Abstract
Background:
Neurocognitive disorders (i.e., dementia) are a leading cause of cognitive decline and loss of independence among older adults. While reported rates are higher among autistic adults, it is unclear whether this disparity persists after accounting for known risk factors.
Objective:
We compared neurocognitive disorder risk between autistic and non-autistic older adults after adjusting for known risk factors and evaluated whether risk factors moderated this disparity. We replicated our analyses among subsets of autistic older adults with and without co-occurring intellectual disability (ID).
Design:
Retrospective longitudinal cohort study
Setting:
National Medicare Standard Analytical Files (2013-2021)
Participants:
The sample included 9,201 autistic and 18,356 non-autistic older adults aged 65 or older, who were matched on demographic and clinical characteristics.
Methods:
Our dependent variable was time to neurocognitive disorder, defined as years between age 65 or older and the date of first diagnosis. Results: Autistic older adults had a 20% higher adjusted risk of neurocognitive disorders than non-autistic older adults (95% CI=14-25%; p<0.001). Risk was highest among autistic adults with co-occurring ID (adjusted subhazard ratio [SHR]=1.46; 95% CI=1.36-1.57). The disparity between groups was amplified in the presence of most known risk factors, notably hypertension (SHR=2.04; 95% CI=1.79-2.32), high cholesterol (SHR=1.60; 95% CI=1.46-1.75), depression (SHR=1.52; 95% CI=1.42-1.62), and type 2 diabetes (SHR=1.45; 95% CI=1.36-1.55).
Conclusions:
Autistic older adults, particularly those with ID, face significantly higher risk of neurocognitive disorders even after adjusting for known risk factors. These findings emphasize that risk factors may impact the autistic population differently and highlight the need for early screening and tailored prevention strategies.
Keywords: dementia, autism, aging, Medicare, older people
Introduction
Autistic individuals experience higher rates of neurocognitive disorders (i.e., dementia)[1] than the general population,[2-4] which has raised healthcare providers’ concerns. A recent study reported neurocognitive disorder diagnoses in 35.12% of autistic older adults (65+ years) without ID and 31.22% with ID.[3] These rates are substantially higher than the 10-11% prevalence reported for the general population of older adults.[5,6] Given the higher prevalence of neurocognitive disorders in autistic individuals and the increasing number of autistic adults aging into older adulthood,[7] more research is needed to identify modifiable risk factors and inform strategies for healthy aging.
Relative to the general population, autistic older adults experience higher rates of many risk factors identified in the 2024 update of the Lancet Commission on dementia[8] such as diabetes, hypertension, hearing loss, and obesity.[2,4] A national Medicare study reported that autistic older adults had significantly higher odds of obesity (odds ratio[OR]=1.4), diabetes (OR=1.6), and hypertension (OR=2.0) compared to controls.[9] A recent narrative review further highlights that these health conditions occur at higher rates in autistic older adults and may contribute to their complex health needs.[10]
Although research on neurocognitive disorders in autism is growing, important gaps remain. Existing studies suggest aging in autism is shaped by both general aging processes and autism-specific characteristics, with additional variability introduced by co-occurring conditions such as ID.[11] However, most studies of neurocognitive disorders in autism include only those without ID, limiting what is known about those with ID.[11] Identifying neurocognitive disorders in people with ID is especially difficult because many standard assessment tools lack specificity and don’t account for pre-existing cognitive differences.[12] Additionally, medical personnel often lack the training needed to assess and support autistic adults with ID who may have differences in communication and cognitive processing.[13] More research is needed to better understand neurocognitive disorder risk in autistic people with and without co-occurring ID to inform support for this population.
Despite higher rates of risk factors among autistic adults, their contribution to higher rates of neurocognitive disorders remains unexamined. Our primary objective was to determine whether controlling for known neurocognitive disorder risk factors[8] moderated the difference in neurocognitive disorder risk between autistic and non-autistic older adults. Our secondary objective was to determine whether stratifying the autistic older adults by the presence or absence of co-occurring ID changed our results. We also explored whether the difference in neurocognitive disorder risk between autistic and non-autistic older adults varied depending on the presence of known risk factors. We hypothesized that autistic older adults continue to have higher risk of neurocognitive disorders than non-autistic older adults after controlling for known risk factors and stratifying by co-occurring ID.
Methods
Data source
We used 2013-2021 Medicare Standard Analytical Files, which is a code-limited dataset including de-identified beneficiary-level claims for 100% of Medicare Fee-For-Service beneficiaries from inpatient and outpatient facilities. Direct identifiers are excluded, but key elements like service dates, diagnoses, procedures, and county-level geography are retained. Outpatient claims capture services provided by institutional providers such as hospital outpatient departments, rural health clinics, dialysis facilities, rehabilitation centers, Federally Qualified Health Centers, and community mental health centers. These claims exclude professional services from non-institutional physicians, physician assistants, clinical social workers, and nurse practitioners, which are billed separately and not captured in outpatient claims.[14]
Study Population
We included autistic adults in this study if they: (1) were aged 65+; (2) were enrolled in Medicare A and B for ≥12 consecutive months between 2013-2021; (3) had at least one inpatient or two outpatient encounters with an autism diagnosis between 2013-2021; and (4) had at least one observed health care encounter during the 12 months before their dementia diagnosis or their last observed inpatient or outpatient encounter, to allow for assessment of risk factors. We identified autism diagnoses using International Classification of Disease (ICD) codes from the 9th edition (ICD-9) and 10th edition (ICD-10) (Appendix 2).[15,16]
Our population comparison (PC) group met the same inclusion criteria as our autistic group, except they did not have an encounter with an autism diagnosis. We used variable ratio propensity score matching to match at most two PC beneficiaries with one autistic older adult; we matched exactly on sex, race, Health Maintenance Organization (HMO) status, and known risk factors (i.e., hypertension, hearing loss, type 2 diabetes, obesity, high cholesterol, traumatic brain injury (TBI), depression, alcohol use disorder, and tobacco use).[8] We performed greedy nearest neighbor matching within a caliper of 0.2[17] on year of birth, Charlson Comorbidity Index (CCI),[18,19] county-level estimated median household income,[20] number of inpatient and outpatient visits, longitude/latitude coordinates of county geometric centroid for patient residence, and duration of observation.
Measures
Our dependent variable was time to neurocognitive disorder diagnosis, measured as the number of years between age 65+ and the date of first diagnosis. We identified ICD-9 and ICD-10 diagnoses of clinically recognized dementia and related neuocognitive disorders using General Equivalence Mapping[21] and the Healthcare Cost and Utilization Project’s (HCUP) Clinical Classifications Software Refined (CCSR)[22] category for neurocognitive disorders (Appendix 2). Beneficiaries alive at the end of the observation period without a neurocognitive disorder diagnosis were censored at the date of their last observed encounter; beneficiaries who died before receiving a neurocognitive disorder diagnosis were coded as having a competing risk event.
Our primary independent variable was autism diagnosis. Our secondary independent variables were known risk factors outlined in the 2024 update of the Lancet Commission on dementia[8] that could be measured using claims data, and included hypertension, hearing loss, type 2 diabetes, obesity, high cholesterol, TBI, depression, alcohol use disorder, and tobacco use. These conditions were identified using diagnosis codes grouped according CCSR (Appendix 2).[22]
To compare subgroups of autistic older adults with and without a co-occurring ID, we identified autistic older adults with at least one inpatient or outpatient encounter of co-occurring ID using ICD-9 and ICD-10 codes (Appendix 2).
Covariates included sex, race, HMO status, U.S. region, rurality, estimated median household income, age and year of enrollment at the beginning of the study period. We used the beneficiary’s Social Security Administration code for their county of residence at the beginning of the study to create categories of rurality based on the United States Department of Agriculture rural-urban continuum codes[23] and to estimate the median per capita household income for individuals aged 65+.[20] For U.S. census counties with missing data, mean imputation was applied to interpolate estimated median household income in that county. All other variables were obtained at the start of their observation period. We used the CCI[18,19] solely for matching purposes to ensure comparable co-occurring condition profiles across individuals, excluding dementia from the score to avoid artificially inflating CCI values among those with the outcome of interest.
Statistical Analysis
We summarized characteristics of the beneficiaries using median [interquartile range, IQR] for continuous variables and frequencies [relative percentage, %] for categorical variables. We used Fine and Gray competing risks regression[24] to compare the instantaneous risk of neurocognitive disorders between autistic older adults and the PC group, while accounting for death as a competing event, and we modeled known dementia risk factors as time-varying covariates. Models included all risk factors as well as the covariates sex, rurality, region, HMO status, race, estimated median household income, and enrollment year and age. For a detailed description of methodology, see Appendix 1.
Our secondary analysis investigated whether the risk of neurocognitive disorders between autistic older adults and the PC group was moderated in the presence of known risk factors using a series of interaction models (e.g., autism×hypertension), while adjusting for all other risk factors and covariates as detailed in the primary analysis.
All analyses were reproduced as a sensitivity analysis, conducted separately among autistic beneficiaries with and without co-occurring ID to assess whether observed differences persisted across these subgroups compared to the PC group. We assessed significance at 0.05 level and performed all analyses using SAS v9.4.
Results
This study included 9,201 autistic older adults and 18,356 matched PC older adults (Table 1). Most beneficiaries were male (68.8%), white (89.3%), or living in a metropolitan area (79.3%). After matching, incidence of neurocognitive disorders was higher in the autistic group (25.7%) compared to the PC group (11.1%), and 16.2% of the overall sample died before receiving a neurocognitive disorder diagnosis (Autistic adults=20.4%; PC=14.1%). Among autistic older adults, 47.1% had a co-occurring ID (Appendix 3).
Table 1.
Demographics, clinical characteristics, and outcomes for matched cohort of autistic and non-autistic Medicare beneficiaries
| PC group (n=18,356) |
Autistic older adults (n=9,201) |
Total (N=27,557) |
|
|---|---|---|---|
| Demographics | |||
| Male, n (%) | 12,626 (68.8%) | 6,329 (68.8%) | 18,955 (68.8%) |
| Race, n (%) | |||
| White | 16,393 (89.3%) | 8,210 (89.2%) | 24,603 (89.3%) |
| Black | 1,240 (6.8%) | 624 (6.8%) | 1,864 (6.8%) |
| Hispanic | 102 (0.6%) | 52 (0.6%) | 154 (0.6%) |
| Other | 621 (3.4%) | 315 (3.4%) | 936 (3.4%) |
| Region, n (%) | |||
| Midwest | 5,255 (28.6%) | 2,193 (23.8%) | 7,448 (27.0%) |
| Northeast | 4,343 (23.7%) | 2,792 (30.3%) | 7,135 (25.9%) |
| South | 5,699 (31.0%) | 2,466 (26.8%) | 8,165 (29.6%) |
| West | 3,059 (16.7%) | 1,750 (19.0%) | 4,809 (17.5%) |
| Rurality, n (%) | |||
| Large Metro Area (>1,000,000) | 8,388 (45.7%) | 4,307 (46.8%) | 12,695 (46.1%) |
| Small Metro Area (<1,000,000) | 5,920 (32.3%) | 3,226 (35.1%) | 9,146 (33.2%) |
| Metro Area Adjacent | 2,267 (12.4%) | 1,004 (10.9%) | 3,271 (11.9%) |
| Non-Metro Area (>2,500) | 1,261 (6.9%) | 526 (5.7%) | 1,787 (6.5%) |
| Rural Area (< 2,500) | 520 (2.8%) | 138 (1.5%) | 658 (2.4%) |
| Estimated Median household income, median (IQR) | 41,926 (35,720-49,943) | 41,782 (35,720-49,104) | 41,877 (35,720-49,510) |
| Clinical Characteristics | |||
| Age at index, median (IQR) | 65 (65-69) | 65 (65-69) | 65 (65-69) |
| Year of index, median (IQR) | 2014 (2013-2017) | 2014 (2013-2017) | 2014 (2013-2017) |
| CCI, median (IQR) | 0 (0-3) | 1 (0-3) | 0 (0-3) |
| Ever enrolled in an HMO, n (%) | 2,150 (11.7%) | 1,084 (11.8%) | 3,234 (11.7%) |
| Co-occurring ID, n (%) | -- | 4,332 (47.1%) | -- |
| Down Syndrome, n (%) | * | 96 (1.0%) | * |
| Known Risk Factors | |||
| Hypertension, n (%) | 14,200 (77.4%) | 7,113 (77.3%) | 21,313 (77.3%) |
| Hearing loss, n (%) | 2,839 (15.5%) | 1,429 (15.5%) | 4,268 (15.5%) |
| Type 2 Diabetes, n (%) | 6,655 (36.3%) | 3,336 (36.3%) | 9,991 (36.3%) |
| Obesity, n (%) | 4,755 (25.9%) | 2,383 (25.9%) | 7,138 (25.9%) |
| High cholesterol, n (%) | 12,938 (70.5%) | 6,481 (70.4%) | 19,419 (70.5%) |
| TBI, n (%) | 840 (4.6%) | 425 (4.6%) | 1,265 (4.6%) |
| Depression, n (%) | 6,708 (36.5%) | 3,367 (36.6%) | 10,075 (36.6%) |
| Alcohol Use Disorder, n (%) | 696 (3.8%) | 355 (3.9%) | 1,051 (3.8%) |
| Tobacco Use, n (%) | 1,877 (10.2%) | 946 (10.3%) | 2,823 (10.2%) |
| Outcome | |||
| Event, n (%) | |||
| Dementia | 2,039 (11.1%) | 2,363 (25.7%) | 4,402 (16.0%) |
| Censored | 13,732 (74.8%) | 4,964 (54.0%) | 18,696 (67.8%) |
| Died (competing risk) | 2,585 (14.1%) | 1,874 (20.4%) | 4,459 (16.2%) |
Unable to report per data use agreement, due to cell sizes <11
TBI = Traumatic Brain Injury; ID = intellectual disability; CCI = Charlson Comorbidity Index; HMO = Health Maintenance Organization; PC = population control
Autistic older adults had a 20% higher adjusted risk of neurocognitive disorders overall compared to the PC group (95%CI: 14%, 25%; p<0.001) (Table 2). Furthermore, in the presence of known risk factors, autistic older adults had significantly greater risk of neurocognitive disorders than the PC group. Among the risk factors examined, autistic older adults with the following risk factors experienced the most elevated neurocognitive disorder risk compared to the PC group: hypertension (SHR:2.04; 95%CI: 1.79, 2.32), high cholesterol (SHR:1.60; 95%CI: 1.46, 1.75), depression (SHR:1.52; 95%CI: 1.42, 1.62), or diabetes (SHR:1.45; 95%CI: 1.36, 1.55).
Table 2.
Adjusted hazard rates of neurocognitive disorders among autistic older adults compared to matched controls, overall and stratified by known risk factors
| Autism vs PC | Autism without ID vs PC | Autism with ID vs PC | ||||
|---|---|---|---|---|---|---|
| SHR (95% CI) | p-value* | SHR (95% CI) | p-value* | SHR (95% CI) | p-value* | |
| Overall | 1.20 (1.14, 1.25) | <0.001 | 1.04 (0.97, 1.11) | 0.29 | 1.46 (1.36, 1.57) | <0.001 |
| Hypertension | ||||||
| Yes | 2.04 (1.79, 2.32) | <0.001 | 1.54 (1.28, 1.84) | <0.001 | 2.72 (2.24, 3.30) | <0.001 |
| No | 1.09 (1.04, 1.15) | 0.98 (0.91, 1.05) | 1.30 (1.21, 1.41) | |||
| Hearing loss | ||||||
| Yes | 1.18 (1.12, 1.24) | 0.12 | 1.00 (0.93, 1.08) | 0.006 | 1.48 (1.37, 1.60) | 0.47 |
| No | 1.31 (1.16, 1.49) | 1.32 (1.10, 1.58) | 1.38 (1.15, 1.65) | |||
| Type 2 Diabetes | ||||||
| Yes | 1.45 (1.36, 1.55) | <0.001 | 1.24 (1.14, 1.36) | <0.001 | 1.76 (1.61, 1.94) | <0.001 |
| No | 0.93 (0.86, 1.00) | 0.83 (0.75, 0.91) | 1.12 (1.00, 1.25) | |||
| Obesity | ||||||
| Yes | 1.28 (1.21, 1.35) | <0.001 | 1.06 (0.98, 1.14) | 0.22 | 1.67 (1.54, 1.82) | <0.001 |
| No | 0.96 (0.87, 1.06) | 0.96 (0.83, 1.10) | 0.99 (0.86, 1.14) | |||
| High cholesterol | ||||||
| Yes | 1.60 (1.46, 1.75) | <0.001 | 1.32 (1.16, 1.49) | <0.001 | 2.23 (1.93, 2.58) | <0.001 |
| No | 1.07 (1.01, 1.13) | 0.94 (0.87, 1.02) | 1.26 (1.16, 1.37) | |||
| TBI | ||||||
| Yes | 1.23 (1.17, 1.29) | <0.001 | 1.08 (1.01, 1.15) | <0.001 | 1.48 (1.38, 1.59) | 0.18 |
| No | 0.76 (0.62, 0.93) | 0.62 (0.47, 0.80) | 1.17 (0.84, 1.63) | |||
| Depression | ||||||
| Yes | 1.52 (1.42, 1.62) | <0.001 | 1.30 (1.19, 1.42) | <0.001 | 1.84 (1.68, 2.03) | <0.001 |
| No | 0.89 (0.83, 0.96) | 0.81 (0.74, 0.9) | 1.06 (0.95, 1.18) | |||
| Alcohol Use Disorder | ||||||
| Yes | 1.23 (1.17, 1.29) | <0.001 | 1.09 (1.02, 1.17) | <0.001 | 1.46 (1.36, 1.57) | 0.58 |
| No | 0.66 (0.53, 0.83) | 0.56 (0.44, 0.73) | 1.70 (0.99, 2.91) | |||
| Tobacco Use | ||||||
| Yes | 1.26 (1.20, 1.32) | <0.001 | 1.15 (1.07, 1.24) | <0.001 | 1.42 (1.32, 1.53) | 0.002 |
| No | 0.82 (0.71, 0.94) | 0.60 (0.51, 0.72) | 2.25 (1.71, 2.96) | |||
Represents p-value from the interaction between the risk factor and autism (e.g., TBI × autism)
Note: all models controlled for sex, rurality, region, HMO status, race, enrollment year, enrollment age, and estimated median household income, as well as all remaining risk factors not included in the interaction
TBI = Traumatic Brain Injury; SHR = subdistribution hazard ratio; CI = confidence interval; PC = population control
In our sensitivity analysis stratifying by co-occurring ID, we found the risk of neurocognitive disorders was significantly higher among autistic older adults with co-occurring ID compared to the PC group (46% higher risk; 95%CI: 36%,57%; p<0.001). In contrast, the risk of neurocognitive disorders was not significantly different between autistic older adults without co-occurring ID and the PC group.
In general, findings from our secondary analysis were consistent between subsamples; that is, in the presence of known risk factors, neurocognitive disorder risk remained elevated among autistic older adults with and without ID relative to non-autistic older adults. Notable exceptions included obesity, hearing loss, TBI, and alcohol use disorder. Obesity (p=0.22) was no longer a significant modifier of neurocognitive disorder risk between autistic older adults without ID and their matched controls. Additionally, among older adults without hearing loss, autistic individuals without ID had a 32% higher risk of neurocognitive disorders compared to the PC group (95%CI:10%,58%), while this increased risk was not observed among those with hearing loss (95%CI:-7%,8%). Furthermore, we found that TBI (p=0.18) and alcohol use disorder (p=0.58) were not significant modifiers of neurocognitive disorder risk between autistic older adults with ID and their matched controls; that is, elevated risk of neurocognitive disorders was present regardless of whether these factors were present or absent.
Discussion
Neurocognitive disorders (i.e., dementia) are a leading cause of cognitive decline and loss of independence among older adults and are increasingly recognized as a critical public health concern as populations age.[25] In the general population, modifiable risk factors such as diabetes and depression have been strongly linked to increased risk.[8] However, much less is known about whether autistic older adults face elevated neurocognitive disorder risk after accounting for these factors. This study offers a unique contribution to the literature by comparing neurocognitive disorder risk between autistic and non-autistic older adults while controlling for a set of established modifiable risk factors. Specifically, we assessed whether autistic older adults experienced an elevated risk of neurocognitive disorders compared to non-autistic peers, the extent to which risk factors modified this relationship, and whether our results were consistent for autistic older adults with and without co-occurring ID.
Autistic older adults experienced an elevated risk of neurocognitive disorders compared to non-autistic peers, primarily driven by those with co-occurring ID. This suggests that ID plays a critical role in the risk of neurocognitive disorder diagnosis. Unlike a prior study which found higher neurocognitive disorder rates in autistic older adults (65+ years) without ID,[3] we observed higher rates in those with ID, possibly due to differences in data sources and longer follow-up in our dataset. Overlapping cognitive and behavioral features of autism, ID, and neurocognitive disorders may make diagnosis more difficult, increasing the risk of misdiagnosis and delayed detection.[26] Screening tools such as the Dementia Questionnaire for People with Learning Disabilities have been developed to detect neurocognitive disorders in individuals with ID by assessing changes in behavior and autonomy rather than performance relative to population norms, however, there remains a lack of standardized and widely validated measures for those with autism.[12] A recent consensus report developed by the Summit’s Autism/Dementia Working Group of 30 international, multidisciplinary experts focused on autism, intellectual disability, and aging highlighted the importance of early and ongoing assessments, use of biomarkers and neuroimaging, and accommodations such as virtual assessments in familiar settings to improve diagnostic accuracy.[27] Our findings underscore the need to integrate these tailored approaches into clinical practice, especially for those with co-occurring ID.
For our secondary objective, we found that the presence of conditions like hypertension, high cholesterol, depression, and diabetes were associated with disproportionately greater neurocognitive disorder risk for autistic older adults. These findings highlight that risk factors may not have the same consequences for all individuals, underscoring the importance of understanding how social and clinical factors influence neurocognitive disorder risk in this population. One potential explanation is that autistic adults may experience more difficulties managing chronic conditions. Prior research has shown that autistic individuals often face challenges communicating with healthcare providers, planning difficulties related to executive dysfunction, and navigating complex healthcare systems, which reduces their access to routine healthcare.[28-30] Sensory sensitivities and mental health concerns can interfere with nutritional choices, medication management, and follow-up care.[2,28] These barriers may reduce the effectiveness of managing chronic conditions such as diabetes, potentially increasing the long-term risk of neurocognitive disorders.[31]
Alternatively, certain risk factors may confer greater harm for autistic adults compared to their non-autistic peers. That is, equal exposure to risk factors such as hypertension or diabetes may not yield equal outcomes for autistic older adults compared to non-autistic peers. These differences may reflect a combination of underlying biological mechanisms and lifelong exposure to chronic stress, which can heighten vulnerability to neurodegenerative processes. For example, prior research has shown that autistic individuals may experience greater exposure to stress[32] and elevated inflammation in the brain [33] and blood [34] which accelerate “inflammaging,” a process associated with neurocognitive disorders.[35] Similar disparities in risk amplification have been documented in other marginalized populations; for instance, racially minoritized groups often experience greater neurocognitive disorder risk in the presence of the same risk factors due to intersecting structural and psychosocial stressors.[36]
Lastly, our sensitivity analysis revealed that certain risk factors had different effects between those with versus without ID. Obesity significantly heightened the disparity in neurocognitive disorder risk between autistic older adults with co-occurring ID and their controls, but not among those without ID. This may reflect challenges in managing chronic conditions like obesity in individuals with ID, who often rely on caregivers for health-related decisions [37] and use multiple psychotropic medications that contribute to obesity and neurocognitive decline.[38] Limited health literacy and difficulty accessing clear health information can further hinder effective chronic condition management in this population.[37,39] Additionally, TBI and alcohol use disorder did not significantly modify neurocognitive disorder risk among autistic adults with ID compared to their matched controls. Both TBI and alcohol use disorder rely heavily on caregiver reporting or direct communication with healthcare providers, which may be limited or unreliable in this population.[37] Moreover, alcohol use disorder was rare among autistic adults with ID (1.2%) and thus may have lacked sufficient power for this analysis. Finally, among autistic older adults without ID, neurocognitive disorder risk in the presence of hearing loss was similar to their matched controls. This may reflect better detection and management of hearing loss in this subgroup, who may be more likely to recognize symptoms, receive treatment, and reduce their risk of cognitive decline.[36,40] Conversely, hearing loss diagnosis may be more challenging in individuals with ID, which relies on self-report and the ability to complete standardized assessments.[41]
Limitations and Future Directions
This study has several limitations. Our data are based on Medicare claims, which only capture diagnoses recorded during healthcare encounters. Therefore, individuals with less healthcare access may be underdiagnosed for neurocognitive disorders and risk factors, potentially biasing our estimates. While propensity score matching helped reduce some confounding related to healthcare access by matching on variables such as HMO, income, patient residence, and healthcare utilization, residual bias may remain from unmeasured factors. For example, several known risk factors, including untreated vision loss, social isolation, air pollution exposure, physical inactivity, and educational attainment, were not available in Medicare claims.
Our sample included diagnosed autistic Medicare beneficiaries aged 65+, who were born before autism was formally recognized in the Diagnostic and Statistical Manual of Mental Disorders[42] and for whom validated diagnostic protocols for neurocognitive disorders are lacking.[27] Thus, our findings may not generalize to younger or undiagnosed autistic populations. Our findings may be biased by a survival effect, as autistic adults surviving to age 65 may represent a heathier subset of all autistic adults. Additionally, underdiagnosis of autism in older adults has been reported,[43] potentially resulting in misclassification of controls and attenuating differences between groups.
Observed differences in time to recorded neurocognitive disorder diagnosis may reflect reporting patterns rather than true differences in disease onset. Time to diagnosis may not equate to time of disease onset, as autistic older adults without ID may notice cognitive changes earlier and seek evaluation sooner than non-autistic peers due to heightened self-awareness or clinician misinterpretation.[44] In contrast, those with ID may experience delayed recognition of decline and later recorded diagnoses due to communication challenges and existing cognitive impairments.[45] Additionally, neurocognitive disorder diagnosis and risk factors rely heavily on caregiver reporting for those with ID. Future research should examine how living arrangements, such as living in residential care facilities, affect the disparity between populations.
Conclusions
This study compared the risk of neurocognitive disorders (i.e., dementia) among autistic older adults and matched non-autistic peers using a national Medicare sample. After adjusting for known risk factors, autistic older adults had a significantly higher risk of neurocognitive disorders compared to their non-autistic peers. Higher risk was most pronounced among those with co-occurring ID, though elevated risk was also observed across the broader autistic population. Several modifiable risk factors, including hypertension, high cholesterol, depression, and diabetes, were associated with elevated neurocognitive disorder risk. These findings highlight the urgent need for tailored intervention strategies to improve risk factor management among autistic individuals and to strengthen early identification efforts, especially for those with intellectual disability.
Supplementary Material
Figure 1.
SHRs and 95% CIs for the overall effect of neurocognitive disorders between autistic and PC older adults, and effects from sensitivity analyses comparing autistic older adults with (autism + ID) and without co-occurring ID (autism – ID) and their matched controls
Keypoints:
Autistic older adults had 20% higher risk of neurocognitive disorders than controls after adjusting for known risk factors.
Risk was especially elevated among those with co-occurring intellectual disability.
Findings highlight the importance of tailored screening and prevention strategies in this underserved and aging population.
Acknowledgements
We wish to acknowledge the valuable contributions of the autistic adults and family members of autistic adults who serve on our community advisory board.
Declaration of Sources of Funding
Research reported in this publication was supported by the National Institute on Aging of the National Institutes of Health under Award Number R01AG082873 (Hand and Bishop, MPIs). The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health. The funder/sponsor had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.
Footnotes
Declaration of Conflicts of Interest
None.
Contributor Information
Melica Nikahd, The Ohio State University - Center for Biostatistics, Columbus, Ohio, United States.
Madison Hyer, The Ohio State University - Center for Biostatistics, Columbus, Ohio, United States.
Beth Wolf, Medical University of South Carolina - Division of Biostatistics, Charleston, South Carolina, United States.
Brian Patterson, University of Wisconsin-Madison - Berbee Walsh Department of Emergency Medicine, Madison, Wisconsin, United States.
Lauren Bishop, University of Wisconsin-Madison - Waisman Center, Madison, Wisconsin, United States; University of Wisconsin-Madison - Sandra Rosenbaum School of Social Work, Madison, Wisconsin, United States.
Brittany Hand, The Ohio State University - Division of General Internal Medicine, Columbus, Ohio, United States.
Data Availability
The data used in this study are not publicly available due to restructions under our data use agreement (DUA). However, the analytic code will be submitted to a public repository (Dryad DOI: 10.5061/dryad.3j9kd51zv) within one year of publication, and investigators with their own Centers for Medicare & Medicaid Services (CMS) DUAs will be able to reproduce the analytical dataset.
References
- 1.American Psychiatric Association. DSM-5-TR Neurocognitive Disorders Supplement. American Psychiatric Association Publishing. https://psychiatryonline.org/pb-assets/dsm/update/DSM-5-TR_Neurocognitive-Disorders-Supplement_2022_APA_Publishing.pdf [Google Scholar]
- 2.Croen LA, Zerbo O, Qian Y, et al. The health status of adults on the autism spectrum. Autism. Oct 2015;19(7):814–23. doi: 10.1177/1362361315577517 [DOI] [PubMed] [Google Scholar]
- 3.Vivanti G, Lee W-L, Ventimiglia J, Tao S, Lyall K, Shea LL. Prevalence of Dementia Among US Adults With Autism Spectrum Disorder. JAMA Network Open. 2025;8(1):e2453691–e2453691. doi: 10.1001/jamanetworkopen.2024.53691 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Gilmore D, Harris L, Longo A, Hand BN. Health status of Medicare-enrolled autistic older adults with and without co-occurring intellectual disability: An analysis of inpatient and institutional outpatient medical claims. Autism. Jan 2021;25(1):266–274. doi: 10.1177/1362361320955109 [DOI] [PubMed] [Google Scholar]
- 5.Manly JJ, Jones RN, Langa KM, et al. Estimating the Prevalence of Dementia and Mild Cognitive Impairment in the US: The 2016 Health and Retirement Study Harmonized Cognitive Assessment Protocol Project. JAMA Neurology. 2022;79(12):1242–1249. doi: 10.1001/jamaneurol.2022.3543 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Stroup TS, Olfson M, Huang C, et al. Age-Specific Prevalence and Incidence of Dementia Diagnoses Among Older US Adults With Schizophrenia. JAMA Psychiatry. 2021;78(6):632–641. doi: 10.1001/jamapsychiatry.2021.0042 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Dietz PM, Rose CE, McArthur D, Maenner M. National and State Estimates of Adults with Autism Spectrum Disorder. J Autism Dev Disord. Dec 2020;50(12):4258–4266. doi: 10.1007/s10803-020-04494-4 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Livingston G, Huntley J, Liu KY, et al. Dementia prevention, intervention, and care: 2024 report of the Lancet standing Commission. Lancet. Aug 10 2024;404(10452):572–628. doi: 10.1016/s0140-6736(24)01296-0 [DOI] [PubMed] [Google Scholar]
- 9.Hand BN, Angell AM, Harris L, Carpenter LA. Prevalence of physical and mental health conditions in Medicare-enrolled, autistic older adults. Autism. Apr 2020;24(3):755–764. doi: 10.1177/1362361319890793 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Klein CB, Klinger LG. Aging Well and Autism: A Narrative Review and Recommendations for Future Research. Healthcare (Basel). Jun 17 2024;12(12)doi: 10.3390/healthcare12121207 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Bessé M, Morel-Kohlmeyer S, Houy-Durand E, et al. Cognitive and Cerebral Aging Research in Autism: A Systematic Review on an Emerging Topic. Autism Research. 2025;18(6):1122–1145. doi: 10.1002/aur.70031 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Zeilinger EL, Stiehl KAM, Weber G. A systematic review on assessment instruments for dementia in persons with intellectual disabilities. Research in Developmental Disabilities. 2013/November/01/ 2013;34(11):3962–3977. doi: 10.1016/j.ridd.2013.08.013 [DOI] [PubMed] [Google Scholar]
- 13.Stringfellow MK, Fields NL, Lee K, Anderson KA, Brokaw E. Healthy Aging and Older Adults With Autism: A Scoping Review. The Gerontologist. 2024;64(11)doi: 10.1093/geront/gnae026 [DOI] [PubMed] [Google Scholar]
- 14.Research Data Assistance Center (ResDAC). Carrier (Fee-for-Service). ResDAC. https://resdac.org/cms-data/files/carrier-ffs [Google Scholar]
- 15.Benevides TW, Carretta HJ, Graves KY. Case Identification and Characterization of Autistic Young Adults in 2010 Medicare Fee-for-Service Claims. Autism Adulthood. Sep 1 2019;1(3):210–218. doi: 10.1089/aut.2018.0036 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Krantz M, Dalmacy D, Bishop L, Hyer JM, Hand BN. Mortality rate and age of death among Medicare-enrolled autistic older adults. Res Autism Spectr Disord. Feb 2023;100doi: 10.1016/j.rasd.2022.102077 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Austin PC. Optimal caliper widths for propensity-score matching when estimating differences in means and differences in proportions in observational studies. Pharm Stat. Mar-Apr 2011;10(2):150–61. doi: 10.1002/pst.433 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Quan H, Li B, Couris CM, et al. Updating and validating the Charlson comorbidity index and score for risk adjustment in hospital discharge abstracts using data from 6 countries. Am J Epidemiol. Mar 15 2011;173(6):676–82. doi: 10.1093/aje/kwq433 [DOI] [PubMed] [Google Scholar]
- 19.Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis. 1987;40(5):373–83. doi: 10.1016/0021-9681(87)90171-8 [DOI] [PubMed] [Google Scholar]
- 20.U.S. Census Bureau. Median household income by age of householder (Table B19049): American Community Survey 5-year estimates, 2013-2021. https://data.census.gov/cedsci/table?q=B19049
- 21.National Bureau of Economic Research (NBER). ICD-9-CM and ICD-10-CM and ICD-10-PCS Crosswalk or General Equivalence Mappings. https://www.nber.org/research/data/icd-9-cm-and-icd-10-cm-and-icd-10-pcs-crosswalk-or-general-equivalence-mappings
- 22.Agency for Healthcare Research and Quality (AHRQ). Clinical Classification Software Refined (CCSR). Healthcare Cost and Utilization Project (HCUP). 2025. https://www.hcup-us.ahrq.gov/toolssoftware/ccsr/ccs_refined.jsp [Google Scholar]
- 23.U.S. Department of Agriculture Economic Research Service. Rural-Urban Continuum Codes. https://www.ers.usda.gov/data-products/rural-urban-continuum-codes/
- 24.Fine JP, Gray RJ. A Proportional Hazards Model for the Subdistribution of a Competing Risk. Journal of the American Statistical Association. 1999;94(446):496–509. doi: 10.2307/2670170 [DOI] [Google Scholar]
- 25.World Health Organization. Global status report on the public health response to dementia. 2021. Accessed 2025/07/07/. http://www.jstor.org/stable/resrep48405
- 26.Larson EB. Evaluation of cognitive impairment and dementia. UpToDate Waltham, MA: UpToDate. 2016; [Google Scholar]
- 27.Janicki MP, McCallion P, Jokinen N, et al. Autism, Diagnostics, and Dementia: A Consensus Report From the 2nd International Summit on Intellectual Disabilities and Dementia. Int J Geriatr Psychiatry. Jun 2025;40(6):e70110. doi: 10.1002/gps.70110 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Mason D, Ingham B, Urbanowicz A, et al. A Systematic Review of What Barriers and Facilitators Prevent and Enable Physical Healthcare Services Access for Autistic Adults. J Autism Dev Disord. Aug 2019;49(8):3387–3400. doi: 10.1007/s10803-019-04049-2 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Nicolaidis C, Raymaker DM, Ashkenazy E, et al. “Respect the way I need to communicate with you”: Healthcare experiences of adults on the autism spectrum. Autism. 2015;19(7):824–831. doi: 10.1177/1362361315576221 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Raymaker DM, McDonald KE, Ashkenazy E, et al. Barriers to healthcare: Instrument development and comparison between autistic adults and adults with and without other disabilities. Autism. 2017;21(8):972–984. doi: 10.1177/1362361316661261 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Wang K, Zhao S, Lee EK-P, et al. Risk of Dementia Among Patients With Diabetes in a Multidisciplinary, Primary Care Management Program. JAMA Network Open. 2024;7(2):e2355733–e2355733. doi: 10.1001/jamanetworkopen.2023.55733 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Bishop-Fitzpatrick L, Minshew NJ, Mazefsky CA, Eack SM. Perception of Life as Stressful, Not Biological Response to Stress, is Associated with Greater Social Disability in Adults with Autism Spectrum Disorder. Journal of Autism and Developmental Disorders. 2017/January/01 2017;47(1):1–16. doi: 10.1007/s10803-016-2910-6 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Vargas DL, Nascimbene C, Krishnan C, Zimmerman AW, Pardo CA. Neuroglial activation and neuroinflammation in the brain of patients with autism. Annals of Neurology. 2005;57(1):67–81. doi: 10.1002/ana.20315 [DOI] [PubMed] [Google Scholar]
- 34.Masi A, Quintana DS, Glozier N, Lloyd AR, Hickie IB, Guastella AJ. Cytokine aberrations in autism spectrum disorder: a systematic review and meta-analysis. Mol Psychiatry. Apr 2015;20(4):440–6. doi: 10.1038/mp.2014.59 [DOI] [PubMed] [Google Scholar]
- 35.Ferrucci L, Fabbri E. Inflammageing: chronic inflammation in ageing, cardiovascular disease, and frailty. Nat Rev Cardiol. Sep 2018;15(9):505–522. doi: 10.1038/s41569-018-0064-2 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Yeo BSY, Song H, Toh EMS, et al. Association of Hearing Aids and Cochlear Implants With Cognitive Decline and Dementia: A Systematic Review and Meta-analysis. JAMA Neurol. Feb 1 2023;80(2):134–141. doi: 10.1001/jamaneurol.2022.4427 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Geukes C, Bröder J, Latteck ÄD. Health Literacy and People with Intellectual Disabilities: What We Know, What We Do Not Know, and What We Need: A Theoretical Discourse. Int J Environ Res Public Health. Feb 5 2019;16(3)doi: 10.3390/ijerph16030463 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Dwyer M, Peklar J, McCallion P, McCarron M, Henman MC. Factors associated with polypharmacy and excessive polypharmacy in older people with intellectual disability differ from the general population: a cross-sectional observational nationwide study. BMJ Open. 2016;6(4):e010505. doi: 10.1136/bmjopen-2015-010505 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Literacy and Misunderstanding Prescription Drug Labels. Annals of Internal Medicine. 2006;145(12):887–894. doi: 10.7326/0003-4819-145-12-200612190-00144%m17135578 [DOI] [PubMed] [Google Scholar]
- 40.Bucholc M, McClean PL, Bauermeister S, et al. Association of the use of hearing aids with the conversion from mild cognitive impairment to dementia and progression of dementia: A longitudinal retrospective study. Alzheimers Dement (N Y). 2021;7(1):e12122. doi: 10.1002/trc2.12122 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41.Danyluk A, Jacob R. Hearing Loss Diagnosis and Management in Adults with Intellectual and Developmental Disabilities. Advances in Medicine. 2023;2023(1):6825476. doi: 10.1155/2023/6825476 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42.Rosen NE, Lord C, Volkmar FR. The Diagnosis of Autism: From Kanner to DSM-III to DSM-5 and Beyond. J Autism Dev Disord. Dec 2021;51(12):4253–4270. doi: 10.1007/s10803-021-04904-1 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 43.O'Nions E, Petersen I, Buckman JEJ, et al. Autism in England: assessing underdiagnosis in a population-based cohort study of prospectively collected primary care data. Lancet Reg Health Eur. Jun 2023;29:100626. doi: 10.1016/j.lanepe.2023.100626 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 44.Torenvliet C, Groenman AP, Radhoe TA, Agelink van Rentergem JA, Van der Putten WJ, Geurts HM. A longitudinal study on cognitive aging in autism. Psychiatry Res. Mar 2023;321:115063. doi: 10.1016/j.psychres.2023.115063 [DOI] [PubMed] [Google Scholar]
- 45.Dekker AD, Wissing MBG, Ulgiati AM, et al. Dementia in people with severe or profound intellectual (and multiple) disabilities: Focus group research into relevance, symptoms and training needs. J Appl Res Intellect Disabil. Nov 2021;34(6):1602–1617. doi: 10.1111/jar.12912 [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
The data used in this study are not publicly available due to restructions under our data use agreement (DUA). However, the analytic code will be submitted to a public repository (Dryad DOI: 10.5061/dryad.3j9kd51zv) within one year of publication, and investigators with their own Centers for Medicare & Medicaid Services (CMS) DUAs will be able to reproduce the analytical dataset.