Abstract Objective To identify parent, child, community, and health care provider characteristics associated with early intervention (EI) referral and multidisciplinary evaluation (MDE) by EI. Methods We conducted a mixed methods secondary analysis of data from a randomized controlled trial of a developmental screening program in 4 urban primary care practices. Children <30 months of age not currently enrolled in EI and their parents were included. Using logistic regression, we tested whether parent, child, community, and health care provider characteristics were associated with EI referral and MDE completion. We also conducted qualitative interviews with 9 pediatricians. Interviews were recorded, transcribed, and coded. We identified themes using modified grounded theory. Results Of 2083 participating children, 434 (21%) were identified with a developmental concern. A total of 253 children (58%) with a developmental concern were referred to EI. A total of 129 children (30%) received an MDE. Failure in 2 or more domains on developmental assessments was associated with EI referral (adjusted odds ratio AOR 3.15, 95% confidence interval CI 1.89–5.24) and completed MDE (AOR 2.16, 95% CI 1.19–3.93). Faxed referral to EI, as opposed to just giving families a phone number to call was associated with MDE completion (AOR 2.94, 95% CI 1.48–5.84). Pediatricians reported that office processes, family preference, and whether they thought parents understood the developmental screening tool influenced the EI referral process. Conclusions In an urban setting, one third of children with a developmental concern were not referred to EI, and two thirds of children with a developmental concern were not evaluated by EI. Our results suggest that practice-based strategies that more closely connect the medical home with EI such as electronic transmission of referrals (eg, faxing referrals) may improve completion rates of EI evaluation.