|Home | About | Journals | Submit | Contact Us | Français|
Cleveland physicians with in-depth knowledge about hospitals in their community were asked to rank general hospitals, on a 5-point scale, in an attempt to evaluate quality of patient care. Nine physicians showed reasonable agreement in their judgments. In a broader sample, six additional physicians ranked the same hospitals, to provide a check on overall inter-rater reliability. Average hospital rank assigned by the first group correlated +.89 with average rank for the broader sample, indicating good criterion reliability.
Later, published data about the hospitals, such as number of residency programs offered, number of beds, and average length of stay, were correlated with the rankings provided by all 15 physicians. Resulting correlations were substantial (e.g., +.86 for number of approved residency programs), and the best combination of three variables predicted the criterion with a multiple correlation of +.94. Cross validation in four other metropolitan areas in Ohio and Pennsylvania, where 75 physician raters provided ratings of 36 general hospitals, showed that the three variables isolated for the Cleveland sample predicted the overall quality of patient care in the four other communities, with a correlation of +.82.
The study was made by a team of physicians and psychologists as part of a larger investigation of practicing physicians and the doctor-patient relationship. The work reported here was directed at establishing ratings of the hospitals in which these physicians work and were trained.