Physician-to-physician coaching is beneficial in shaping the culture in a practice and addressing and sometimes altering specific behaviors. This approach is useful for physicians who have an increased number of patient-based complaints,(1) who need help with other interpersonal interactions,(2) who are dealing with burnout,(3,4) and to guide physicians in addressing and improving patient experience.(5,6)
We were interested in whether we could elevate patient experience scores in a safety net public hospital system and boost communication skills in a cohort of individuals. We also wanted to help individuals self-identify a path for improvement using a pre-shadow coaching self-assessment with a post-shadow coaching written summary that included observations and individualized tactics from the shadow coach as a surrogate for long-term repeat shadow coaching. We used ambulatory Press Ganey survey results as a measure of the impact of the self-assessment and coaching process.
METHODS
We identified all primary care physicians and advanced practice professionals (APPs) (n = 111) at a public health system who had active patient experience scores in the previous 12 months. We divided the group into segments based on patient experience scores and clinical department (internal medicine or family medicine). The list of potential physicians and advanced practice professionals was reviewed by the physician coach and the leader of medical group practice.
We asked clinicians who had patient experience scores at least once in the second quartile in the previous 12 months (n = 39) if they would be willing to participate in one-on-one coaching with a physician. We titled this program a Clinician Mastery Program and focused the approach on professional development to reduce the potential perceived stigma of a clinician participating in a coaching program.
Each candidate received an invitation to participate (email and phone call) from the physician coach. The email was concise, and the phone call was designed to explain the reason for the ask, explain what was anticipated in terms of time and engagement from the physician or advanced practice professional, and determine if the physicians were interested in participating.
Eleven clinicians volunteered to participate. This cohort included one specialist physician — a hematologist-oncologist. The clinical chairs for these clinicians were informed so the clinical leaders were aware and supportive of these individuals’ participation in the program.
Physician Coaching
The physician coach was the health system’s top-scoring physician in patient experience scores. This individual was a family medicine physician who had a full-time practice and had not previously engaged in coaching training. To allow this physician to participate as a coach, his responsibilities were altered to 60% clinical and 40% other duties to include one-on-one coaching.
The coach contacted the organizational development segment of the organization and created a unique self-assessment designed for each individual to categorize themselves across several domains relevant to patient interactions (Figure 1).
Figure 1. Self-Assessment Completed by All Individuals in the Cohort
Clinicians took approximately 20 minutes of non-patient-facing time to complete the self-assessment. They returned the assessment to the physician coach before any formal teaching, coaching session, or observation session.
Each individual had one-on-one scheduled time with the physician coach to review the self-assessment and to schedule the shadow coaching time during a regularly scheduled clinic. At this session, the physician coach reviewed the self-assessment and asked the clinicians specific questions as to their concerns and desired areas of focus. The physician coach also explained the intent of the at-the-elbow clinic session and outlined a script so that the clinician could explain to the patient the reason for the coach’s presence.
The physician coach scheduled two half-day clinic sessions with each clinician for a period of two to three hours each. In the clinic room, the physician coach sat to the side as the clinician provided an introduction and description to the patient as to the purpose of the physician coach.
After each session, the physician coach compiled a set of notes that included comments related to introduction, communication, service, and transition to checkout. Various sub-categories included language, pace of conversation, teaching and teach back, eye contact, personalization, meeting patient needs, and explaining next steps and ensuring following up.
The physician coach set up follow-up time with the physician or APP to review the qualitative assessment of the clinic interactions and discuss potential strategies that the clinician could use to further improve patient engagement during the clinic visit. This discussion also included tips for the physician or APP to practice in subsequent clinic sessions with regard to the observations from the coach.
Based on these parameters, this work was designed and presented as a pilot quality improvement project and as such institutional review board review was not required.
The organization had recently transitioned to Press Ganey as a vendor for patient experience. Each clinician’s patient experience scores for seven domains (concern for questions and worries, discussed treatment, made efforts to include in decisions, explained problem and conditions, extent to which provider listened, likelihood of recommending the provider, time spent with patient) were evaluated as a percent in the particular domain before and then at 30 days and at 90 days post-observation and coaching.
Statistical Methods
Patient Experience Scores Across Domains
Data were analyzed using Microsoft Excel. Data are reported as mean ± standard deviation (SD). Composite percentile scores for each clinician were created by summing all domain percentile scores and calculating an average ± SD.
Individual domain scores were calculated summing all of the physician and APP scores pre-intervention, at 30 days post-in clinic observation, and at 90 days post-in clinic observations for each domain and calculating an average ± SD.
A paired t-test was used to assess whether there was a significant change in scores from pre- to post-intervention. Statistical significance was set at p<0.05.
Calculation of Return on Investment (ROI)
Reduction in no-show rate was chosen as the parameter to base an ROI calculation.(7) A conservative percentage reduction (3.5%) for clinicians who demonstrated an increase in patient experience scores over the intervention timeframe was imputed for nine months subsequent to the intervention timeframe.
The standard Medicare clinic charge for a 15-minute visit and uncovered amount published for the health system was used as the financial parameter for the calculation. This clinic charge value was less than the values provided online for the health system for other payers. The compensation offset for the physician coach was also noted.
Given that the self-assessment and non-clinic visit coaching interactions occurred outside of patient-facing clinic time, there were no additional costs incurred from these activities. Also, the physician coach worked to increase clinic volume on the remaining clinic days and to distribute other potential patients to other clinicians to avoid any reduction because of lost patient visits.
The ROI calculation was the aggregate of the change in volume based on the decreased no-show rate multiplied by the base clinic charge minus the cost of physician compensation.
RESULTS
The seven women and four men in the cohort practiced at different clinic sites; all had been in practice for five years or more. Each completed the self-assessment before their coaching experience.
Based on the self-assessment review, there was no significant aggregation of self-defined areas of excellence or opportunity among the cohort. The pre-coaching and observation patient experience scores in the various domains were totaled and evaluated across 111 surveys in the month before the observation and coaching session. These included:
Clinician concern about questions and worries.
Clinician discussed treatment.
Clinician made efforts to include the patient in decisions.
Clinician provided explanation of problem/conditions.
The extent to which the clinician listened.
Likelihood of recommending the clinician.
Time the clinician spent with the patient.
The percent scores ranged from a low of 63.05 ± 13.1% (mean ± SD) for time spent with the patient to a high of 70.23 ± 13.6% for clinician concern for questions and worries.
The percent for each domain increased at 30 days post-in clinic observation (N = 132 survey results). The average increase in percent was 4.43 ± 2.2% across all domains and the majority of individuals in the group showed an increase in composite scores.
Interestingly, the change at 30 days for discussed treatment (pre: 65.3 ± 12.2; 30 days post: 75.6 ± 16.2, p = 0.021), efforts to include in decisions (pre: 67.4 ± 13.1, 30 days post: 77.6 ± 11.96, p = 0.024), and time spent with patient (pre: 63.05 ± 14.09; 30 days post: 74.33 ± 11.1, = 0.046) all were significantly different. The other parameters were not significantly different at 30 days.
The change at 90 days post-review and coaching (N = 159 survey results) compared to pre-review and observation was an average increase of 6.32 ± 1.83% (Figure 2). The parameters that had been significantly different at 30 days were not significantly different at 90 days.
Figure 2. Cohort Scores for Each Domain (mean ± SD) Pre-and Post-90 Days Observation and Coaching
Given the size of the cohort, it was also possible to examine intra-individual changes across their percent scores between pre- and 90 days post-observation and coaching. Sixty-four percent (seven of the 11 in the cohort) had an increase in aggregate percent scores when collating all domain scores for each individual (Table 1). Five demonstrated a significant improvement while notably three individuals demonstrated a significant decline, including the only two APPs in the cohort.
We also examined the potential ROI for this intervention. There has been a noted association between improvements in patient experience and reductions in no-show rates over time.(7) Given a 3.5% reduction in no-shows over the nine months subsequent to the study timeframe, a conservative estimate based on the published data, and an average of $210 per visit, the potential benefit of the program in clinic visit was calculated for just the clinicians who demonstrated an increase in patient experience scores at $127,008. This was more than the $108,000 cost of the physician coaching and survey completion time.
DISCUSSION
A one-on-one physician coaching intervention in a safety net ambulatory environment improved patient experience scores among a subset of physicians and APPs. For the group that improved, this effect was evident early and appeared to be consistent through 90 days after in-clinic observation and coaching.
Interestingly, a subset of individuals, including the two APPs in the cohort, had a decline in patient experience scores, suggesting that the intervention was not effective for them, possibly because of a different emphasis in terms of physician-patient versus APP-patient interactions.
For the individuals with sustained improvement, the effect appeared to be sustained over 90 days even though no particular domain in the patient experience scores stood out over that time frame. Some of this may be attributed to the use of the self-assessment before the observation and coaching session.
This was a novel feature of this study, as is the estimate of return on investment of this approach in this health system. The self-assessment could have allowed the cohort to be more thoughtful with regard to personally defined areas for improvement and to do so repeatedly following the observation and coaching session. Arguably, this contributes to the sustainability of the impact of coaching.(8)
At least one study has recognized that the impact of a coaching intervention to improve patient experience scores wanes over time.(9) While it would have been helpful to examine more longitudinal data in this cohort, as well as follow-up observation and coaching sessions, this was beyond the scope of this pilot as constructed.
The coaching intervention through observation and specific notes reinforced components of interacting with patients: communication, service, and meeting patients’ needs. This was important since this approach was undertaken in a public health system with a diverse patient population and a diverse physician and APP population. Indeed, to our knowledge, this was only the second time that such an approach has been applied to an underserved population(9) — within which it is traditionally more difficult to achieve positive patient experience scores.(10)
This inherently introduced differential needs for cultural literacy, an important variable in achieving higher-level patient experience scores.(11) The data suggest a possible need to understand how to cross cultures better, given the diversity of patient and clinician groups and may have contributed to the lack of improvement by some individuals.
Several features of this pilot make it feasible to consider implementing it more broadly. First, the self-assessment and in-clinic observation differed from simulation,(12) making this approach easier to replicate in any healthcare setting. By identifying a key practicing physician leader, it was easier to gain alignment among physicians and other clinicians because the observer and coach is “one of their own.” Finally, the ratio of one physician leader to 11 practicing clinicians worked well and was not burdensome.
There certainly are considerations when looking at the results of this pilot. This was a small cohort and thus susceptible to type I and II error. Cohort members were selected to be part of this project and this in turn could lead to a potential participation bias. The data are examined point-in-time. The intervention has a set of components: a self-assessment, observation, and coaching. It is difficult to delineate a discrete effect from each specific aspect of this composite intervention. Impact could have been different among the cohort of individuals who did not elect the coaching process. This type of analysis is important to do in any follow-up study.
Finally, there was one physician coach for this cohort. The results when using other physician coaches could potentially differ. Nonetheless, even with these caveats, this pilot approach appears to work for a subset of clinicians as a template for practice-based learning and education around patient experience and professional activity with the added benefit of a modest return on investment.
References
Webb LE, Dmochowski RR, Moore IN, Pihert JW, Catron TF, Troyer M, Martinez W, Cooper WO, Hickson GB. Using Coworker Observations to Promote Accountability for Disrespectful and Unsafe Behaviors by Physicians and Advanced Practice Professionals. Jt Comm J Qual Patient Saf/ 2016;42(4):149–64. https://doi.org/10.1016/S1553-7250(16)42019-2 .
Egener B. Addressing Physicians’ Impaired Communication Skills. J Gen Intern Med. 2008;23:1890–1895. https://doi.org/10.1007/s11606-008-0778-7 .
Kiser SB, Sterns JD, Lai PY, Horick NK, Palamara K. Physician Coaching by Professionally Trained Peers for Burnout and Well-Being: A Randomized Clinical Trial. JAMA Network Open 2024;7(4):e245645. https://doi.org/10.1001/jamanetworkopen.2024.5645 .
Boet S, Etherington C, Dion P-M, Desjardins C, Kaur M, Ly V, Denis-LeBlanc M, Andreas C, Sriharan A. Impact of Coaching on Physician Wellness: A Systematic Review. PLoS One. 2023;18(2):e0281406. https://doi.org/10.1371/journal.pone.0281406 .
Sharieff GQ. MD to MD Coaching: Improving Physician-Patient Experience Scores: What Works, What Doesn’t. J Patient Exp. 2017;4(4):210–212. https://doi.org/10.35680/2372-0247.1303 .
Quigley DD, Qureshi N, Slaughter ME, Kim S, Talamantes E, Hays RD. Provider and Coach Perspectives on Implementing Shadow Coaching To Improve Provider-Patient Interactions. J Eval Clin Prac. 2021;27:1381–1389. https://doi.org/10.1111/jep.13575 .
Aysola J, Xu C, Huo H, Werner RM. The Relationships Between Patient Experience and Quality and Utilization of Primary Care Services. J Patient Exp. 2020;7(6):1678–1684. https://doi.org/10.1177/2374373520924190 . Epub 2020 May 26. PMID: 33457630; PMCID: PMC7786755.
Quigley DD, Elliott MN, Slaughter ME, Talamantes E, Hays RD. Follow- Up Shadow Coaching Improves Primary Care Provider-Patient Interactions and Maintains Improvements When Conducted Regularly: A Spline Model Analysis. J Gen Intern Med. 2022;38(1);221–227. https://doi.org/10.1007/s11606-022-07881-y .
Quigley DD, Elliott MN, Slaughter ME, Burkhart Q, Chen AY, Talamantes E, Hays RD. Shadow Coaching Improves Patient Experience with Care, But Gains Erode Later. Med Care. 2021;59(11):950–960. https://doi.org/10.1097/MLR.0000000000001629 .
Chatterjee P, Joynt KE, Orav J, Jha AK. Patient Experience in Safety-Net Hospitals Implications for Improving Care and Value-Based Purchasing. JAMA Intern Med. 2012;172:1204–1210. https://doi.org.10.1001/archinternmed.2012.3158 .
Paez KA, Allen JK, Beach MC, Carson KA, Cooper LA. Physician Cultural Competence and Patient Ratings of the Patient-Physician Relationship. J Gen Intern Med. 2009;24:495–498. https://doi.org/10.1007/s11606-009-0919-7 .
Seiler A, Knee A, Shaaban R, Bryson C, Paadam J, Harvey R, Igarashi S, LaChance C, Benjamin E, Lagu T. Physician Communication Coaching Effects on Patient Experience. PLoS One. 2017;12(7):e0180294. https://doi.org/10.1371/journal.pone.0180294