Introduction
Autism spectrum disorder (ASD, hereafter referred to as autism) is a frequent and often serious neurodevelopmental disorder (Lai, Lombardo, & Baron-Cohen, Reference Lai, Lombardo and Baron-Cohen2014). Co-occurring mental and physical health conditions are frequent in autistic persons – herein we will use identity-first language to refer to autistic people – and are associated with poor functional outcomes and a generally poor quality of life (Lai et al., Reference Lai, Kassee, Besney, Bonato, Hull, Mandy and Ameis2019; Mason et al., Reference Mason, McConachie, Garland, Petrou, Rodgers and Parr2018; Weir, Allison, Warrier, & Baron-Cohen, Reference Weir, Allison, Warrier and Baron-Cohen2021).
Having socially valued roles (e.g. living independently and being employed or studying, being married or a parent) is often used to define ‘good’ adult outcome in autistic people (Mason et al., Reference Mason, Capp, Stewart, Kempton, Glaser, Howlin and Happé2021a). According to Mason et al. (Reference Mason, Capp, Stewart, Kempton, Glaser, Howlin and Happé2021a) meta-analysis, one in five autistic persons will have a ‘good’ adult outcome (i.e. living independently and being employed/studying) and one in four people a ‘fair’ adult outcome (i.e. living independently or being employed/student).
Although for many having meaningful social roles could have positive effects on wellbeing and quality of life (Mason et al., Reference Mason, McConachie, Garland, Petrou, Rodgers and Parr2018; McCauley, Pickles, Huerta, & Lord, Reference McCauley, Pickles, Huerta and Lord2020b; Moss, Mandy, & Howlin, Reference Moss, Mandy and Howlin2017), some autistic persons may find social roles challenging as they can result in increased negative experiences (e.g. loss of employment, social isolation, failed relationships, sexual abuse), poorer mental health and a reduced quality of life (Howlin, Reference Howlin2021; Lounds Taylor, Reference Lounds Taylor2017; Moss et al., Reference Moss, Mandy and Howlin2017; Pecora, Hancock, Mesibov, & Stokes, Reference Pecora, Hancock, Mesibov and Stokes2019). Besides, some of the typical adult experiences – based on socio-cultural norms, roles and expectations in the general population such as finding a job, living in relationship or becoming parent – are not necessarily central or meaningful for every autistic individual and may therefore not be adequate markers of adult outcome in this population (Howlin, Reference Howlin2021).
Good mental health generally contributes to objective (e.g. impact of health status on functional outcome) and subjective (e.g. self-perceived health status and perceived impact on wellbeing) aspects of adult outcome (Lounds Taylor, Adams, & Bishop, Reference Lounds Taylor, Adams and Bishop2017; McCauley et al., Reference McCauley, Pickles, Huerta and Lord2020b; Park et al., Reference Park, Song, Demetriou, Pepper, Norton, Thomas and Guastella2019). However, these associations could depend on the level of social support (Bishop-Fitzpatrick et al., Reference Bishop-Fitzpatrick, Hong, Smith, Makuch, Greenberg and Mailick2016). Aligning with this, depressive symptoms in autistic people diagnosed during early childhood often peak in early adulthood in relation to increased social expectations and challenges but also less support from providers (McCauley, Elias, & Lord, Reference McCauley, Elias and Lord2020a). This could be particularly true for late-diagnosed autistic young adults who achieved positive outcomes without receiving any formal support from transition or mental health services.
Self-reported gender is a prominent facet of a person's identity influencing the experience of autism. According to the ‘female autism phenotype’ (i.e. a behavioral expression of autism that could be more common in women; Hull et al., Reference Hull, Lai, Baron-Cohen, Allison, Smith, Petrides and Mandy2020), while autistic women have similar core autistic symptoms compared with men, they also tend to report poorer mental and physical health and a reduced quality of life (Brugha et al., Reference Brugha, Spiers, Bankart, Cooper, McManus, Scott and Tyrer2016; Dubreucq & Dubreucq, Reference Dubreucq and Dubreucq2021; Hull et al., Reference Hull, Lai, Baron-Cohen, Allison, Smith, Petrides and Mandy2020; Taylor & DaWalt, Reference Taylor and DaWalt2020; Tint et al., Reference Tint, Hamdani, Sawyer, Desarkar, Ameis, Bardikoff and Lai2018). Given autistic women tend to report more often than men to be involved in a relationship or to be a parent (Power et al., Reference Power, Kyaga, Uher, MacCabe, Långström, Landen and Svensson2013; Schöttle, Briken, Tüscher, & Turner, Reference Schöttle, Briken, Tüscher and Turner2017), this could be related to the impact of gendered social, norms, roles, pressures and expectations on a person's wellbeing and sense of identity (Dubreucq & Dubreucq, Reference Dubreucq and Dubreucq2021; Kanfiszer, Davies, & Collins, Reference Kanfiszer, Davies and Collins2017).
According to Lai, Anagnostou, Wiznitzer, Allison, and Baron-Cohen (Reference Lai, Anagnostou, Wiznitzer, Allison and Baron-Cohen2020), support for autistic adults should be strengths-based, person-centered and should aim at maximizing the individual's potential, minimizing barriers and optimizing the person-environment fit. This includes early diagnosis, improving social communication, coping with sensory experiences, treating co-occurring physical health and mental health conditions, encouraging family support and reducing stigma (Howlin, Reference Howlin2021; Lai et al., Reference Lai, Anagnostou, Wiznitzer, Allison and Baron-Cohen2020; Mason et al., Reference Mason, Capp, Stewart, Kempton, Glaser, Howlin and Happé2021a).
In France, based on the above findings, the need to improve early diagnosis, to treat co-occurring health conditions and to offer personalized strengths-based care for autistic individuals led the Ministry of Research and the Ministry of Health to support the development of a national network of Expert Centers for ASD, coordinated by the foundation FondaMental since 2012 (http://www.fondation-fondamental.org; see Henry et al. (Reference Henry, Etain, Godin, Dargel, Azorin and Gard2015) and Schürhoff et al. (Reference Schürhoff, Fond, Berna, Bulzacka, Vilain and Capdevielle2015) for a description of Expert Centers in other conditions such as bipolar disorder or schizophrenia).
This paper presents data on objective (i.e. general functioning and socially valued roles) and subjective (i.e. self-perceived health status) aspects of adult outcome and co-occurring health conditions from a large national multi-centric non-selected sample of 383 autistic adults and no co-occurring intellectual disability (ID). The main aims of this study were to examine the relationship between health status and general functioning and potential gender differences in adult outcome. We made the following research hypotheses:
1) Although good mental health usually contributes to better adult outcome (Brugha et al., Reference Brugha, Spiers, Bankart, Cooper, McManus, Scott and Tyrer2016; Lord et al., Reference Lord, Charman, Havdahl, Carbone, Anagnostou, Boyd and McCauley2022; McCauley et al., Reference McCauley, Elias and Lord2020a, Reference McCauley, Pickles, Huerta and Lord2020b; Park et al., Reference Park, Song, Demetriou, Pepper, Norton, Thomas and Guastella2019), we hypothesized that general functioning would be positively associated with depression in a sample composed of late-diagnosed autistic adults (i.e. who may have been confronted to challenges in early adulthood but received less support from providers; Bishop-Fitzpatrick et al., Reference Bishop-Fitzpatrick, Hong, Smith, Makuch, Greenberg and Mailick2016; McCauley et al., Reference McCauley, Elias and Lord2020a). We also hypothesized that general functioning would be negatively associated with impairment in social communication.
2) According to the literature reviewed above (Brugha et al., Reference Brugha, Spiers, Bankart, Cooper, McManus, Scott and Tyrer2016; Dubreucq & Dubreucq, Reference Dubreucq and Dubreucq2021; Hull et al., Reference Hull, Lai, Baron-Cohen, Allison, Smith, Petrides and Mandy2020; Taylor & DaWalt, Reference Taylor and DaWalt2020; Tint et al., Reference Tint, Hamdani, Sawyer, Desarkar, Ameis, Bardikoff and Lai2018), we hypothesized that there would be gender differences in aspects of adult outcome related to health status (i.e. comparable objective outcomes but poorer mental and physical health and poorer self-reported health status in autistic women).
3) While associations between verbal ability or female gender and co-occurring health conditions can vary according to sample characteristics (e.g. age and sample composition; Day, McNaughton, Naples, & McPartland, Reference Day, McNaughton, Naples and McPartland2020; McCauley et al., Reference McCauley, Pickles, Huerta and Lord2020b), given the nature and age of our sample (i.e. late-diagnosed autistic adults) we made the hypothesis that female gender and verbal ability – two factors linked to compensation (Bargiela, Steward, & Mandy, Reference Bargiela, Steward and Mandy2016; Hull et al., Reference Hull, Lai, Baron-Cohen, Allison, Smith, Petrides and Mandy2020) – would be associated with more co-occurring physical and mental health conditions.
4) Finally, we hypothesized that people reporting first contact with psychiatric services at the visit at the expert center would have fewer co-occurring health conditions than those with a psychiatric history (Fusar-Poli, Brondino, Politi, & Aguglia, Reference Fusar-Poli, Brondino, Politi and Aguglia2022).
Material and methods
Study population
The FondaMental Advanced Centers of Expertise for ASD (FACE-ASD) cohort is made up of participants assessed in a French network of seven Expert Centers for ASD without co-occurring ID (Créteil, Bordeaux, Grenoble, Versailles, Marseille, Caen, Strasbourg). Participants are referred to these centers by their general practitioner or psychiatrist who remains in charge of routine care and treatment, or they are self-referred. The criteria for inclusion in the cohort were: (i) diagnosis of autism (DSM-5; APA, 2013); (ii) age above 18 years; (iii) without co-occurring ID (i.e. IQ > 70). The exclusion criteria were: (i) inability to read or speak French; (ii) non-willingness to give informed consent. However, the present study was restricted to data collected at baseline for participants with a confirmed diagnosis of autism.
During the evaluation, a systematic and comprehensive clinical, functional and cognitive assessment is performed to confirm the diagnosis and establish the individual's strengths and weaknesses, autonomy, and occupational level (online Supplementary Table S1). Follow-up is planned to last for 3 to 5 years. At the end of each evaluation, a detailed evaluation report is sent to the patient and the referring physician along with a personalized care program including the rationale for pharmacological and psychosocial treatment recommendations. Psychosocial treatment includes psychoeducation, cognitive remediation, cognitive behavioral therapy, social skills training, peer-support interventions, supported employment and supported parenting. The relevant Ethical Review Board (CPP- Est IV) approved the appraisal protocol on 18 June 2019. All participants gave their informed consent.
Site selection and training
All seven sites in France were already actively involved in the treatment of autistic people, as well as in research on autism or psychiatric rehabilitation. Each Expert Center has agreed and been trained to use the same package of assessment tools. Clinical team members have monthly group meetings to monitor quality control, to ensure a good inter-rater reliability and to provide training in new therapeutic interventions. During these meetings, clinical cases of potential autism with or without potential co-occurring health conditions (e.g. depression) are also discussed to minimize case ascertainment differences between sites. Case ascertainment procedures are described in online Supplementary Table S4.
Data collection
Screening visit
Social responsiveness is evaluated with the parent-reported and self-reported Social Responsiveness Scale- 2nd version (Constantino and Gruber, Reference Constantino and Gruber2012; Cronbach's alpha = 0.94–0.96; Stordeur, Boele, Peyre, Delorme, and Acquaviva, Reference Stordeur, Boele, Peyre, Delorme and Acquaviva2019). Self-reported autistic symptoms are measured using Ritvo Autism-Asperger Diagnostic Scale (Picot et al., Reference Picot, Michelon, Bertet, Pernon, Fiard, Coutelle and Baghdadli2021).
Diagnostic evaluation
The diagnostic evaluation includes: (ii) the Autism Diagnostic Interview-Revised (ADI-R; Lord, Rutter, & Le Couteur, Reference Lord, Rutter and Le Couteur1994; Rutter, Lecouteur, & Lord, Reference Rutter, Lecouteur and Lord2011) to investigate the developmental history and the informant-reported past and current autistic symptoms; (ii) the Autism Diagnostic Observation Schedule- 2nd version (ADOS-2; Lord et al., Reference Lord, Rutter, Goode, Heemsbergen, Jordan, Mawhood and Schopler1989; Lord, Rutter, DiLavore, & Risi, Reference Lord, Rutter, DiLavore and Risi2015) for current autistic symptoms. When ADI-R is not realizable (e.g. parents not available to participate in the evaluation), the Asperger Syndrome Diagnostic Interview (Gillberg, Gillberg, Råstam, & Wentz, Reference Gillberg, Gillberg, Råstam and Wentz2001) is used alternatively to investigate informant-reported autistic symptoms with one of the person's close relatives other than parents (e.g. the partner).
Medical and cognitive evaluation
General information on education, social outcome, co-occurring mental and physical health conditions and healthcare services utilization is recorded. A medical interview by a trained psychiatrist (DSM-5 criteria) and the Diagnostic Interview for Genetic Studies (DIGS, Nurnberger et al., Reference Nurnberger, Blehar, Kaufmann, York-Cooler, Simpson, Harkavy-Friedman and Reich1994) are used to record co-occurring mental health and physical health conditions. Intellectual quotient is measured using the Wechsler Adult Intelligence Scale-4th edition (WAIS-IV; full scale and subscales; verbal comprehension index for verbal ability and perceptual reasoning index for non-verbal ability; Wechsler, Reference Wechsler2011).
Adult outcome in relation to health status
Health is defined as a ‘state of complete physical, mental and social well-being, and not merely the absence of disease and infirmity’ (WHO, 2014). Health encompasses both objective (e.g. impact of health status on functional outcome) and subjective outcomes (e.g. self-perceived health status and perceived impact on wellbeing). We therefore measured two aspects of health status that can be associated with adult outcome: clinician-rated general functioning and self-perceived health status.
General functioning: General functioning was measured using the Global Assessment of Functioning scale (GAF; Startup, Jackson, and Bendix, Reference Startup, Jackson and Bendix2002). While GAF is not commonly used in autism, its simplicity, its generic nature (i.e. not related to a specific diagnostic group) and its good psychometric properties – in particular in contexts where raters are adequately trained – supported its selection as a measure of functional outcome. As socially valued roles may not be appropriate benchmarks in autism (Howlin, Reference Howlin2021), raters were also told to consider impairment only in outcomes participants identified as desirable for them (e.g. work for some people but not for others). Given concerns have been raised about GAF validity and reliability in routine clinical contexts (Aas, Reference Aas2010), regular meetings and training sessions for raters were organized to ensure a good inter-rater reliability.
Self-perceived health status: Self-perceived health status was measured using the self-report EQ-5D-DL scale (score ranging from 0 to 100; five domains including self-care, mobility, pain/discomfort, anxiety/depression and ability to perform everyday activities; Johnson, Coons, Ergo, and Szava-Kovats, Reference Johnson, Coons, Ergo and Szava-Kovats1998).
Statistical analysis
Data are presented as the mean and s.d. for continuous variables and number and percentage for categorical variables. For comparison between groups, χ2 or Fisher exact test was used for categorical variables Spearman's correlation test, Kruskal-Wallis rank sum test or linear model ANOVAS for continuous variables. Univariate significance p values were computed, and covariates significant at the 10% level were included in a multivariable logistic regression model. Age, gender, education and verbal ability were forced in the final model as confounding factors. The first question was to investigate the correlates of adult outcome. Psychosocial function (GAF continuous score) was retained as the predictor for the first multivariate analysis and self-perceived health status as the predictor for the 2nd multivariate analysis. The second question was to determine whether there would be gender differences in adult outcomes in univariable analyses. The third question was to investigate the correlates of co-occurring health conditions. Considering that most of the sample (>80%) had at least one co-occurring health condition, we retained respectively having three or more physical health conditions and having three or more mental health conditions as the predictors for the 3rd and the 4th multivariable analyses. The fourth question was to determine whether people reporting first contact with psychiatric services at the visit at the expert center (>70%) differed from those with past psychiatric history (i.e. inpatient or outpatient care). Thus, we retained reporting first contact with psychiatric services at the visit at the Expert Center as the predictor for the fifth multivariate analysis. p values < 0.05 were considered significant. All statistical analyses were performed using R (R Foundation for Statistical Computing, Vienna, Austria; https://www.R-project.org/; R Core Team, 2015).
Results
Three hundred and eighty-three autistic persons were recruited from the FACE-ASD network. They had been included in this cohort study between 2013 and 2020. All participants were diagnosed with autism in adulthood, during the evaluation at the expert centers. Participants were mostly men (262; 68.4%) with a mean age and a mean age of diagnosis of 28.1 (s.d. = 10.9) years. The overall IQ was 108.4 (s.d. = 19.6) and the mean ADOS-2 social communication score was 10.9 (s.d. = 4.5). Baseline sample characteristics are presented in Table 1 (see online Supplementary Table S2 for a presentation by gender).
Table 1. Sample characteristics
Overall functioning
Two hundred and sixty-nine participants (82.5%) had high school education and 202 (64.3%) had tertiary education. Participants were mostly single (208; 79.7%), unemployed (232; 72.5%) and had no children (n = 221; 86.3%). They had mild to moderate functional impairment (GAF-F = 57.6; s.d. = 9.4). Eighty-five participants (33.5%) met the criteria for good functional outcome.
Co-occurring health conditions
Participants had on average 2.9 [s.d. = 2.4; (0–10); median = 3.00] co-occurring mental health and 2.5 [s.d. = 1.9; (0–9)]; median = 2.00) physical health conditions. Approximately half of the sample had three or more co-occurring mental health or physical health conditions (respectively 53.3 and 49.4%). The most frequent co-occurring conditions were anxiety disorder (n = 227; 73.2%), depression (lifetime; n = 151; 49%; current, n = 87; 27.3%), sleep disorders (53.4%), gastro-intestinal problems (n = 150; 45.7%) and rheumatology/autoimmune conditions (n = 137; 41.5%). The visit at the Expert Center was the first contact with mental health services for most participants (232; 73.4%). Seventy-three participants (23%) reported a history of psychiatric hospitalization. Data on co-occurring health conditions and healthcare services utilization are shown on Table 1.
Correlates of clinician-report general functioning
Table 2 shows a multivariable analysis of the correlates of psychosocial function. Psychosocial function was best predicted by the absence of depression, impairment in social communication and verbal ability.
Table 2. Linear regression on the correlates of general functioning
GAF, Global Assessment of Functioning.
Bold values refer to values significant at p < 0.05.
Gender differences in adult outcome
Table 3 presents the univariable gender differences in adult outcome. Compared to men, autistic women reported more often to be a parent, to have previous contact with mental health services and to have co-occurring mental health conditions (eating disorders, self-harm behaviors and history of suicide attempt) and a general poorer quality of life.
Table 3. Gender differences in adult outcome (univariate)
Bold values refer to values significant at p < 0.05.
Correlates of self-perceived health status
Table 4 shows a multivariable analysis of the correlates of self-perceived health status. Self-perceived health status was negatively associated with the female gender.
Table 4. Linear multivariable regression on the correlates of self-perceived health status (EQ5D)
Bold values refer to values significant at p < 0.05.
Correlates of co-occurring health conditions
Table 5 presents the results of a multivariate analysis on the correlates of having three or more co-occurring physical health conditions. Having three or more co-occurring physical health conditions was best predicted by female gender, higher verbal abilities and social communication impairments. Table 6 presents the results of a multivariate analysis on the correlates of having three or more co-occurring mental health conditions. Having three or more co-occurring mental health conditions was only predicted by verbal ability.
Table 5. Logistic regression on the correlates of having three or more co-occurring physical health conditions
GAF, Global Assessment of Functioning.
Bold values refer to values significant at p < 0.05.
Table 6. Logistic regression on the correlates of having three or more co-occurring mental health conditions
GAF, Global Assessment of Functioning.
Bold values refer to values significant at p < 0.05.
Correlates of having first contact with psychiatric services
Those reporting that the visit to the expert center was the first contact with psychiatric services did not differ significantly from the other participants. The results are shown on online Supplementary Table S3.
Discussion
Main findings
To our knowledge, this study is the first to examine the correlates of two aspects of adult outcome related to health status, i.e. general functioning and self-perceived health status, in a national multi-center sample of late-diagnosed autistic adults and no co-occurring ID. One third of the sample reported ongoing depression and half of the sample reported three or more co-occurring mental and physical health conditions. Contrary to expectations from our first hypothesis, general functioning was negatively associated with co-occurring depression. We found however as expected a negative association between general functioning and impairment in social communication. Our 2nd hypothesis was supported by the results. Although we found no gender differences in objective markers of outcome (e.g. social communication impairment or verbal ability), autistic women in this study were more likely to have socially valued roles than men (significant association for being parent; trends for significance for tertiary education and intimate relationships). Female gender was associated with poorer physical and mental health (e.g. a 7-fold risk for having three or more co-occurring physical health conditions) and poorer self-perceived health status. Supporting our 3rd hypothesis, the verbal ability was associated with higher risk for having three or more co-occurring physical health conditions. Contrasting with our 4th hypothesis, autistic persons reporting first contact with psychiatric services did not differ from those with past psychiatric history.
Interpretation of the results
Adult outcome
Participants' GAF mean score was also comparable to those reported by Cederlund, Hagberg, Billstedt, Gillberg, and Gillberg (Reference Cederlund, Hagberg, Billstedt, Gillberg and Gillberg2008) and Helles, Gillberg, Gillberg, and Billstedt (Reference Helles, Gillberg, Gillberg and Billstedt2017). One in four participants in this study were employed, one in five were in a romantic relationship and one in seven were parents. While the heterogeneity of previous samples, methods and reported outcomes limits the possibilities of comparisons between studies, the low proportion of people reporting to have socially valued roles in the present study concur with findings from other studies (Farley et al., Reference Farley, Cottle, Bilder, Viskochil, Coon and McMahon2018; Helles et al., Reference Helles, Gillberg, Gillberg and Billstedt2017; Howlin & Moss, Reference Howlin and Moss2012; Power et al., Reference Power, Kyaga, Uher, MacCabe, Långström, Landen and Svensson2013; Strunz et al., Reference Strunz, Schermuck, Ballerstein, Ahlers, Dziobek and Roepke2017).
Gender differences in adult outcome
Compared with autistic men, autistic women in this study had similar impairments in social communication but had poorer subjective adult outcomes (i.e. increased self-harm and suicidal history; more co-occurring physical health conditions and a worse self-perceived health status). This pattern of outcome concurs with previous findings (Hull et al., Reference Hull, Lai, Baron-Cohen, Allison, Smith, Petrides and Mandy2020; Kassee et al., Reference Kassee, Babinski, Tint, Lunsky, Brown, Ameis and Einstein2020; Kõlves, Fitzgerald, Nordentoft, Wood, & Erlangsen, Reference Kõlves, Fitzgerald, Nordentoft, Wood and Erlangsen2021; Weir et al., Reference Weir, Allison, Warrier and Baron-Cohen2021). Autistic women in this study were more likely than men to have socially valued roles. While this concurs with some studies (Strunz et al., Reference Strunz, Schermuck, Ballerstein, Ahlers, Dziobek and Roepke2017; Tint et al., Reference Tint, Hamdani, Sawyer, Desarkar, Ameis, Bardikoff and Lai2018), this is in contrast with other studies that reported no gender differences in adult social outcome (DaWalt et al., Reference DaWalt, Taylor, Bishop, Hall, Steinbrenner, Kraemer and Odom2020; McCauley et al., Reference McCauley, Pickles, Huerta and Lord2020b; Myers, Davis, Stobbe, & Bjornson, Reference Myers, Davis, Stobbe and Bjornson2015) or greater difficulties for autistic women to maintain tertiary education positions over time (Taylor, Henninger, & Mailick, Reference Taylor, Henninger and Mailick2015).
Several factors could contribute to this pattern of gender differences. First, gendered social norms roles, expectations and pressures can have an impact on a person's sense of identity and wellbeing (e.g. social pressures related to motherhood considered by some autistic women as stress-triggering and a daunting life experience; Kanfiszer et al., Reference Kanfiszer, Davies and Collins2017). Second, while autistic women are more likely to be involved in a relationship or parent than autistic men, they also report challenges related to these socially valued roles (e.g. increased risk for sexual or domestic abuse; Bargiela et al., Reference Bargiela, Steward and Mandy2016; Pecora et al., Reference Pecora, Hancock, Mesibov and Stokes2019; peripartum depression, feelings of social isolation in the postnatal period, or communication difficulties with perinatal health providers; Pohl, Crockford, Blakemore, Allison, & Baron-Cohen, Reference Pohl, Crockford, Blakemore, Allison and Baron-Cohen2020). Third, compared with autistic men, autistic women often report more unmet needs related to their mental health concerns, to vocational services and to reproductive health issues; Brugha et al., Reference Brugha, Spiers, Bankart, Cooper, McManus, Scott and Tyrer2016; Dubreucq & Dubreucq, Reference Dubreucq and Dubreucq2021; Taylor & DaWalt, Reference Taylor and DaWalt2020; Tint & Weiss, Reference Tint and Weiss2018). Aligning with this, autistic women are more likely to be undiagnosed or misdiagnosed than autistic men even in epidemiological studies (Brugha et al., Reference Brugha, Spiers, Bankart, Cooper, McManus, Scott and Tyrer2016; Loomes, Hull, & Mandy, Reference Loomes, Hull and Mandy2017).
Another potential explanation to these gender differences could be that autistic men, presumed to have poorer social function, are more often referred to expert centers, whereas only autistic women with severe social communication impairments are referred. Autistic women in this study were more likely to report previous contact with psychiatric services compared with autistic men. Aligning with this, some studies suggested that autistic women may be more likely than men to receive alternative prior psychiatric diagnoses that are no longer present after autism diagnosis, possibly delaying their access to diagnosis and adequate support from mental health services (Brugha et al., Reference Brugha, Spiers, Bankart, Cooper, McManus, Scott and Tyrer2016; Fusar-Poli et al., Reference Fusar-Poli, Brondino, Politi and Aguglia2022; Kentrou, Oostervink, Scheeren, & Begeer, Reference Kentrou, Oostervink, Scheeren and Begeer2021; Leedham, Thompson, Smith, & Freeth, Reference Leedham, Thompson, Smith and Freeth2020; Loomes et al., Reference Loomes, Hull and Mandy2017).
Factors associated with general functioning
Aligning with previous studies on adult outcome (Lounds Taylor et al., Reference Lounds Taylor, Adams and Bishop2017; McCauley et al., Reference McCauley, Pickles, Huerta and Lord2020b; Park et al., Reference Park, Song, Demetriou, Pepper, Norton, Thomas and Guastella2019), good mental health, low impairments in social communication and verbal ability (i.e. verbal knowledge and verbal reasoning) were associated with better general functioning.
Contrasting with Fusar-Poli et al. (Reference Fusar-Poli, Brondino, Politi and Aguglia2022), autistic adults who reported first contact with mental health services at the visit at the expert center – the majority of the sample – did not differ from those with previous psychiatric history. This suggests that autism remains under-detected and that the current healthcare service provision does not meet the needs of autistic people (Lai et al., Reference Lai, Kassee, Besney, Bonato, Hull, Mandy and Ameis2019; MacManus, Bebbington, Jenkins, & Brugha, Reference MacManus, Bebbington, Jenkins and Brugha2016; Weir et al., Reference Weir, Allison, Warrier and Baron-Cohen2021). Efforts should therefore be made to improve the access to early diagnosis and to adequate healthcare, i.e. the provision of the standardized diagnostic evaluation, treatment of co-occurring health conditions and strengths-based psychosocial treatment to optimize the person-environment fit and to improve social inclusion (Lai et al., Reference Lai, Kassee, Besney, Bonato, Hull, Mandy and Ameis2019).
Co-occurring health conditions
In line with Lai et al. (Reference Lai, Lombardo and Baron-Cohen2014), most of the participants (76.2%) reported at least one co-occurring mental health condition and more than 50% reported multiple co-occurring mental health conditions. The mean number of co-occurring mental health conditions was comparable with the findings of Joshi et al. (Reference Joshi, Wozniak, Petty, Martelon, Fried, Bolfek and Biederman2013) in a middle-aged sample of autistic people without co-occurring ID. Having three or more co-occurring mental health conditions was positively associated with verbal knowledge and reasoning – two domains related to higher insight into autism-related challenges and poorer mental health (Livingston & Happé, Reference Livingston and Happé2017; Livingston, Colvert, Bolton, & Happé, Reference Livingston, Colvert, Bolton and Happé2019; McCauley et al., Reference McCauley, Elias and Lord2020a).
The high frequency of co-occurring physical health conditions in autistic people in this study concurs with recent meta-analyses (Rydzewska, Dunn, & Cooper, Reference Rydzewska, Dunn and Cooper2021; Weir et al., Reference Weir, Allison, Warrier and Baron-Cohen2021). In line with these studies, we found that the female gender was associated with a 7-fold risk for having three or more co-occurring physical health conditions. Aligning with Kõlves et al. (Reference Kõlves, Fitzgerald, Nordentoft, Wood and Erlangsen2021), we found no association between suicidal history and the number of co-occurring physical health conditions.
Clinical implications
With replication, this study may have clinical implications. First, we found distinct patterns of associations when considering gender differences in objective (i.e. general functioning) and subjective aspects of health-related adult outcome (i.e. co-occurring mental health conditions and self-perceived health status). Aligning with findings from other studies (Brugha et al., Reference Brugha, Spiers, Bankart, Cooper, McManus, Scott and Tyrer2016; Dubreucq & Dubreucq, Reference Dubreucq and Dubreucq2021; Hull et al., Reference Hull, Lai, Baron-Cohen, Allison, Smith, Petrides and Mandy2020; Tint et al., Reference Tint, Hamdani, Sawyer, Desarkar, Ameis, Bardikoff and Lai2018), while autistic women in this study were more likely to have social roles than autistic men, they also reported poorer health outcomes. This further supports the need to improve autistic women's access to diagnosis, to adequate treatment and more generally to service provision that could address their unique challenges and needs for care (e.g. provider training to the female autism phenotype or extra support in the domains of romantic relationships and reproductive health; Dubreucq & Dubreucq, Reference Dubreucq and Dubreucq2021; Hull et al., Reference Hull, Lai, Baron-Cohen, Allison, Smith, Petrides and Mandy2020; Strang et al., Reference Strang, van der Miesen, Caplan, Hughes, daVanport and Lai2020).
Second, co-occurring physical health conditions are common and are associated with poorer adult outcomes, but remain often under-detected and under-treated (Rydzewska et al., Reference Rydzewska, Dunn and Cooper2021; Weir et al., Reference Weir, Allison, Warrier and Baron-Cohen2021). The barriers to seeking care include impairment in social communication, negative experiences with healthcare providers and healthcare providers' lack of knowledge about autism/about the female autism phenotype (Calleja, Islam, Kingsley, & McDonald, Reference Calleja, Islam, Kingsley and McDonald2020; Dubreucq & Dubreucq, Reference Dubreucq and Dubreucq2021; Mason et al., Reference Mason, Ingham, Urbanowicz, Michael, Birtles, Woodbury-Smith and Parr2019; Tint & Weiss, Reference Tint and Weiss2018). Information and training on how to provide adequate care to autistic persons (e.g. improving providers' knowledge about autism, reducing barriers to communication of one's health needs) directed to healthcare providers are therefore needed (Calleja et al., Reference Calleja, Islam, Kingsley and McDonald2020; Mason et al., Reference Mason, Ingham, Birtles, Michael, Scarlett, James and Parr2021b, Reference Mason, Taylor, Ingham, Finch, Wilson, Scarlett and Parr2022; Weir et al., Reference Weir, Allison, Warrier and Baron-Cohen2021).
Limits
This study has also a number of limitations. Although our network of Expert Centers for autism covers a large proportion of the French territory, it cannot be definitively asserted that it constitutes a representative sample of the French population of autistic people. Besides, given the discrepancy between highly frequent co-occurring health conditions in autistic adults and a healthcare utilization that does not differ from the general population (MacManus et al., Reference MacManus, Bebbington, Jenkins and Brugha2016), non-epidemiological studies may be only representative of people who present to or are known to mental health services (Brugha et al., Reference Brugha, Spiers, Bankart, Cooper, McManus, Scott and Tyrer2016). Moreover, the present sample is constituted of autistic persons diagnosed during adulthood, which may limit the extension of some results to other populations. Considering the criteria for defining adult outcome in autism, the cross-sectional nature of this study may be a limitation as the subjective aspects of adult outcome (e.g. self-perceived health status) refer to a process rather to an outcome and thus may vary over time. The use of the DIGS- an interview which has not been designed to guide the examiner on how to disentangle symptoms that could be due to autism from symptoms related to other neurodevelopmental conditions- to screen for potential co-occurring health conditions is a substantial limitation as it may lead to artificial inflation of these conditions rates (Clark, Cuthbert, Lewis-Fernández, Narrow, & Reed, Reference Clark, Cuthbert, Lewis-Fernández, Narrow and Reed2017; Reed et al., Reference Reed, First, Kogan, Hyman, Gureje, Gaebel and Saxena2019). This limitation could be addressed through the implementation of the forthcoming revised WHO-SCAN interview. However, the monthly organization of meetings between clinical team members from all centers and the attention given to this issue in the case ascertainment procedure used by all sites might have reduced this risk. The absence of autistic individuals with co-occurring ID is another limitation, as it cannot be asserted that our results are also valid for this population. While socially valued roles may not be appropriate benchmarks in autism, our choice to define adult outcome using a measure of general functioning is a limitation as it limits the possibility to make direct comparisons with the results of other adult follow-up studies. While it cannot be asserted this definition, which relies mainly on a restricted health-related measure of functional outcome, is more valid than the usual definition of good adult outcome based on socially valued roles, raters were told to consider functional impairment only in outcomes participants identified as desirable for them. While concerns were raised about GAF's validity and reliability in routine clinical settings (Aas, Reference Aas2010), the organization of regular training sessions/inter-rater meetings might have reduced this limitation. However, the implementation of more general measures of quality of life, commonly used and validated in autism (e.g. the WHO-QoL-Bref; Ayres et al., Reference Ayres, Parr, Rodgers, Mason, Avery and Flynn2018) is needed.
Supplementary material
The supplementary material for this article can be found at https://doi.org/10.1017/S0033291722002902
Acknowledgements
The authors are grateful to the reviewers of a previous version of the manuscript for their thoughtful comments that were very helpful. We thank the FondaMental Foundation (www-fondationfondamental.org), which is a non-profit foundation supporting research in psychiatry in France and coordinating the infrastructure of Expert Centers for ASD. We express all our thanks to the patients who have accepted to be included in the present study. We thank the team of FondaMental foundation, Hakim Laouamri and his team (Seif Ben Salem, Karmène Souyris, Victor Barteau and Mohamed Laaidi) for the development of the FACE-ASD computer interface, data management, quality control and regulatory aspects.
Author contributions
The authors had full access to the data in the study and take the responsibility for the integrity of the data and the accuracy of the data analysis. Dr Julien Dubreucq drafted the article and made the literature review. Mr Mohamed Lajnef conducted the statistical analysis. Dr Romain Coutelle, Dr Ophélia Godin, Pr Mario Speranza and Pr Marion Leboyer contributed to the interpretation of data and critically revised the article. All the authors were involved in the collection and analysis of the data. All authors contributed to and have approved the final manuscript.
Financial support
This work was funded by AP-HP (Assistance Publique des Hôpitaux de Paris), Fondation FondaMental (RTRS Santé Mentale), by the GIS (Groupement d'Intérêt Stratégique) Autisme by the Investissements d'Avenir program managed by the ANR under reference ANR-11-IDEX-0004-02 and ANR-10-COHO-10-01, and by INSERM (Institut National de la Santé et de la Recherche Médicale). The funding source had no role in the study design, data collection, analysis, preparation of the manuscript, or decision to submit the manuscript for publication.
Conflict of interest
None.
