Association Between Standardized Screening and Identification of Developmental Problems Among Children <3 Years of Age

Model DescriptionLogistic Regression, Odds Ratio for >10% Identification (95% Confidence Interval)Linear Regression, Increase in Mean Proportion of Children Identified (95% Confidence Interval)
Bivariate model (standardized universal screening only)1.90 (1.11–3.25)1.04 (−0.09 to 2.18)
Multivariate model (standardized universal screening plus Medicaid participation)1.80 (0.98–3.32)0.73 (−0.49 to 1.94)
Backward stepwise model (standardized universal screening plus gender, medical school/university affiliation, and Medicaid participation)1.78 (0.95–3.32)0.83 (−0.42 to 2.08)
Full theoretical model (standardized universal screening plus gender, age, managed care and Medicaid participation, proportion of time in general pediatrics, medical school/university affiliation, and proportion of patients 0–3 y of age)1.71 (0.90–3.25)0.95 (−0.33 to 2.22)