Costs of autism spectrum disorders and access to services in Spain

Abstract

Background

Autism spectrum disorder (ASD) is a heterogeneous neurodevelopmental condition with significant impact on families and society. The objective of this article is to present the results of the first cost-of-illness study of ASD in Spain.

Methods

An observational, cross-sectional study was designed to estimate the annual cost per individual by means of an anonymous survey and a bottom-up approach to collect use of services and costs. A societal perspective was adopted, including direct costs related to education, health care, and social care sectors, and productivity loss.

Results

Complete responses of 334 individuals on the spectrum were obtained (mean age 12.3 years, 82% male, 46.4% with autistic disorder). The most used health care services were general practitioners (35.6%), accident and emergency services (15.3%), and community nurses (10.8%). Among specialized health care services, the most frequently used were psychologists, psychiatrists and other types of individual therapy. The most used drugs were antipsychotics. There was a variety of social services available and accessed. The most often reported service accessed by carers was additional visits to general practitioner (15%). The average annual direct cost of ASD in Spain was €10,974 per individual (95%CI 8717–13879). Logistic regression found that individuals with anxiety, intellectual disability, and an autistic disorder were more likely to access services. The average cost of productivity losses per carer was €2186.4.

Conclusions

The average annual direct cost of ASD in Spain was €10,974 per individual. Individuals with anxiety or intellectual disability co-occurring with autistic disorder were more likely to access services. Findings from this study would be useful to inform service planning and future research.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12913-026-14879-3.

Introduction

Individuals with autism spectrum disorder (ASD) share traits such as deficits in social communication, and repetitive and restricted patterns of behaviour, often from early childhood [1]. ASD is one of the most prevalent neurodevelopmental conditions, with a median male-to-female ratio of 4.2 [2]. The prevalence of ASD in Europe varies from 0.48% in France to 3.13% in Iceland [3–6]. In the north of Spain the prevalence was estimated at 0.59% (95% CI: 0.48–0.73) in a study performed as part of the ‘Autism Spectrum Disorders in the European Union’ (ASDEU) project [3]. The prevalence, the diagnostic delays [7], and the complex management of diverse clinical trajectories of people on the spectrum challenge health systems globally by early, intense and continuous requirements of multidisciplinary interventions addressed by specialized teams [8]. However, the overall social impact associated with diagnostic uncertainties for ASD, lack of effective treatments, and its economic consequences, rests largely on families. Families of children on the spectrum incur higher health and social costs compared to those of children with other special needs, with parents taking substantial time off work [4].

Cost of illness (COI) studies highlight the burden associated with a condition in a specific country or setting and can identify the drivers of costs, such as co-occurring conditions, sex, socio-economic and demographic characteristics. This information can be used by authorities to make decisions or set policies. After the first cost estimates for ASD was published by Järbrink and Knapp for the United Kingdom [9], other studies have followed suit [10–20].

A recent review on costs of ASD including approximately 50 publications from several countries identifies six types of costs: healthcare costs; therapeutic costs; special-education costs; costs of accommodation, respite care and out-of-pocket expenses; costs associated with productivity loss for adults on the spectrum; and costs of informal care and lost productivity for family/carers [21].

Despite the significant impact of ASD on families and society [22], there are no previous studies on the cost of ASD in Spain. The main aims of the ASDEU project were to estimate the prevalence of ASD and their socio-economic impact in several European countries. The aim of this paper is to report the detailed results of the COI study of ASD in Spain, and specifically: (1) use of services by individuals on the spectrum and their carers; (2) costs of ASD by sector (education, health care and social care); and (3) main cost-drivers.

Methods

An observational, cross-sectional study was designed to estimate the annual cost per individual with a diagnosis of ASD in Spain. A bottom-up approach was adopted by adding individual data that were obtained by means of an anonymous web-based questionnaire. The subjects were children and adults on the spectrum. The survey could be answered by carers on behalf of the individual on the spectrum. Despite the fact that the classification of ASD is no longer recognized, we used it for comparison purposes and because many individuals still identify themselves with the diagnosis they received before the changes in terminology. To be precise, individuals were asked about their subtype of ASD (autistic disorder, Asperger’s disorder, pervasive developmental disorder not otherwise specified (PDD-NOS), childhood disintegrative disorder (CDD) or other).

Subjects were recruited by convenience sampling [23]. This cost-of-illness study was approved by the ethics committee from Nuestra Señora de la Candelaria University Hospital, Tenerife, Spain (reference number PI 28/17).

Devising of the questionnaire and procedure

An anonymous web-based questionnaire was used to collect data on use of services related to ASD and productivity losses related to ASD by persons and carers during the last six months. Initially drawn up in English for a study in Scotland [16], the questionnaire was translated into Spanish and adapted according to the specific features of the Spanish education, healthcare and social care sectors. The types of data collected included qualitative information aimed at characterizing persons on the spectrum and carers, and quantitative data on service use (including medication) and cost. Individuals were also questioned whether they paid for the service. The English and Spanish versions of the questionnaire are available upon request to the authors.

The survey was conducted on LimeSurvey (http://www.limesurvey.org) between February 2017 and February 2018. It was disseminated by means of autism-related organizations (parents, person on the spectrum and professional associations and private care providers), national and regional press releases, newsletters and social media posts.

Direct and indirect costs

Direct and indirect costs are shown separately and from the societal perspective. To estimate the additional direct costs related to ASD, the questionnaire collected quantitative data on use of services related to ASD (education services, outpatient and inpatient hospital services, medication, residential respite care and other health and social care services). Only uses and costs directly related to ASD were collected, for instance education costs did not include ordinary education but only special education, that is, education in special schools and those special services provided in mainstream schools.

Unit costs were collected with the help of ASDEU partners in Spain (see Table S1 in supplementary material). The unit costs were adapted in some services according to length of the contacts reported by individuals. When salaries per month were provided, data were adjusted using the average number of weekly work hours in a main job in Spain (EUROSTAT). When no public unit costs were available or when the services were mainly provided privately, the mean unit cost reported by the respondents was used. In the absence of local estimates, the cost of special education was taken from foreign sources and adapted to Euros [16]. When needed, costs were adjusted by inflation using national data on consumer price index [24] to be expressed in Euros from 2017.

As individuals were asked about their use of services for the last six months, the costs were multiplied by two to estimate the annual cost, except for special education as the unit cost was already in annual terms.

The cost of services used by carers and productivity loss due to cessation or reduction of working time by parents or other relatives were also estimated. Productivity loss was measured by means of the human capital approach [16, 25], as the product of hours lost as reported by carers multiplied by the mean hourly wage in Spain (€21.17) obtained from monthly wages [26] and assuming four weeks per month and 37.7 weekly hours worked [27]. If a carer worked less than the mean usual weekly hours worked on the main job, we estimated the six-month productivity loss using the following formula: (average usual weekly hours worked - number of hours worked per week) * hourly labour costs * 26 weeks. It was not possible to estimate the productivity loss by individuals on the spectrum due to the small sample of adults.

Analyses

Analyses were conducted for those individuals who completed the survey and had residence in Spain. Firstly, descriptive statistics were used to present demographic characteristics of individuals on the spectrum and their carers, use of services and their costs: mean and standard deviation (SD) for continuous data and proportions on categorical data. Non-parametric bootstrapping (bias-corrected, accelerated [BCa] percentile interval, with 1000 replications) was used to estimate the confidence intervals of the costs.

Secondly, a multivariate analysis was performed to identify the drivers of total direct costs and three sector-related costs (education, healthcare and social care sector). A two-part model was adopted. In the first model we explored the probability of having used any service by means of a logistic regression (binary variable where value ‘1’ means ‘use of service’ [or cost > €0] and ‘0’ means ‘no use of service’ [or cost = €0]). In the second model we explored the effect of potential drivers in direct costs by means of ordinary least squares (OLS) model and bootstrap. In this regression only those individuals that used services were taken into account. Only logistic regressions were conducted by type of disorder (autistic disorder, Asperger’s disorder) and by age (< 16 and ≥ 16 years) as the number of individuals that used services was not sufficient and prevented OLS models for total costs. Statistically significant differences were considered when P-value < 0.05. These analyses were performed only on direct cost and for individuals on the spectrum as the proportion of carers who answered the survey was limited. For the same reason, both direct and indirect costs were not included in the estimation of total costs.

Thirdly, deterministic sensitivity analysis was performed on descriptive results assuming alternative values of unit costs: costs reported by the individuals if they paid for the service and the unit costs used in the base case when individuals did not pay for the service (Scenario 1); costs reported by the individuals when they were lower (Scenario 2) or higher (Scenario 3) than the cost used in the base case.

All analyses were conducted using STATA version 15.0 (StataCorp LLC, TX, USA) [28].

Results

Sample characteristics

984 subjects accessed the Spanish questionnaire, of whom 334 completed the survey and had residence in Spain and were included in the analysis (Table 1). Almost 90% of responses were submitted by the parents of individuals on the spectrum; 3.9% of responses were submitted by individuals who were on the spectrum. Most (82%) of the individuals on the spectrum were male; the mean age was 12.3 years with a range 2 to 55 years. Responses were obtained from all regions within Spain with the exception of one small region, La Rioja. The subtype of ASD, diagnosed by a doctor according to the respondents, was autistic disorder in 46.4% of respondents with 32.9% and 17.7% diagnosed with Asperger’s disorder and PDD-NOS, respectively. The most commonly reported co-occurring conditions diagnosed by a doctor were intellectual disability, attention deficit hyperactivity disorder (ADHD), mood disorders and sleep disturbances. Most individuals (97.3%) had lived in a private household with parents or relatives over the last six months.

Table 1.

Characteristics of individuals with ASD (N = 334)

Person who answered the survey	Individual with ASD	3.9%
	Parent	89.5%
	Other family member	2.4%
	Not a family member	4.2%
Age (years) of the person on the spectrum	Mean, SD	12.3, 9.1
	Median; Range	10.0; 2–55
Gender of the person on the spectrum	Male	82.0%
	Female	17.7%
Nationality	Spanish	92.8%
	Double nationality, Spanish and other	4.2%
	Other nationality	3.0%
Self-reported type of ASD*	Autistic disorder	46.4%
	Asperger’s disorder	32.9%
	PDD-NOS	17.7%
	CDD	0.6%
	Other/No answer	2.4%
Other co-occurring conditions ever diagnosed by a doctor of the person with ASD	Sleep disturbance	13.8%
	Anxiety	15.3%
	ADHD	20.4%
	Intellectual disability	20.1%
	None	33.5%

ASD: Autism spectrum disorder; ADHD: Attention deficit hyperactivity disorder; CDD: Childhood disintegrative disorder; PDD-NOS: Pervasive developmental disorder-not otherwise specified; SD: Standard deviation

* Despite the fact that the classification of ASD is no longer recognized, we use it for comparison purposes and because many individuals still identify themselves with the diagnosis they received before the changes in terminology

Among the 83 individuals who had left school and were older than the minimum working age in Spain (16 years), six had a paid job and one was employed in an unpaid position.

Use of services

Table 2 shows the percentage of respondents who reported the use of each service. The proportion of individuals on the spectrum who have not left education and who attended a mainstream school was 55.4%. A notable proportion of individuals attended secondary schools or technical colleges (12%). Among other special education professionals, the services most used were educational psychologists (61.3%), speech and language therapists (49.2%) and classroom assistants (39.4%). A specialist teacher was reported to be used by one third of individuals. Tuition or tutorial at school, college or university was accessed by 18.3%.

Table 2.

Individuals using services over the past six months, per sector (N = 334)

	Service	%	Mean hours, days or visits
Education sector
Educational services accessed by individuals with ASD who have not left education (*)	None	2.7%	-
	Mainstream school	55.4%	-
	Secondary education / Technical college	12.0%	-
	University	1.8%	-
	Special Unit/Resource in mainstream school	11.7%	-
	Special day school (general)	10.5%	-
	Special day school for autism	7.5%	-
	Special residential school	2.4%	-
	Home education (as an alternative to school)	1.2%	-
	Other	6.9%	-
Other professionals at school, college or university, hours per week	Educational psychologist	61.3%	2.7
	School family worker/Education social worker	13.5%	4.3
	Classroom assistant	39.4%	10.5
	Specialist teacher	33.7%	14.5
	Speech and language therapist	49.2%	2.6
	Occupational therapist	17.9%	3.3
	Physiotherapist	10.1%	1.9
	Disability service advisor	7.2%	0.7
Tuition/Tutorial support, hours per week	Individual tuition at home	13.0%	3.3
	Individual tuition elsewhere (e.g. school/college/university)	18.3%	4.6
	Tuition in a small group (e.g. school/college/university)	9.6%	10.1
Health care sector
Inpatient hospital care, days	Psychiatric hospital	1.2%	9.8
	Psychiatric ward in a general hospital	0%
	General medical ward	3.4%	3.1
	Hospital care in prison/ secure/semi-secure unit	0.3%	140
Outpatient hospital care, times	Psychiatric outpatient visit	13.8%	3.0
	Accident and Emergency (A&E)	15.3%	2.8
Medication	Any medication among several groups (**)	32.0%	-
Other health services, visits	Psychiatrist	33.2%	2.6
	Psychologist	36.2%	18.8
	Individual counselling/therapy	20.1%	28.3
	Group counselling/therapy	12.3%	15.5
	Speech therapist	15.6%	30.4
	Physiotherapist	4.8%	15.4
	General practitioner	35.6%	4.4
	Nurse / community nurse	10.8%	2.3
Social care sector
Residential respite care services, days	Residential care-home for children/adolescents	5.8%	7
	Residential care-home for adults	3.0%	28.1
	Foster care	1.5%	4.2
Other social services, visits	Other community learning disability team member	5.1%	28.3
	Community challenging behaviour team member	0.9%	24.7
	Child development centre	1.5%	127.3
	Occupational therapist	9.3%	30.6
	Social worker	4.8%	3.7
	Home help/home care worker	2.7%	47.1
	Outreach worker/family support	3.9%	33.1
	Befriender	3.00%	13.3
	Day care centre	1.8%	106.8
	Social club	3.0%	16.6
	After-school club	5.4%	20.6
	Play-schemes	23.7%	26.3
	Sheltered workshop	1.2%	71
	Individual placement and support	0.3%	5
	Holiday schemes (1 day; similar to social clubs, play schemes, or day care)	4.5%	13.1
	Holiday schemes (> 1 day; similar to holiday trips)	10.2%	1.2
	Baby-sitter	7.2%	41.7

*Total does not sum up to 100% because individuals could choose more than one option

**The groups of medications reported and considered in the analysis were: anticholinergic agents, antidepressants, antiepileptic, antihistamines, antimigraine, antipsychotics, anxiolytics, hypnotics and sedatives, antiadrenergic, psychostimulants, agents used for ADHD and nootropics

Among healthcare services, the most used were those targeted at the general population such as accident and emergency services (15.3%), general practitioners (35.6%) or community nurses (10.8%). Among specialized healthcare the most found were psychologists (36.2%), psychiatrists (33.2%) and other types of individual therapy (20.1%). The most frequently used type of drugs was antipsychotics, by 19% and 28% of children and adults, respectively. Risperidone was the most commonly used drug in both groups.

There was a variety of social services available and accessed but the most used were play-schemes (23.7%) and holiday schemes lasting more than one day (10.2%). Residential respite care was not used by a high number of individuals: for example, only 8.8% used a residential care-home.

Some services accessed by a large number of individuals on the spectrum were paid for directly by carers, such as educational psychologists or speech therapists at school, college or university and psychologists and other therapists in the community among other healthcare professionals (see Table S2 in supplementary material). Other services accessed by a low number of individuals were also paid for directly by carers, in particular several social care services. For example, all individuals (9.3%) that accessed the services provided by occupational therapists paid out-of-pocket. Residential respite care services were accessed by less than 2% of the sample, but more than 70% paid for these services directly.

Direct costs related to ASD

The average annual direct cost of ASD in Spain was €10,974 per individual (95%CI €8716.69 – €13,879.31) (Table 3). This figure does not include the use of services by carers and indirect costs (see below). By sectors, the cost of education represented 80% of the cost, while healthcare costs (17%) and social care costs (3%) were a lesser proportion. Despite not being accessed very often, inpatient care and residential respite care services were very costly compared to other services.

Table 3.

Annual direct costs per individual with ASD (€, 2017 prices)

	Mean	SD	95% CI*
Special educational services	6359.3	20521	4572.455–9019.461
Other professionals at school, college or university	1980.6	4766.3	1569.184–2586.405
Tuition/Tutorial support	462.7	2060.3	258.449–699.489
Subtotal: Education sector	8795.6	20974.3	6840.178–11556.98
Inpatient hospital care	644.9	8585.2	108.382–2022.813
Outpatient hospital care	216.5	686.4	154.727–295.411
Medication	209.7	982.1	124.494–335.975
Other health services	807.8	4050.6	502.447–1331.923
Subtotal: Health care sector	1879.0	9617.8	1147.954–3385.651
Residential respite care services	111.1	1347.4	19.868–317.630
Other social services	188.3	927.6	98.333–305.512
Subtotal: Social care sector	299.4	1702.3	159.443–535.072
Total direct cost	10973.9	23745.4	8716.69–13879.31
Cost of health and social care services accessed by carers	162.6	373.90	127.640–204.744

*Estimated by bootstrap, bias-corrected, accelerated (BCa) percentile interval, with 1000 replications

ASD: Autism spectrum disorders; CI: Confidence interval; SD: Standard deviation

Figure S1 (Supplementary material) illustrates the differences in the weight of each sector depending on the type of population. In general, the cost of social care was small in comparison to the total cost, which was comprised of special education costs mainly in the younger population, while social care was more accessed by adults. The proportion of education cost over the total cost among people with autistic disorder also seemed higher than in people with Asperger’s disorder, especially among older individuals. These findings were further explored in the multivariate analysis.

Drivers of direct costs

The results of the two-part multivariate model are shown in Table S3 (supplementary material) for the whole sample and total direct cost as the dependent variable. The first part of the model, logistic regression, revealed that individuals with anxiety, intellectual disability or an autistic disorder were more likely to access services (odds ratio > 1). The second part of the model (regression) revealed that virtually all the factors considered affected the costs, except for sex, anxiety, ADHD and intellectual disability. The results of the exploratory multivariate analysis by sector and subgroups are included in the supplementary material (Tables S4 and S5).

There were few factors that would account for the access to services in each sector (Table S4). Special education costs were affected by age, sex, epilepsy, type of ASD and type of school. Healthcare access was affected by sex, anxiety and ADHD, although the relationship with costs was not statistically significant. Although the sample with costs in social care was small, age and anxiety were identified as factors that accounted for the access to social services.

The subgroup analysis revealed that anxiety, studying in a mainstream school and being an active worker were cost drivers in individuals with autistic disorder; ADHD was the only significant variable to account for costs in individuals with Asperger’s disorder; in children younger than 16 years the cost drivers were co-occurring ADHD, autistic disorder and studying in a mainstream school; and finally, having autistic disorder or working/job seeking was associated to costs in individuals aged 16 years old or older (Table S5).

Costs related to carers of ASD individuals

Carers who completed the survey on behalf of the person on the spectrum were asked about their autism-related use of services and their employment status (see Table S6 in Supplementary material). Eighty out of 321 people, all of them parents, reported using at least one service. The service most often accessed was additional visits to the general practitioner (15%). Counselling, psychiatric services and social services were also used by approximately 9% to 10% of carers. In terms of costs the use of these services accounted for an average of €162 per year.

Two hundred and four carers (64% of all carers) were employed. The mean number of hours worked in a week was 36.2 (SD: 10.53). One hundred carers had to have at least one absence from work, with a mean of 9.1 (SD: 13.69) days absence in six months. Carers reported caring for the person on the spectrum for an average of 68.7 (SD: 48.78) hours per week. The cost of productivity losses per carer was estimated at €1093.2 and €2186.4 over six and 12 months, respectively.

Sensitivity analysis

The results of the three scenarios drawn can be found in Table S7 in supplementary material. The total cost in Scenario 1 (costs reported by the individuals if they paid for the service) was higher than in the base case (€12547 vs. €10974). This was due to fees paid for professional services at school, college or university and other health and social services. However, the cost of inpatient and outpatient hospital care was much lower because in some cases the costs paid by carers could be affected by copayments. By sector, the cost of education represented 78% of the cost, while healthcare costs (10%) and social care costs (12%) represented a higher proportion of the direct cost than in the base case. In the other two scenarios, with lower or higher unit costs than the cost used in the base case, a range for direct costs was estimated, for example, total direct cost ranged from €10,627 in Scenario 2 to €17,277 in Scenario 3.

Discussion

To the best of our knowledge this is the first cost-of-illness study of ASD in Spain. This paper reports the use of services by individuals on the spectrum and their carers and estimated the sector-related costs of ASD. We also aimed to identify the factors that affect the costs of ASD. The methodology reflects the COI study by MacKay et al. for Scotland [16]. In the framework of the ASDEU project, we translated the questionnaire used in the MacKay study in Scotland into Spanish, and adapted this according to the particular features of the education, healthcare and social care systems in Spain to collect services available in Spain and not included in the original questionnaire.

The Spanish sample was comprised of 334 individuals (82% male, mean age 12 years). As the sample was very young most individuals were at school. Consequently, there was a large number of individuals accessing educational services and education costs were an important part of the total costs. The average annual direct cost of ASD was estimated at €10,974 (95%CI €8716.69 – €13,879.31) per individual, with education, healthcare and social care representing 80%, 17% and only 3% of all direct costs, respectively. These proportions are not very different from those estimated for Scotland, where education sector costs make up almost three quarters and social care approximately 10% of total costs [16]. While some individuals on the spectrum may be employed (e.g., individuals with Asperger’s Syndrome or ASD associated with less disability), most encounter significant barriers to employment [29]. Consequently, and given the small sample of people reporting having a job, lost productivity due to co-occurring conditions in persons on the spectrum was reported separately.

Multivariate analyses highlighted the cost drivers for ASD. As the models’ goodness of fit were low and the models do not seem good enough to fit the observation and predict results, we only interpreted the sign of the model coefficients, that is, the direction of the change in the parameters. By sectors, special education costs were affected by age, type of ASD and other co-occurring conditions. Non-higher education in Spain comprises compulsory primary education between the ages of 6 and 12, compulsory secondary education between 12 and 16 (or alternatively, basic training cycles), and upper secondary education between 16 and 18 [30]. For children with special needs there are both special education centres and resources to support children integrated in mainstream schools. Some aims of the Spanish strategy for autism are individualized education, adaptation of teaching methods and continuous support [31]. The young population in this study appears representative of the Spanish population going to school as 83.4% of children with special needs attend a mainstream school in Spain [32]. Consequently, the results in this group could be considered accurate and generalizable.

Accounting for costs related to health care and social care services was more difficult given the low proportion of adults in the sample and the rare use of health care and social care services compared to education services. The most accessed health care services, apart from the general practitioner, were therapists. Antipsychotics, and especially Risperidone, were the most reported drugs used by both children and adults, as shown by previous studies [33].

The main limitation of this study is sample size and based on a convenience sample. It was not possible to conduct exactly the same analysis as in MacKay et al. [16] because of some differences in social and demographic data, sample size and the distribution of samples. The study in Scotland had a very large sample size (N = 950) where most possible strata according to age and type of disorder had a sufficient size to enable further regression analysis. ASD is a heterogeneous group of disorders that can present a wide range of co-occurring conditions [34], as shown by the different services accessed by individuals with different conditions, ages and co-occurring conditions. Therefore, the representativeness of some of these subgroups (adults for example) would be limited. We believe the sample of children with autistic disorder may be large enough to represent this group.

Other study limitations are the following. Firstly, there is no control group in the study that would enable a more accurate estimate of the incremental cost due to ASD [35], although participants were asked to report on ASD-related costs only. This is a very common limitation in cost-of-illness studies as a comparative prospective study is costlier than a non-comparative prospective study. Secondly, there is uncertainty around the unit costs due to the fact that a large number of services are provided by several public and private institutions across the country. To address this, sensitivity analyses were conducted using unit costs reported by people on the spectrum or their carers when they paid directly for the service. An alternative approach for future studies should be the estimation of unit costs according to standardized methods [36, 37], although this would be a difficult task depending on the sector and institutions as there are services that could be provided by different bodies, sectors and even by two sectors in collaboration, such is the case of some social & health services for people with special needs. Thirdly, costs are expressed in Euros from 2017. These are valid for the period up to 2021, as in 2022 there was a notable increase in inflation in Spain [24] as in the entire EU area. Nevertheless, the information is relevant to know the available and used services in Spain before the COVID-19 pandemics and will be useful for future comparison. Fourthly, service use was collected over the past six months and multiplied by two to estimate the annual costs. This enabled us to mitigate potential recall bias. Fifthly, although the study is based on an anonymous survey, we do not have any reason to believe that the answers were not genuine. Finally, the estimation is a prevalence-based estimation of costs, not an incidence approach, so no lifetime cost can be estimated. Future studies could expand this work estimating the annual total costs at national level using a modular approach and data (prevalence by subpopulations) from other sources such as in Knapp et al. [38]. All in all, these limitations do not hamper the results presented in this manuscript.

This study does not take into account the health burden or the quality of life of individuals on the spectrum and their carers, nor is it possible to conclude whether access of services meets the needs. Nevertheless, the results revealed that a notable proportion of individuals pays for some services and that for most services the prices paid out-of-pocket were higher than the public unit cost used in the analysis, leading to equity concerns. There was also a variety of social care services and most were not frequently used. It is unclear whether the reasons for reduced use are lack of demand, providers and services and/or affordability. Although Spanish ministries coordinate and regulate the public sectors, to avoid inequities for example, the provision of public health care and public education is the responsibility of the regional authorities in the 17 autonomous community regions in Spain. This could account for differences in access to services across participants but it was not possible to assess the impact of the region of residence. Moreover, even in countries with publicly funded health care and social systems, low quality services have been associated with greater financial burdens and dissatisfaction with care [39]. Results from surveys conducted in the ASDEU project [40] highlighted that needs are not always met with access to services and evidence-based recommendations. Qualitative studies would help understand the needs, barriers and factors explaining the access to services [41] and health technology assessment would contribute to the public provision of effective and cost-effective services [42, 43]. Other studies could be conducted to comprehend the range of services actually available in each Spanish region, privately and publicly provided and/or funded.

The results of the ASDEU cost-of-illness study provide current information on the needs of individuals on the spectrum and their families. The results will be useful to characterize and monitor the evolution of these needs, establishing a basis for evaluating both the potential cost-effectiveness and overall economic impact of future educational, social and healthcare policies or interventions in Spain. Ultimately, this will inform service planning and research based on person-centred services for people on the spectrum according to intellectual status, level of disability and co-occurring conditions [16].

Supplementary Information

Below is the link to the electronic supplementary material.

Abbreviations

ASD

Autism spectrum disorders

ADHD

Attention deficit hyperactivity disorder

CDD

Childhood disintegrative disorder

CI

Confidence interval

COI

Cost of illness

OLS

Ordinary least squares

PDD-NOS

Pervasive developmental disorder-not otherwise specified

SD

Standard deviation

Author contributions

L.G.P., V.R.G., R.L., C.V.N, B.S., R.C.B., V.I., M.K., M.L.S., D.S., M.P., P.S.A. and all authors contributed to the study conception and design. Material preparation and data collection were performed by L.G.P., V.R.G., R.L., V.I., R.C., M.P. The analysis was conducted by B.S., L.G.P., V.R.G., C.V.N. The first draft of the manuscript was written by L.G.P. and all authors (L.G.P., V.R.G., R.L., C.V.N, B.S., R.C.B., V.I., M.K., M.L.S., D.S., M.P., P.S.A.) commented on previous versions of the manuscript. M.P. and P.S.A. supervised the project. All authors read and approved the final manuscript.

Funding

The ASDEU (Autism Spectrum Disorder in Europe) Project was funded by the European Parliament, managed by the Directorate-General of Health and Consumers of the European Commission (DG-SANCO), and implemented in Spain by the Carlos III Institute of Health (Ref./SANCO/2014/C2/035).

Data availability

The datasets used during the current study are available from the corresponding author on reasonable request.

Declarations

Ethics approval and consent to participate

The need for written informed consent was waived by the ethics committee from Nuestra Señora de la Candelaria University Hospital, Tenerife, Spain (reference number PI 28/17) due to the anonymity of the study. The online survey was self-administered, voluntary and anonymous. We did not collect any personal information that could identify the participants. The participants were informed about this in the consent information before accessing the survey. The data was stored securely on a password-protected computer and only accessible by the researchers. Written informed consent was not applicable as the data was collected anonymously online. We followed ethical guidelines and Spanish regulations at the time (2017) that allowed anonymous surveys with implied or tacit consent by answering the questionnaire. The study, including the questionnaire and the consent information, received approval by the ethics committee from Nuestra Señora de la Candelaria University Hospital, Tenerife, Spain (reference number PI 28/17).

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.