Undiagnosed Psychotic Disorder in Autistic Individuals with Intellectual Disabilities and Suspected Obsessive-Compulsive Disorder: An Explorative, Clinical Study

ABSTRACT Introduction Mental health assessment in autistic individuals with intellectual disabilities may be challenging. Findings from non-autistic samples indicate overlap between obsessive-compulsive disorder (OCD) and psychotic disorder/schizophrenia, but little is known about this potential overlap in autistic people. Methods Explorative, retrospective chart study involving comprehensive, multimodal assessments for 18 autistic individuals with suspected OCD in a specialized mental health department. All participants had an intellectual disability (12) or significant impairments in adaptive behavior (6). Results While no participants had been diagnosed with psychotic disorder at referral, 7/18 participants were diagnosed with a psychotic disorder following assessment (schizophrenia or unspecified psychotic disorder). Conclusion OCD symptoms may overshadow psychotic symptoms in autistic individuals with intellectual disabilities. A combination of conventional assessment tools and assessment tools developed for autistic people may be helpful in differentiating OCD and psychosis in this population, as well as prototypical symptom considerations and exploring the developmental trajectories of symptoms.


Introduction
Autism spectrum disorder (American Psychiatric Association, 2013;World Health Organization, 2018) is a heterogeneous set of early-appearing neurodevelopmental conditions characterized by social interaction and communication difficulties, and focused/repetitive behaviors, activities, or interests (Lord et al., 2018;Lai et al., 2014). Atypical cognitive profiles are common, including difficulties involving social cognition and perception, as well as atypical perception and processing of sensory input (Crane et al., 2009;Lai et al., 2014). Between 18-47% of autistic individuals have been estimated to have a co-occurring intellectual disability (Lai et al., 2014;Postorino et al., 2016;Rivard et al., 2015;Roman-Urrestarazu et al., 2021), usually involving more severe difficulties in communication and social interaction (Lord et al., 2018;Lai et al., 2014). Intellectual disability is diagnosed when an individual displays impairment in intellectual functioning (IQ < 70), impairments in adaptive behavior, and these impairments have been present during the developmental period (American Psychiatric Association, 2013;World Health Organization, 2018). The diagnosis refers to an extremely heterogeneous set of conditions, etiologically and phenotypically (Burack et al., 2021).
Challenges in assessment include difficulties in distinguishing autismrelated characteristics from symptoms of mental disorder, as well as distinguishing between symptoms of different mental disorders (Helverschou et al., 2011;Kerns & Kendall, 2012;Postorino et al., 2017;Underwood et al., 2011). Adapted assessment tools for autistic people have been developed for specific disorders (e.g., for anxiety; Rodgers et al., 2016), but there is a lack of general and comprehensive assessment tools that include symptoms from across the range of mental disorders Underwood et al., 2011). The only diagnostic scale available for autistic people, the Schedule for the Assessment of Psychiatric Problems Associated with Autism (SAPPA; Bolton & Rutter, 1994; see also, Battaglia et al., 2016;Bradley & Bolton, 2006), has unclear psychometric properties and has been criticized because it does not take sufficiently account of potential atypical symptom manifestations . The Autism Spectrum Disorder-Comorbidity for Adults (ASD-CA; Matson & Boisjoli, 2008) is a screening tool for mental disorder in autistic people, but findings indicate that its sensitivity may be poor in people with co-occurring intellectual disabilities (LoVullo & Matson, 2009;Underwood et al., 2011). The Psychopathology in Autism Checklist (PAC; Helverschou et al., 2009), originally developed as a screening tool for mental disorder in autistic adults with intellectual disabilities, seems to have adequate psychometric properties in this population (Helverschou et al., 2009(Helverschou et al., , 2021 but remains a screening tool and has limitations with regard to diagnostic use (Helverschou et al., 2021). Finally, the sensitivity of other existing assessment tools may be poor in individuals with co-occurring intellectual disabilities Kerns et al., 2021;Underwood et al., 2011).
Thus, misdiagnoses of mental disorders may be common, including: a) intellectual disability-or autism-related behaviors being misattributed to mental disorder; b) symptoms of a co-occurring mental disorder being misattributed to intellectual disability and/or autism ("diagnostic overshadowing"; Reiss et al., 1982;see also, Au-Yeung et al., 2019;Bakken & Høidal, 2014;Bakken et al., 2016;Helverschou et al., 2011;Jopp & Keys, 2001). For autistic individuals with limited verbal language skills, there may be a particular risk of overlooking or misinterpreting symptoms of mental disorders for which diagnosis depends on recognition of subjective/experiential symptoms, such as trauma-related disorders or psychosis (Bakken, 2021;Kildahl et al., 2017;Kildahl, Bakken et al., 2019).

Autism and Obsessive-Compulsive Disorder
Obsessive-compulsive disorder (OCD) involves recurrent intrusive thoughts/ worries and/or compulsive/repetitive behaviors (American Psychiatric Association, 2013). To meet criteria for OCD, obsessions and/or compulsions must be time consuming and cause marked distress or impairment. Typically, intrusive thoughts or images are accompanied by anxiety or distress, which is temporarily relieved by performing intentional compulsive/repetitive behaviors (American Psychiatric Association, 2013). The prevalence of OCD seems to be increased in autistic individuals (Lai et al., 2019;Postorino et al., 2017;Scahill & Challa, 2016), including individuals with co-occurring intellectual disabilities (Bakken et al., 2010;Bradley et al., 2011).
While OCD symptoms and autism-related repetitive/restrictive behaviors seem to constitute distinct phenomena (Bedford et al., 2020;Scahill & Challa, 2016), distinguishing compulsive/repetitive behaviors in OCD from autismrelated repetitive/restrictive behaviors may be challenging (Bedford et al., 2020;Kannabiran & McCarthy, 2009;Kerns & Kendall, 2012;Postorino et al., 2017;Scahill & Challa, 2016), particularly in autistic individuals with intellectual disabilities (Postorino et al., 2017). Recent findings from Santore et al. (2020) indicate that parent appraisal of whether a specific behavior is autism-related or OCD-related may not always be in line with the subjective appraisals of the autistic children/adolescents themselves (see also, Postorino et al., 2017). Thus, distinguishing between autism-related repetitive behaviors and OCD symptoms does not seem to be possible by relying on informant observations of behavior alone, as these findings indicate that it is necessary to take account of the mental state and level of discomfort/distress of the individual to adequately distinguish OCD-related and autism-related repetitive behaviors (Postorino et al., 2017;Santore et al., 2020;Scahill & Challa, 2016). This is likely to be challenging in autistic individuals with co-occurring intellectual disabilities (Bradley et al., 2011).
Autism-related repetitive/inflexible behaviors may constitute attempts of increasing predictability and reducing anxiety, but the specific behavior itself seems less likely to be associated with distress or anxiety (Scahill & Challa, 2016). In other words, behavioral patterns may overlap for OCD symptoms and autism-related repetitive or inflexible behaviors, while their underlying motivations are unlikely to overlap to the same degree (Postorino et al., 2017;Santore et al., 2020;Scahill & Challa, 2016). In addition, autism is conceptualized as an early-appearing neurodevelopmental condition (Lord et al., 2018;Lai et al., 2014), while debut of OCD typically occurs during late childhood, adolescence, or early adulthood (Solmi et al., 2021;Taylor, 2011). In addition, some routine-related symptoms and behaviors may be more "prototypical" of OCD (e.g., checking, cleaning, counting) while some may be more "prototypical" of autism (e.g., focused interests, tapping, touching; Kannabiran & McCarthy, 2009;Scahill & Challa, 2016). However, increases or changes to repetitive/restrictive behavior may occur for mental disorders other than OCD (Ghaziuddin et al., 2002;Helverschou et al., 2011;Kildahl et al., 2017;, indicating a potential risk of misinterpreting signs of other mental disorders as OCD in autistic individuals.

Autism and Schizophrenia
Symptoms of schizophrenia and other psychotic disorders include positive symptoms such as hallucinations, delusions, and disorganized thought/ speech/behavior, as well as negative symptoms such as social withdrawal and loss of interest/emotionality (American Psychiatric Association, 2013).
Autistic individuals appear to be at increased risk of schizophrenia (Lai et al., 2019;De Giorgi et al., 2019;Marín et al., 2018;Varcin et al., 2022;Zheng et al., 2018), and this may include autistic individuals with intellectual disabilities (Bakken et al., 2010;De Giorgi et al., 2019;Guinchat et al., 2015). In the general population, longer duration of untreated psychosis prior to initial treatment is associated with poorer long-term outcomes (Penttilä et al., 2014;Qin et al., 2014), underlining the importance of early recognition and treatment of psychotic disorders.
Differentiating autism and schizophrenia may be challenging (Bakken, 2021;Bakken & Høidal, 2014;Kannabiran & McCarthy, 2009;Kildahl et al., 2017;Oliver et al., 2021;Palucka et al., 2008;Trevisan et al., 2020;Underwood et al., 2015). A recent review found similar levels of social cognitive difficulties in autistic people and people with schizophrenia (Oliver et al., 2021), and these conditions may overlap with regard to apparent social withdrawal and repetitive behaviors (Bakken & Høidal, 2014;Varcin et al., 2022). Trevisan et al. (2020) found autism and schizophrenia to be differentiated by positive but not apparent negative symptoms, recommending attention to presence or absence of positive symptoms to distinguish these conditions. Bakken and Høidal (2014) reached a similar conclusion, emphasizing presence of hallucinations and disorganized speech/behavior, while further cautioning that some autismrelated phenomena, such as idiosyncratic speech patterns, may be misinterpreted as delusions unless they are sufficiently explored (see also, Kannabiran & McCarthy, 2009;Palucka et al., 2008).
Furthermore, autism is an early-appearing neurodevelopmental condition (Lai et al., 2014), while schizophrenia spectrum disorders typically have onset in adolescence or early adulthood (Solmi et al., 2021). In line with this, Palucka et al. (2008) emphasized the developmental and historical perspective as important to adequately distinguish these conditions, while Bakken and Høidal (2014) highlighted age of onset as well as relapses as important markers. However, increased prevalence of autism spectrum disorder has been reported among individuals with early onset schizophrenia (Driver et al., 2020;Kannabiran & McCarthy, 2009), further highlighting the potential complexity

Schizophrenia and OCD
While OCD and schizophrenia are conceptualized in different ways and traditionally have been considered distinct conditions (Rasmussen & Parnas, 2022), recent findings from non-autistic samples indicate they may share substantial overlap (Buchholz et al., 2021;Mawn et al., 2020;Rasmussen et al., 2020;Swets et al., 2014). In their meta-analysis, Swets et al. (2014) found mean prevalence of OCD in individuals with schizophrenia to be 12.3%, higher than the 1-3% prevalence found in the general population (Ruscio et al., 2010). In addition, obsessive-compulsive symptoms may occur in up to 30% of individuals with schizophrenia (Buchholz et al., 2021;Mawn et al., 2020;Rasmussen et al., 2020;Swets et al., 2014). Presence of these symptoms of seem to require adjustments of treatment strategies for schizophrenia (Buchholz et al., 2021), indicating that individuals should routinely be assessed for OCD or schizophrenia when presenting with symptoms of either condition (Mawn et al., 2020).
Traditionally, obsessions have been understood as intrusive thoughts, occurring in the context of intact resistance and insight, while obsessivecompulsive phenomena in schizophrenia have been understood to be associated with symptoms such as disorganized thinking or catatonia (Rasmussen & Parnas, 2022). However, recent perspectives suggest that these phenomena cannot be clearly distinguished (Rasmussen & Parnas, 2022). It has also been suggested that "obsessive psychosis" may constitute a distinct clinical category (e.g., Robinson et al., 1976). Hallucinatory experiences, however, is associated with schizophrenia and not OCD (American Psychiatric Association, 2013), and mean age of onset seems to be higher for schizophrenia than for OCD (Solmi et al., 2021).
In sum, distinguishing OCD and schizophrenia seem to be challenging in non-autistic individuals with average or above intellectual abilities (Mawn et al., 2020;Rasmussen et al., 2020;Rasmussen & Parnas, 2022). Despite increased prevalence of schizophrenia and OCD in autistic individuals (Lai et al., 2019;Scahill & Challa, 2016;Varcin et al., 2022), differential diagnosis of these potentially co-occurring mental disorders has been sparsely explored in this population, including whether such differential diagnosis is possible and what may be helpful to distinguish these conditions.

Aim
The aim of the current study is to explore the assessments and final diagnostic conclusions in clinical cases of autistic individuals with suspected OCD, including the possibility of overlaps between OCD and psychotic disorders, focusing on individuals with co-occurring intellectual disabilities or significant impairment in adaptive behavior: (1) Following comprehensive, specialized mental health assessments in autistic adults and older adolescents, what are the final diagnoses in cases where a clinical suspicion of OCD has been raised prior to or during assessment? (2) Are the PAC psychosis and PAC OCD scales helpful in differentiating OCD and psychosis in a clinical sample of autistic adults and older adolescents?
(3) What strategies do mental health professionals use to differentiate OCD, psychosis and autism-related characteristics, and what symptoms or behaviors do they emphasize?

Design and Setting
The study was designed as a retrospective, explorative study, using clinical data collected during comprehensive, interdisciplinary, multimodal assessments in a specialized psychiatric department. The department in question accepts referrals for individuals older than 16 from other hospital-level services in mental health or habilitation (see, Bakken et al., 2018), and functions as a tertiary mental health service for people with intellectual disabilities in an area comprising approximately 3.1 million inhabitants. Referrals for autistic people without intellectual disability diagnoses are accepted if individuals, at referral, have prolonged and significant impairments in adaptive behavior that are clinically assessed to be comparable to those seen in intellectual disabilities. The department consists of two inpatient wards and a specialized outpatient clinic, and cases usually involve a high degree of complexity. Inpatient assessments are conducted by teams including a clinical psychologist, a psychiatrist, mental health and intellectual disability nurses, as well as other ward staff. Outpatient assessments are carried out by a clinical psychologist and a psychiatrist, sometimes in collaboration with an intellectual disability nurse. The duration of assessments is typically around 6 months and these are multimodal, including interviews with the patient (when feasible) and multiple informants (typically parents and care staff working in patients' homes), use of multiple structured assessment tools (see Materials), plus direct observation and/or interactions with the patient in various contexts. Assessments also include reviews of patients' developmental histories and complete medical charts. For examples, see, Kildahl, Berg et al. (2020;inpatient), or Kildahl, Engebretsen et al. (2019;outpatient).

Ethics
The study was approved by the Data Protection Official at the Oslo University Hospital (#19/22,199). Written, informed consent was obtained from participants themselves and/or their legal guardians. The study was conducted in compliance with the American Psychological Association ethical standards concerning the treatment of the sample and in obtaining informed consent.

Participants
All inpatients and outpatients under assessment or treatment at the department between January 2017 and December 2020 were eligible for participation. Inclusion criteria were: (1) autism spectrum disorder diagnosis according to ICD-10 (World Health Organization, 1992), and (2) previously diagnosed OCD, or a clinical suspicion of OCD by department staff sometime during the assessment. Participants were mostly referred for comprehensive mental health assessment, and not necessarily due to OCD alone. Twenty-two former and current patients met the inclusion criteria.
Consent was obtained from 18 participants (4 females, 14 males; 12 inpatient, 6 outpatient), whose demographic data are presented in Table 1. Age of participants ranged from 16 to 46 (M = 24.44, SD = 9.59). All participants had been diagnosed with autism spectrum disorder prior to referral, and these diagnoses were reevaluated and confirmed during the assessment. One participant had a co-occurring severe intellectual disability, four had moderate intellectual disabilities, and seven had mild intellectual disabilities. Six had no co-occurring intellectual disability, but had significant impairments in adaptive behavior and often assumed IQs in the 70-85 area. Due to the difficulties obtaining accurate information about intellectual abilities in individuals with complex mental and/or behavioral difficulties, more specified IQ measures were not available. Intellectual disability diagnoses, reevaluated and confirmed according to the ICD-10 during the assessments, were based on a combination of medical charts/previous assessments, which included previous Wechsler tests for a majority of participants, clinical judgment, and evaluation of adaptive behavior. Adaptive behavior composite standard scores from the Vineland Adaptive Behavior Scales, second edition, expanded version (Sparrow et al., 2008), using the original American norms, ranged from 20 to 72 (M = 41.61, SD = 17.13). Socioeconomic data were not collected, but all participants received services and support from their local municipalities and lived in a group home or similar.

Materials
Data were collected from the comprehensive reports from each assessment, including clinical case formulations, diagnoses at referral, final diagnoses, psychopharmacological treatment, results from assessment tools, and demographic data. Scores from two instruments used in all assessments (Psychopathology in Autism Checklist, Aberrant Behavior Checklist) were retrieved for statistical analysis, while scores from a measure of psychotic symptoms (Positive and Negative Syndrome Scale) were retrieved where available. Information about other assessment tools was retrieved to describe   the basis for the clinical diagnoses, but results from these were not retrieved for further analyses.

Assessments and Diagnostic Formulations
All assessments included at least one general diagnostic instrument such as the Mini International Neuropsychiatric Interview (MINI; Sheehan et al., 1998;Mordal et al., 2010; et al., 1985), at least one measure of autism characteristics, typically the Autism Diagnostic Interview-Revised (Lord et al., 1994) or the Social Communications Questionnaire (Rutter et al., 2003), and a measure of adaptive behavior (Vineland Adaptive Behavior Scales, second edition, expanded form; Sparrow et al., 2008). All instruments were used in their Norwegian versions or translations. All assessments concluding with OCD, or involving the discontinuation of a previously given OCD diagnosis, included either the Yale-Brown Obsessive Compulsive Scale (Goodman et al., 1989) or the Children's Yale-Brown Obsessive Compulsive Scale (Scahill et al., 1997); all assessments involving a conclusion of psychotic disorder included the Positive and Negative Syndrome Scale (PANSS; Kay et al., 1987). Further instruments were used in several of the assessments according to the specific concerns and co-occurring difficulties of each specific individual, e.g., measures for depression, sensory issues, anxiety, executive functioning, motor difficulties, etc.
Diagnostic formulations were made jointly by the responsible psychiatrist and clinical psychologist involved in each case. In addition, all diagnostic conclusions were discussed, and consensus usually reached, in interdisciplinary meetings involving all the department's psychiatrists and psychologists.

Psychopathology in Autism Checklist (PAC)
The PAC (Helverschou et al., 2009), originally developed as a screening tool for mental disorder in autistic adults with co-occurring intellectual disabilities, has been validated for this specific population in its Norwegian version (Helverschou et al., 2009). Recent findings suggest it may be helpful in differentiation of OCD and psychosis in autistic individuals (Helverschou et al., 2009(Helverschou et al., , 2021. The PAC consists of 42 items rated by informants on a 4-point scale (not a problem = 1, minor problem = 2, moderate problem = 3, severe problem = 4). Scores for 30 items are reduced to four scales: psychosis, depression, anxiety, and OCD. The 12 remaining items are reduced to a score for general adjustment problems. Though the sensitivity and specificity varies somewhat for the individual scales, recent findings suggest that the PAC is well suited to differentiate autistic individuals with intellectual disabilities who meet criteria for a co-occurring psychiatric disorder from those who do not, and adequate to good psychometric properties have been reported (Helverschou et al., 2009(Helverschou et al., , 2021. The PAC is a proxy measure and was completed by parents and/or professional caregivers. Participants' individual scores on the PAC scales are presented in Table 1.

Positive and Negative Syndrome Scale (PANSS)
The PANSS is a 40-item assessment scale developed for measurement of symptom severity in schizophrenia (Kay et al., 1987). It includes a structured clinical interview (Opler et al., 1999) but is scored by the clinician based on information also from other sources (Kay et al., 1987). All items are scored on a seven-point Likert scale. Seven items are reduced to a scale of positive psychotic symptoms (range 0-49), seven items are reduced to a scale of negative psychotic symptoms (range 0-49), while the remaining sixteen items are reduced to a scale of general psychopathology (range 0-112). The PANSS has shown good psychometric properties (Kay et al., 1987;Peralta & Cuesta, 1994), also in individuals with mild intellectual disabilities (Hatton et al., 2005). While not validated for autistic individuals with intellectual disabilities, previous case studies have described it as helpful in assessment of psychotic symptoms in this population, including when used primarily with informant/proxy report and clinical observation (e.g., Kildahl et al., 2017;Kildahl, Berg et al., 2020;Rysstad et al., 2022). According to the assessment reports in the current study, the structured clinical interview (Opler et al., 1999), i.e. self-report, had been attempted or completed with participants in all the assessments that included the PANSS. Scoring of the PANSS, however, was based on these interviews in combination with proxy report and clinical observations from inpatient or group home settings.

Procedure
Data were collected from participants' records by KCD and ANK. Frequency analyses were conducted for diagnoses at referral and final diagnoses. Sensitivity and specificity of the PAC scales for OCD and psychosis were calculated for final, clinical diagnoses, using cutoff values from the validation study by Helverschou et al. (2009).
The current sample is small, limiting statistical analysis and generalizability. We therefore chose to use a nonparametric test, the Mann-Whitney U test, to explore group differences for PAC psychosis and PAC OCD. The same test was used for further exploration of group differences on the remaining PAC and ABC scales. Due to the small sample size and the explorative nature of the study, the alpha level was set at .05. SPSS Statistics 26 was used for the statistical analyses. The small sample size indicates that findings from these analyses should be interpreted with caution.
For the clinical case formulations, a thematic content analysis (Green & Thorogood, 2018) was conducted. All case formulations were printed and read individually by all authors. They were then re-read focusing on research question 3 (What strategies do mental health professionals use to differentiate OCD, psychosis and autism-related characteristics, and what symptoms or behaviors do they emphasize?), using a three-step approach: (1) strategies/symptoms/behaviors differentiating psychosis and autism, (2) strategies/symptoms/behaviors differentiating OCD from repetitive or ritualistic behaviors associated with autism, (3) strategies/symptoms/behaviors differentiating psychosis and OCD. All quotes relating to each step were extracted verbatim from the case formulations. Lists of quotes compiled individually by the authors were then compared and discussed, before the authors collaborated in grouping these examples into final categories. Quotes substantiating each category were chosen by all authors. Included quotes were translated from Norwegian and minimally edited for clarity by KS. KCD and ANK checked all translations.

Frequencies
At referral, 12/18 participants (67%) had been diagnosed with OCD, while none had been diagnosed with a psychotic disorder. Following assessment, the number of participants diagnosed with OCD decreased to 8/18 (44%), while the number diagnosed with a psychotic disorder increased to 7/18 (39%). Of these, 5/18 participants (28%) were diagnosed with co-occurring psychosis and OCD. Three participants (3/18; 17%) were diagnosed with OCD but not psychosis, while 2/18 (11%) were diagnosed with psychosis but not OCD. Psychotic disorder diagnoses included schizophrenia (5) and unspecified nonorganic psychosis (2). Remaining participants were diagnosed with at least one co-occurring psychiatric disorder, including anxiety disorders or co-occurring anxiety and depression (6/18; 33%), bipolar disorder (1), or personality disorder (1). Psychotic disorders were only identified in participants with mild or no intellectual disability, while OCD was identified in participants with mild or moderate intellectual disabilities, as well as in participants without intellectual disabilities.
Several participants had been treated with antipsychotics prior to their assessment, despite not having been diagnosed with a psychotic disorder, which is not uncommon in this population (Posey et al., 2008). Cases were generally complex, some involving co-occurring conditions such as epilepsy, attention deficit/hyperactivity disorder, other psychiatric disorders, sensory disabilities, as well as self-injurious and other "challenging" behaviors. PANSS scores were primarily available for participants who had been diagnosed with a psychotic disorder, see, Table 1.

PAC sensitivity/specificity
For the PAC psychosis scale, 8/18 scored above cutoff, five of whom were diagnosed with a psychotic disorder. There were two false negatives, i.e. two participants were diagnosed with a psychotic disorder but scored below the cutoff on PAC psychosis, both scoring just below the cutoff (2.2, 2.1; cutoff = 2.3); and three false positives, i.e. participants who scored above the cutoff but were not diagnosed with a psychotic disorder. Sensitivity for the PAC psychosis scale in the current sample was .71 and specificity was .73.
For the PAC OCD scale, 12/18 scored above cutoff, seven of whom were diagnosed with OCD. There was one false negative, and five false positives. Sensitivity for the PAC OCD scale in the current sample was .88 and specificity was .45.
For the combination of the psychosis and OCD scales, 3/5 were correctly identified. For the combination of these scales in the current sample, sensitivity was .60 and specificity was .85.

Group Differences
Scores on PAC psychosis were higher for participants who were diagnosed with a psychotic disorder (Mdn = 2.50) than for those who were not (Mdn = 2.00). A Mann-Whitney U-test indicated that this difference was statistically significant U(N psychotic = 7, N nonpsychotic = 11) = 15.50, z = −2.09, p = .037. Further exploration showed no other significant differences between these groups on PAC or ABC scales.
Scores on PAC OCD were higher for participants who were diagnosed with OCD (Mdn = 2.80) than for those who were not (Mdn = 2.35). A Mann-Whitney U-test indicated that this difference was not statistically significant U(N OCD = 8, N nonOCD = 10) = 22.00, z = −1.61, p = .107. Further exploration showed no other significant differences between these groups on PAC or ABC scales.

Qualitative Analyses of Clinical Case Formulations
Results from the content analysis of the clinical case formulations are presented in Table 2. Most of the reports explicitly stated that these conditions are challenging to differentiate, and that the final diagnostic conclusions had been reached with some degree of uncertainty. Several reports stated that behavioral level differentiation was challenging, and diagnoses were based on patients displaying observable symptoms belonging to certain diagnostic categories. In addition to these "prototypical" examples, most patients displayed symptoms or behaviors that were difficult to attribute to any single, diagnostic category. Several clinicians warned against attempts at "sorting" all the respective patients' symptoms/characteristics into either of the diagnostic categories, because of potential interactions or overlaps between these categories.
To differentiate characteristics associated with autism from symptoms of psychotic disorder, clinicians appeared primarily to emphasize four aspects: a) patients' developmental histories and symptom trajectories (whether these were in line with typical trajectories in psychotic disorders, involving substantial changes to or loss of functioning during adolescence or early adulthood), b) disorganized behavior or speech, c) positive symptoms of psychosis (delusions/hallucinations), and d) treatment response (effects of treatment with antipsychotics, or effects of discontinuing antipsychotic treatment previously prescribed for "challenging" behaviors or anxiety).
To differentiate OCD symptomology from repetitive/ritualistic behaviors associated with autism, clinicians appeared to emphasize two main aspects: a) whether patients experienced these symptoms or behaviors as ego-dystonic or ego-syntonic (i.e. whether patients themselves seemed bothered by the behaviors or the thoughts triggering them), and b) whether the thought content or  (Continued) behavioral expressions were prototypical of OCD (e.g., numbers, fear of bacteria, hand washing).
To differentiate OCD symptoms from psychotic symptoms, clinicians seemed to emphasize similar aspects to those used to differentiate autism from psychosis and OCD, in particular a) patients' developmental histories and symptom trajectories, b) whether thought content or behavioral expressions were more typical of OCD or psychosis -or both (e.g., whether the patient was hearing a voice telling them to do something or whether they experienced it as their own thoughts), and c) treatment response. For treatment response, two patterns were described: one in which OCD symptoms appeared to subside along with psychotic symptoms, and one in which OCD symptoms appeared to increase as psychotic symptoms decreased and the patient's behavior became less disorganized.

Discussion
In a clinical sample of autistic adults/older adolescents with suspected OCD and intellectual disabilities or significant impairments in adaptive behavior, comprehensive multimodal mental health assessments revealed that some of the participants met criteria for a previously undiagnosed psychotic disorder. While the sample is not representative and the generalizability of findings may be limited, this nevertheless suggests a risk that underlying psychotic disorders may go unrecognized in autistic individuals with intellectual disabilities presenting with apparent OCD symptomology. As psychotic disorders and OCD have different implications for treatment and care, these findings highlight the importance of thorough, multimodal assessments when co-occurring OCD is suspected in autistic individuals. Furthermore, the current study highlights the complexity of mental health assessment in this population, as well as the importance of expert clinical judgment and not relying on checklists alone. Comprehensive, multimodal assessments conducted over a period of time in many of the cases revealed symptoms that previously had been overlooked or misinterpreted. This highlights the importance of these approaches in assessment, including clinical observation. Use of conventional assessment tools such as MINI/Kiddie-SADS seemed helpful, even if these instruments have not been adapted for autistic people and were often used in a nonstandardized way (with informants). Clinicians with specific knowledge concerning mental disorders in autistic individuals scored the PANSS, and scores for participants receiving a diagnosis of psychotic disorder were similar to scores seen in non-autistic patients with psychotic disorders (Leucht et al., 2005). In line with previous findings (Helverschou et al., 2021), the PAC may be helpful in these assessments, particularly with regard to uncovering psychotic symptoms in autistic individuals with suspected OCD. However, these findings highlight that any of these methods alone are unlikely to be sufficient in cases involving this degree of complexity, and that their results need to be interpreted and judged by expert clinicians.
Behaviors typically associated with OCD, such as repeated checking, excessive hand washing or doing something a certain number of times, may be easily observable to others. Symptoms of psychotic disorder may primarily manifest as distressing, internal, subjective experiences that are not directly observable. Thus, symptoms of an underlying psychotic disorder may be masked or overshadowed by apparent OCD symptoms or increased repetitive/ritualistic behaviors. To differentiate these conditions, clinicians in the current study relied on participants' developmental histories and symptom trajectories, particularly looking for loss of functioning (including loss of verbal language skills and adaptive behavior), signs of disorganized behavior, or positive psychotic symptoms such as hallucinations (see also, Bakken & Høidal, 2014;Trevisan et al., 2020). In line with this, Rasmussen et al. (2020) have emphasized the importance of exploring the developmental aspects of mental disorders in differential diagnostic assessments (see also, Bakken & Høidal, 2014;Palucka et al., 2008), as well as including prototypical considerations.
As suggested by Underwood et al. (2015), identification of psychotic disorder in autistic individuals may in some cases depend on monitoring response to attempted treatment. For some participants in the current study, this involved trials of discontinuing antipsychotics previously prescribed due to "challenging" behaviors. Varying effects of antipsychotic treatment on OCD symptomology were described, including cases where apparent OCD symptoms were alleviated with antipsychotic treatment, as well as cases where OCD symptoms apparently worsened as psychotic symptoms subsided and participants' behaviors became less disorganized. It is possible that these effects reflect two differing phenomenologies, which may be challenging to differentiate at the behavioral level: one where apparent OCD symptomology is secondary to psychotic symptoms (as in "obsessional psychosis"), and one involving a "true" co-occurrence of OCD and psychosis with these disorders following more distinct developmental trajectories. However, some antipsychotics may exacerbate OCD symptoms (Schirmbeck & Zink, 2012), and more research is needed to disentangle the potentially different underlying processes in the overlap between psychotic disorder and OCD.

Limitations
The current study was conducted using a small sample from a specialized psychiatric department and therefore has limited generalizability. Findings from the statistical analyses are unlikely to be representative and cannot be generalized beyond the current sample. Moreover, these patients may have been referred because other services found their symptoms peculiar or unusual and suspected presence of other mental health issues. Recent findings suggest that informants' reports and understanding of apparent OCD symptoms may not always be in line with the experience of autistic individuals themselves (Santore et al., 2020). It is therefore possible that clinicians also in the current study misinterpreted OCD symptoms as autism-related repetitive/ritualistic behaviors in assessments which primarily relied on informant report, and that OCD was under-recognized in the current sample. The sample also had a wide age-range. While these limitations reduces the study's likely generalizability, the current sample included individuals who are often excluded from larger-scale studies due to the complexity and severity of their difficulties and are therefore rarely represented in mental health research.
Clinicians were aware of participants' PAC scores when writing their case formulations and may have been affected by them. However, none of the clinicians appeared to emphasize these explicitly in their diagnostic formulations. Finally, the current study was a retrospective, clinical study, and while all assessments appeared to include similar approaches and assessment tools, they may have varied in their approaches and structures.

Conclusion
These findings indicate that psychotic disorders may be overlooked in autistic individuals displaying OCD-like symptomology. Adequate differential diagnostic assessment and identification of these mental disorders is important, as recommended pharmacological and psychological treatments for OCD and psychosis differ considerably. Psychotic disorders going unrecognized, and lack of appropriate treatment of such disorders in autistic individuals with intellectual disabilities, may have severe consequences for affected individuals' mental health and prognoses. Mawn et al. (2020) have recommended that individuals presenting with either OCD or psychosis should be routinely assessed for both disorders, and the current findings indicate that this recommendation applies also to autistic individuals, including those with cooccurring intellectual disabilities.
Complex and severe symptom presentations such as the ones described in the current study may be disentangled by focusing on the developmental aspects of these presentations and symptom trajectories, as well as prototypical considerations concerning the different mental disorders. Though further research is needed, well-established assessment tools developed for the general population (e.g., PANSS) may be helpful in these assessments, particularly if combined with more specific tools developed for autistic individuals (e.g., PAC). However, interdisciplinary and multimodal approaches including expert clinical observation and judgment is likely to be necessary to disentangle these complex symptom presentations.