The longitudinal relation between childhood autistic traits and psychosexual problems in early adolescence: The Tracking Adolescents’ Individual Lives Survey study

Abstract

Individuals with autistic traits are considered to be prone to develop psychosexual problems due to their limited social skills and insight. This study investigated the longitudinal relation between autistic traits in childhood (T1; age 10-12 years) and parent-reported psychosexual problems in early adolescence (T2; age 12-15 years). In a general population cohort study (n = 1873; the Tracking Adolescents’ Individual Lives Survey (TRAILS)), autistic traits and psychosexual problems were determined. Logistic regression analyses were used to investigate whether childhood autistic traits, in individuals displaying no psychosexual problems in childhood, predicted the presence of psychosexual problems in adolescence, while controlling for pubertal development and conduct problems. Higher levels of autistic traits at T1 significantly predicted mild psychosexual problems at T2, above and beyond pubertal development and conduct problems. Particularly two dimensions of autistic traits at T1 were significant predictors; i.e. ‘reduced contact/social interest’ and ‘not optimally tuned to the social situation’. Children with autistic traits – especially those with limited social interest and social regulation problems – showed to have a higher risk to develop psychosexual problems, albeit mild, in early adolescence as reported by parents. Although we showed that autistic traits predict psychosexual problems, it is only one of multiple predictors.

Keywords

adolescence autistic traits childhood longitudinal survey psychosexual

Introduction

Early adolescence (ages 12–15 years) is marked by physical and psychosocial changes and challenges (Dahl, 2004), including the accelerated development of sexuality. Healthy sexual development requires a wide variety of skills and knowledge, including the awareness of social and cultural rules concerning sexuality. Social skills seem especially important in a healthy sexual development (Maniglio, 2012; T Hart-Kerkhoffs et al., 2009).

A group of adolescents who have difficulty with social skills are adolescents with autistic traits (Constantino and Todd, 2005; Stokes et al., 2007). These individuals have difficulty with social skills such as reading the signals others show and being finely tuned to other people. With autistic traits, we mean individuals who have some features of autism, for example, difficulty with communication, even though they do not meet the full diagnostic criteria for an autism spectrum disorder (ASD). Autistic features have shown to be a continuously distributed trait across the population (e.g. Constantino, 2009), where many individuals exhibit autistic traits to some extent.

During adolescence, sexual and intimate situations are progressively likely to occur, which makes complex social skills, including significant interest in others, peer-relations and very finely adapted behaviour to often implicit rules increasingly important. In light of their impaired social functioning (Jobe and Williams White, 2007), the high demand on social skills makes adolescence and a healthy sexual development, especially challenging for individuals with autistic traits (Bailey et al., 1998; Bolton et al., 1994). The discrepancy between the needed social skills for a healthy sexual development and shortcomings in social skills may lead to psychosexual problems. In this article, we define psychosexual problems as cognitions and behaviours that fall outside the typical standards of society (Chaffin et al., 2006), entailing, for instance, excessive masturbation, public masturbation and excessive interest in sexuality (Bonner et al., 1999).

A potential mechanism behind the development of psychosexual problems may be that the difficulties initiating and maintaining social relationships with others leads to little exposure to social situations and/or social clumsiness, for example, because the individual with autistic traits does not initiate but also does not react to social interaction. Difficulties with adapting one’s behaviour to social situations in turn may cause socially inappropriate behaviours, for example, because behaviour is not adapted to the situation and/or because the individual quickly overreacts. These inappropriate behaviours can cause awkward or difficult social interactions. The limited social relationships may also lead to less learning (opportunities) on appropriate psychosexual behaviours; individuals with ASDs obtain less of their sexual knowledge from social sources than typically developing individuals (Brown-Lavoie et al., 2014; Stokes et al., 2007). Parents of autistic adolescents report about their child’s limited knowledge and skills (e.g. distinguishing between public and private behaviour and tuning behaviour to privacy rules; Nichols and Blakeley-Smith, 2009). Even in young adults, non-clinical autistic traits have been related to negative social outcomes due to social and communication impairments (Jobe and Williams White, 2007).

To our knowledge, no research has been performed on the longitudinal relation between autistic traits and psychosexual problems. Although research has shown that problems in adequate social skills in the long run result in problem behaviour and maladjustment in general (Jenson, 2010), little is known on whether the limited social skills of adolescents with autistic traits result in psychosexual problems.

Apart from the absence of longitudinal studies, only few studies have investigated the concurrent relation between autistic traits and psychosexual problems. Some studies have found a relationship between severe psychosexual problems (e.g. sexual delinquency) and diminished social competence in interpersonal relationships that appear to be similar to the behaviours which individuals with autistic traits display (Kumagami and Matsuura, 2009; Maniglio, 2012; T Hart-Kerkhoffs et al., 2009). While illustrating that individuals with severe psychosexual problems may display autistic traits, this cannot be taken to mean that individuals with autistic traits often have psychosexual problems. Studies in clinical samples with ASD diagnoses found evidence for psychosexual problems. For example, Stokes and Kaur (2005) found more psychosexual problems in cognitively able autistic adolescents with poorer social skills than in those with better social skills. Also, institutionalized cognitively able adolescents with ASD showed psychosexual problems such as touching of the genitals in public and masturbating while others are present (Hellemans et al., 2007). Although individuals with autistic traits may not have identical difficulties, they may have psychosexual problems, similar to individuals with a clinical diagnosis of ASD. This needs further investigation.

Taken together, the studies available have indicated a cross-sectional association between autistic traits and psychosexual problems. However, previous studies did not explore the longitudinal relation between autistic traits in childhood and psychosexual problems in early adolescence. Such research is necessary to clarify whether autistic traits are predictive of the development of psychosexual problems. The current study therefore investigated the potential relation between childhood autistic traits and psychosexual problems in adolescence in a large general population sample by means of a commonly used parent-reported questionnaire (see section ‘Measures’). We hypothesized that higher levels of autistic traits in childhood were related to parent-reported psychosexual problems in adolescence. We additionally examined whether (a) particular dimensions of autistic traits were related to psychosexual problems, (b) the change in the level of autistic traits from childhood to adolescence was associated with psychosexual problems in adolescence and (c) putative covariates (i.e. pubertal development, intelligence and conduct problems) influenced the relation between autistic traits in childhood and psychosexual problems in adolescence.

Methods

Sample

This study used data from the Tracking Adolescents’ Individual Lives Survey (TRAILS), an ongoing longitudinal cohort study of Dutch youth focused on the development of mental ill-health and health from childhood to adulthood which started in 2000 (for the design of this study, see De Winter et al., 2005; Huisman et al., 2008). The study was approved by the Dutch national ethical committee; the ‘Centrale Commissie Mensgebonden Onderzoek’. To consider the longitudinal relation between autistic traits and psychosexual problems, we used data from the first assessment wave (T1, ages 10–12 years, M = 11.11 years, standard deviation SD = 0.55 years) and the second assessment wave (T2, ages 12–15 years, M = 13.55 years, SD = 0.53 years). Average time between the two measurements was 2.46 years (SD = 0.45 years; range: 1.39–4.01 years). As a sample, 1687 adolescents from the originally 2230 adolescents who participated in the TRAILS study met our inclusion and exclusion criteria.

Although 2230 adolescents took part in TRAILS at T1, 379 (16.9%) adolescents were excluded from the current analyses due to missing data on either the Children’s Social Behaviour Questionnaire (CSBQ; used for measurement of autistic traits, see section ‘Measures’) at T1 (n = 62), or on the Child Behavior Checklist (CBCL; used for measurement of psychosexual problems, see section ‘Measures’) at T1 (n = 131) or at T2 (n = 186). There were no significant differences between the two groups (missing data versus no missing data) on gender (χ²(1, N = 2230) = 1.11, p = 0.29). However, there were small but significant differences in pubertal development – those with missing data were more advanced (M = 1.97, SD = 0.86; t(2114) = 2.67, p < 0.01) than those without missing data (M = 1.85, SD = 0.73) – and age – those with missing data were significantly older (M = 11.16 years, SD = 0.56 years; t(2230) = 1.99, p < 0.05) than those without missing data (M = 11.10 years, SD = 0.55 years). Also, there was a moderate difference regarding total Intelligence Quotient (IQ) which was significantly lower in those with missing data (M = 90.42, SD = 15.63) than in those without missing data (M = 97.82, SD = 14.78; t(2221) = −6.57, p < 0.001).

It is important to note that as the current study was intended to investigate the predictive value of autistic traits at T1 on psychosexual problems at T2, all children who already showed psychosexual problems at T1 (n = 164 (7.35%)) were excluded from the analyses (CBCL; used for measurement of psychosexual problems, see section ‘Measures’). Total IQ was significantly lower in the group with psychosexual problems at T1 (M = 94.61, SD = 14.48) than in the group without psychosexual problems at T1 (M = 98.03, SD = 14.81; t(2048) = −2.85, p < 0.01). Gender also significantly differed between these two groups (group with psychosexual problems at T1 = 69% male; group without psychosexual problems at T1 = 48% male; (χ²(1, N = 2054) = 27.07, p < 0.001). There were no significant differences regarding pubertal development (t(1993) = −0.82, p = 0.41) or age (t(2054) = −0.17, p = 0.86).

Exclusion resulted in a final sample of n = 1687 (75.7% of 2230), of which 53% (n = 889) was female. The mean age was 11.10 years (SD = 0.55 years) at T1 and 13.54 years (SD = 0.52 years) at T2.

Measures

Psychosexual problems at T2

The ‘Sex Problems Scale’ of the parent-report CBCL (Achenbach, 1991; Letourneau et al., 2004) was used to determine the occurrence of psychosexual problems. The Sex Problems Scale, as described in the CBCL manual (Achenbach, 1991) as well as in the Dutch CBCL manual (Verhulst et al., 1996), has been identified by means of principal component analysis. The manual describes that internal consistency (by means of Cronbach’s alpha) in the Dutch norm-sample is moderate (α = 0.39), which is similar to the internal consistency we found in our sample (α = 0.35). This moderate internal consistency may in part be explained by the low prevalence rate of the items and the somewhat diverse problems, and in part by the relatively few items (Achenbach, 1991). Research that included the CBCL Sex Problems Scale for the purpose of measuring psychosexual problems by means of parent-report has nonetheless supported the content and concurrent validity of this scale (Friedrich, 2003; Friedrich et al., 1992; Letourneau et al., 2004, 2008) as well as the discriminant validity (Bonner et al., 1999; Pithers et al., 1998). Friedrich et al. (1992), for example, investigated the association between the CBCL Sex Problems Scale and the more extensive Child Sexual Behavior Inventory (CSBI); they found that both measures were better than chance level to distinguish between sexually abused children and a normative sample, and although the sensitivity of the CSBI was slightly better, no significant difference was found regarding specificity. Two studies found that children with sexual problem behaviour (i.e. children exhibiting sexual behaviours which were repetitive, unresponsive to adult intervention and supervision, equivalent to adult criminal violations, pervasive, occurring across time and situations or developmentally unexpected sexual acts) scored significantly higher on the CBCL Sex Problems Scale than children without such problems (Bonner et al., 1999; Pithers et al., 1998). Two more recent studies by Letourneau and colleagues (2004, 2008) divided individuals into groups with and without psychosexual problems based on the CBCL Sex Problems Scale. The intervention study showed that the individuals with psychosexual problems improved significantly more with regard to internalizing and externalizing problems as compared to individuals without psychosexual problems (Letourneau et al., 2008).

Parents rate each item on the CBCL on a 3-point scale (0 = not true, 1 = somewhat or sometimes true, 2 = very true or often true). It was decided to use parent reports, as the reliability of self-report in individuals with autistic traits has been questioned (e.g. Cederlund et al., 2010), we expected the adolescents would also underestimate psychosexual problems. The CBCL Sex Problems Scale consists of items regarding exposing behaviour, sexual problems, thinking of sex too much and playing with own genitals in public. In the current study, psychosexual problems were dichotomized, where if parents scored at least a ‘1’ on the items of the Sex Problems Scale, their child was considered to have psychosexual problems (scoring 1 = psychosexual problems, or 0 = no psychosexual problems), similar to the outcome measure used by Letourneau et al. (2008). This dichotomous variable was based on the scores on the following 4 items: ‘plays with own sex parts in public’, ‘plays with own sex parts too much’, ‘sexual problems’ and ‘thinks about sex too much’.

Autistic traits at T1 and T2

The Dutch parent-report CSBQ (Hartman et al., 2007) was used to assess the level of autistic traits. The measure was developed to quantify both the various problem dimensions on which children with autistic traits differ as well as to include the milder as well as the more severe autistic behaviours in one instrument. Parents rate their children on 49 items, using a 3-point scale (0 = not true, 1 = somewhat or sometimes true, 2 = very true or often true). The CSBQ has six empirically derived dimensional subscales: ‘orientation problems in time, place, or activity’ (e.g. has trouble doing two things at once and sees no danger), ‘not optimally tuned to the social situation’ (e.g. keeps pushing a topic to get his or her way and makes a problem out of little things), ‘difficulties in understanding social information’ (e.g. tells stories inconsistently and takes things literally), ‘reduced contact and social interest’ (e.g. makes little eye contact and has little or no need for contact with others), ‘fear of and resistance to changes’ (e.g. panics in new situations and resists changes) and ‘stereotyped behaviours’ (e.g. rocks body and smells objects). These scales are combined into a total score. For all of the analyses, the mean score of the total or subscale scores were used. The CSBQ has shown good psychometric properties (e.g. Hartman et al., 2006). The internal consistency (Cronbach’s α) in our sample of the total scale was good (α = 0.91 for 49 items). This also held for the subscales: Subscale 1 – ‘reduced contact and social interest’ (α = 0.72); Subscale 2 – ‘not optimally tuned to the social situation’ (α = 0.85); Subscale 3 – ‘orientation problems in time, place, or activity’ (α = 0.77); Subscale 4 – ‘difficulties in understanding social information’ (α = 0.73); Subscale 5 – ‘stereotyped behavior’ (α = 0.66); and Subscale 6 – ‘fear of and resistance to changes’ (α = 0.73). In our sample, the range of the total mean score at T1 was 0–1.29 (M = 0.24, SD = 0.21). On the total CSBQ T1 score, 22% of our sample fell within the ‘high–very high’ range based on population norms. At T2, the range of the total mean score was 0–1.27 (M = 0.20, SD = 0.20).

Pubertal development at T1

Physical maturation is a good predictor of sexual activity (Flannery et al., 1993; Halpern et al., 1993). Autistic traits have also been related to pubertal development (Whitehouse et al., 2011). Therefore, pubertal development was regarded as a putative covariate. The participants’ pubertal development was measured using parent-reported Tanner stages (Tanner, 1962). This is a staging system which uses the five standard Tanner stages of pubertal development (Marshall and Tanner, 1969, 1970) in the form of schematic drawings of secondary sexual characteristics (i.e. pubic hair growth, breast growth in females and penis growth in males). The parent (usually the mother) was provided with gender-appropriate sketches and asked which of the sketches ‘looked most like their child’. These ratings have been widely used and have demonstrated good reliability and validity (Dorn et al., 1990). Mean pubertal development at T1 was 1.86 (range: 1–5).

IQ at T1

IQ was also regarded a putative covariate, as it may be related to both psychosexual problems (Seto and Lalumiere, 2010) and autistic traits (Hoekstra et al., 2010). IQ was estimated based on the Vocabulary and Block Design subtests (Sattler, 1992) of the Revised Wechsler Intelligence Scales for Children (WISC-R; Van Haassen et al., 1986; Wechsler, 1974). IQ estimates ranged between 45 and 149, with a mean of 98.12 (SD = 14.80).

Conduct problems at T1

As conduct problems have been related to autistic traits (e.g. Geluk et al., 2012; Totsika et al., 2011) as well as psychosexual problems (e.g. Fergusson et al., 2005; T Hart-Kerkhoffs et al., 2009; Wu et al., 2010), it was also considered as a putative covariate. Conduct problems were assessed by means of the CBCL (Achenbach, 1991) Diagnostic and Statistical Manual of Mental Disorders (DSM)-oriented Conduct Problems Scale. This scale consists of 12 items (rated on a 3-point scale: 0 = not true, 1 = somewhat or sometimes true, 2 = very true or often true) related to conduct disorder. The internal consistency of the conduct problems scale has been established as good in previous literature (e.g. Nakamura et al., 2009; Verhulst and Van der Ende, 2013) and was also good in the current sample (α = 0.78).

Statistical analyses

First, we compared the group with psychosexual problems at T2 to the group without psychosexual problems at T2 with regards to their mean scores on the main predictors and putative covariates (i.e. pubertal development, IQ and conduct problems) by means of independent t-tests. To check for a significant difference in the distribution of gender in the two groups, we ran a chi-square test. Any significant differences on the putative covariates (i.e. pubertal development, IQ and CBCL conduct problems) between the groups with and without psychosexual problems at T2 resulted in the inclusion of these variables as covariates in the subsequent logistic regression analyses. Furthermore, we explored the item-endorsement frequencies of psychosexual problems at T2. In addition, we assessed what percentage of cases scored high to very high on the CSBQ in each group (with or without sexuality problems). In this subsample, we also explored the item-endorsement frequencies of psychosexual problems at T2.

Second, the main analyses consisted of two logistic regressions, which were to investigate the longitudinal relation between autistic traits and psychosexual problems, while controlling for covariates. Model 1 used the mean total score on the CSBQ at T1 as main predictor and psychosexual problems at T2 as the outcome. Model 2 used the separate mean scores on the six subscales of the CSBQ at T1 to predict psychosexual problems at T2. The second logistic regression analysis was performed in order to investigate whether specific autistic traits and/or associated features predicted psychosexual problems.

Third, as autistic traits may also change over time (e.g. Gotham et al., 2012; Holmboe et al., 2014; Woolfenden et al., 2012), we investigated whether the change in the level of autistic traits from T1 to T2 was related to having psychosexual problems at T2. For this purpose, we computed a difference score (T2 CSBQ − T1 CSBQ) and included this variable as the predictor in an additional logistic regression model together with the relevant covariates.

Results

Preliminary data inspection

For descriptive purposes, we ran a comparison (see Table 1) between the group with psychosexual problems at T2 and the group without psychosexual problems at T2. In both groups, we only included individuals who were free from psychosexual problems at T1. The results show that the group with psychosexual problems at T2 scores significantly higher on the total mean score as well as on the subscales mean scores of the CSBQ. With respect to the putative covariates, the groups only differed significantly in their pubertal development (p = 0.05) and conduct problems (p < 0.01).

Table 1.

Comparison of group without (N=1620) and with (N=67) psychosexual problems at T2.^a

Variables	Without psychosexual problems (N=1620)		With psychosexual problems (N=67)		Significance
	n	Mean (SD)	n	Mean (SD)	p
T1 mean Total score CSBQ	1620	0.23 (0.20)	67	0.39 (0.27)	<0.001*
T1 ‘Not optimally tuned’ scale (CSBQ)	1620	0.41 (0.36)	67	0.66 (0.47)	<0.001*
T1 ‘Reduced soc. Interest’ scale (CSBQ)	1620	0.12 (0.17)	67	0.25 (0.27)	<0.001*
T1 ‘Orientation problems’ scale (CSBQ)	1620	0.19 (0.28)	67	0.30 (0.38)	=0.02*
T1 ‘Difficulties in understanding’ scale (CSBQ)	1620	0.30 (0.32)	67	0.48 (0.38)	<0.001*
T1 ‘Stereotyped behavior’ scale (CSBQ)	1620	0.11 (0.20)	67	0.20 (0.28)	<0.01*
T1 ‘Fear of change’ scale (CSBQ)	1620	0.17 (0.34)	67	0.32 (0.50)	=0.01*
T2 mean Total score CSBQ	1611	0.19 (0.19)	67	0.40 (0.32)	<0.001*
T2 ‘Not optimally tuned’ scale (CSBQ)	1612	0.32 (0.33)	67	0.64 (0.47)	<0.001*
T2 ‘Reduced soc. Interest’ scale (CSBQ)	1613	0.14 (0.21)	67	0.30 (0.34)	<0.001*
T2 ‘Orientation problems’ scale (CSBQ)	1613	0.14 (0.22)	67	0.31 (0.39)	<0.001*
T2 ‘Difficulties in understanding’ scale (CSBQ)	1613	0.30 (0.32)	67	0.52 (0.44)	<0.001*
T2 ‘Stereotyped behavior’ scale (CSBQ)	1612	0.07 (0.15)	67	0.17 (0.28)	<0.01*
T2 ‘Fear of change’ scale (CSBQ)	1608	0.13 (0.30)	67	0.37 (0.46)	<0.001*
T1 Pubertal development	1587	1.84 (0.74)	65	2.03 (0.81)	=0.05*
T1 IQ	1619	98.85 (14.47)	66	99.70 (15.93)	=0.64
T1 Conduct problems	1620	0.12 (0.14)	67	0.19 (0.18)	<0.01*
T1 Gender	1620	46.9% male	67	56.7% male	=0.12
T2 Sex Problems scale item – Plays with own sex parts in public	1617	0% (n=0)	66	3% (n=2)
T2 Sex Problems scale item – Plays with own sex parts too much	1615	0% (n=0)	66	21.2% (n=14)
T2 Sex Problems scale item – Sexual problems	1616	0% (n=0)	66	7.6% (n=5)
T2 Sex Problems scale item – Thinks about sex too much	1616	0% (n=0)	67	79.1% (n=53)

T2 = second assessment wave. Comparison of the main predictor CSBQ total score and subscales and putative covariates (physical development, intelligence, and gender), as well as item-endorsement on the Sex Problems scale items of the Child Behavior Checklist

SD: Standard deviation; CSBQ: Children’s Social Behaviour Questionnaire. * is significant difference.

Table 1 also shows that mainly the items ‘plays with own sex parts too much’ and ‘thinks about sex too much’ were endorsed by parents. Psychosexual problems showed a skewed distribution with most children displaying no psychosexual problems, supporting the decision to dichotomize psychosexual problems at T2 (scoring 1 = having psychosexual problems, or 0 = not having psychosexual problems).

Furthermore, in the group without psychosexual problems, 20% of the adolescents scored in the high–very high range of the CSBQ. This compares to 42% of adolescents in the group with psychosexual problems. Conversely, in the sample with high to very high scores on the CSBQ (n = 348), approximately 8% (n = 27) showed psychosexual problems at T2 (proportion per item: plays with own sex parts in public – n = 1 (4%); plays with own sex parts too much – n = 4 (15%); sexual problems – n = 4 (15%); thinks about sex too much – n = 23 (85%)), illustrating that psychosexual problems are more prevalent in individuals with relatively high levels of autistic traits.

Associations between T1 autistic traits and T2 psychosexual problems

As shown in Table 2, when adding the covariates (i.e. pubertal development and CBCL conduct problems) as predictors in the logistic regression, there was a significant effect of CSBQ T1 total score. In analyses with the CSBQ subscales, the subscales ‘reduced social interest’ and ‘not optimally tuned’ had a significant effect on psychosexual problems after controlling for the covariates.

Table 2.

Results of the logistic regression models 1 and 2 predicting psychosexual problems including covariates.^a

Model 1 ^b
	OR ^d	95% CI	p
T1 CSBQ Total score	10.62	[3.37, 33.50]	<0.00
T1 Tanner stages	1.41	[1.02, 1.94]	0.04
T1 Conduct problems (CBCL scale)	2.08	[0.40, 10.96]	0.39
Model 2^c
	OR ^d	95% CI	p
T1 ‘Reduced social interest’ scale (CSBQ)	3.84	[1.13, 12.99]	0.03
T1 ‘Not optimally tuned’ scale (CSBQ)	2.29	[1.01, 5.20]	0.05*
T1 ‘Orientation problems’ scale (CSBQ)	0.44	[0.15, 1.29]	0.13
T1 ‘Difficulties in understanding’ scale (CSBQ)	1.32	[0.55, 3.18]	0.53
T1 ‘Stereotyped behavior’ scale (CSBQ)	1.97	[0.69, 5.62]	0.21
T1 ‘Fear of change’ scale (CSBQ)	1.37	[0.74, 2.52]	0.31
T1 Tanner stages	1.44	[1.05, 1.99]	0.03
T1 Conduct problems (CBCL scale)	2.58	[0.44, 15.22]	0.29

The results presented are excluding children with psychosexual problems at T1. Model 1 includes the mean total score on the CSBQ as well as the Tanner stages and IQ; Model 2 includes the mean subscale scores on the CSBQ as well as the Tanner stages and IQ.

CSBQ: Children’s Social Behaviour Questionnaire; CBCL: Child Behavior Checklist; OR: odds ratio, CI: confidence interval.

Nagelkerke R² = 0.07

Nagelkerke R² = 0.09

The odds ratios represent the OR per unit of mean score.

exact p value is 0.048

Associations between autistic traits over time and T2 psychosexual problems

The change in the level of autistic traits over time (i.e. the difference score of T2 CSBQ − T1 CSBQ) also significantly predicted psychosexual problems at T2 (OR = 6.41, confidence interval (CI) = (1.63, 25.16), p < 0.01) while controlling for the effects of the covariates (i.e. pubertal development and conduct problems) (see Table 3). Noticeably, the adolescents with psychosexual problems at T2 had a higher mean score at both T1 and T2 on the CSBQ (see Table 1). The model with the difference scores on the subscales of the CSBQ showed that only the scale ‘orientation problems’ has a significant effect on psychosexual problems, above and beyond pubertal development and conduct problems (see Table 3).

Table 3.

Results of the logistic regression models 1 and 2 with difference score predicting psychosexual problems including covariates.^a

Model 1^b
	OR ^d	95% CI	p
T2–T1 CSBQ Total score	6.40	[1.63, 25.13]	<0.01
T1 Tanner stages	1.34	[0.97, 1.84]	0.07
T1 Conduct problems scale(CBCL)	19.26	[5.11, 72.60]	<0.001
Model 2^c
	OR ^d	95% CI	p
T2–T1 ‘Reduced social interest’ scale (CSBQ)	0.79	[0.20, 3.04]	0.73
T2–T1 ‘Not optimally tuned’ scale (CSBQ)	2.05	[0.75, 5.60]	0.16
T2–T1 ‘Orientation problems’ scale (CSBQ)	4.37	[1.28, 14.87]	0.02
T2–T1 ‘Difficulties in understanding’ scale (CSBQ)	0.98	[0.39, 2.48]	0.96
T2–T1 ‘Stereotyped behavior’ scale (CSBQ)	0.57	[0.16, 2.06]	0.40
T2–T1 ‘Fear of change’ scale (CSBQ)	1.12	[0.53, 2.37]	0.77
T1 Tanner stages	1.32	[0.96, 1.81]	0.09
T1 Conduct problems scale (CBCL)	24.28	[6.14, 95.96]	<0.001

CSBQ: Children’s Social Behaviour Questionnaire; CBCL: Child Behavior Checklist; OR: odds ratio, CI: confidence interval.

Nagelkerke R² = 0.05

Nagelkerke R² = 0.07

The odds ratios represent the OR per unit of mean score.

Discussion

The purpose of this study was to investigate whether autistic traits in childhood predicted the occurrence of psychosexual problems in early adolescence. Our study revealed a positive longitudinal association between childhood autistic traits and parent-reported psychosexual problems in early adolescence. The results show that autistic traits in childhood, above and beyond pubertal development and conduct problems, predict psychosexual problems in early adolescence. In the group with psychosexual problems in adolescence, many more adolescents scored high to very high levels of autistic traits in childhood compared to those without problems sexuality. In particular, limited social interest and problems in adapting one’s behaviour were related to thinking too much about sex and playing with own sex parts too much. Especially, those adolescents with a lack of motivation to initiate and reciprocate social contact, overreacting and/or a lack of regulation of emotions and behaviours in social situations were more vulnerable to develop psychosexual problems. In addition, psychosexual problems at T2 were associated with the change in the level of autistic traits from T1 to T2. As shown in Table 1, adolescents with psychosexual problems at T2 already showed higher levels of autistic traits at T1 than those without psychosexual problems at T2, and there was a slight increase in the level of autistic traits in the individuals with psychosexual problems at T2, whereas the adolescents without psychosexual problems at T2 were characterized by a decrease in the level of autistic traits from T1 to T2. Our results thus support the notion that autistic traits are associated with psychosexual problems (Ballan, 2012). It should be noted, however, that although it was shown that autistic traits are a significant predictor of psychosexual problems, they are only one of multiple predictors of problems in the sexual development.

Psychosexual problems may be the result of the fact that individuals with autistic traits have more difficulties with figuring out unwritten social rules (Hénault, 2006) or less knowledge on social boundaries (i.e. privacy) (Nichols and Blakeley-Smith, 2009; Stokes and Kaur, 2005). The difficulties that individuals with ASDs have with deriving the appropriate rules and norms from a situation or context complicates the adaptation of their behaviour (if they do not understand or know the rules, they cannot adapt to these rules either). Individuals who do not meet the full diagnostic criteria for ASD seem to have comparable difficulties. Seeing as the knowledge or the ability of the child to keep behaviours and cognitions private may be limited (Stokes and Kaur, 2005), we speculate that the problems most often reported by the parents in our sample, ‘plays with own sex parts too much’ and ‘thinks about sex too much’, may be result of these difficulties with privacy.

Another putative mechanism explaining the relation we found between autistic traits and psychosexual problems may be that the social impairments and limited social interest and the tuning problems make it unlikely for adolescents with autistic traits to be involved in peer groups and thus to learn from social sources (Brown-Lavoie et al., 2014; Stokes et al., 2007). In such peer groups, they would have valuable learning opportunities regarding sexuality and social interaction (Sullivan and Caterino, 2008). Therefore, a lack of such involvement may ultimately limit the knowledge and skills of the adolescent with autistic traits, potentially resulting in psychosexual problems.

In our sample, mostly mild aspects of psychosexual problems were endorsed (e.g. thinking too much about sex and plays with own sex part too much), which are not alarming behaviours or cognitions for adolescents. In addition, the judgement ‘too much’ is a subjective one, which may lead to differences in the reporting of problems due to a difference in attitudes of the parents. In our large sample, only 67 parents (3.6%) reported these psychosexual problems in adolescence. This may indicate that there is a problem of socially inappropriate behaviour or cognitions rather than variation in openness of the parents concerning sexuality. However, the higher scores could equally be due to other reasons (i.e. scoring tendencies of the parents); therefore, further validation of the CBCL Sex Problems Scale is needed. Although the reported psychosexual problems are relatively mild, this group might be at-risk to develop more severe forms of psychosexual problems, especially in combination with other additional risk factors. There is evidence that adolescents with escalated psychosexual problems (i.e. sexual delinquency) have autistic traits (Maniglio, 2012; T Hart-Kerkhoffs et al., 2009). Although the likelihood of committing actual sexual offences is predicted by multiple factors (Maniglio, 2012), autistic traits may be one of these many predictors.

Although clearly not all adolescents with autistic traits will develop psychosexual problems, as a group, these adolescents are at-risk. Because of the risk and potential problems, it is important to timely assess indicators of psychosexual problems, to be able to provide appropriate support as soon as necessary. Support for individuals at-risk of developing psychosexual problems might be provided in the form of socio-sexual trainings (e.g. Visser et al., 2012). Such a socio-sexual training explicitly teaches children with autistic traits to understand how to develop and maintain social relations and what appropriately tuned behaviour entails. Such a training may increase the likelihood of a healthy sexual development and potentially prevent or minimize the risk of developing psychosexual problems and the potential escalation of psychosexual problems (T Hart-Kerkhoffs et al., 2009; Wieckowski et al., 1998).

Conclusions from our study are mainly applicable to a general population with average IQ, as our sample showed a normal range of intelligence (IQ: M = 98.1, SD = 14.8). Due to missing data, adolescents with lower IQs were excluded from our sample. However, the differences between the adolescents with and without missing data were still within 1 SD (IQ in group with missing data, M = 90.42, SD = 15.63; those without missing data, M = 97.82, SD = 14.78; (t(2221) = −6.57, p < 0.001). The same applies to our selection of individuals who did not have psychosexual problems in childhood (group with psychosexual problems at T1, mean IQ = 94.35, SD = 14.28; group without psychosexual problems at T1, mean IQ = 98.12, SD = 14.79; t(2031) = −3.11, p < 0.01). The exclusion of those with psychosexual problems at T1 (who had a significantly lower IQ than those without psychosexual problems at T1), suggests there may be a relationship between psychosexual problems and IQ. Therefore, the individuals who were excluded due to missing data, who also had significantly lower IQ scores, may be a group with more psychosexual problems. The inclusion of individuals with a normal range of intelligence limits the generalizability of the current findings to clinical samples with lower IQ scores. Future research could investigate the relationship in a sample with lower IQ scores. However, our data do support the view that irrespective of cognitive abilities, autistic traits may be related to psychosexual problems (Nichols and Blakeley-Smith, 2009).

There are a few limitations to our studies. First, only a limited range of psychosexual problems was covered. Although the validity and reliability of the CBCL Sex Problems Scale have been examined in previous research (e.g. Bonner et al., 1999; Letourneau et al., 2008), this scale needs further psychometric research. Moreover, this measure is not exhaustive, since other problematic sexual behaviours, such as touching others inappropriately, as well as frustrations of the adolescent and concerns of parents were not assessed. In future research, it would therefore be valuable to investigate psychosexual problems in relation to autistic problems using a more extensive measure. Second, as the reliability of self-report in adolescents with autistic traits is questioned (e.g. Cederlund et al., 2010), we assessed psychosexual problems by means of parent-report. However, the usage of parent-report may lead to a biased result, in particular, as the scale is based on parental report regarding 4 items which have a subjective element (‘too much’). If we speculate, potentially parents of adolescents with autistic traits report psychosexual problems more often, as they may be more aware of the psychosexual problems. A possible explanation may be that the parent–child relationship in adolescence in a child with autistic traits may be less private, as the parent is more involved in day-to-day care of their child, than in parent–child relationships where the child is typically developing. Therefore, it would be valuable to pursue this line of research including measures from multiple informants. Third, our results are based on data from a general population sample, and our findings thus pertain to mild variation in autistic problems. Future research should assess whether indeed autistic traits in a clinical sample would predict future psychosexual problems, and thus generalize across all levels of severity.

Taken together, the results of this study highlight the importance of assessing autistic traits in childhood, as children with relatively high levels may be at-risk to develop psychosexual problems. Especially when considering that adult sexual outcome is rooted in adolescence (Tolman and McClelland, 2011). That said, we want to emphasize that displaying autistic traits, although a significant predictor of psychosexual problems, is only one of multiple predictors of psychosexual problems. Yet, the current results showed that the (change in) level of autistic traits significantly predicted psychosexual problems in adolescence above and beyond the effects of pubertal stage and conduct problems. In this field, we are only beginning to learn how autistic traits contribute to differences in socio-sexual learning, development and behaviour.

Footnotes

Acknowledgements

We are grateful to all adolescents, their parents and teachers who participated in this research and to everyone who worked on this project and made it possible.

Declaration of conflicting interests

Professor F.C. Verhulst MD, PhD, is a contributing author to the Achenbach System of Empirically Based Assessment (ASEBA) from which he receives remuneration. The other authors declare no conflict of interest.

Funding

This project was financially supported by the Sophia Children’s Hospital Fund (grant number 617, titled ‘Tackling Teenage: A Multicentre Study on Psychosexual Development and Intimacy in Adolescents with Autism Spectrum Disorder’). This research is part of the TRacking Adolescents’ Individual Lives Survey (TRAILS). Participating centres of TRAILS include various departments of the University Medical Center and University of Groningen, the Erasmus University Medical Center Rotterdam, the University of Utrecht, the Radboud Medical Center Nijmegen and the Parnassia Bavo group, all in the Netherlands. TRAILS has been financially supported by various grants from the Netherlands Organization for Scientific Research NWO (Medical Research Council program grant GB-MW 940-38-011, ZonMW Brainpower grant 100-001-004, ZonMw Risk Behavior and Dependence grants 60-60600-97-118, ZonMw Culture and Health grant 261-98-710, Social Sciences Council medium-sized investment grants GB-MaGW 480-01-006 and GB-MaGW 480-07-001, Social Sciences Council project grants GB-MaGW 452-04-314 and GB-MaGW 452-06-004, NWO large-sized investment grant 175.010.2003.005, and NWO Longitudinal Survey and Panel Funding 481-08-013, the Dutch Ministry of Justice (WODC), the European Science Foundation (EuroSTRESS project FP-006), Biobanking and Biomolecular Resources Research Infrastructure BBMRI-NL (CP 32)), and the participating universities.

References

Achenbach

(1991) Manual for the Child Behavior Checklist 2-3, 1992 Profile. Burlington, VT: Department of Psychiatry, University of Vermont.

Bailey

Palferman

Heavey

(1998) Autism: the phenotype in relatives. Journal of Autism and Developmental Disorders 28(5): 369–392.

Ballan

(2012) Parental perspectives of communication about sexuality in families of children with autism spectrum disorders. Journal of Autism and Developmental Disorders 42(5): 676–684.

Bolton

Macdonald

Pickles

(1994) A case-control family history study of autism. Journal of Child Psychology and Psychiatry 35(5): 877–900.

Bonner

Walker

Berliner

(1999) Children with Sexual Behavior Problems: Assessment and Treatment-Final Report (Grant No. 90-CA-1469). Washington, DC: U.S. Deparment of Health and Human Services, National Clearinghouse on Child Abuse and Neglect.

Brown-Lavoie

Viecili

Weiss

(2014) Sexual knowledge and victimization in adults with autism spectrum disorders. Journal of Autism and Developmental Disorders. 1–12. DOI: 10.1007/s10803-014-2093-y.

Cederlund

Hagberg

Gillberg

(2010) Asperger syndrome in adolescent and young adult males. Interview, self-and parent assessment of social, emotional, and cognitive problems. Research in Developmental Disabilities 31(2): 287–298.

Chaffin

Berliner

Block

(2006) Report of the Task Force on Children with Sexual Behavior Problems. Beaverton, OR: Association for the Treatment of Sexual Abusers (ATSA).

Constantino

(2009) How continua converge in nature: cognition, social competence, and autistic syndromes. Journal of the American Academy of Child & Adolescent Psychiatry 48(2): 97–98.

10.

Constantino

Todd

(2005) Intergenerational transmission of subthreshold autistic traits in the general population. Biological Psychiatry 57(6): 655–660.

11.

Dahl

(2004) Adolescent brain development: a period of vulnerabilities and opportunities. Keynote Address. Annals of the New York Academy of Sciences 1021(1): 1–22.

12.

De Winter

Oldehinkel

Veenstra

(2005) Evaluation of non-response bias in mental health determinants and outcomes in a large sample of pre-adolescents. European Journal of Epidemiology 20(2): 173–181.

13.

Dorn

Susman

Nottelmann

(1990) Perceptions of puberty: adolescent, parent, and health care personnel. Developmental Psychology 26(2): 322–329.

14.

Fergusson

John Horwood

Ridder

(2005) Show me the child at seven: the consequences of conduct problems in childhood for psychosocial functioning in adulthood. Journal of Child Psychology and Psychiatry 46(8): 837–849.

15.

Flannery

Rowe

Gulley

(1993) Impact of pubertal status, timing, and age on adolescent sexual experience and delinquency. Journal of Adolescent Research 8(1): 21–40.

16.

Friedrich

(2003) Studies of sexuality of nonabused children. Sexual Development in Childhood 7: 107–120.

17.

Friedrich

Grambsch

Damon

(1992) Child sexual behavior inventory: normative and clinical comparisons. Psychological Assessment 4(3): 303–311.

18.

Geluk

CaML

Jansen

Vermeiren

(2012) Autistic symptoms in childhood arrestees: longitudinal association with delinquent behavior. Journal of Child Psychology and Psychiatry 53(2): 160–167.

19.

Gotham

Pickles

Lord

(2012) Trajectories of autism severity in children using standardized ADOS scores. Pediatrics 130(5): e1278–e1284.

20.

Halpern

Udry

Campbell

(1993) Testosterone and pubertal development as predictors of sexual activity: a panel analysis of adolescent males. Psychosomatic Medicine 55(5): 436–447.

21.

Hartman

Luteijn

Moorlag

(2007) VISK. Vragenlijst voor Inventarisatie van Sociaal gedrag van Kinderen, herziene handleiding 2007. Amsterdam: Hartcourt.

22.

Hartman

Luteijn

Serra

(2006) Refinement of the Children’s Social Behavior Questionnaire (CSBQ): an instrument that describes the diverse problems seen in milder forms of PDD. Journal of Autism and Developmental Disorders 36(3): 325–342.

23.

Hellemans

Colson

Verbraeken

(2007) Sexual behavior in high-functioning male adolescents and young adults with autism spectrum disorder. Journal of Autism and Developmental Disorders 37(2): 260–269.

24.

Hénault

(2006) Asperger’s Syndrome and Sexuality: From Adolescence through Adulthood. London: Jessica Kingsley Publishers.

25.

Hoekstra

Happé

Baron-Cohen

(2010) Limited genetic covariance between autistic traits and intelligence: findings from a longitudinal twin study. American Journal of Medical Genetics Part B: Neuropsychiatric Genetics 153(5): 994–1007.

26.

Holmboe

Rijsdijk

Hallett

(2014) Strong genetic influences on the stability of autistic traits in childhood. Journal of the American Academy of Child & Adolescent Psychiatry 53(2): 221–230.

27.

Huisman

Oldehinkel

De Winter

(2008) Cohort Profile: the Dutch ‘TRacking Adolescents’ Individual Lives Survey’; TRAILS. International Journal of Epidemiology 37(6): 1227–1235.

28.

Jenson

(2010) Advances in preventing childhood and adolescent problem behavior. Research on Social Work Practice 20(6): 701–713.

29.

Jobe

Williams White

(2007) Loneliness, social relationships, and a broader autism phenotype in college students. Personality and Individual Differences 42(8): 1479–1489.

30.

Kumagami

Matsuura

(2009) Prevalence of pervasive developmental disorder in juvenile court cases in Japan. The Journal of Forensic Psychiatry & Psychology 20(6): 974–987.

31.

Letourneau

Chapman

Schoenwald

(2008) Treatment outcome and criminal offending by youth with sexual behavior problems. Child Maltreatment 13(2): 133–144.

32.

Letourneau

Schoenwald

Sheidow

(2004) Children and adolescents with sexual behavior problems. Child Maltreatment 9(1): 49–61.

33.

Maniglio

(2012) The role of parent–child bonding, attachment, and interpersonal problems in the development of deviant sexual fantasies in sexual offenders. Trauma, Violence, & Abuse 13(2): 83–96.

34.

Marshall

Tanner

(1969) Variations in pattern of pubertal changes in girls. Archives of Disease in Childhood 44(235): 291–303.

35.

Marshall

Tanner

(1970) Variations in the pattern of pubertal changes in boys. Archives of Disease in Childhood 45(239): 13–23.

36.

Nakamura

Ebesutani

Bernstein

(2009) A psychometric analysis of the Child Behavior Checklist DSM-Oriented Scales. Journal of Psychopathology and Behavioral Assessment 31(3): 178–189.

37.

Nichols

Blakeley-Smith

(2009) ‘I’m not sure we’re ready for this …’: working with families toward facilitating healthy sexuality for individuals with autism spectrum disorders. Social Work in Mental Health 8(1): 72–91.

38.

Pithers

Gray

Busconi

(1998) Children with sexual behavior problems: identification of five distinct child types and related treatment considerations. Child Maltreatment 3(4): 384–406.

39.

Sattler

(1992) Assessment of Children’s Intelligence and Special Abilities. 3rd ed. Boston, MA: Allyn & Bacon.

40.

Seto

Lalumiere

(2010) What is so special about male adolescent sexual offending? A review and test of explanations through meta-analysis. Psychological Bulletin 136(4): 526–575.

41.

Stokes

Kaur

(2005) High-functioning autism and sexuality: a parental perspective. Autism 9(3): 266–289.

42.

Stokes

Newton

Kaur

(2007) Stalking, and social and romantic functioning among adolescents and adults with autism spectrum disorder. Journal of Autism and Developmental Disorders 37(10): 1969–1986.

43.

Sullivan

Caterino

(2008) Addressing the sexuality and sex education of individuals with autism spectrum disorders. Education and Treatment of Children 31(3): 381–394.

44.

T Hart-Kerkhoffs

Jansen

Doreleijers

(2009) Autism spectrum disorder symptoms in juvenile suspects of sex offenses. Journal of Clinical Psychiatry 70(2): 266–272.

45.

Tanner

(1962) Growth at Adolescence. Oxford: Blackwell.

46.

Tolman

McClelland

(2011) Normative sexuality development in adolescence: a decade in review, 2000–2009. Journal of Research on Adolescence 21(1): 242–255.

47.

Totsika

Hastings

Emerson

(2011) A population-based investigation of behavioural and emotional problems and maternal mental health: associations with autism spectrum disorder and intellectual disability. Journal of Child Psychology and Psychiatry 52(1): 91–99.

48.

Van Haassen

De Bruyn

Pijl

(1986) Wechsler Intelligence Scale for Children–Revised (Dutch Version, Manual). Lisse: Swets, Zeitlinger BV.

49.

Verhulst

Van Der Ende

(2013) Handleiding ASEBA: Vragenlijsten voor leeftijden 6 tot en met 18 jaar. Rotterdam: ASEBA Nederland.

50.

Verhulst

Van Der Ende

Koot

(1996) Handleiding voor de CBCL/4-18: (Nederlandse versie). Rotterdam: Afdeling Kinder- en Jeugdpsychiatrie, Sophia Kinderziekenhuis/Academisch Ziekenhuis Rotterdam/Erasmus Universiteit Rotterdam.

51.

Visser

Greaves-Lord

Dekker

(2012) ‘Ik Puber’; Een onderzoek naar de effectiviteit van de Ik puber-training. Wetenschappelijk Tijdschrift Autisme 11(2): 64–67.

52.

Wechsler

(1974) Manual for the Wechsler Intelligence Scale for Children. Rev. ed. New York: Psychological Corporation.

53.

Whitehouse

AJO

Maybery

Hickey

(2011) Brief report: autistic-like traits in childhood predict later age at menarche in girls. Journal of Autism and Developmental Disorders 41(8): 1125–1130.

54.

Wieckowski

Hartsoe

Mayer

(1998) Deviant sexual behavior in children and young adolescents: frequency and patterns. Sexual Abuse: A Journal of Research and Treatment 10(4): 293–303.

55.

Woolfenden

Sarkozy

Ridley

(2012) A systematic review of the diagnostic stability of autism spectrum disorder. Research in Autism Spectrum Disorders 6(1): 345–354.

56.

Witkiewitz

Mcmahon

(2010) A parallel process growth mixture model of conduct problems and substance use with risky sexual behavior. Drug and Alcohol Dependence 111(3): 207–214.