Thomas F. Lyons

The relationship of physicians' medical recording performance to their medical care performance.

Within eight diagnostic categories on an individual case level of analysis, separate indices of recorded physician conformance to predetermined criteria for optimal care were derived. One index for each diagnosis was composed of measures that depended upon a physicians medical record recording and another was composed of measures recorded independently of the physician. Correlation coefficients between indices averaged +0.23 with a range from -0.03 to +0.53. The magnitudes of the correlations were related to the estimated reliabilities of the subindices. These data, congruent with previous research, indicate that measures of good medical recording performance and good medical care performance are related, that the overall relationship is not at all perfect, and that the reliabilities of the measures may attenuate the degree of obtained relationships.

Domain of practice and the quality of physician performance

This study has attempted to determine the relationship between physician domain of practice and their quality of medical care. The study examined whether there was a quality difference in practice either 1) between general practitioners and specialists, or 2) between specialists practicing within and outside their specialty domains. The sample consisted of 454 physicians of Hawaii, involving 18 specialty categories: general practitioners (133) and specialists (321). The study finds that when the general practitioners in this study practiced without limitation, their quality of care was not as good as that of the specialists. The study also found that when the specialists practiced outside their specialty areas, the relative quality of their performance declined. The latter finding was sustained even when other important practice and background variables were held constant. The study also discusses its generalizability and policy implications.

The Use of Item Importance Weights in Assessing Physician Performance with Predetermined Criteria Indices

Conditions for using single measures or combinations of measures of physician performance are discussed. Studies reporting combination indices of proficiency are reviewed for their use of equally or differentially weighted components and for the use of these with many or with single diagnoses. Comparisons of diagnostic-specific measures are made using differential and equal item weights. Under the general conditions of this research, few psychometric differences existed. The use and acceptability of differential item weights in this research are also discussed.

Interrelationships of physician performances: technical quality and utilization and implications for quality and utilization controls.

The study determined the intrarelationships and interrelationships of two sets of physician performance measures: technical quality and utilization of medical resources. The performance measures were based on 2,431 episodes in 15 diagnostic categories involving 454 physicians. A series of correlations were used to determine these relationships. The technical performance of physicians for prehospitalization, hospitalization, and posthospitalization care were found to be positively related (&OV0401;= .38). The interrelationships between efficient utilization and technical quality were also positive (&OV0401;= .22). There appears to be some homogeneity in the performance in the areas of technical quality and utilization patterns, but this is not strong enough to allow substitution one for another. It is therefore recommended that efforts to control or measure quality of care and utilization patterns need to be directed to each aspect of physician performances.

Interdiagnosis relationships of physician performance measures in hospitals.

Correlations among ten diagnostic categories of hospital care physician performance measures are reported. Using measures of conformance to predetermined criteria for optimal performance, mean scores of individual physician performances within diagnostic categories were calculated and correlated. Measurement reliability estimates were computed and it was suggested that a minimum of four cases in each diagnostic category he used for measurement of performance on the individual physician level of analysis. There appeared to he homogeneity of performance measures among some but not all diagnoses studied. This finding reinforces the need to examine interdiagnostic correlations before attempting to measure overall individual physician performance by combining measures from separate diagnostic categories.