Evaluating Long-Term Effects of an Early Detection Program for Autism Spectrum Disorder in the Netherlands

Background: The importance of early detection of autism spectrum disorder (ASD) followed by early intervention has been increasingly emphasized for children’s adaptive, social and cognitive functioning. Early detection programs have demonstrated to lower the age at diagnosis directly after implementation. However, long-term effects are largely unknown. This study aimed to evaluate the long-term effects of an early detection program, training primary care providers and using a systematic screening procedure, applied by Oosterling et al. (2010), after a freeze of active investment.

Objectives: To investigate the long-term effects of an early detection program for ASD on the age at referral compared to non-ASD diagnoses.

Methods: The early detection program encompassed (a) training of primary care providers to recognize early signs of autism, (b) use of a systematic screening protocol including the Early Screening of Autistic Traits questionnaire, and (c) formation of a multidisciplinary diagnostic team. The effectiveness of the program was evaluated by a controlled study involving ASD (N=513) and non-ASD (N=722) referrals (aged 0-6 years) across three periods: pre-implementation (PRE; status quo), implementation (IMPL; actual implementation), and post-implementation (POST). During POST the multidisciplinary team continued to provide highly specialized mental health care service for infants and toddlers, but there was no specific effort put into training of primary care providers and use of the screening protocol. The proportion of children referred < 36 months of age versus the proportion of children 3-6 years was used as the outcome measure in binary logistic regression modeling. Predictors included diagnosis (ASD vs. non-ASD) and time point (PRE, IMPL, POST).

Results: There was a significant diagnosis by time point interaction effect (Wald chi-square (2) = 7.898, p= .019). The odds of being referred before 36 months for ASD versus non-ASD was stronger during implementation, but not before or after implementation (IMPL vs. PRE: 3.1, 95% CI 1.2-7.6; IMPL vs. POST: 1.7, 95% CI 1.0-3.0; PRE vs. POST: non-significant), with 33.3% (vs. 66.7%), 58.7% (vs. 41.3%) and 42.9% (vs. 57.1%) of the cases being referred for ASD during PRE, IMPL and POST, respectively. Post-hoc analyses revealed that the ASD referrals were 3.8 times (95% CI 1.8-7.9) and 4.0 times (95% CI 2.7-6.0) more likely to be referred before 36 months during IMPL as compared to PRE and POST, respectively. No significant differences were found between PRE and POST. In contrast, non-ASD referrals were only 2.3 times (95% CI 1.6-3.3) and 1.9 times (95% CI 1.1-3.3) more likely to be referred before 36 months during IMPL and PRE, respectively, as compared to POST. No significant differences were found between PRE and IMPL.

Conclusions: The early detection program led to earlier referral of children with ASD when corrected for other referrals during that time period, but this effect was not sustained when active investment was ended. This study highlights the importance of ongoing investment and encourages policy makers and health care managers to overcome barriers after implementation.