Predicting emotional valence in autism: a preregistered study in the Bayesian Brain framework.

Over the past few decades, improvements in autism awareness and diagnosis have resulted in an increasing number of individuals with autism diagnoses in older adulthood, with 1 in 45 adults having an autism diagnosis in the US [,]. Yet despite this demographic shift, we know very little about aging in autistic people, particularly regarding long-term health outcomes like mortality risk. Prospective longitudinal studies are essential for uncovering mechanistic pathways and providing in-depth clinical insights, but it is often cost prohibitive to obtain large, nationally representative samples. In contrast, administrative claims data offer a valuable opportunity to study aging in autism at a population-level, enabling efficient analyses of co-occurring condition patterns and mortality outcomes. While claims data cannot replace the depth of prospective studies, they provide a powerful tool for identifying population-level risks and informing future targeted research. Previous research shows that autistic older adults have a higher rate and risk of mortality than their nonautistic counterparts [,]. This may be attributable to the unique health risks and co-occurring condition patterns seen in autistic older adults. For example, autistic individuals have higher rates of diabetes, dementia, and cardiovascular conditions, which may contribute to mortality differently among autistic individuals compared with the general population [,]. Consequently, researchers may underestimate the mortality risk for autistic older adults if they rely on tools developed for the general population. A commonly used tool in observational claims data research to predict mortality is the Charlson Comorbidity Index (CCI). The CCI was developed in 1987 to predict one-year mortality of breast cancer patients based on 19 weighted conditions, providing a valuable tool for assessing patient risk in clinical and research settings []. Researchers developed a weighted index to account for the number and seriousness of co-occurring conditions. The weighted sum of these conditions contributes to the CCI score, which can be utilized in regression models to control for mortality risk. The CCI has been updated over the years, including accommodating changes in coding systems, such as the International Classification of Disease (ICD) from the 9th and 10th editions (ICD-9, ICD-10), and improving predictive accuracy [,]. In 2011, Quan and colleagues used administrative records from inpatient hospitalizations to update and validate the CCI; this resulted in 12 weighted conditions with strong predictive ability for 30-day (area under the curve [AUC] = 0.88) and 1-year (AUC: 0.90) mortality for the general population []. The CCI provides important foundational work for the healthcare field by offering a standardized method to quantify the burden of co-occurring conditions and predict mortality risk []. Its widespread use across clinical research and health services has enabled comparisons of patient outcomes, informed risk adjustment in large datasets, and guided resource allocation [].