Implementing preventive services: to what extent can we change provider performance in ambulatory care? A review of the screening, immunization, and counseling literature.

Strategies to improve the delivery of preventive care often consist of office-based interventions, which are designed to modify provider behaviors or practice patterns. We report on a meta-analysis of 117 behavioral outcomes extracted from 43 studies. Meta-analytic techniques were used to express the results in a common metric, which allowed quantitative comparisons across outcomes. Studies were examined by domains of preventive care (screening, immunization, and counseling) and divided into two groups based on unit of analysis (provider or patient categories). The mean effect size reflects the difference in proportion of physicians providing the targeted behavior between the experimental and comparison groups. In the provider category, the weighted mean effect size for screening was .14, for immunization was .18, and for counseling was .28. In the patient category, the weighted means for screening and immunization were .12 and .15, respectively, but were smaller for the counseling (.08). Because tests for homogeneity of effect sizes were rejected in the patient category, caution in interpreting mean effect sizes is warranted because of variability across individual values. In summary, office-based interventions were found to have positive effects on providers' adherence to preventive recommendations. We discuss the methodological issues and needs for future work to enhance the delivery of preventive services.