Objective: To examine the effect of changing age at diagnosis on the diagnosed prevalence of autism among different birth cohorts.
Design: Population-based cohort study.
Setting: Children were identified in the Danish Medical Birth Registry and psychiatric outcomes were obtained via linkage with the Danish National Psychiatric Register.
Participants: All children born in Denmark from January 1, 1994, through December 31, 1999 (N = 407 458).
Main outcome measures: The age-specific prevalence, hazard ratio, and relative risk by age.
Results: Statistically significant shifts in age at diagnosis were observed for autism spectrum disorder; children diagnosed before age 9 years in the cohorts born between January 1, 1994, and December 31, 1995, between January 1, 1996, and December 31, 1997, and between January 1, 1998, and December 31, 1999, were on average diagnosed at ages 5.9 (95% confidence interval [CI], 5.8-6.0), 5.8 (95% CI, 5.7-5.9), and 5.3 (95% CI, 5.2-5.4) years, respectively. The relative risk comparing the 1996-1997 birth cohort with the 1994-1995 birth cohort at age 3 years was 1.20 (95% CI, 0.86-1.67), which decreased to 1.10 (95% CI, 1.00-1.20) at age 11 years. Similarly, the relative risk comparing the 1998-1999 birth cohort with the 1994-1995 birth cohort at age 3 years was 1.69 (95% CI, 1.24-2.31), which decreased to 1.23 (95% CI, 1.11-1.37) at age 11 years. Similar results were observed for childhood autism.
Conclusions: Shifts in age at diagnosis inflated the observed prevalence of autism in young children in the more recent cohorts compared with the oldest cohort. This study supports the argument that the apparent increase in autism in recent years is at least in part attributable to decreases in the age at diagnosis over time.