Table 4.

Multivariable logistic model for primary care physician participation among all recruited physicians (n = 324)

		Odds Ratio	95% Confidence Interval
Office characteristics
No. eligible patients ≥ 90th percentile*		0.55	0.21	1.42
Chlamydia screening rate**
	Lower tertile	-	-	-
	Middle tertile	1.29	0.73	2.31
	Upper tertile	2.26	1.23	4.16
No. eligible physicians per office ≥ 90th percentile†		1.46	0.66	3.22
Physician characteristics
Age > 55 years		2.31	1.09	4.93
Female, mean (%)		1.57	0.92	2.70
	Ethnicity
	White, mean (%)	-	-	-
	African American, mean (%)	1.82	0.43	7.71
	Asian, mean (%)	0.85	0.35	2.02
	Hispanic, mean (%)	1.22	0.43	3.45
	Specialty
	Internal medicine, mean (%)	-	-	-
	Family medicine/general practice, mean (%)	1.09	0.64	1.85
	Pediatrics, mean (%)	0.46	0.20	1.03
Doctor of osteopathy, mean (%)		0.65	0.33	1.28
International medical graduate, mean (%)		0.57	0.28	1.16

Based on generalized estimation equations with logit link accounting for clustering of physicians within offices.

Recruited physicians provided their email address for subsequent contact. Participating physicians logged on to the study Internet site. The number is reduced due to missing data.

^* Dichotomous variable indicating whether number of eligible patients for a given office was ≥ 90th percentile for number of eligible patients for all offices. Patient eligibility for Chlamydia screening was defined by HEDIS Technical Specifications, 2000.

^** Offices classified according to Chlamydia screening rate tertiles. Chlamydia screening rates were determined from HEDIS Technical Specifications, 2000.

^† Dichotomous variable indicating whether number of physicians in the office of primary care physician was ≥ 90th percentile for number of physicians in all offices. Eligible physicians had at least 20 eligible female patients.