Validity of Physician and Administrator Peer Review for Selecting Specialist Narrow Network Participants

Increasingly, insurers control costs by selectively contracting with high-performing, low-cost providers through "narrow networks." Health care systems are faced with selecting specialists for inclusion in these shared-risk contracts. One approach is to use peer ratings, but it remains uncertain whether physicians can identify which of their peers are quantitatively high-quality, highly efficient care providers. If the results of administrator and physician peer ratings were a reliable stand-in for empirical scores, then they might substitute for the resource- and time-intensive task of quantitative performance ratings. This study examines whether peer ratings correlate with empirical ratings from a multisource, quantitative performance score derived from electronic health record (EHR) data. The authors evaluated 74 specialists' quantitative performance using a composite of metrics derived from EHR data. A single-item survey was administered separately that asked 32 primary care physicians (PCPs) to rate the specialists on their orientation to value-based health care. Bivariate and multivariate relationships were assessed between mean PCP ratings and quantitative performance scores, and logistic regression with receiver operating characteristic curve analysis was employed to evaluate the accuracy of peer review in classifying specialists scoring in the top quartile of quantitative performance. Correlations between PCP ratings and quantitative scores were positive and significant (r = 0.70; P < 0.0001). Multivariable linear regression explained 52% of the variance in quantitative scores. Peer ratings predicted top-quartile performance with an accuracy of 82%. Physician and administrator peer ratings identified top-performing specialists fairly accurately, suggesting that PCPs are well positioned as both stewards and assessors of specialist efficiency and quality.