Using Quality Improvement Collaboratives to Improve Early Identification of Autism By Primary Care Providers Serving Low Resource Urban Communities
Almost a decade ago, the AAP (2006, 2007) published a groundbreaking policy statement on surveillance and screening of autism spectrum disorder (ASD). These guidelines recommended that all children be screened for ASD using formal screeners during pediatric well-child visits at 18 and 24 months. Although limited, available data suggest that only a minority of primary care providers (PCPs) has successfully implemented the AAP’s recommendations (Volkmar & Reichow, 2014).
To investigate the preliminary efficacy of quality improvement collaboratives (QICs) for improving (1) ASD-specific screening rates during well-child visits at 18 and 24 months, and (2) self-reported knowledge/attitudes about early identification of ASD among primary care physicians.
This research was conducted in collaboration with two urban PCPs in the Eastern US: a network of five clinics (PCP#1), and a clinic affiliated with a hospital center (PCP#2). The communities served by these PCPs were primarily African American (28% and 41%, respectively) and Hispanic (31% and 49%, respectively). The QICs were implemented over a 4-month period. Activities were guided by a multi-disciplinary leadership team comprised of the researchers and three individuals from within the clinic. The team met monthly to develop, evaluate, and refine an office-wide implementation system that addresses clinic-specific barriers. In addition, the leadership team determined the content of monthly training workshops offered to the clinic staff. Both PCPs chose to implement the M-CHAT (Robins et al., 2001).
To evaluate increases in screening practices and physician knowledge/attitudes, two kinds of outcome measures were collected: 1) Monthly chart reviews were completed for a period of about 6 months, starting one month prior to the QIC. Using a random sampling approach, medical charts of children from 17-25 months at the time of well-child visit were reviewed for documentation about ASD-specific screening and referrals. For PCP#1 and PCP#2, we abstracted data from 250 and 140 charts, respectively. 2) Before and after the QIC period, physicians completed a questionnaire evaluating knowledge/attitudes about screening, communicating with families, and referral practices.
A total of 15 physicians completed the pre- and post-questionnaire measures (female: 72%; racial/ethnic minority: 80%; years in practice: M=26, SD=12.5). Results revealed significant increases in physicians’ knowledge/attitudes about screening measures, t(14)=2.7, p < .05. No significant increases were found with regard to physicians’ knowledge/attitudes about early diagnosis, strategies for communicating with families, and effective referral practices. In addition, chart review results indicated that ASD-specific screening rates increased from 16% to 76% at PCP#1, and from 24% to 39% at PCP#2. Across both PCPs, ASD-specific screening measures identified a total of 10 children at high risk for ASD. Four of these were referred by the treating physician to the local early intervention program to determine eligibility for services.
Despite advances in early identification of ASD, significant challenges persist in moving screeners from academic centers to the “real world”. The current research demonstrates the preliminary efficacy of QICs for implementing ASD-specific screening measures by PCPs in ethnically diverse and low resource communities. Future directions for optimizing this learning system will be discussed.