Extracted text

OCR Page 1 of 63
QUALITY AND COST OF DISEASE MANAGEMENT 20 management program, but developmental costs would have to be spread over a larger program and several years to be adequately addressed. In addition, the development of the communication skills curriculum would be spread over several diseases. Even with conservative assumptions, the development costs would have a small effect on the cost-effectiveness ratio spread over more diseases, patients, and years of program operation. Effects of the Program at the Community Level Only a fraction of the physicians in the intervention communities participated in the program. About 200 physicians in the Intervention communities accounted for over 85 percent of all the patients with any service claims for asthma. However, the program reached only 65 of these physicians, and they were the ones with fewer emergency visits for asthma per 1,000 patients at the baseline. Their patients experienced relatively less severe illness when compared with the patients of untrained physicians in both the intervention and comparison communities. Thus, there was a bias against favorable outcomes for trained physicians. Despite this bias and the relatively small number of physicians trained, the program had an effect on the use of emergency services and drugs at the community level. Since the effect on emergency service visits was spread among all (trained and untrained) physicians in the intervention communities, we did not observe large, consistent changes at the community level for emergency visits. However. the effect on visits was large in the first quarter, when the program was heavily marketed to all 200 physicians by mail, faxing, and phone calls. There is an explanation for the sustained effect at the community level on drug use but not on emergency service visits. Trained physicians retained a direct link with and control over drug prescribing. For every asthma patient with a service claim, there are roughly two others without a