Prevalence of physical and mental health conditions in Medicare-enrolled,autistic older adults

Data source

Data used for this study were derived from Medicare Standard Analytic Files (SAF) for the years 2016 and 2017, which included Limited Data Set information on 100% of Medicare beneficiaries for these years. De-identified beneficiary-level healthcare claims data for inpatient and outpatient records were used in this analysis. The outpatient records used in this study contained medical billing claims from institutional outpatient providers such as hospital outpatient departments, rural health clinics, renal dialysis facilities, outpatient rehabilitation facilities, Federally Qualified Health Centers, and community mental health centers. These outpatient records do not include professional service claims from non-institutional professional providers such as physicians, physician assistants, clinical social workers, or nurse practitioners.

Study population

Autistic adults were included in this study if they (1) were aged 65 years or older, (2) were enrolled in Medicare Fee for Service for at least 6 months in 2016 or 2017, and (3) had at least one inpatient or outpatient encounter with an autism spectrum diagnosis at any time during 1 January 2016–31 December 2017. Autism spectrum diagnoses were identified using International Classification of Diseases, 10th edition (ICD-10) codes F84.0, F84.1, F84.5, or F84.9. Consistent with practices in existing literature (e.g. Demiralp et al., 2019; Lankiewicz et al., 2018; Mahr et al., 2018), we excluded beneficiaries enrolled in Medicare Advantage Plans due to concerns regarding the availability of accurate, complete data in these managed care programs. A PC group of beneficiaries without autism diagnoses was selected at a 10:1 ratio to autistic beneficiaries using group frequency matching for 5-year age category and sex. The PC group met all the same inclusion criteria, except without a medical encounter containing a diagnosis of autism at any time during the study period.

Measures

We identified mental and physical health conditions from inpatient and outpatient medical claims using the Healthcare Cost and Utilization Project (HCUP; 2016) Beta Multilevel Clinical Classification Software (CCS) for ICD-10, which identifies conditions based on the diagnosis codes included in the medical billing record and groups them into a smaller number of clinically relevant categories. Supplemental Table 1 provides a list of all physical and mental health conditions examined and how these were defined in the present study.

Statistical analysis

Demographic characteristics were summarized descriptively. To maintain beneficiary confidentiality, only conditions where frequency counts were >10 for all groups are shown in the tables. Separate logistic regression models were performed to compare autistic and PC older adults on the odds of each condition while controlling for sex, race/ethnicity, age, rural residence, and estimated household income. Rural residence was defined by living in a non-metropolitan statistical area. Household incomes were estimated for each beneficiary using the median per capita household income during 2017, among individuals over 65 years of age, for the beneficiary’s county of residence. We also performed logistic regression analyses stratified by sex to compare the odds of having a medical encounter for each condition among males and females separately while controlling for race/ethnicity, age, rural residence, and estimated household income. Odds ratios were interpreted according to Agresti (2013). We used SAS statistical software, version 9.4 for all analyses.

Ethical approval

The Institutional Review Board (IRB) of The Ohio State University reviewed this study and determined it to be IRB-exempt due to the use of limited datasets.

Implications for healthcare delivery

Nonetheless, this study offers a valuable contribution to the literature and is applicable to the medical care of autistic adults over the age of 65 today. Many of our findings are consistent with what is known about the increased risk of physical health conditions (e.g. heart disease, epilepsy, and gastrointestinal conditions) and mental health conditions (e.g. schizophrenia, psychotic disorders, personality disorders, and attention-deficit disorders) in younger autistic individuals (Croen et al., 2006, 2015). By emphasizing that these conditions continue to be highly prevalent in older populations, our results may inform specialized approaches to screening and management of co-occurring conditions for older autistic adults. For example, autistic adults in our sample had significantly greater odds of metabolic disorders. Evidence from younger autistic populations suggests that multiple factors increase risk for metabolic disorders including prescription psychotropic drugs (Shedlock et al., 2016), sensory motor challenges that can reduce physical activity (Lawson & Foster, 2016), sensory preferences that can limit food choices (Polfuss et al., 2016), and sleep disturbances (Zuckerman et al., 2014). However, healthcare providers who treat older autistic adults may not be aware of the various factors contributing to obesity in this population, which likely necessitate a specialized approach to weight management.

This study also offers insights into the autism-specific prevalence of conditions that tend to be more common in the general older adult population such as osteoporosis, cognitive decline, heart disease, cancer, cerebrovascular disease, and osteoarthritis. Autistic older adults were significantly more likely to have each of these diagnoses. By alerting clinicians to conditions for which autistic older adults are at greater risk, our results may ultimately improve the medical management of these conditions, improve health related quality of life for autistic older adults, and provide opportunities to reduce the likelihood of premature death due to unaddressed, or under-addressed, health conditions.

While our results revealed that older autistic adults may be over eight times more likely to experience cognitive conditions such as delirium, dementia, and amnesia, other studies have found that older autistic adults are equally as likely or less likely to experience age-related cognitive changes (Lever & Geurts, 2016a). However, there are a number of methodologic differences between this study and that of Lever and Guerts, which may account for this discrepancy. For example, Lever and Guerts had a younger population of autistic older adults, which included those aged 55+ years. Thus, it is possible that the older age range used in this study (65+ years) may, in part, explain the increased prevalence of cognitive conditions. In addition, Lever and Guerts used self-report measures to identify cognitive conditions, while we used medical billing records. As such, our study included those individuals who may not have been able to participate in survey methods and detected physician-diagnosed conditions, as opposed to self-perceived/self-reported conditions.

It is also important to note the increased likelihood of suicide and intentional self-inflicted injury we observed among autistic older adults. Our results add to a growing body of literature that indicates suicidality (Cassidy et al., 2014; Hirvikoski et al., 2016; Kato et al., 2013; Kirby et al., 2019; Zahid & Upthegrove, 2017) and intentional self-inflicted injury (Maddox et al., 2018; Moseley et al., 2019) are significantly more common in autistic adults of all ages. Distinguishing between suicidal and non-suicidal self-inflicted injury was beyond the scope of the present study, but this may be considered in future work. These results, as well as findings of increased odds of other mental health conditions, highlight the need for proactive screening to identify older autistic adults with unmet mental healthcare needs.

Implications for policy

With regard to policy implications, increased awareness of physical and mental health condition prevalence among older autistic adults may inform policies on eligibility for health care and community-based services, as well as policies for redesigning health benefits tailored to the needs of this patient population. In addition, our findings address a priority area identified by the IACC, which advises the Secretary of Health and Human Services on Federal activities related to autism. One of the objectives identified in the IACC’s (2017) 2016–2017 strategic plan was to support research to better understand and meet the needs of autistic individuals as they age, including research to “reduce disabling co-occurring physical and mental health conditions in adults with [autism], with the goal of improving safety, reducing premature mortality, and enhancing quality of life.” By characterizing the prevalence of these conditions in a large, national sample of older autistic adults, our results can inform future funding priorities to develop innovative solutions to better manage these conditions.

Methodologic considerations and future directions

We acknowledge several limitations to this work. There were a number of variables for which we could not control that may be strongly associated with condition prevalence (e.g. socioeconomic factors, perceived met/unmet healthcare needs, social support). This study relied solely on diagnostic codes found in medical claims data, the selection of which may have been influenced by clinician bias and/or challenges with obtaining an accurate diagnosis due to communication difficulties (Nicolaidis et al., 2015). Examining longitudinal change in the occurrence of physical and mental health conditions was beyond the scope of the present study, but will be important to address in future work. Also beyond the scope of the present study, but an important consideration for future studies, was obtaining a finer degree of resolution for certain types of physical and mental health conditions experienced by autistic older adults (e.g. prevalence of specific types of anxiety disorders). Of note, this study focused exclusively on condition prevalence, which is one indicator of health outcomes. Future studies examining other health outcomes, such as met and unmet healthcare needs, satisfaction with care, healthcare access, service utilization, and cost, are warranted to provide a more holistic picture of the health status and healthcare needs of this population.

It is also important to acknowledge that this study identified autistic adults based on the medical billing records of institutional providers, which include hospitals, hospital outpatient departments, rural health clinics, renal dialysis facilities, and outpatient rehabilitation facilities that provide service to Medicare beneficiaries. Professional service claims from non-institutional ³ professional providers including physicians, physician assistants, clinical social workers, nurse practitioners were not included in this study. Therefore, it is possible that some autistic older adults who saw non-institutional providers or who were undiagnosed may have been missed. Similarly, physical and mental health conditions for which beneficiaries saw non-institutional providers or did not seek treatment would not have been captured. However, as this limitation applies to both autistic and PC adults in this study, comparisons of relative prevalence are valid and offer a valuable contribution to the literature.