Self and informant reports of mental health difficulties among adults with autism findings from a long-term follow-up study

Abstract

Data on psychiatric problems in adults with autism are inconsistent, with estimated rates ranging from around 25% to over 75%. We assessed difficulties related to mental health in 58 adults with autism (10 females, 48 males; mean age 44 years) whom we have followed over four decades. All were of average non-verbal intelligence quotient when diagnosed as children. Overall ratings of mental health problems were based on data from the Family History Schedule (Bolton et al., 1994). Informant reports indicated that many of the cohort (44%) had experienced no mental health problems in adulthood; 28% had experienced mild to moderate difficulties, 23% had severe and 5% very severe problems. Depression was the most commonly reported problem. Among those adults (n = 22) able to report on their own mental state, again many (45%) reported no mental health problems, although 27% reported very severe mental health problems related to anxiety, depression and/or obsessive–compulsive symptoms. Informant ratings of poor mental health were not associated with gender, severity of autism in childhood, or child or adult intelligence quotient, but there were small correlations with overall social functioning (rho = 0.34) and current autism severity (rho = 0.37). The findings highlight the difficulties of assessing mental health problems in adults with autism and the need for appropriately validated measures.

Keywords

adults autism spectrum disorders mental health psychiatric comorbidity

Data on rates of mental health problems in adults with autism are often contradictory (see reviews by Davis et al., 2011; Levy and Perry, 2011; Mannion et al., 2014; Mazzone et al., 2012), with figures for major psychiatric disorder ranging from 20%–25% (Hutton et al., 2008) to over 75% (Joshi et al., 2013). Although the higher figures tend to be based on samples specifically referred for psychiatric assessment, 69% of adults in the epidemiological sample of Buck et al. (2014) were identified as meeting lifetime criteria for psychiatric disorder. In contrast, Hutton et al. (2008) found that rates of adult-onset psychiatric disorders were comparable with population norms and Brugha et al. (2011) found no evidence of increased use of mental health services among adults with autism.

There is greater consistency with regard to the types of psychiatric problems associated with autism. Depressive and anxiety-related disorders, including phobias and obsessive–compulsive disorder (OCD), are among the most frequently reported; attention deficit hyperactivity disorder (ADHD) is also common, while rates of schizophrenia or severe psychosis appear relatively low (Buck et al., 2014; Lainhart, 1999; Lugnegård et al., 2011; Mazzone et al., 2012; Sterling et al., 2008). However, estimates of specific disorders also vary widely. Thus, Mazzone et al. (2012), in a review of psychiatric comorbidities, reported rates of depression ranging from 17% to 70%; for bipolar disorder the range was 8% to 27%, for ADHD symptoms 28% to 65% and for anxiety 42% to 56%.

There are several reasons for these inconsistencies. Sampling is a particular issue, since mental health problems will be more frequent among psychiatric referrals than in community-based samples. Estimates are also affected by the assessment measures employed. For example, the 24 studies reviewed by Mazzone et al. (2012) included a total of 26 different assessment measures. Buck et al. (2014) found that parents reported 34% of individuals as having a clinically diagnosed psychiatric disorder whereas assessments using the Mini Psychiatric Assessment Schedules for Adults with Developmental Disabilities (PAS-ADD) (Prosser et al., 1998) identified 69% as having a lifetime psychiatric disorder. Buck et al. (2014) also found little agreement between carers’ descriptive reports of anxiety (11% of cases identified) and their responses on the Mini PAS-ADD (53% of cases identified with anxiety). The situation is further complicated by the lack of validated mental health measures for use with adults with autism. In the general population, diagnosis typically relies on self-report, but this may not always be possible in autism. On the other hand, informant reports may not accurately represent the extent of mental health problems experienced by the individual himself or herself, and there is little consensus regarding the relative value of informant versus self-report data.

Factors associated with mental health difficulties in adults with autism

It has been suggested that individuals with autism of average or above intelligence are more likely to experience mental health problems, particularly depression, than individuals of lower intelligence (e.g. Hill et al., 2004; Sterling et al., 2008), possibly because they are more exposed to social demands and/or because they are more aware of their social limitations. However, again, this finding is inconsistent. Hutton et al. (2008) and Farley et al. (2009) noted that mental health problems occurred among individuals of all ability levels and Buck et al. (2014) found that the association between intelligence quotient (IQ) and anxiety varied according to the measures used.

The impact of other factors remains unclear. The role of gender has not been systematically studied, mainly because of the very small numbers of females in most studies. Regarding age, Esbensen et al. (2009) recorded increased use of psychotropic medication after the age of 20 years and Hutton et al. (2008), noted that onset of new psychiatric disorders often occurred in the early to mid-20’s. However, as in the general population (Kessler et al., 2007), Totsika et al. (2010) found that rates of psychiatric disorder in adults with developmental disabilities tended to decrease with age. Data on the relationship with autism severity are also inconsistent. There are some reports of a negative association between autism severity and mental health (Sterling et al., 2008), but Simonoff et al. (2012) found no association between these variables in their community cohort of adolescents.

Finally, despite a strong association between mental health and social functioning among adults in the general population (e.g. Baumgartner and Burns, 2014), there is relatively little research on the relationship between mental health problems and outcome in autism. Nevertheless, several studies (e.g. Cederlund et al., 2008; Farley et al., 2009; Hofvander et al., 2009) note that adults with autism and mental health difficulties were functioning particularly poorly.

Background to this study

The present investigation is the latest stage in a 40-year longitudinal study of individuals diagnosed with autism as children at the Maudsley Hospital, London. Members of this cohort were involved in earlier follow-up studies of cognitive, social and mental functioning (e.g. Howlin et al., 2004; Hutton et al., 2008). The current focus is on individuals who, when diagnosed in childhood, did not have intellectual disability (see Howlin et al., 2013, 2014).

Aims

Our aim was to explore mental health difficulties among an adult cohort initially referred for a diagnosis of autism in childhood. As children all had a non-verbal IQ ⩾ 70. The principal research questions addressed were as follows.

What are the rates and types of mental health problems reported in this group?

What is the consistency in reported rates of problems between different measures/different informants?

Which characteristics (age, gender, IQ and severity of autism in child and adulthood, current social functioning) are associated with adult mental health?

Methods

Participants

From clinical case records of individuals seen between 1950–1979, 90 consecutive cases were identified who met the following criteria: confirmed childhood diagnosis of autism (made between 2–13 years), childhood non-verbal IQ ⩾ 70 and current age ⩾ 21 years. In 2008–2011, we traced 82 of these individuals, of whom 60 (67% of the eligible sample) agreed to participate. Insufficient data were available for two individuals (one male and one female); both had limited communication skills and neither had a family member or other informant to report for them. The remaining sample comprised 48 males and 10 females. Average age when first seen was 6 years 9 months (standard deviation (SD) = 2 years 9 months; range = 2–13 years); current age: mean 43 years 3 months (SD = 9 years 1 month; range = 29–64 years).

Average time between initial assessment and follow-up was 37 years. There were no significant differences in childhood characteristics when originally assessed between non-participants and current participants (non-verbal IQ, mean 88.9 vs 88.7; language level at initial assessment: spontaneous phrase speech: 62% vs 57%; total Autism Diagnostic Interview (ADI) score, mean 39.2 vs 39.7).

Measures

Autism severity and diagnostic confirmation

For individuals seen in the early years, initial clinical diagnoses were made/confirmed by M. Rutter, using International Classification of Diseases (ICD)-9 or ICD-10 criteria; diagnosis was subsequently re-confirmed retrospectively using the ADI (Le Couteur et al., 1989) when this became available. Current diagnosis was re-confirmed using the Autism Diagnostic Interview–Revised (ADI-R), (Rutter et al., 2003). Assessments of autism severity in childhood were based on ADI algorithm scores for Communication, Reciprocal Social Interaction and Restricted Repetitive and Stereotyped Behaviour and, in adulthood, on ‘Current’ scores on the same ADI-R domains. In confirming diagnosis, we used the criteria of Risi et al. (2006) for autism spectrum disorder (ASD) (i.e. meeting criteria on at least two core ADI-R domains plus age of onset ⩽36 months).

Cognitive and language assessments

Because of the very variable abilities of participants, and differences in IQ tests with age, assessment measures varied over time. In both childhood and adulthood, a ‘best estimate’ IQ score was derived from the best quality, age appropriate test completed. Ratings for overall use of language were based on the ADI-R; language comprehension was assessed on the British Picture Vocabulary Scale (BPVS, 2nd ed), Dunn et al., 1997). Although all participants completed standard IQ tests (Wechsler, Raven’s, Leiter or Merrill Palmer) when first seen, in adulthood 15 individuals could not be directly assessed; for them an estimate of mental age was obtained from the Vineland Adaptive Behavior Scales (VABS; Sparrow et al., 2005) (all cognitive details in Howlin et al. (2014); see also Supplemental Text, Note 1).

Social outcomes

Social functioning in adulthood was assessed primarily on the Family History Schedule (FHS) (Bolton et al., 1994; Fombonne et al., 1997), a semi-structured interview developed for multi-site genetic studies (e.g. Bolton et al., 1998; Pickles et al., 2013; Pinto et al., 2010). From FHS scores for residential status, employment, friendships and intimate relationships (each scored 0 – no problems to 3 – major difficulties) we derived a composite social outcome rating (0 – very good to 4 – very poor functioning) as used in previous studies (Clegg et al., 2005; Farley et al., 2009; Howlin et al., 2004, 2013; see Supplemental Text, Note 2 for scoring examples and Supplemental Table 1(a) and (b) for rating details).

Mental health throughout adulthood – self and informant ratings

The FHS also provided information on the occurrence of major mental health difficulties since the age of 16 years. These included symptoms of OCD, depression (episodic and chronic), bipolar disorder, anxiety disorder and ‘other’ (e.g. schizophrenia, mania and paranoia). Each disorder is rated on a 4-point scale (0 = no mental health difficulty, 3 = hospitalised because of the disorder). (See Supplemental Text, Note 2 for examples of FHS questions and scoring). In addition, the presence of self-injurious behaviour was rated from the ADI-R. The decision to include this was based on the fact that 11 individuals showed levels of self-injury, often accompanied by physical aggression to others, that severely restricted their social activities and/or others’ contact with them (for fear of being attacked or causing an outburst). It seemed inappropriate to rate these individuals as having ‘No disorder’ as the difficulties were clearly associated with some form of stress, although whether due to external or internal causes was unknown (adult IQ in all but two of these individuals was ⩽60).

On the basis of their total score, participants were assigned an overall mental health rating from 0 (no difficulties) to 4 (very severe difficulties; see Table 1). Informants for the FHS and ADI-R comprised 40 mothers, 5 fathers, 2 siblings and 10 non-relatives (one general practitioner (GP), two psychiatrists and the remainder residential care staff who had daily or almost daily contact with the individuals involved); one individual had no informant data. Self-report data on the FHS were obtained from those individuals (n = 22) who were able to report on their own social and mental health situations (all had a BPVS language comprehension age > 13 years and adult IQ > 70).

Table 1.

Summary mental health ratings.^a

Overall rating
No problems	No difficulties in any area (all items scored 0)
Mild problems	Mild symptoms (i.e. score of 1) in one area)
Moderate problems	Mild difficulties (score of 1) in >1 area
Severe problems	Persistent problems and/or requiring treatment (score of 2) in 1 area
Very severe	Severe difficulties (score of 2) in >1 area
	Or In-patient hospital treatment required.
	Or Severe difficulties in one area and mild difficulties in 2 areas

0 = none; 1 = symptoms lasted < 1 month, or no significant distress or impairment; 2 = symptoms lasting ⩾1 month, associated with distress or impairment and/or received treatment; 3 = hospitalised for at least 2 days, or required electroconvulsive therapy (ECT).

Scoring for individual disorders (see Supplemental Text, Note 2 for examples of FHS probes and codings).

Current mental health – self-report

Individuals able to report on their own mental health were also asked to complete a number of standardised self-report measures.

The General Health Questionnaire-12 (GHQ-12), Goldberg and Williams, 1988) is a 12-item (0–3 point scale) psychiatric screen; a cut-off score of 11 indicates clinical ‘caseness’.

OCDs were assessed using a modified version of the Yale–Brown Obsessive Compulsive Scale (Y-BOCS), Goodman et al., 1989). In this, the ASD-related items included in the Children’s Yale–Brown Obsessive Compulsive Scale modified for pervasive developmental disorders (CY-BOCS-PDD) (Scahill et al., 2006, for example, lining up objects) were added to the standard Y-BOCS; unlike the CY-BOCS-PDD, the obsessive section was retained.

Anxiety and Depression were assessed with the Beck Anxiety Inventory (BAI)-II (Beck, 1993) and Beck Depression Inventory (BDI) (Beck et al., 1996). Items are rated from 0 to 3 (high severity). Both scales have extensive data on validity and reliability.

ADHD symptoms were assessed using the Adult ADHD Self-Report Scale (ASRS-v1.1; Kessler et al., 2005), an adult specific checklist comprising 18 items (each rated 0 – no symptoms to 4 – very frequent symptoms). The scale has high internal consistency and good concurrent validity with the clinician-administered ADHD rating scale (Adler et al., 2006).

A total of 21 adults (who also had self-report and informant data on the FHS) completed the BAI, BDI and ASRS; 20 completed the GHQ but only 17 completed the Y-BOCS. One individual refused all standardised tests, considering them inappropriate for people with ASD.

Procedures

Assessments were conducted in participants’ own homes/residential placements or in the clinic depending on personal choice. Ethical approval was obtained from the Joint South London and Maudsley/Institute of Psychiatry National Health Service (NHS) Research Ethics Committee (reference: 07/H0807/65; 19/12/07).

Inter-rater reliability

Scripts of 20 ADI-R interviews (all identification data and original scores removed) were randomly selected and independently rated (by S.S., P.M., P.H.). Item-level, quadratic-weighted Kappa’s ranged from 0.57 to 1.00 (89% ⩾ 0.73). Intra Class Correlations for ADI-R total and domain scores ranged from 0.93 to 0.99. Item-level agreement for adult social outcome scores (for 50 randomly selected participants) ranged from k = 0.77 to k = 0.99; intraclass correlation coefficients (ICCs) for composite scores ⩾ 0.99. For mental health, item-level agreement ranged from k = 0.65 to k = 0.93; ICCs for informant and self-report summary ratings = 0.82–0.89.

Statistics

Most analyses were conducted using non-parametric analyses (Spearman rho, Fisher Exact, or Mann–Whitney U tests), as ratings for mental and social outcomes were ordinal and non-normally distributed. All tests were two-tailed with alpha level set at p < 0.01 to compensate for multiple computations. Missing data were dealt with by step-wise deletion. Data on the individuals who could/could not provide self-report information are provided below.

Results

Participant characteristics

Tables 2 and 3 summarise IQ and ADI-R scores in child and adulthood, and adult social outcome ratings for the total cohort. Participants (n = 22) able to report on their own mental health on the FHS had significantly lower current ADI-R scores than individuals unable to self-report (M = 30.8 (SD = 6.3) vs 20.3 (SD = 7.6), t = −5.61; p < 0.001), adult IQ scores were significantly higher (M = 97.4 (SD = 19.6) vs 53.6 (SD = 30.5) t = 5.90; p = < 0.001) and age did not differ (M = 45.1 years (8.5) vs 42.2 (9.4) t = 1.17, p = 0.25). None of the 15 adults who could be assessed only on the Vineland could self-report.

Table 2.

Cohort characteristics.^a

	Child assessment	Current assessment
	Mean (SD) [range]	Mean (SD) [range]
Age	6.7 years (2.2) [2–13]	43.3 years (9.1) [29–64]
‘Best estimate’ IQ	88.7 (14.8) [70–133]	69.7^b,c (34.3) [20–139]
	ADI at diagnostic confirmation^d	ADI-R ‘current’ scores^e
Total	39.7 (8.4) [15–56]	26.9 (8.5) [7–47]
Reciprocal Social Interaction	18.8 (5.7) [2–28]	13.6 (5.1) [0–20]
Communication	13.6 (3.9) [6–22]	9.8 (4.1) [2–20]
Repetitive and Stereotyped Behaviours	7.4 (2.1) [2–11]	3.5 (2.2) [0–9]

SD: standard deviation; IQ: intelligence quotient; ADI: Autism Diagnostic Interview; ADI-R: Autism Diagnostic Interview–Revised.

See Howlin et al. (2013, 2014) for full details.

One individual refused all standardised tests, considering them inappropriate for people with ASD.

Mean IQ excluding 15 individuals assessed on Vineland only in adulthood = 87.3 (20.1).

53 participants ⩾ADI cut-off on three domains at diagnostic re-confirmation; 5 ⩾ cut-off on two domains).

All adults meeting ADI-R criteria in ⩾ two core domains.

Table 3.

Adult social outcomes.^a

Outcome measure^a	n (%)
Employment
Professional/highly skilled	3 (5)
Non/manual skilled	6 (10)
Part/unskilled	9 (16)
Sheltered employment/long-term unemployed	40 (69)
Living situation
Independent	7 (12)
Semi-independent	18 (31)
Residential home	31 (53)
Hospital care	2 (3)
Close relationships
Marriage/long-term relationship	4 (7)
Short-term relationships only	6 (10)
No close reciprocal relationships	48 (83)
Social outcome rating
Very good	4 (7)
Good	5 (9)
Fair	14 (24)
Poor	15 (26)
Very poor	20 (34)

See Howlin et al. (2013) for full details.

Mental health outcomes

Table 4 summarises FHS informant and self-report data on OCD, anxiety and depression (no other disorders were reported). FHS informant data were available for 9 females and 48 males; only 2 females were able to self-report.

Table 4.

FHS mental health: self and informant ratings.

Number of informants = 57 and number of self-report = 22	OCD, n (%)	Depression, n (%)	Anxiety, n (%)
Never any symptoms
Informant	43 (75)	41 (72)	42 (73)
Self	12 (55)	14 (64)	12 (55)
Mild symptoms present for less than 1 month/no significant interference in life
Informant	4 (7)	2 (3)	4 (7)
Self	1 (5)	2 (9)	3 (14)
Symptoms lasted at least one month/received treatment for the disorder
Informant	2 (3)	13 (23)	8 (14)
Self	6 (27)	6 (27)	6 (27)
Hospitalised because of the disorder
Informant		1 (2)	–
Self		–	1 (4)
Don’t know/unable to rate severity
Informant	9 (16)	1 (2)	4 (7)
Self	3 (14)	–	–

OCD: obsessive–compulsive disorder.

FHS informant ratings

Depression, lasting for several weeks at a time and/or resulting in the need for treatment, was the most frequent specific problem reported, occurring in 14 individuals (25%; 11 males, 3 females), one of whom had required in-patient admission; anxiety was reported in 8 adults (14%; 7 males, 1 female); 2 males had significant OCD problems, requiring treatment (see Table 4). In all, 5 individuals (8%; 4 males, 1 female) were described as having marked problems in two or more areas. In addition, 11 (19%, 9 males, 2 females) were reported to show moderate to severe self-injurious behaviours (one resulting in severe self damage); 5 of these also exhibited physically aggressive behaviours to others. Five individuals described as showing severe aggression to self or others had no other reported mental health problems; in the remainder, these behaviours seemed to be associated with anxiety or depression.

On the basis of overall informant mental health ratings, 25 individuals (44%; 21 males, 4 females) were described as having no problems, 16 (28%; 14 males, 2 females) were rated as having mild to moderate difficulties and 13 (23%; 11 males, 2 females) and 3 (5%; 2 males, 1 female) as having severe or very severe problems respectively. Of the 15 adults who could be assessed only on the Vineland, 12 were described as having at least moderate problems (3 moderate, 5 severe and 4 very severe).

FHS self-ratings

Among the individuals who provided self-reports on the FHS, anxiety was the most common problem (31%) and more than one-quarter had experienced significant depression and obsessive–compulsive difficulties (see Table 4). The two females reported only mild problems (one with anxiety and one with depression). In total, 10 individuals (45%) were rated as having no mental health problems, 5 (23%) had mild-moderate difficulties, 1 had severe problems and 6 (27%) very severe problems. Six reported moderate to severe problems in more than one area.

Other self-report screening measures

Among the 21 individuals (19 males, 2 females) completing the Beck and ADHD scales, two were at or above the clinical cut-off scores for moderate to severe problems on the BAI and BDI; the number scoring in the clinical range on the ASRS-v1.1 was also low (see Table 5). Eight individuals scored above cut-off on the GHQ. Only 17 individuals completed the Y-BOCS (one refused; two others had IQ ⩽ 75); of these 3 (18%) scored at or above threshold for moderate problems; two had extreme problems.

Table 5.

Self-reported mental health difficulties.

Measure	At or above clinical cut – off/ moderate to severe disorder, n (%)
^aGHQ-12 (n = 20)	8 (40)
^bY-BOCS (n = 17)	5 (29)
^cBDI (n = 21)	2 (10)
^dBAI (n = 21)	2 (10)
^eASRS-v1.1 (n = 21)	2 (10)

GHQ: General Health Questionnaire (maximum score = 36; cut off = 11).

Y-BOCS: Yale–Brown Obsessive Compulsive Scale (maximum = 40; moderate disorder 16–23; severe 24–31; extreme > 31).

Beck Anxiety Inventory (BAI): maximum = 63; mild disorder = 8–15; moderate = 16–25; severe > 26).

Beck Depression Inventory (BDI): maximum = 63; mild = 11–20; moderate = 21–30; severe > 30).

ASRS: Adult ADHD Self-Report Scale (maximum = 72; clinical cut-off = 4; positive scores on core items = 1–6).

Consistency between measures

FHS self versus informant

Correlations between self and informant FHS ratings for the 21 individuals with both sources of information were moderate (Spearman’s rho = 0.54, p = 0.01). There were no differences in ratings for overall mental health (self and informant medians = 0; range 0 to 4; z = 1.27; p = 0.20) and although more individuals rated themselves as having very severe problems than their informant the difference was non-significant (ns = 6 vs 1; Fisher Exact test, p = 0.09). Self-reports of anxiety, OCD and depression also indicated higher levels than described by caregivers although the differences were not significant (Self vs Informant: anxiety, ns = 7 vs 4; OCD, ns = 6 vs 2; depression ns = 6 vs 3; Fisher Exact test, p values all ⩾0.23). There were no reports, from either source, of diagnoses of mania, bipolar disorder or schizophrenia.

We also explored whether informant ratings differed for individuals who were, or were not, able to report on their own mental health. No significant differences were found. (Informant ratings for those with self-report vs without self-report: median 0; range 0 to 5 vs 1.0; range 0 to 5; Mann–Whitney z = 1.23; p = 0.22.)

FHS self-report versus other self-report measures

Among the 21 individuals with both FHS and other self-report data the proportion rated with moderate to severe OCD problems was similar for the Y-BOCS (29%) and FHS (28%; Fisher Exact test p = 1.00) The proportions scoring within the clinical range for anxiety and depression on the Beck questionnaires were considerably lower than indicated by FHS self-reports (10% moderate-severe scores on both BDI and BAI; 28% and 33%, respectively, reporting significant depression/anxiety on the FHS), although the difference was not statistically significant (Fisher exact tests, ps .13 and .24).

Factors associated with adult mental health

Because of the small number of self-report questionnaires completed, only informant ratings were used in the following analyses.

Childhood variables

There were no significant correlations (p < 0.01) between informant-based total mental health scores in adulthood and childhood measures (IQ; ADI total or domain scores; all correlations rho ⩽.22; p values 0.10 to 0.50).

Adult variables

There were no gender differences. Three out of the nine females with informant data were described as having severe or very severe problems compared with 13 out of 48 males (Fisher Exact test p = 0.70); median scores did not differ (median for both = 1.0; range = 0–5; Mann–Whitney z = −0.07, p = 0.95).

Correlations between total mental health and other adult scores (age, IQ and ADI-R domain scores) were also non-significant (at p < 0.01), (all correlations rho = 0.23–0.29; p values = 0.09–0.02). There was, however, a small but significant positive correlation with total social outcome scores (rho = 0.34; p = 0.01) and with total ADI-R score (rho = 0.37; p = < 0.01), such that those with the highest mental health scores also had the poorest social outcomes and the highest number of current symptoms on the ADI-R.

Discussion

This study investigated the frequency and correlates of mental health problems in 58 individuals with autism in middle adulthood. The findings support other reports of relatively high rates of mental health problems, especially related to depression, anxiety or OCD, in autism (cf. Mazzone et al., 2012), Nevertheless, it is important to note that outcomes were extremely varied, from individuals who had required in-patient treatment to those who showed no psychopathology, and almost half the sample had experienced no mental health problems during adulthood. Exploration of factors related to mental health revealed no strong association with age, gender or childhood autism severity. There were no significant correlations with child or adult IQ; among those with severe/very severe problems, half currently had an average or above average IQ; half scored below 70. Other studies (e.g. Farley et al., 2009; Hutton et al., 2008) have also failed to find an association with IQ. For example, Hutton et al. (2008), in their study of younger adults (n = 135) found that the onset of new psychiatric problems occurred in individuals of normal intelligence as well as in those with severe intellectual impairments. As in many other studies, the effects of gender are unclear because of the small numbers of females involved, although Lai et al. (2011) also found no male/female differences in rates of clinically significant mental health symptoms.

Unlike some studies (e.g. Billstedt et al., 2005; Ghaziuddin and Zafar, 2008), we found only weak correlations between informant reports of psychiatric problems and adults’ current social functioning or severity of autistic symptoms. Several individuals who were very socially isolated and highly dependent on others showed no sign of depression, anxiety or other overt psychiatric disorder. In contrast, others who were working and living independently and had close relationships, had serious mental health problems. Hutton et al. (2008) also reported widely varying mental health outcomes and little relationship with social functioning. However, while the correlations between mental health, social functioning and current autism severity, were small, they were significant, suggesting that, although these are separate to an important extent, they are not totally unrelated.

Direct comparison with other studies is difficult due to many issues that are relevant to research on psychopathology in autism more generally. First, even comparisons with normative data (e.g. Andrews et al., 2001; Brugha et al., 2011; Kessler et al., 2007) present problems because national estimates vary widely, depending on sampling, time frame, assessments used, disorders studied and diagnostic criteria employed. Comparisons with other autism studies are compromised for similar reasons.

Second, differing measures result in inconsistencies both within and between studies. In the present cohort, although correlations between self and informant reports on the FHS were moderately high, there were more self-reports of very severe mental health problems. For example, five individuals, who were described by parents as having no or very mild problems, themselves reported significant difficulties related to anxiety, depression or OCD. Moreover, Beck self-ratings of anxiety and depression tended to indicate lower rates of problems than self-report FHS data – probably because the FHS allows for more direct prompts and probes. As yet there is no ‘gold standard’ psychiatric assessment that has been specifically validated for adults with autism (Gotham et al., 2015). Concerns about the most appropriate mental health assessments to use, the relevant time frame (recent or lifetime) and the impact of self versus informant report are well documented (cf. Brugha et al., 2011; Underwood et al., 2010) but remain unresolved.

Third, assessment of mental health is complicated by the often atypical presentation of psychiatric problems in autism (Hutton et al., 2008; Kerns et al., 2014). Even when detailed clinical psychiatric assessments are conducted, differential diagnosis can be compromised by the overlap between core autism symptoms and symptoms associated with other disorders, particularly OCD (e.g. McDougle et al., 1995; Russell et al., 2005; Rydén and Bejerot, 2008). For example, in the present cohort, just over one-quarter of adults who were able to report on their own mental health indicated that they had experienced significant OCD-type difficulties. In contrast, information provided by parents indicated that these symptoms had often been present since early childhood and more accurately matched DSM-IV/DSM-5 criteria (American Psychiatric Association (APA), 1994, 2013) for autistic-type repetitive and stereotyped behaviours/interests.

These and many other issues explain why there is such difficulty in establishing accurate estimates of mental health problems. This study has the advantage of being one of the few to examine both child and current correlates of mental health in older adults who were part of an on-going longitudinal study (i.e. not-referred because of psychiatric difficulties). With the exception of the study by Buck et al. (2014), sample size is considerably larger, length of follow-up (>35 years) longer and the age of the cohort much older than in most other adult research (Magiati et al., 2014). Autism diagnosis was also re-confirmed during adulthood. Nevertheless, because it was not possible, within the scope of the study, to complete detailed clinical psychiatric assessments, it was not possible formally to confirm mental health diagnoses. The time frame was also broad, covering problems from age 16 years onwards, and again, the period of focus (last month, last year and lifetime) greatly influences estimates of psychopathology (Kessler et al., 2007). As in most other autism studies, too, the range of conditions assessed was limited to major groups of disorders. There was no breakdown of anxiety disorders, for example, into different subtypes and no focus on other problems such as drug or alcohol misuse, or post-traumatic stress disorder (PTSD) (e.g. as in Andrews et al., 2001; Kessler et al., 2007), although we did include data on severe self-injury and aggression.

A further problem was that fewer than half of the participants with autism were able to complete self-report measures. Those unable to self-report tended to be of significantly lower intellectual ability and have more severe autism symptoms, and although there was no significant difference in informant ratings for individuals who could, or could not, self-report, assessment of psychiatric problems in less cognitively able individuals presents many difficulties. The representativeness of the sample could also be questioned, since all participants were of average non-verbal IQ as children. Because this sample was identified at a time when only more ‘classical’ cases were likely to be diagnosed, it is uncertain how far the findings would be replicated in more heterogeneous cohorts reaching adulthood today. Finally, as in many other adult studies, assessments of social functioning relied on global, categorical measures related to work, independent living and social relations. There was no broader exploration of the extent of social and community participation (cf. Shattuck et al., 2011), or systematic analysis of social or environmental factors that may have been associated with the onset or maintenance of mental health problems.

Clinical implications

The very high rates of psychiatric comorbidity reported in some studies (Mazzone et al., 2012) can be of great concern to parents and to individuals with autism themselves. However, in the present cohort, whom we have followed up over several decades, both self and informant ratings indicated good mental health in 40%–45%. Nevertheless, as in other studies, if problems did occur these were frequently associated with anxiety, depression and OCD. The small but significant correlations between mental health and social functioning also indicate the need for adult autism services to provide diagnostic assessment and interventions for psychiatric problems as well as focussing on core autism features. Despite this, only a minority of our participants had ever received specialist psychiatric help.

Although the moderate but significant association between self and informant reports indicates that family members, or care staff in close contact with individuals, can provide useful information on mental health for adults with autism unable to report on their own difficulties, commonly used mental health screens may underestimate rates of problems. Findings from this and other studies highlight the lack of appropriately validated measures to assess mental health problems in adults with autism (Brugha et al., in press; Gotham et al., 2015) and, without these, affected individuals are less likely to receive appropriate diagnostic and intervention services. However, the scarcity of professionals who have expertise in diagnosing and treating mental health problems in individuals with autism remains a major challenge (Underwood et al., 2010).

Footnotes

Acknowledgements

We are most grateful to all the families who gave so generously of their time during the course of the study and to the National Autistic Society who assisted with tracing.

Funding

This study was generously funded by the UK Nuffield Foundation.

References

Adler

Spencer

Faraone

. (2006) Validity of pilot Adult ADHD Self-Report Scale (ASRS) to rate adult ADHD symptoms. Annals of Clinical Psychiatry 18: 145–148.

American Psychiatric Association (APA) (1994) Diagnostic and Statistic Manual of Mental Disorders (DSM-IV). 4th ed. Washington, DC: APA.

American Psychiatric Association (APA) (2013) Diagnostic and Statistic Manual of Mental Disorders (DSM-5). 5th ed. Washington, DC: APA.

Andrews

Henderson

Hall

(2001) Overview of the Australian National Mental Health Survey: prevalence, comorbidity, disability and service utilization. British Journal of Psychiatry 178: 145–153.

Baumgartner

Burns

(2014) Measuring social inclusion – a key outcome in global mental health. International Journal of Epidemiology 43(2): 354–641.

Beck

(1993) Beck Anxiety Inventory (BAI). San Antonio, TX: The Psychological Corporation.

Beck

Steer

Brown

(1996) Beck Depression Inventory-II (BDI-II). San Antonio, TX: The Psychological Corporation.

Billstedt

Gillberg

(2005) Autism after adolescence: population-based 13- to 22-year follow-up study of 120 individuals with autism diagnosed in childhood. Journal of Autism and Developmental Disorders 35: 351–360.

Bolton

Macdonald

Pickles

. (1994) A case-control family history study of autism. Journal of Child Psychology and Psychiatry 35: 877–900.

10.

Bolton

Pickles

Murphy

. (1998) Autism, affective and other psychiatric disorders: patterns of familial aggregation. Psychological Medicine 28: 385–395.

11.

Brugha

McManus

Bankart

. (2011) Epidemiology of autism spectrum disorders in adults in the community in England. Archives of General Psychiatry 68(5): 459–466.

12.

Brugha

Tempier

Doos

. (in press) Outcome measures in intervention trials for adults with autism spectrum disorders; a systematic review of assessments of core autism features and associated emotional and behavioural problems. International Journal of Methods in Psychiatric Research.

13.

Buck

Viskochil

Farley

. (2014) Psychiatric comorbidity and medication use in adults with autism spectrum disorder. Journal of Autism and Developmental Disorders 44: 3063–3071.

14.

Cederlund

Hagberg

Billstedt

. (2008) Asperger syndrome and autism: a comparative longitudinal follow-up study more than 5 years after original diagnosis. Journal of Autism and Developmental Disorders 38: 72–85.

15.

Clegg

Hollis

Mawhood

. (2005) Developmental language disorders – a follow-up in later adult life. Cognitive, language and psychosocial outcomes. Journal of Child Psychology and Psychiatry 46: 128–149.

16.

Davis

TEI

Moree

Dempsey

. (2011) The relationship between autism spectrum disorders and anxiety: the moderating effect of communication. Research in Autism Spectrum Disorders 5: 324–329.

17.

Dunn

Whetton

. (1997) British Picture Vocabulary Scale. 2nd ed. Windsor: NFER-Nelson.

18.

Esbensen

Greenberg

Seltzer

. (2009) A longitudinal investigation of psychotropic and non-psychotropic medication use among adolescents and adults with autism spectrum disorders. Journal of Autism and Developmental Disorders 39(9): 1339–1349.

19.

Farley

McMahon

Fombonne

. (2009) Twenty-year outcome for individuals with autism and average or near-average cognitive abilities. Autism Research 2: 109–118.

20.

Fombonne

Bolton

Prior

. (1997) A family study of autism: cognitive patterns and levels in parents and siblings. Journal of Child Psychology and Psychiatry 38: 667–683.

21.

Ghaziuddin

Zafar

(2008) Psychiatric comorbidity of adults with autism spectrum disorders. Clinical Neuropsychiatry 5: 9–12.

22.

Goldberg

Williams

(1988) A User’s Guide to the General Health Questionnaire. Windsor: NFER-Nelson.

23.

Goodman

Price

Rasmussen

. (1989) The Yale–Brown Obsessive Compulsive Scale. I. Development, use, and reliability. Archives of General Psychiatry 46: 1006–1011.

24.

Gotham

Unruh

Lord

(2015) Depression and its measurement in verbal adolescents and adults with autism spectrum disorder. Autism 19(4): 491–504.

25.

Hill

Berthoz

Frith

(2004) Brief report: cognitive processing of own emotions in individuals with autistic spectrum disorder and in their relatives. Journal of Autism and Developmental Disorders 34: 229–235.

26.

Hofvander

Delorme

Chaste

. (2009) Psychiatric and psychosocial problems in adults with normal-intelligence autism spectrum disorders. BMC Psychiatry 9: 35.

27.

Howlin

Goode

Hutton

. (2004) Adult outcome for children with autism. Journal of Child Psychology and Psychiatry 45: 212–229.

28.

Howlin

Moss

Savage

. (2013) Social outcomes in mid to later adulthood among individuals diagnosed with autism and average non-verbal IQ as children. Journal of the American Academy of Child and Adolescent Psychiatry 52(6): 572–581.

29.

Howlin

Moss

Savage

. (2014) Cognitive and language skills in adults with autism: a 40-year follow-up. Journal of Child Psychology and Psychiatry 55(1): 49–58.

30.

Hutton

Goode

Murphy

. (2008) New-onset psychiatric disorders in individuals with autism. Autism 12: 373–390.

31.

Joshi

Wozniak

Petty

. (2013) Psychiatric comorbidity and functioning in a clinically referred population of adults with autism spectrum disorders: a comparative study. Journal of Autism and Developmental Disorders 43: 1314–1325.

32.

Kerns

Kendal

Berry

. (2014) Traditional and atypical presentations of anxiety in youth with autism spectrum disorder. Journal of Autism and Developmental Disorders 44: 2851–2861.

33.

Kessler

Adler

Ames

. (2005) The World Health Organization Adult ADHD Self-Report Scale (ASRS): a short screening scale for use in the general population. Psychological Medicine 35: 245–256.

34.

Kessler

Angermeyer

Anthony

.; The WHO World Mental Health Survey Consortium (2007) Lifetime prevalence and age-of-onset distributions of mental disorders in the World Health Organization’s World Mental Health Survey Initiative. World Psychiatry 6: 168–176.

35.

Lai

M-C

Lombardo

Pasco

. (2011) A behavioral comparison of male and female adults with high functioning autism spectrum conditions. PLoS ONE 6(6): e20835. DOI: 10.1371/journal.pone.0020835.

36.

Lainhart

(1999) Psychiatric problems in individuals with autism, their parents and siblings. International Review of Psychiatry 11: 278–298.

37.

Le Couteur

Rutter

Lord

. (1989) Autism diagnostic interview: a standardized investigator-based instrument. Journal of Autism and Developmental Disorders 19: 363–387.

38.

Levy

Perry

(2011) Outcomes in adolescents and adults with autism: a review of the literature. Research in Autism Spectrum Disorders 5: 1271–1282.

39.

Lugnegård

Hallerbäck

Gillberg

(2011) Psychiatric comorbidity in young adults with a clinical diagnosis of Asperger syndrome. Research in Developmental Disabilities 32(5): 1910–1917.

40.

McDougle

Kresch

Goodman

. (1995) A case-controlled study of repetitive thoughts and behavior in adults with autistic disorder and obsessive–compulsive disorder. American Journal of Psychiatry 152: 772–777.

41.

Magiati

Tay

Howlin

(2014) Cognitive, language, social and behavioural outcomes in adults with autism spectrum disorders: a systematic review of longitudinal follow-up studies in adulthood. Clinical Psychology Review 34(1): 73–86.

42.

Mannion

Brahm

Leader

(2014) Comorbid psychopathology in autism spectrum disorder. Review Journal of Autism and Developmental Disorders 1: 124–134.

43.

Mazzone

Ruta

Reale

(2012) Psychiatric comorbidities in Asperger syndrome and high functioning autism: diagnostic challenges. Annals of General Psychiatry 11: 16.

44.

Pickles

Parr

Rutter

. (2013) New interview and observation measures of the Broader Autism Phenotype: impressions of interviewee measure. Journal of Autism and Developmental Disorders 43(9): 2082–2089.

45.

Pinto

Pagnamenta

Klei

. (2010) Functional impact of global rare copy number variation in autism spectrum disorders. Nature 466: 368–372.

46.

Prosser

Moss

Costello

. (1998) Reliability and validity of the Mini PAS-ADD for assessing psychiatric disorders in adults with intellectual disability. Journal of Intellectual Disability Research 42(4): 264–272.

47.

Risi

Lord

Gotham

. (2006) Combining information from multiple sources in the diagnosis of autism spectrum disorders. Journal of the American Academy of Child and Adolescent Psychiatry 45: 1094–1103.

48.

Russell

Mataix-Cols

Anson

. (2005) Obsessions and compulsions in Asperger syndrome and high-functioning autism. British Journal of Psychiatry 186: 525–528.

49.

Rutter

Le Couteur

Lord

(2003) The Autism Diagnostic Interview-Revised (ADI-R). Los Angeles, CA: Western Psychological Services.

50.

Rydén

Bejerot

(2008) Autism spectrum disorders in an adult psychiatric population. A naturalistic cross-sectional controlled study. Clinical Neuropsychiatry 5: 13–21.

51.

Scahill

McDougle

Williams

. (2006) Children’s Yale-Brown Obsessive Compulsive Scale modified for pervasive developmental disorders. Journal of American Academy of Child and Adolescent Psychiatry 45: 1114–1123.

52.

Shattuck

Orsmond

Wagner

. (2011) Participation in social activities among adolescents with an autism spectrum disorder. PLoS ONE 6(11): e27176.

53.

Simonoff

Jones

Pickles

. (2012) Severe mood problems in adolescents with autism spectrum disorder. Journal of Child Psychology and Psychiatry 53(11): 1157–1166.

54.

Sparrow

Cicchetti

Balla

(2005) Vineland Adaptive Behavior Scales (Vineland-II). Circle Pines, MN: AGS Publishing.

55.

Sterling

Dawson

Estes

. (2008) Characteristics associated with presence of depressive symptoms in adults with autism spectrum disorder. Journal of Autism and Developmental Disorders 38(6): 1011–1018.

56.

Totsika

Felce

Kerr

. (2010) Behavior problems, psychiatric symptoms, and quality of life for older adults with intellectual disability with and without autism. Journal of Autism and Developmental Disorders 40(10): 1171–1178.

57.

Underwood

McCarthy

Tsakanikos

(2010) Mental health of adults with autism spectrum disorders and intellectual disability. Current Opinion in Psychiatry 23: 421–426.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.09 MB