing programs and digital improvements are accompanied by appropriate technological training and monitoring.”

Limitations and Conclusions

The limitations identified by the authors (lack of generalizability, attendance rates, coaching variability) are highly appropriate.

1. The authors must explicitly add the lack of long-term follow-up as a limitation. Surveys were completed immediately post-intervention. In public health and occupational medicine, it is vital to know if the cognitive improvements in workload perception and self-efficacy persist 6 or 12 months after the coaching program ends.

Thank you, we agree, and have added the following to the paper in the limitations section, “Additionally, the lack of long-term follow up was another limitation, as the surveys were conducted soon after coaching, and the coaching sessions were over a relatively short period. Thus, we cannot determine if coaching over a longer time would increase effectiveness, or if any potential cognitive improvements in workload perception or self-efficacy may persist for longer periods of time following coaching.”

2. Given the single-arm design, potential selection bias, and the fact that the final sample size fell slightly below the calculated power, the conclusions must be phrased more conservatively. The authors should tone down any causal assertions and frame their positive findings as preliminary trends that warrant confirmation in future multi-center randomized controlled trials.

Thank you for this suggestion, we have added the following to the conclusion: “. The trends we noted following a coaching intervention require additional confirmation in future multi-center randomized controlled trials in different practice settings.”

Round 2

Reviewer 2 Report

Comments and Suggestions for Authors

Excellent work on the edits as I believe it has contributed to a stronger manuscript. Minor comments below:

Materials and Methods

Line 164-165: It appears the sentence "Only participants who completed the pre- and post-coaching..." is a duplicate as it is also written later (lines 179-180). Can remove one of them.

Results

Line 211: Regarding the mean number of sessions attended, I would suggest using the overall mean (4.17) SD (1.75) instead of the range based on level of experience.

Discussion

Line 353: I don't necessarily agree that it's an interesting finding that there were high baseline levels of PA for all profiles. It is quite well documented in burnout literature that healthcare providers have a higher baseline PA, which is why PA is not typically used (instead we use high EE and/or high DP) for overall burnout assessment for healthcare providers/physicians.
Excellent work with addressing all limitations- this section is very comprehensive

Author Response

Excellent work on the edits as I believe it has contributed to a stronger manuscript. Minor comments below:

Materials and Methods

Line 164-165: It appears the sentence "Only participants who completed the pre- and post-coaching..." is a duplicate as it is also written later (lines 179-180). Can remove one of them.

Thank you, the duplicate was removed.

Results

Line 211: Regarding the mean number of sessions attended, I would suggest using the overall mean (4.17) SD (1.75) instead of the range based on level of experience.

Thank you for this suggestion, the paper was updated to: “On average, participants attended three to four of the six coaching sessions mean (4.17) SD (1.75)”

Discussion

Line 353: I don't necessarily agree that it's an interesting finding that there were high baseline levels of PA for all profiles. It is quite well documented in burnout literature that healthcare providers have a higher baseline PA, which is why PA is not typically used (instead we use high EE and/or high DP) for overall burnout assessment for healthcare providers/physicians.

Thank you for this observation, we removed the comment that this was an interesting finding. We also note that there is a wide amount of variation in the different subscales of the MBI in physicians (https://jamanetwork.com/journals/jama/fullarticle/2702871_, and have left the rest of the discussion points in the manuscript:

“We found high levels of personal accomplishment on the LPA for all 3 profiles, including the profile with high levels of EE and DP (Figure 2). Though we cannot determine the exact cause, we postulate that this finding may be related to a unique sociocultural aspect of serving in military medicine, or perhaps selection bias with highly motivated individuals self-selecting to participate in a coaching program.”

Excellent work with addressing all limitations- this section is very comprehensive

Thank you -we agree that this has been strengthened and is comprehensive.

Reviewer 3 Report

Comments and Suggestions for Authors

Comments to the Authors

Dear Authors,

Thank you for your careful revision of the manuscript and for the effort devoted to addressing the comments raised during the previous round of review.

I have reviewed the revised version of the manuscript, “Coaching the Caregivers: A Burnout Intervention for Military Physicians and Medical Students.” The revisions have satisfactorily addressed my previous concerns and have improved the overall clarity and quality of the manuscript.

In particular, I appreciate the authors’ efforts to strengthen the presentation and interpretation of the findings and to provide a clearer discussion of their implications. The revised manuscript is now well structured and the conclusions are appropriately supported by the results.

I have no further substantive comments. In my opinion, the manuscript is suitable for publication in its current form.

I would like to thank the authors again for their thorough work in revising the manuscript and for their constructive response to the reviewers’ comments.

Author Response

Comments to the Authors

Dear Authors,

Thank you for your careful revision of the manuscript and for the effort devoted to addressing the comments raised during the previous round of review.

I have no further substantive comments. In my opinion, the manuscript is suitable for publication in its current form.

I would like to thank the authors again for their thorough work in revising the manuscript and for their constructive response to the reviewers’ comments.

-- Thank you for your comments. We appreciate your thorough review and agree that the presentation and interpretation of the findings provide a clearer discussion of study implications. No changes were made to the manuscript.