A Multifaceted Organizational Physician Assessment Program: Validity Evidence and Implications for the Use of Performance Data

Objective

To provide validity evidence for a multifaceted organizational program for assessing physician performance and evaluate the practical and psychometric consequences of 2 approaches to scoring (mean vs top box scores).

Participants and Methods

Participants included physicians with a predominantly outpatient practice in general internal medicine (n=95), neurology (n=99), and psychiatry (n=39) at Mayo Clinic from January 1, 2013, through December 31, 2014. Study measures included hire year, patient complaint and compliment rates, note-signing timeliness, cost per episode of care, and Likert-scaled surveys from patients, learners, and colleagues (scored using mean ratings and top box percentages).

Results

Physicians had a mean ± SD of 0.32±1.78 complaints and 0.12±0.76 compliments per 100 outpatient visits. Most notes were signed on time (mean ± SD, 96%±6.6%). Mean ± SD cost was 0.56±0.59 SDs above the institutional average. Mean ± SD scores were 3.77±0.25 on 4-point and 4.06±0.31 to 4.94±0.08 on 5-point Likert-scaled surveys. Mean ± SD top box scores ranged from 18.6%±16.8% to 90.7%±10.5%. Learner survey scores were positively associated with patient survey scores (r=0.26; P=.003) and negatively associated with years in practice (r=−0.20; P=.02).

Conclusion

This study provides validity evidence for 7 assessments commonly used by medical centers to measure physician performance and reports that top box scores amplify differences among high-performing physicians. These findings inform the most appropriate uses of physician performance data and provide practical guidance to organizations seeking to implement similar assessment programs or use existing performance data in more meaningful ways.

Measuring reflection on participation in quality improvement activities for maintenance of certification

PURPOSE

To validate a measure of reflection on participation in quality improvement (QI) activities and to identify associations with characteristics of QI projects, participants, and teams.

METHOD

This was a prospective validation study of all Mayo Clinic team participants who submitted QI projects for maintenance of certification (MOC) credit from 2010 to 2012. The authors developed a measure of reflection on participation in QI activities and explored associations between participants' overall reflection scores and characteristics of projects, participants, and teams.

RESULTS

A total of 922 participants (567 physicians) on 118 teams completed QI projects and reflections. Factor analysis revealed a two-dimensional model with good internal consistency reliabilities (Cronbach alpha) for high (0.85) and low (0.81) reflection. Reflection scores (mean [standard deviation]) were associated with projects that changed practice (yes: 4.30 [0.51]; no: 3.71 [0.57]; P < .0001), changed the health care system (yes: 4.25 [0.54]; no: 4.03 [0.62]; P < .0001), and impacted patient safety (P < .0001). Physicians' reflection scores (4.27 [0.57]) were higher than support staff scores (4.07 [0.55]; P = .0005). A positive association existed between reflection scores and the number of QI roles per participant (P < .0001). There were no associations with participant gender, team size, or team diversity.

CONCLUSIONS

The authors identified associations between participant reflection and the impact of QI projects, participants' professional roles, and participants' involvement with projects. With further study, the authors anticipate that the new measure of reflection will be useful for determining meaningful engagement in MOC.