Table 2.

Multiple linear regression analysis of the effect of academic detailing on (1) knowledge of risk factors at post-test and (2) barriers to breast cancer screening (N = 168 for both)*

	Effect of Academic Detailing on Knowledge of Risk Factors at Post-test				Effect of Academic Detailing on Barriers to Breast Cancer Screening
	Beta^†	95% CI^‡	R²Δ^§	P	Beta^†	95% CI^‡	R²Δ^§	P
No. years of medical practice	−0.10	−0.03, 0.007		.25	−0.02	−2.75, 2.65		.81
Medical school	0.08	−0.21, 0.62		.33	−0.01	−0.90, 0.78		.89
Percentage of patients insured by Medicaid or Medicare	−0.18	−0.02, 0.00		.04	0.05	−0.009, 0.02		.52
Percentage of patients insured by managed care	0.03	−0.005, 0.008		.70	−0.09	−0.02, 0.005		.24
Baseline knowledge of risk factors for breast cancer^‖	0.35	0.24, 0.68	0.18	<.00001	n/a
Baseline barriers to breast cancer screening^¶	n/a				0.60	0.44, 0.77	0.2	<.0001
Intervention^**^†† ^f,g	0.23	−0.14, 1.06	0.04	.01	−0.48	−1.62, 3.75	0.13	<.00001

* Using self-report data, after completion of academic detailing intervention.
† Beta is the standardized regression coefficient, ie, a standardized measure of the change in outcome attributable to one predictor with the remaining predictors held constant.68
‡ 95% confidence intervals
§ R²Δ is the change in the percentage of variation in the outcome explained by all of the predictors in the model. R²Δ is for the full model including all of the listed factors except intervention.
‖ Range, 1 to 7.
¶ Range, 0 to 8.
** Intervention model with knowledge of breast cancer risk factors as outcome adjusted for baseline knowledge of breast cancer risk factors, number of years of practice, whether attended US medical school, follow-up barriers to breast cancer screening, percentage of patients enrolled in managed care, percentage of patients enrolled in Medicare or Medicaid, percentage of patients with other insurance, overall model, P < .00001.
†† Intervention model with barriers to breast cancer screening as outcome adjusted for baseline barriers to breast cancer screening, follow-up knowledge of breast cancer risk factors, number of years of practice, whether attended US medical school, percentage of patients enrolled in managed care, percentage of patients enrolled in Medicare or Medicaid, percentage of patients with other insurance, overall model, P < .00001