Making a Case for an Autism-Specific Multimorbidity Index: A Comparative Cohort Study

Autistic people experience challenges in healthcare, including disparities in health outcomes and multimorbidity patterns distinct from the general population. This study investigated the efficacy of existing multimorbidity indices in predicting COVID- 19 mortality among autistic adults and proposes a bespoke index, the ASD-MI, tailored to their specific health profile. Using data from the CVD-COVID-UK/COVID-IMPACT Consortium, encompassing England's entire population, we identified 1,027 autistic adults hospitalized for COVID- 19, among whom 62 died due to the virus. Predictors were selected using logistic regression with fivefold cross-validation, comparing AUCs amongst multimorbidity indices. Diabetes, coronary heart disease, and thyroid disorders were selected as predictors for the ASD-MI, outperforming the Quan Index, a general population-based measure, with an AUC of 0.872 versus 0.828, respectively. Notably, the ASD-MI exhibited better model fit (pseudo-R2 0.25) compared to the Quan Index (pseudo-R2 0.20). These findings underscore the need for tailored indices in predicting mortality risks among autistic individuals. However, caution is warranted in interpreting results, given the limited understanding of morbidity burden in this population. Further research is needed to refine autism-specific indices and elucidate the complex interplay between long-term conditions and mortality risk, informing targeted interventions to address health disparities in autistic adults. This study highlights the importance of developing healthcare tools tailored to the unique needs of neurodivergent populations to improve health outcomes and reduce disparities.