OBJECTIVE. Data on the current costs of medical services for children with autism spectrum disorders are lacking. Our purpose for this study was to compare health care utilization and costs of children with and without autism spectrum disorders in the same health plan.
PATIENTS AND METHODS. Participants included all 2- to 18-year-old children with autism spectrum disorders (n = 3053) and a random sample of children without autism spectrum disorders (n = 30529) who were continuously enrolled in the Kaiser Permanente Medical Care Program in northern California between July 1, 2003, and June 30, 2004. Data on health care utilization and costs were derived from health plan administrative databases.
MAIN OUTCOME MEASURES. Outcome measures included mean annual utilization and costs of health services per child.
RESULTS. Children with autism spectrum disorders had a higher annual mean number of total clinic (5.6 vs 2.8), pediatric (2.3 vs 1.6), and psychiatric (2.2 vs 0.3) outpatient visits. A higher percentage of children with autism spectrum disorders experienced inpatient (3% vs 1%) and outpatient (5% vs 2%) hospitalizations. Children with autism spectrum disorders were nearly 9 times more likely to use psychotherapeutic medications and twice as likely to use gastrointestinal agents than children without autism spectrum disorders. Mean annual member costs for hospitalizations ($550 vs $208), clinic visits ($1373 vs $540), and prescription medications ($724 vs $96) were more than double for children with autism spectrum disorders compared with children without autism spectrum disorders. The mean annual age- and gender-adjusted total cost per member was more than threefold higher for children with autism spectrum disorders ($2757 vs $892). Among the subgroup of children with other psychiatric conditions, total mean annual costs were 45% higher for children with autism spectrum disorders compared with children without autism spectrum disorders; excess costs were largely explained by the increased use of psychotherapeutic medications.
CONCLUSIONS. The utilization and costs of health care are substantially higher for children with autism spectrum disorders compared with children without autism spectrum disorders. Research is needed to evaluate the impact of improvements in the management of children with autism spectrum disorders on health care utilization and costs.
Autism spectrum disorders (ASDs) are relatively common neurodevelopmental disorders characterized by impairments in reciprocal social interaction and communication and restricted and repetitive behaviors and interests.1 Among individuals with ASDs, which include autistic disorder, Asperger disorder, and pervasive developmental disorder not otherwise specified, impairments range from mild to severe. The reported prevalence of ASDs has increased dramatically over recent decades,2 but it is currently not clear how much of this increase is explained by the expansion of the diagnosis, increased awareness, or a true increase in incidence.3
Although a growing body of evidence suggests that early intensive behavioral therapy results in improvements in social and language skills in autistic children,4,5 and biomedical therapies (ie, psychiatric medications) are efficacious for the treatment of symptoms frequently co-occurring with ASDs, such as hyperactivity, anxiety, or aggressive outbursts,6,7 there are currently very few evidence-based biomedical treatments for the core symptoms defining autism.6,7 As knowledge about the biological underpinnings of the core deficits increases, it is likely that new biomedical interventions will become available. To evaluate the cost-effectiveness of these new therapies, there is a need for baseline data on current costs of medical services for this population, data which are currently lacking.
In this study, we compared the utilization and costs of medical services for children with and without ASDs insured within the same large, group-model, integrated health service delivery system. Comparisons between children with and without ASDs were also made among the subset of children with other psychiatric conditions.
We studied children with continuous enrollment in the Kaiser Permanente Medical Care Program (KP) in northern California during the period July 1, 2003, through June 30, 2004, who were 2 to 18 years old in June 2003. The KP is an integrated, group-model, nonprofit health plan serving >3 million residents of northern California. Members of the KP are demographically similar to the California population except that the very poor and very wealthy are underrepresented.8 During the study year, there were 601538 members who were 2 to 18 years of age. All children with at least 1 ASD diagnosis (autism [International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) 299.0]; Asperger disorder or pervasive developmental disorder not otherwise specified [ICD-9-CM 299.8]) recorded in the KP clinical databases by the start of the 12-month study period (July 1, 2003) were included in the study and composed the ASD group (n = 3053). For comparison, we randomly selected a 10:1 sample of the remaining children who did not have an ASD diagnosis by the end of the study period (June 30, 2004) (n = 30529). Thus, our study focused on the utilization and cost of health services for children already diagnosed with ASDs, not on services and costs experienced in the process of obtaining an ASD diagnosis.
We investigated all medical costs incurred by the KP during the study period, including costs for services received at KP plan facilities as well as outside approved facilities. Dental services and custodial nursing care are not provided directly by the KP; therefore, costs associated with these services were not included in our analyses. Also not included were patient out-of-pocket expenses, including copayments (typically $5–$15) and certain overhead and administrative costs associated with KP membership.
Costs for services provided directly by the KP were obtained from the northern California region's Cost Management Information System, an automated system that integrates hospital, laboratory, radiology, outpatient, and home health utilization databases with the program's financial ledger. Costs, which include facility and program overhead, were generated according to departments, medical centers, and patients using standard accounting methods and program-specific relative-value units for each service. Costs for KP-approved services provided by non-KP vendors (hospitals, individual providers) were also included. Because the cost of non-KP emergency department (ED) visits could not be distinguished from any immediately subsequent hospitalization, those costs were counted as hospital costs instead of ED costs. However, non-KP ED use represents a small proportion of overall ED use. Pharmacy costs were obtained directly from the KP's Pharmacy Information Management System, an automated, region-wide clinical database that records each prescription dispensed at an outpatient KP pharmacy.
Health care utilization data were obtained from northern California's Cost Management Information System and the KP's Pharmacy Information Management System and inpatient and outpatient databases. Utilization was defined in terms of outpatient clinic visits (pediatrics, psychiatry, neurology, other), inpatient hospitalizations (psychiatric, nonpsychiatric), outpatient hospitalizations (such as same-day surgeries), ED visits, and prescription medications (psychotherapeutic agents, antiinfective agents, analgesics/antirheumatic agents, gastrointestinal agents, respiratory/allergic disorder agents). Psychiatric visits at which medication management occurred were also identified.
The primary outcome measure for cost was mean annual cost per member. We compared total mean costs and service-type–specific mean costs for children with and without ASDs with a 2-tailed t test. We conducted age- and gender-adjusted ordinary least-squares regression with age categorized into the following 5 mutually exclusive groups: 2 to 4, 5 to 9, 10 to 14, and 15 to 18 years. Differences in health services utilization between children with and without ASDs were compared with a χ2 statistic for categorical measures (eg, percentage of members with service encounters) and a t test for continuous measures (eg, mean number of service encounters).
We identified 3053 children with a previous diagnosis of ASDs who were continuously enrolled during the study period. Compared with the 30529 randomly sampled controls, a larger proportion of the children with ASDs were male and slightly younger (Table 1). A significantly higher percentage of children with ASDs had a comorbid psychiatric condition (attention-deficit/hyperactivity disorder [ADHD], behavioral disorder, adjustment disorder, anxiety disorder, impulse disorder, mood disorder, psychosis, axis II disorder) diagnosed during the study period compared with children without ASDs (33% vs 6%; P < .0001), although among this subset of study children, the distribution of specific psychiatric conditions was similar between the 2 groups (Table 1).
Utilization of Health Services
Outpatient Clinic Visits
During the study year, a larger proportion of children with ASDs had outpatient visits to pediatrics (75% vs 64%; P < .0001), psychiatry (38% vs 6%; P < .0001), and neurology (5.1% vs 0.4%; P < .0001) compared with children without ASDs (Table 2). On average, children with ASDs had twice as many outpatient clinic visits as children without ASDs (Table 2). Within each age and gender strata, children with ASDs had ∼40% more visits to a pediatrician. Psychiatric visits occurred ∼7 times more often among children with ASDs for all age and gender categories except for 2- to 4-year-olds, for whom psychiatric visits occurred 13 times more frequently than controls of the same age (mean: 0.49 [among cases] and 0.04 [among controls]; P < .0001). The percentage of all KP psychiatric outpatient visits that were completed for the purpose of medication management was 2.5 times higher for children with ASDs compared with children without ASDs (31% vs 13%; P < .0001). Approximately one third (35%) of all neurology visits were related to seizures for children with and without ASDs.
Inpatient (Overnight) Hospitalizations
A significantly higher percentage of children with ASDs had at least 1 inpatient hospitalization during the study year (3% vs 1%; P = .0001), and this was true of hospitalizations for psychiatric as well as nonpsychiatric reasons (Table 2). On average, children with ASDs had 4 times as many inpatient hospital days (P < .0001) and 12 times as many psychiatric inpatient hospital days (P < .0001) as children without ASDs. However, there was no difference in the mean number of inpatient hospital days for nonpsychiatric reasons between the 2 groups of children (P = .84) (Table 2). Among children with ASDs, the most common discharge diagnoses for psychiatric hospitalizations were mood disorders (59%), psychotic disorders (20%), and impulse disorders (17%); the most common discharge diagnoses for nonpsychiatric hospitalizations were respiratory illness (17%), epilepsy/seizure (15%), congenital anomaly (13%), and digestive illness (13%).
Differences in inpatient hospitalizations between children with and without ASDs were most dramatic among children who were 15 to 18 years of age. On average, 15- to 18-year-olds with ASDs were 3.5 times more likely to have overnight hospitalizations (7% vs 2%) and had 10 times as many inpatient hospital days (0.77 vs 0.07) compared with 15- to 18-year-old children without ASDs. This excess was almost entirely a result of psychiatric hospitalizations, for which the primary discharge diagnoses in boys were mood disorder (50%), psychotic disorder (28%), or impulse disorder (28%) and in girls was mood disorder (71%).
Outpatient (Same-Day) Hospitalizations
The percentage of children who had outpatient (same-day) hospitalizations was 2.5 times higher among children with ASDs than children without ASDs (5% vs 2%; P < .0001) (Table 2), and this was true for every age and gender group. Among those with same-day hospitalizations, children with ASDs were more likely than those without ASDs to have discharge diagnoses related to diseases of the digestive system (34% vs 15%; P < .0001), mental disorders (17% vs 4%; P < .0001), and congenital anomalies (9% vs 6%; P = .09). In addition, twice as many children with ASDs had same-day hospitalizations resulting from injury or poisoning (0.6% vs 0.3%; P = .001).
The percentage of children who had a visit to the ED was similar for children with and without ASDs (11% vs 10%; P = .05); however, the average number of ED encounters was ∼30% higher for children with ASDs (Table 2). This trend was more pronounced among 15- to 18-year-olds, of whom the children with ASDs had 70% more ED encounters than the children without ASDs.
The percentage of children who had any medication prescriptions during the study year was ∼40% higher among children with ASDs than children without ASDs (Table 2). This was true for boys and girls and children of all ages except those under 5, for whom the percentage receiving medication prescriptions was similar for children with and without ASDs (68% vs 65%; P = .14). Most drug classes were prescribed more frequently for children with ASDs, especially psychotherapeutic agents (43% vs 5.0%; P < .0001) and gastrointestinal agents (5% vs 2%; P < .0001).
Among the subgroup of study children with other psychiatric conditions, children with ASDs were more likely than children without ASDs to receive at least 1 prescription for psychotherapeutic medication (84% vs 54%; P < .0001). Furthermore, among those children prescribed psychotherapeutic agents, children with ASDs received, on average, a greater total number of prescriptions (8.7 vs 4.5; P < .0001), more types of psychotherapeutic agents (2.0 vs 1.4; P < .0001), and more prescriptions per medication type than children without ASDs.
Cost of Health Services
Among children with ASDs, the average annual cost of health care increased with age (from $1962 for 2- to 4-year-olds to $4256 for 15- to 18-year-olds), and costs were ∼16% higher for girls than boys ($3066 vs $2633; P = .16) and 3 times higher for those with psychiatric comorbidities than those without ($4849 vs $1682; P < .0001). The ratio of ASD/non-ASD costs was somewhat lower for boys (ratio: 2.9) than girls (ratio: 3.4) and increased with age (Fig 1). After adjusting for age group and gender, the average annual estimated cost of health care for children with ASDs was 3 times the cost for children without ASDs ($2757 vs $892; P < .0001). Total costs remained over 2 times higher after additional adjustment for psychiatric comorbidities ($2103 vs $957; P < .0001).
Differences in the average annual health care costs between children with and without ASDs paralleled differences observed for utilization of specific health care services. For example, children with ASDs had 2.5 times higher outpatient clinic costs ($1373 vs $540; P < .0001), 2.5 times higher hospital care costs ($550 vs $208; P < .0001), 30% higher ED costs ($67 vs $52; P = .0001), and 7.5 times higher medication costs ($724 vs $96; P < .0001) (Table 3). The ratio of ASD/non-ASD costs was fairly stable across age for outpatient clinic visits, outpatient hospitalizations, and ED encounters but rose steadily with age for inpatient hospitalizations (Table 4). The cost of medications was double for 2- to 4-year-olds with ASDs but 8 times higher for 5- to 9-, 10- to 14-, and 15- to 18-year-olds with ASDs compared with children without ASDs (Table 4).
The pattern of health care costs differed for children with and without ASDs (Fig 2). Outpatient clinic costs were the largest component of total costs for both groups of children; however, medications represented 27% of total costs for children with ASDs but only 11% for children without ASDs.
Fifty-three percent of the total annual cost of medical care for children with ASDs was accounted for by only 10% of the children. For this upper decile, the average annual cost of care was $14289, compared with $2715 for all children with ASDs. Children in the upper decile were more likely to be 15- to 18-year-olds (29% vs 16%) and female (28% vs 18%) compared with the remaining 90% of children with ASDs. They were also more likely to have inpatient hospitalizations (24.0% vs 0.4%), outpatient hospitalizations (21% vs 3%), ED visits (35% vs 8%), and outpatient visits to psychiatry (75% vs 34%) and neurology (12% vs 4%) and to have a prescription for psychotherapeutic medications (84% vs 39%). Inpatient hospital costs accounted for 28% of total costs for this upper decile of children with ASDs, compared with 0.6% of total costs for the remaining 90% of children with ASDs. Among children with ASDs with inpatient hospitalizations for psychiatric reasons (n = 41), all but 1 was in the upper decile of total cost. Excluding children in the upper decile of cost, children with ASDs still had significantly higher mean costs than children without ASDs ($1430 vs $393; P < .0001).
Among children with other psychiatric conditions, total average annual health care costs were 45% higher ($4849 vs $3353; P < .0001), outpatient clinic costs were 20% higher ($2462 vs $2059; P = .001), inpatient hospitalization costs were 14% higher ($693 vs $607; P = .65), outpatient hospitalization costs were 77% higher ($221 vs $125; P = .03), and psychotherapeutic medication costs were 4 times higher ($1232 vs $303; P < .0001) for children with ASDs compared with children without ASDs. Prescriptions for antipsychotic agents accounted for approximately half of the excess psychotherapeutic medication costs for children with ASDs. In contrast, ED-visit costs were 15% lower for children with ASDs ($104 vs $123; P = .18). Among children with no comorbid psychiatric conditions, total average annual health care costs were similarly elevated for children with ASDs compared with children without ASDs ($1682 vs $735; P < .0001).
Within the same integrated health care delivery system, utilization of all service categories was significantly higher for children with ASDs compared with children without ASDs. Outpatient clinic visits and hospitalizations occurred at least twice as often for children with ASDs; the largest observed differences were for psychiatric clinic visits, psychiatric hospitalizations, and psychotherapeutic medications. Total health care costs were 3 times higher for children with ASDs because of increased costs for hospitalizations, medications, and outpatient clinic visits. Cost differentials were particularly high for children between 15 and 18 years of age. Findings were similar for the subgroup of children with other psychiatric conditions, as well as for the subgroup of children without other psychiatric conditions.
These results are consistent with findings from the few available published studies of health care costs among children with behavioral and developmental disorders. In a study of health care financing for severe developmental disabilities, Birenbaum et al9 compared data from 1985 to 1986 for children and young adults under the age of 25 with autism (n = 308) or severe or profound mental retardation (n = 326) to national data from the 1980 National Medical Care Utilization and Expenditure Survey and the 1987 National Medical Expenditure Survey. They reported that children with autism had 2.4 times higher average annual health care expenditures and double the hospitalization rate compared with all American children and that hospitalizations accounted for one third of the total health care expenditures.9 In a comparison of health services use and health care expenditures in children under 18 years of age with and without disabilities, Newacheck et al10 analyzed data from the Medical Expenditure Panel Survey from 1999 and 2000. They reported that children with disabilities (defined as limited in or unable to conduct their major activity or require early intervention or special education services) had >4 times the number of hospitalizations, 8 times as many days in the hospital, twice as many physician visits, 5 times as many nonphysician visits, and twice as many ED visits and received ∼3 times the number of prescribed medications as children without disabilities.10 Total health care expenditures were ∼4 times higher for children with disabilities, and differences in expenditures for specific types of services paralleled differences in utilization between the 2 groups of children.10 Similar findings were also reported for children with special health care needs,11 behavioral disorders (affective psychoses, anxiety, miscellaneous behavioral disorders, depression, conduct disorder, oppositional-defiant disorder, and ADHD),12 mental handicap (mental retardation, mental illness),13 depression,12 or psychosocial morbidity defined by the Pediatric Symptom Checklist.14 A significant proportion (40%–70%) of children with ASDs function in the mental retardation range2 and/or have co-occurring psychiatric conditions such as ADHD, depression, and anxiety.15–18
Similar to other studies,11,12 we found that the majority of total health care costs for children with ASDs were accounted for by a small minority of children, for whom hospitalizations (psychiatric as well as nonpsychiatric) accounted for a large proportion of the excess costs. In both girls and boys, a majority of hospitalizations were for nonpsychiatric reasons in children under 10 years and for psychiatric reasons in children 10 to 18 years of age. However, there was still a significant excess cost for children with ASDs among the bottom 90% of the population.
Our finding of significantly increased rates of hospitalizations and medication prescriptions for gastrointestinal disturbances among children with ASDs lends support to the many anecdotal reports and limited scientific evidence of gastrointestinal dysfunction in this patient population.19–22
Among children with ASDs, we found that those with psychiatric comorbidities had significantly higher health care costs. Interestingly, total health care costs were also significantly higher for children with ASDs with co-occurring psychiatric conditions compared with children with similar psychiatric conditions but no ASD diagnoses. Thus, the presence of ASDs seems to raise health care costs, primarily psychotherapeutic medication and hospitalization costs, above and beyond costs usually incurred to treat non-ASD psychiatric conditions. These results suggest that the presence of behavioral and social impairments complicate the management of co-occurring psychiatric conditions, which may lead to increased use of psychotherapeutic medications in children with ASDs. It is also possible that, because of the absence of evidence-based biomedical therapies for treating the core symptoms of ASDs, providers may be more likely to prescribe psychotherapeutic agents that target the co-occurring psychiatric conditions. These findings underscore the urgent need to develop and evaluate biological agents for the treatment of the core symptoms of ASDs. The possibility that increased health service utilization and costs result from the lack of integration of biomedical and behavioral services also deserves evaluation in future studies.
Our study had several limitations that deserve mention. First, we focused exclusively on health services and costs incurred within the health plan. We could not estimate costs associated with educational services, lost productivity of parents (employment and income), lost leisure time, out-of-pocket expenses, behavioral interventions, or other nonmedical treatments and services. In a small pilot study conducted in the United Kingdom designed to test a research instrument for calculating the cost of support for children with ASDs, families reported that the costs for health and social services, which largely comprised inpatient hospital care, social services, respite care, and clinical psychologist appointments, were small compared with the costs for education, early intervention therapy, and income losses.23 A recent report by the US Government Accountability Office24 estimated that the average per-pupil expenditure for educating a child with autism was almost 3 times the average per-pupil expenditure of educating a child who does not receive any special education services ($18790 vs $6556). Second, our study included insured children only, and there is some evidence that families with insurance spend 51% less of their income on health care for their disabled child than families without insurance.10 Third, data were derived from only 1 health plan. Although actual costs in this health plan may not be representative of costs in other health plans or among the uninsured, the relative costs of children with ASDs compared with children without ASDs should reflect these differences in other settings. The KP population is very similar demographically to the general population in the catchment area, and the demographic characteristics and overall prevalence of ASDs within the KP match those reported in other populations.25,26 Fourth, we did not have data on potentially important predictors of utilization (parental level of education, income, number of siblings in the household); thus, our findings could be biased as a result of uncontrolled confounding. Fifth, our study focused on costs to families whose children were already diagnosed with ASDs and did not include costs incurred by families for obtaining the ASD diagnosis. These costs could be substantial. Thus, our results may significantly underestimate the actual costs associated with health care utilization for this population.
The utilization and costs of health care are significantly higher for children with ASDs compared with children without ASDs, underscoring the need to find more appropriate treatment options including biomedical approaches that target the core ASD symptoms. It remains to be seen if improvements in the management of children with ASDs will result in more appropriate health care service utilization and reduction in overall costs.
- Accepted May 4, 2006.
- Address correspondence to Lisa A. Croen, PhD, Division of Research, Kaiser Permanente, 2000 Broadway, Oakland, CA 94612. E-mail:
The authors have indicated they have no financial relationships relevant to this article to disclose.
- ↵American Psychiatric Association. Diagnostic and Statistical Manual of Mental Disorders. 4th ed. Washington, DC: American Psychiatric Association; 1994
- ↵Tsai LY. Psychopharmacology in autism. Psychosom Med.1999;61 :651– 665
- ↵Birenbaum A, Guyot D, Cohen HJ. Health care financing for severe developmental disabilities. Monogr Am Assoc Ment Retard.1990;(14) :1– 150
- ↵Newacheck PW, Inkelas M, Kim SE. Health services use and health care expenditures for children with disabilities. Pediatrics.2004;114 :79– 85
- ↵Guevara JP, Mandell DS, Rostain AL, Zhao H, Hadley TR. National estimates of health services expenditures for children with behavioral disorders: an analysis of the medical expenditure panel survey. Pediatrics.2003;112(6) . Available at: www.pediatrics.org/cgi/content/full/112/6/e440
- ↵Gillott A, Furniss F, Walter A. Anxiety in high-functioning children with autism. Autism.2001;5 :277– 286
- Fombonne E, Chakrabarti S. No evidence for a new variant of measles-mumps-rubella–induced autism. Pediatrics.2001;108(4) . Available at: www.pediatrics.org/cgi/content/full/108/4/e58
- ↵Taylor B, Miller E, Lingam R, Andrews N, Simmons A, Stowe J. Measles, mumps, and rubella vaccination and bowel problems or developmental regression in children with autism: population study. BMJ.2002;324 :393– 396
- ↵US Government Accountability Office. Report to the Chairman and Ranking Minority Member, Subcommittee on Human Rights and Wellness, Committee on Government Reform, House of Representatives: Special Education—Children With Autism. GAO-05-220. Available at: www.gao.gov/new.items/d05220.pdf. Accessed July 25, 2006
- Copyright © 2006 by the American Academy of Pediatrics