Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2026 Jul 1.
Published in final edited form as: Am J Med Genet B Neuropsychiatr Genet. 2025 Mar 7;198(5):e33026. doi: 10.1002/ajmg.b.33026

Association between polygenic risk and symptom severity change after cognitive behavioral therapy for obsessive-compulsive disorder

Julia Bäckman 1, John Wallert 1, Matthew Halvorsen 1,2, Bjorn Roelstraete 3, Elles de Schipper 1, Nora I Strom 1,4,5,6, Thorstein Olsen Eide 7,8,9, Kira D Höffler 7,10,11, Manuel Mattheisen 5,6,12,13, Bjarne Hansen 7,9, Gerd Kvale 7, Kristen Hagen 7,14,15, Jan Haavik 7,16; Nordic OCD and Related Disorders Consortium (NORDiC), David Mataix-Cols 1,15,#, Christian Rück 1,#, James J Crowley 1,2,#
PMCID: PMC12129668  NIHMSID: NIHMS2063255  PMID: 40052195

Abstract

A large proportion of patients undergoing cognitive behavior therapy (CBT) for obsessive-compulsive disorder (OCD) do not respond sufficiently to treatment. Identifying predictors for change in symptom severity after treatment could inform clinical decision-making, allow for better-tailored interventions, and avoid treatment failure. Prior research on predictors for treatment response has however yielded inconsistent findings with limited clinical utility. Here, we investigated the predictive power of nine polygenic risk scores (PRSs) for psychiatric and cognitive traits in 1,598 OCD patients (1,167 adults and 431 children/adolescents) treated with CBT in Sweden and Norway. We fitted linear mixed models adjusted for age, sex, genotyping batch, and the first five ancestry PCs to estimate associations between PRS and symptom severity change from pre- to post-treatment. The PRS for schizophrenia showed a modestly significant association with symptom change (β = 0.013, p = 0.04, R2=0.10), indicating that a higher PRS for schizophrenia was associated with a smaller decrease in symptom severity. No other PRS were significantly associated with the outcome. While these results await replication and expansion, current PRS for psychiatric and cognitive phenotypes do not seem to contribute meaningfully to symptom severity change in CBT for OCD.

Keywords: polygenic risk score, cognitive behavior therapy, obsessive-compulsive disorder, symptom severity

INTRODUCTION

Obsessive-compulsive disorder (OCD) is a disabling psychiatric disorder that affects around 1–2% of the population (Fawcett et al., 2020; Ruscio et al., 2010) and is associated with an increased risk of suicide and overall mortality (De La Cruz et al., 2016; Meier et al., 2016). Cognitive behavioral therapy (CBT) for OCD (Öst et al., 2015; Rosa-Alcázar et al., 2008) has strong evidence for effectiveness in treating OCD and is recommended by guidelines as first-line treatment for the disorder (National Collaborating Centre for Mental Health (UK), 2006). Still, 30–42% of patients do not respond sufficiently to treatment (Mataix-Cols et al., 2022; Öst et al., 2015). The impact of therapeutic non-response is substantial, including profound personal and societal costs (World Health Organization, 2008).

Given the non-negligible individual variability in treatment outcome and limited healthcare resources, it is essential to better understand who will and will not respond to CBT for OCD. Identifying predictors of symptom change could inform clinical decision-making and allow for better-tailored interventions for patients. Several reviews have investigated predictors of outcomes of CBT for individuals with OCD (Keeley et al., 2008; Knopp et al., 2013; McDonald et al., 2023). Overall, few stable predictors have been observed across studies, presenting a challenge for both researchers and clinicians. Predictors that have been found in more than one study include baseline symptom severity (Keeley et al., 2008; Knopp et al., 2013; McDonald et al., 2023), OCD subtype (Keeley et al., 2008; Knopp et al., 2013), and therapeutic alliance (Keeley et al., 2008; McDonald et al., 2023). Some evidence also indicates higher effect sizes for children than adults (Olatunji et al., 2013). There has long been a lack of consensus for defining treatment response and remission in treatment studies (Mataix-Cols et al., 2016), which may have contributed further to result heterogeneity. In conclusion, our knowledge about predictors of response to CBT for OCD, i.e., symptom severity change between pre- and post-treatment, remains insufficient for use in clinical decision-making.

The heritability of OCD is around 50% (Blanco-Vieira et al., 2023; Mataix-Cols et al., 2024). The most extensive genome-wide association study (GWAS) of OCD to date, comprising 53,660 OCD cases and 2,044,417 controls, has identified 30 risk loci associated with the disorder (Strom et al., 2024). As with all psychiatric disorders, each single genetic variant (or single nucleotide polymorphisms, SNPs) usually contribute very little to the risk of developing the disorder. Consequently, polygenic risk scores (PRS) constitute a potentially valuable option to capture the cumulative effects of multiple genetic risk variants. A polygenic risk score (PRS) is a numeric estimate of an individual’s genetic predisposition to a particular trait or disease, calculated by summing the weighted effects of multiple SNPs associated with the target condition or trait. In the somatic field, PRS has successfully been used to identify individuals at elevated risk for coronary artery disease, atrial fibrillation, type 2 diabetes, inflammatory bowel disease, and breast cancer (Khera et al., 2018) Using genetic variables as predictors for psychotherapy-related targets, known as therapygenetics (Lester & Eley, 2013), has also gained traction. Notable findings include that a higher PRS for autism predicted worse outcomes in depression patients treated with internet-based CBT (ICBT) (Andersson et al., 2019) and associations between PRSs for intelligence and depression with symptom status post-ICBT for depressed patients (Wallert et al., 2022). Additionally, PRS for educational attainment has been associated with symptom changes in patients with depression and anxiety treated with ICBT (Bäckman et al., 2024) as well as drop-out and remission status in CBT for dental anxiety (Wannemüller et al., 2021). On the other hand, the largest GWAS of CBT treatment outcome (n = 2,724) was underpowered to find associated SNPs (Rayner et al., 2019).

The present study investigated the association between PRS from nine psychiatric and cognitive phenotypes and OCD symptom change in a sample of well characterized 1,598 adults and children/adolescents with a DSM-5 diagnosis of OCD that underwent specialist CBT for OCD in Sweden and Norway. Psychiatric traits (including ADHD, autism, bipolar disorder, depression, OCD, schizophrenia, and cross-disorder PRS) were selected based on the hypothesis that genetic propensities for both specific and general psychopathology could influence an individual’s ability to benefit from CBT. Given that CBT requires substantial cognitive effort, including understanding its rationale as well as independently planning and completing homework assignments, IQ and educational attainment were also hypothesized to play a role in the process.

METHODS

Participants and intervention

Study participants were sampled from the Nordic OCD & Related Disorders Consortium (NORDiC), a genetic study of OCD in Norway and Sweden (Mataix-Cols et al., 2020). Of the 3,931 participants in NORDiC, 1,598 were included in the analysis (see Figure 1). For baseline characteristics, see Table 1.

Figure 1.

Figure 1.

Open in a new tab

Flowchart of study participants from the NORDiC study, stratified by country. Participants were excluded if not passing genetic quality control filters, and if not having complete pre- and post-measurements on symptom severity C/Y-BOCS (Yale-Brown Obsessive Compulsive Scale and (Children’s) Yale-Brown Obsessive Compulsive Scale (Goodman et al., 1989; Scahill et al., 1997). The full study sample included 1,598 participants.

Table 1.

Sociodemographic and clinical characteristics at baseline and post-treatment for complete cases

Norway (n = 765) Sweden (n = 833) Total (n=1,598)
Adults (n = 765) Children (n=431) Adults (n=402)
Sex, female 530 (69.3%) 259 (60.1%) 251 (62.4%) 1,039 (65.0%)
Age (SD) 31.5 (10.1) 13.3 (2.71) 32.3 (10.5) 26.8 (12.1)
On psychotropic medication† 146 (19.1%) 152 (35.3%) 233 (58.0%) 531 (33.2%)
 Missing 619 (80.9%) 32 (7.4%) 44 (10.9%) 695 (24.5%)
Psychiatric comorbidity 455 (59.5%) 204 (47.3%) 205 (51.0%) 864 (54.1%)
 Depression 295 (38.6%) 52 (12.1%) 91 (22.6%) 438 (27.4%)
 Anxiety 272 (35.6%) 48 (11.1%) 107 (26.6%) 427 (26.7%)
 Schizophrenia 5 (0.7%) 0 (0.0%) 0 (0.0%) 5 (0.3%)
 Bipolar disorder 5 (0.7%) 0 (0.0%) 7 (1.7%) 12 (0.8%)
 ADHD NA 76 (17.6%) 30 (7.5%) 106 (6.6%)
 Autism spectrum disorder NA 50 (11.6%) 37 (9.2%) 87 (5.4%)
 Other NA 36 (1.1%) 28 (0.7%) 64 (0.4%)
 Missingness 6 (0.8%) 0 (0.0%) 24 (0.6%) 24 (1.5%)
Baseline symptom severity level†
 Subclinical (0–13) 1 (0.1%) 8 (1.9%) 9 (2.2%) 18 (1.1%)
 Mild (14–21) 123 (16.1%) 143 (33.2%) 125 (31.1%) 391 (24.5%)
 Moderate (22–29) 521 (68.1%) 242 (56.1%) 210 (52.2%) 973 (60.9%)
 Severe (30–40) 120 (15.7%) 38 (8.8%) 58 (14.4%) 216 (13.5%)
C/Y-BOCS total score pre-treatment (SD) 25.5 (4.11) 23.1 (4.41) 23.9 (5.22) 24.4 (4.62)
C/Y-BOCS total score post-treatment (SD) 11.1 (5.49) 10.6 (6.45 14.1 (6.88) 11.7 (6.28)
Number of CBT sessions (SD) B4DTb 14.5 (7.15) 14.6 (7.42)
Treatment responders (>35% symptom decrease) 639 (83.5%) 334 (77.5%) 231 (57.5%) 1204 (75.3%)
Open in a new tab

Data are integer count (%) or decimal mean (SD). CBT=Cognitive Behavior Therapy, C/Y-BOCS=Yale-Brown Obsessive Compulsive Scale and (Children’s) Yale-Brown Obsessive Compulsive Scale (Goodman et al., 1989; Scahill et al., 1997). Subclinical (0–13), Mild (14–21), Moderate (22–29), Severe (30–40)(Cervin et al., 2022).

†

Assessed before treatment start

b

Bergen 4-day treatment (Hansen et al., 2018; Kvale et al., 2018).

Swedish participants (n = 833) were recruited between 2014–2020 (Mataix-Cols et al., 2020) from OCD specialist clinics (14 sites) and diagnosed by trained clinicians with extensive experience working with the patient group. All had a confirmed diagnosis of OCD and available symptom severity scores at both baseline and post-treatment. Approximately half of the sample were children/adolescents (51.7%). Most individuals were assessed to have mild to moderate OCD at baseline (86.5%) and received treatment in the form of either CBT alone or in combination with medication. All individuals received protocol-driven CBT, including exposure and response prevention (ERP), delivered in either individual, group, or ICBT formats, depending on the practices of the treating clinic.

Norwegian participants (n = 765) were all adults and were recruited between 2016 and 2019 at OCD specialist clinics across Norway (10 sites). Norwegian cases had similar diagnostic procedure, inclusion, and exclusion criteria as those in Sweden, except that all medicated participants in the Norwegian sample were required to be on a stable dosage before commencing CBT. Most individuals had mild to moderate OCD at baseline (84.2%), and all received treatment in the form of either CBT alone or in combination with medication. All Norwegian participants received the Bergen 4-day treatment (B4DT), a concentrated form of ERP (Hansen et al., 2018; Kvale et al., 2018).

Participants in both countries provided either blood or saliva samples, which were genotyped on the Illumina Global Screening Array (GSA) at LIFE&BRAIN in Bonn, Germany.

Controls

Controls were recruited in both countries to check for differences in the PRSs between cases and controls, in order to validate the PRSs in our study sample (see Supplementary methods for more details). Swedish controls were sampled from the LifeGene study, a large prospective population-based study in Sweden (Almqvist et al., 2011). The controls were unaffected by OCD and unrelated to any OCD case to the third degree. All samples were genotyped on the Illumina Global Screening Array.

Norwegian controls were sampled from the Trøndelag Health Study (HUNT), an extensive population-based health survey in Norway (Krokstad et al., 2013). Controls were defined as those without ICD-10 code F42 and ICD-9 code 300 (“anxiety, dissociative and somatoform disorders”). All were genotyped on DeCODE genetics arrays that represent modified versions of Illumina Global Screening arrays.

Quality Control (QC) and analysis of NORDiC genetic data

See the Supplementary methods for details on NORDiC genotype array data processing. Sample QC was performed on the merged genotype data (3,712 cases, 4,227 controls), utilizing metrics on missingness, sex, and relatedness calculated using PLINK v1.90b4.9r (Purcell et al., 2007). We retained samples with low genotype missingness (<0.02), no discordance between observed and reported sex, no cryptic relatedness, and high-confidence European ancestry. After this quality control procedure, we were left with a total of 2957 cases and 3787 controls. We put samples through the standard Ricopili pre-imputation quality control pipeline, but no additional samples beyond these were pruned from the analysis. No filter was set for imputation quality scores.

Selection of post-imputation genotypes for PRS calculation

The imputation protocol for the NORDiC case/control dataset was identical to (Strom et al., 2024), using the Haplotype Reference Consortium (HRC (McCarthy et al., 2016)) as reference panel. We selected a subset of post-imputation genotypes from the NORDiC case/control dataset (NORDiC-SWE: 1882 cases, 3125 controls. NORDiC-NOR: 1075 cases, 662 controls) that, based on the frequency and lack of technical issues, were appropriate for PRS calculation. Using PLINK, we subsetted on autosomal post-imputation genotypes (7,326,955 SNPs total) and kept the subset of variants that meet the following criteria: MAF > 0.05; Hardy-Weinberg equilibrium p-value > 1×10−6; no SNPs in the MHC region (hg19 chr6:28477797–33448354); no variants with missingness call rate > 0.1; no samples with missingness call rate > 0.1; all samples were retained. A total of 3,925,513 variants in NORDiC-SWE and 3,932,691 variants in NORDiC-NOR met these criteria and were usable for PRS calculation.

GWAS discovery datasets

We utilized a total of ten different GWAS discovery datasets for PRS calculation in our study. As a ‘negative control’ we used summary statistics from a GWAS of standing height (UK Biobank, 2018) under the assumption that it is uncorrelated with treatment response. We included two GWASs of psychiatry-relevant traits: educational attainment (Okbay et al., 2022) and IQ (Savage et al., 2018). Next, we included six datasets from GWAS of single psychiatric diagnoses: OCD (Strom et al., 2024); ADHD (Demontis et al., 2019); Autism spectrum disorder (Grove et al., 2019); Bipolar Disorder (Mullins et al., 2021); Depression (Howard et al., 2019); Schizophrenia (European subset, (Trubetskoy et al., 2022). Finally, we included a GWAS from the PGC Cross-disorder Working Group that included multiple psychiatric conditions, including all single diagnoses already listed (Cross-Disorder Group of the Psychiatric Genomics Consortium, 2019). This discovery dataset is the most well-powered for capturing contributions from pleiotropic common risk variants that confer generalized risk for psychiatric diagnoses. See Table S1 for more detailed information. Both cases and controls in this study were included in GWAS discovery datasets.

We chose traits from well-powered GWASs of psychiatric disorders that were available and potentially could affect treatment outcome, and cognitive traits were thought to be of importance since CBT involves active engagement, homework, and independent work from the patient. To validate the PRSs in our sample, we investigated the differences of PRSs in cases vs healthy controls, using linear mixed model.

PRS calculation

PRSs was calculated using PRS-CS (Ge et al., 2019; getian107, 2018). This PRS calculation approach uses known genotypes from separate individuals of the same ancestry (1000 genomes phase 3, European ancestry) to keep as many SNPs as possible in PRS calculation while still considering LD structure present in the reference population. Critically, it can define SNP effect sizes in computing PRS without manually defining p-value thresholds for variant exclusion. We utilized the author-provided 1000 genomes phase 3 European-ancestry genotypes to determine the LD backbone. PRSs were scaled (z-transformed, M = 0, SD =1).

Clinical measures

All clinical assessments were done as part of routine care by skilled clinicians at each of the OCD specialist clinics.

Main outcome

OCD symptom severity was assessed using the Yale-Brown Obsessive Compulsive Scale (Y-BOCS, (Goodman et al., 1989) or the corresponding version for children/adolescents, the Children’s Yale-Brown Obsessive Compulsive Scale (CY-BOCS, (Scahill et al., 1997). The scale consists of 10 items, assessing the time consumed by, extent of distress and interference caused by, and the control over and resistance of, obsessions and compulsions. The two measures will be referred to as C/Y-BOCS henceforth. The main outcome was percentage change in OCD symptom burden (C/Y-BOCS scores) from pre- to post-treatment.

C/Y-BOCSpost−C/Y-BOCSpreC/Y-BOCSpre

Only participants with a main outcome value were included in the main analysis, leaving a total of 1,598 individuals (765 Norwegian and 833 Swedish).

Association between PRSs and symptom severity change

Linear mixed models were used to estimate associations between each PRS and symptom severity change. The exposure were the scaled PRSs, and outcome was percentage change in OCD symptom burden (C/Y-BOCS scores) from pre- to post-treatment.

First, we fitted crude models with site (six Swedish, ten Norwegian) used as random slope and intercept. For the main analysis, we adjusted models for the covariates age, sex, genotyping batch, and the first five ancestry PCs as fixed variables. All inference tests were performed at a 95% confidence level. P-values for modeled coefficients were approximated, assuming a normal distribution. We used the same modelling procedure to do post hoc analysis of a sub sample of Swedish adults (n=402). All analyses were conducted in R (R Core Team, 2023), along with analysis-specific packages lme4 (Bates et al., 2015) and broom.mixed (Bolker et al., 2019).

Multiple imputation

As there was a large amount of missing data on the C/Y-BOCS (46.9%) in the original dataset, we tried to assess the robustness of our main analysis estimates by using imputation. We applied multiple imputation using only the Swedish dataset, which included additional variables that could be used as indicators for phenotype imputation. This included both demographic and clinical variables (in addition to variables from Table 1; Clinical Global Impression Scale-Improvement, and - Severity (CGI-I and CGI-S, (Busner & Targum, 2007)), Montgomery Åsberg Depression Rating Scale-Self rate (MADRS-S, (Svanborg & Asberg, 1994)), and Obsessional Compulsive Inventory-Revised (OCI-R, (Foa et al., 2002). Variables used in multiple imputation had missingness between 0–55%, and variables with the highest proportion missing were CGI-I post (55.1%), CGI-S post (52.5%), MADRS-S pre (48.3%), OCI-R pre (46.1%) and psychotropic medication post treatment (33.2%). Imputation of missing values with multiple variables, and multiple plausible values, provides a quantification of the uncertainty in estimating what the missing values might be, avoiding creating false precision (as can happen with single imputation). Multiple imputation was performed using the MICE package (van Buuren & Groothuis-Oudshoorn, 2011) in R version 4.3.2, which applied chained equations to create 50 imputed datasets.

The imputed dataset was not used in the main analysis. Instead, the same regression models as in the main analysis were then fitted to each dataset, and estimates were pooled using Rubin’s rules. Default settings were used unless otherwise specified. This was done in order to compare the model output in the imputed dataset and main analysis. See Supplementary methods for more details.

RESULTS

The primary analysis, adjusting for age, sex, genotyping batch, and the first five ancestry PCs, showed that only the schizophrenia PRS had a significant association (B=0.013, p=0.04) with symptom change (see Table 2 and Figure 2). This suggests that higher schizophrenia genetic risk was associated with a smaller decrease in symptom severity from pre- to post-treatment. In the crude models, none of the studied PRS were statistically associated with symptom severity change (See Supplementary results). The effect was also observed in the adults only, while none of the other PRS predicted changes in symptom values. Similarly, the association was only statistically significant in the Swedish sample but not the Norwegian sample (see Figure 3 and Supplementary results). The post hoc analysis of the subsample of Swedish adults showed a stronger association for schizophrenia PRS (B=0.038, p=0.01) than the main analysis. As anticipated, PRS for standing height showed no significant difference between cases and controls. In contrast, all other PRS, except for autism, were significantly higher in cases, while cases had lower PRS for IQ and educational attainment (see Supplementary results).

Table 2.

Associations of different PRS with C/Y-BOCS symptom change in OCD

Total sample (n=1,598) Adults (n=1,167) Children (n=431)
β CI (95%) p β CI (95%) p β CI (95%) p
Standing height .004 −.008, .017 .50 .008 −.006, .022 .30 −.003 −.027, .021 .81
ADHD 4×10−4 −.012, .013 .95 .002 −.013, .015 .82 −.007 −.032, .020 .62
Autism −9×10−4 −.014, .011 .88 −.007 −.021, .007 .35 .016 −.010, .042 .23
Bipolar disorder .004 −.008, .017 .50 .001 −.015, .014 .90 .017 −.008, .042 .17
Cross disorder PRS .006 −.007, .019 .36 .002 −.013, .016 .79 .016 −.010, .042 .24
Educational att. .003 −.010, .015 .68 .004 −.008, .020 .49 −.004 −.031, .021 .75
IQ .003 −.009, .016 .62 .009 −.004, .024 .18 −.012 −.037, .013 .36
Depression −.008 −.021, .004 .19 −.010 −.025, .003 .13 .002 −.024, .028 .90
OCD .002 −.011, .014 .79 .003 −.011, .017 .65 −.003 −.028, .023 .84
Schizophrenia .013* 6×10−4, .026 .04 .016* .002, .030 .03 .005 −.020, .030 .69
Open in a new tab

Estimates per each PRS for the total sample, adults, and children, respectively. Models were adjusted for age, sex, genotyping batch, and the first five PCs. CI=confidence interval,

*

significance level p<.05.

PRS=Polygenic Risk Score, C/Y-BOCS=Combination of Yale-Brown Obsessive Compulsive Scale and (Children’s) Yale-Brown Obsessive Compulsive Scale (Goodman et al., 1989; Scahill et al., 1997), OCD=Obsessive Compulsive Disorder, ADHD=Attention-Deficit Hyperactivity Disorder

Figure 2.

Figure 2.

Open in a new tab

The full sample was stratified using a quantile-split approach for the Schizophrenia PRS. Participants from the 1st (n=399) and 4th (n=400) quantile are compared in the figure, where the 1st quantile had the lowest PRS value, and the fourth quantile had the highest PRS value.

Figure 3.

Figure 3.

Open in a new tab

Association of each PRS with symptom change on C/Y-BOCS stratified by country. Estimates (β) on the x-axis with 95% confidence intervals marked by line. N=1,598. Models were adjusted for age, sex, genotyping batch, and the first five PCs. PRS=Polygenic Risk Score. C/Y-BOCS=Combination of Yale-Brown Obsessive Compulsive Scale and (Children’s) Yale-Brown Obsessive Compulsive Scale (Goodman et al., 1989; Scahill et al., 1997), OCD = Obsessive Compulsive Disorder, ADHD = Attention-Deficit Hyperactivity Disorder.

Multiple imputation

The imputations algorithm showed adequate convergence when assessed by plotting the mean and SD of each variable across the iterations. The distributions of imputed values were similar to the observed values for all imputed variables. For further details, see Supplementary methods.

DISCUSSION

We investigated the relationship between PRSs for nine psychiatric and cognitive phenotypes and symptom severity change from before to after completed CBT in 1,598 adults and children/adolescents with OCD sampled from Sweden and Norway. A higher PRS for schizophrenia was associated with a smaller symptom change after CBT for OCD, while no other PRS was significantly related to symptom change. Significant associations for the schizophrenia PRS were found in adjusted models, in the Swedish and adult stratified analysis and in a subsample of Swedish adults. A possible explanation could be that a higher schizophrenia PRS might be associated with reduced insight in people with OCD, leading to a stronger belief in the content of obsessions, which would in turn reduce chances of benefiting from treatment. However, this hypothesis will require empirical testing in future studies .

A somewhat unexpected finding was the lack of association of other psychiatric PRS, particularly OCD, with symptom change during OCD treatment. Similarly, we were unable to replicate previous findings in participants with major depression and anxiety disorders, such as associations between symptom change and PRSs for autism (Andersson et al., 2019) and educational attainment (Bäckman et al., 2024) possibly because the treated diagnoses were different. Our sample size was small for a genetic study of a complex trait like symptom change, which might explain our predominantly null findings .One interpretation of our negative findings is that if there are genuine associations between these PRSs and symptom change during psychotherapy, they are very small. In addition, the genetic propensity for a trait such as OCD is not necessarily the same - or even similar - to the genetic propensity associated with symptom change in CBT.

A strength of this study is the high quality of data. All participants were sourced from specialized OCD clinics and structured diagnostic procedures and symptom ratings were done by expert clinicians, and CBT interventions were following gold-standard treatment protocols. Another strength is the PRS comparison of cases vs controls, which supports the validity of the PRS in our sample. The sample was heterogeneous, including two countries, adults and children in the same analysis which could be seen as both a limitation and a strength, since it increases ecological validity but might decrease statistical power. Limitations include a relatively small sample size, and high missingness in the original data, including the outcome variable. This could introduce attrition bias, as participants who did not complete the post-treatment measurements may have dropped out for various reasons. These reasons could include being too unwell to attend, feeling fully recovered and no longer perceiving a need for treatment, or other unknown factors. Such dropout patterns could systematically skew the results, affecting the validity of the findings. However, the sensitivity analysis after multiple imputation corroborated our main findings. Another limitation was the differences in the GWASs used to calculate PRSs. The GWASs differed in number of cases and controls (see Supplementary methods, Table S1 for more information), which will affect the possibility to find significant results. The Schizophrenia GWAS is well-powered, which might have increased the chance of a significant results for the schizophrenia PRS. An additional limitation was the proportion of treatment responders (75.3%) in this sample was higher than in earlier studies, which could have had effect on both the results and the generalizability of the results. This should be interpreted with caution, since the response definition in this paper was based on one measures (C/Y-BOCS, (Goodman et al., 1989; Scahill et al., 1997), instead of C/Y-BOCS+CGI-I (Busner & Targum, 2007), as suggested by (Mataix-Cols et al., 2016). In addition, although the Norwegian and Swedish treatments were based on the same theory and principals, the treatment setup differed and was much more standardized in the Norwegian subsample.

There are several limits with the use of PRSs to disentangle the links between genetic predisposition and symptom severity change. The utility of PRSs in clinical psychiatry should be understood within the context of realistic expectations regarding its capabilities and limitations. Single PRSs alone will likely not be used to guide future clinical decision-making but could be helpful when combined in a multi-PRS approach (Albiñana et al., 2023; Krapohl et al., 2018) and in combination with other types of data (Boberg et al., 2023; Wallert et al., 2022). For example, PRS variables could be integrated with clinical and healthcare register-based data for prediction of individuals at elevated risk for suboptimal treatment outcomes (Ikeda et al., 2021) and may subsequently guide clinicians in further tailoring of patient treatment, management, and follow-up.

To conclude, current PRS for psychiatric and cognitive phenotypes do not seem to contribute meaningfully to the association to OCD symptom severity change in CBT. In addition, an adequately powered GWAS of symptom change per se would be needed to calculate a potentially more meaningful PRS of symptom chnage to be used in future studies For PRS to be of clinical use, more research is needed.

Supplementary Material

Supinfo

Acknowledgements

We are grateful for the contributions of our study participants and all involved clinicians in Sweden and Norway.

Funding

This study was supported by NIH (R01MH110427 JC); the Swedish Research Council (2015-02271 DMC; 2020-01343 DMC; 2018-02487 CR), Centre for Innovative Medicine (CIMED, 96328 JW; 1003477 JW; 954440 CR), The Söderström König Foundation (SLS-941192 JW; SLS-994792 JW), Trond Mohn Foundation (BFS2019TMT01 BH), and Clinical therapy research in the specialist health services (2017204 CH). The computation was performed on resources provided by SNIC through the Uppsala Multidisciplinary Center for Advanced Computational Science (UPPMAX) under project sens2018605.

CONSORTIA

Members of the Nordic OCD & Related Disorders Consortium (NORDiC):

Julia Bäckman, Long Long Chen, James J. Crowley, Elles de Schipper, Diana Pascal, Jan Haavik, Kristen Hagen, Matthew W. Halvorsen, Bjarne Hansen, Kira D. Höffler, Fredrik Johansson, Anna K. Kähler, Elinor K. Karlsson, Gerd Kvale, Paul Lichtenstein, Kerstin Lindblad-Toh, Manuel Mattheisen, David Mataix-Cols, Kathleen Morrill, Christian Rück, Thorstein Olsen Eide, Nora I. Strom, John Wallert.

Footnotes

Conflict of interest

Prof Mataix-Cols receives royalties from UpToDate, Inc., and is part owner of Scandinavian E-Health AB, all outside the current work.

Ethical approval

This study was approved by the Norwegian Regional Committee for Medical and Health Research Ethics (REK: 2018/52) and by the Regional Ethics Board in Stockholm, Sweden (REPN: 2014/1897–31).

Patient consent

All participants over 18 years provided a written informed consent. For participants under 18 years, written informed consent was obtained from a parent/legal guardian with an assent from the participant.

Data availability

Patient data are deemed sensitive personal information and therefore the dissemination of individual level data is prohibited by Swedish and Norwegian law. Free third-party access is, therefore, not possible. Given obtained ethical approval, data may be used in further analyses.

REFERENCES

  1. Albiñana C, Zhu Z, Schork AJ, Ingason A, Aschard H, Brikell I, Bulik CM, Petersen LV, Agerbo E, Grove J, Nordentoft M, Hougaard DM, Werge T, Børglum AD, Mortensen PB, McGrath JJ, Neale BM, Privé F & Vilhjálmsson BJ (2023). Multi-PGS enhances polygenic prediction by combining 937 polygenic scores. Nature Communications, 14(1), 4702. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Almqvist C, Adami H-O, Franks PW, Groop L, Ingelsson E, Kere J, Lissner L, Litton J-E, Maeurer M, Michaëlsson K, Palmgren J, Pershagen G, Ploner A, Sullivan PF, Tybring G & Pedersen NL (2011). LifeGene--a large prospective population-based study of global relevance. European Journal of Epidemiology, 26(1), 67–77. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Andersson E, Crowley JJ, Lindefors N, Ljótsson B, Hedman-Lagerlöf E, Boberg J, El Alaoui S, Karlsson R, Lu Y, Mattheisen M, Kähler AK, Svanborg C, Mataix-Cols D, Mattsson S, Forsell E, Kaldo V, Schalling M, Lavebratt C, Sullivan PF & Rück C (2019). Genetics of response to cognitive behavior therapy in adults with major depression: a preliminary report. Molecular Psychiatry, 24(4), 484–490. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Bäckman J, Wallert J, Halvorsen M, Crowley J, Mataix-Cols D & Rück C (2024). Educational attainment polygenic risk score and symptom severity change after Internet-delivered cognitive behaviour therapy for depression and anxiety. In Research Square. 10.21203/rs.3.rs-4246791/v1 [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Bates D, Mächler M, Bolker B & Walker S (2015). Fitting Linear Mixed-Effects Models Using lme4. Journal of Statistical Software, 67, 1–48. [Google Scholar]
  6. Blanco-Vieira T, Radua J, Marcelino L, Bloch M, Mataix-Cols D & do Rosário MC (2023). The genetic epidemiology of obsessive-compulsive disorder: a systematic review and meta-analysis. Translational Psychiatry, 13(1), 230. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Boberg J, Kaldo V, Mataix-Cols D, Crowley JJ, Roelstraete B, Halvorsen M, Forsell E, Isacsson NH, Sullivan PF, Svanborg C, Andersson EH, Lindefors N, Kravchenko O, Mattheisen M, Danielsdottir HB, Ivanova E, Boman M, Fernández de la Cruz L, Wallert J & Rück C(2023). Swedish multimodal cohort of patients with anxiety or depression treated with internet-delivered psychotherapy (MULTI-PSYCH). BMJ Open, 13(10), e069427. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Bolker B, Robinson D, Menne D, Gabry J & Buerkner P (2019). Package “broom. mixed.” https://cran.ms.unimelb.edu.au/web/packages/broom.mixed/broom.mixed.pdf
  9. Busner J & Targum SD (2007). The clinical global impressions scale: applying a research tool in clinical practice. Psychiatry , 4(7), 28–37. [PMC free article] [PubMed] [Google Scholar]
  10. Cervin M, OCD Severity Benchmark Consortium & Mataix-Cols D (2022). Empirical severity benchmarks for obsessive-compulsive disorder across the lifespan. World Psychiatry: Official Journal of the World Psychiatric Association (WPA), 21(2), 315–316. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Cross-Disorder Group of the Psychiatric Genomics Consortium. (2019). Genomic Relationships, Novel Loci, and Pleiotropic Mechanisms across Eight Psychiatric Disorders. Cell, 179(7), 1469–1482.e11. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. De La Cruz F, Rydell L, Runeson M, Onofrio B, Brander BM, Rück G, Lichtenstein C, Larsson P & Mataix-Cols H (2016). Suicide in obsessive-compulsive disorder: a population-based study of 36 788 Swedish patients. Molecular Psychiatry, 22(11), 1626–1632. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Demontis D, Walters RK, Martin J, Mattheisen M, Als TD, Agerbo E, Baldursson G, Belliveau R, Bybjerg-Grauholm J, Bækvad-Hansen M, Cerrato F, Chambert K, Churchhouse C, Dumont A, Eriksson N, Gandal M, Goldstein JI, Grasby KL, Grove J, … Neale BM (2019). Discovery of the first genome-wide significant risk loci for attention deficit/hyperactivity disorder. Nature Genetics, 51(1), 63–75. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Fawcett EJ, Power H & Fawcett JM (2020). Women Are at Greater Risk of OCD Than Men: A Meta-Analytic Review of OCD Prevalence Worldwide. The Journal of Clinical Psychiatry, 81(4). 10.4088/JCP.19r13085 [DOI] [PubMed] [Google Scholar]
  15. Foa EB, Huppert JD, Leiberg S, Langner R, Kichic R, Hajcak G & Salkovskis PM (2002). The Obsessive-Compulsive Inventory: development and validation of a short version. Psychological Assessment, 14(4), 485–496. [PubMed] [Google Scholar]
  16. Ge T, Chen C-Y, Ni Y, Feng Y-CA & Smoller JW (2019). Polygenic prediction via Bayesian regression and continuous shrinkage priors. Nature Communications, 10(1), 1776. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. getian107. (2018). PRSsc. Git Commit 071680a2aa3354a03a67d95dede7560a9899c5c6. https://github.com/getian107/PRScs
  18. Goodman WK, Price LH, Rasmussen SA, Mazure C, Delgado P, Heninger GR & Charney DS (1989). The Yale-Brown Obsessive Compulsive Scale. II. Validity. Archives of General Psychiatry, 46(11), 1012–1016. [DOI] [PubMed] [Google Scholar]
  19. Grove J, Ripke S, Als TD, Mattheisen M, Walters RK, Won H, Pallesen J, Agerbo E, Andreassen OA, Anney R, Awashti S, Belliveau R, Bettella F, Buxbaum JD, Bybjerg-Grauholm J, Bækvad-Hansen M, Cerrato F, Chambert K, Christensen JH, … Børglum AD (2019). Identification of common genetic risk variants for autism spectrum disorder. Nature Genetics, 51(3), 431–444. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Hansen B, Hagen K, Öst L-G, Solem S & Kvale G (2018). The Bergen 4-Day OCD Treatment Delivered in a Group Setting: 12-Month Follow-Up. Frontiers in Psychology, 9, 639. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Howard DM, Adams MJ, Clarke T-K, Hafferty JD, Gibson J, Shirali M, Coleman JRI, Hagenaars SP, Ward J, Wigmore EM, Alloza C, Shen X, Barbu MC, Xu EY, Whalley HC, Marioni RE, Porteous DJ, Davies G, Deary IJ, … McIntosh AM (2019). Genome-wide meta-analysis of depression identifies 102 independent variants and highlights the importance of the prefrontal brain regions. Nature Neuroscience, 22(3), 343–352. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Ikeda M, Saito T, Kanazawa T & Iwata N (2021). Polygenic risk score as clinical utility in psychiatry: a clinical viewpoint. Journal of Human Genetics, 66(1), 53–60. [DOI] [PubMed] [Google Scholar]
  23. Keeley ML, Storch EA, Merlo LJ & Geffken GR (2008). Clinical predictors of response to cognitive-behavioral therapy for obsessive-compulsive disorder. Clinical Psychology Review, 28(1), 118–130. [DOI] [PubMed] [Google Scholar]
  24. Khera AV, Chaffin M, Aragam KG, Haas ME, Roselli C, Choi SH, Natarajan P, Lander ES, Lubitz SA, Ellinor PT & Kathiresan S (2018). Genome-wide polygenic scores for common diseases identify individuals with risk equivalent to monogenic mutations. Nature Genetics, 50(9), 1219–1224. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Knopp J, Knowles S, Bee P, Lovell K & Bower P (2013). A systematic review of predictors and moderators of response to psychological therapies in OCD: do we have enough empirical evidence to target treatment? Clinical Psychology Review, 33(8), 1067–1081. [DOI] [PubMed] [Google Scholar]
  26. Krapohl E, Patel H, Newhouse S, Curtis CJ, von Stumm S, Dale PS, Zabaneh D, Breen G, O’Reilly PF & Plomin R (2018). Multi-polygenic score approach to trait prediction. Molecular Psychiatry, 23(5), 1368–1374. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Krokstad S, Langhammer A, Hveem K, Holmen TL, Midthjell K, Stene TR, Bratberg G, Heggland J & Holmen J (2013). Cohort profile: The HUNT study, Norway. International Journal of Epidemiology, 42(4), 968–977. [DOI] [PubMed] [Google Scholar]
  28. Kvale G, Hansen B, Björgvinsson T, Børtveit T, Hagen K, Haseth S, Kristensen UB, Launes G, Ressler KJ, Solem S, Strand A, van den Heuvel OA & Öst L-G (2018). Successfully treating 90 patients with obsessive compulsive disorder in eight days: the Bergen 4-day treatment. BMC Psychiatry, 18(1), 323. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Lester KJ & Eley TC (2013). Therapygenetics: Using genetic markers to predict response to psychological treatment for mood and anxiety disorders. Biology of Mood & Anxiety Disorders, 3(1), 4. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Mataix-Cols D, Andersson E, Aspvall K, Boberg J, Crowley JJ, de Schipper E, Fernández de la Cruz L, Flygare O, Ivanova E, Lenhard F, Lundström L, Rück C, Serlachius E & Cervin M (2022). Operational definitions of treatment response and remission in obsessive-compulsive disorder capture meaningful improvements in everyday life. Psychotherapy and Psychosomatics, 91(6), 424–430. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Mataix-Cols D, Fernández de la Cruz L, Beucke JC, De Schipper E, Kuja-Halkola R, Lichtenstein P & Pol-Fuster J (2024). Heritability of clinically diagnosed obsessive-compulsive disorder among twins. JAMA Psychiatry (Chicago, Ill.), 81(6), 631–632. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Mataix-Cols D, Fernández de la Cruz L, Nordsletten AE, Lenhard F, Isomura K & Simpson HB (2016). Towards an international expert consensus for defining treatment response, remission, recovery and relapse in obsessive-compulsive disorder. World Psychiatry: Official Journal of the World Psychiatric Association (WPA), 15(1), 80–81. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Mataix-Cols D, Hansen B, Mattheisen M, Karlsson EK, Addington AM, Boberg J, Djurfeldt DR, Halvorsen M, Lichtenstein P, Solem S, Lindblad-Toh K, Nordic OCD and Related Disorders Consortium (NORDiC), Haavik J, Kvale G, Rück C & Crowley JJ (2020). Nordic OCD & Related Disorders Consortium: Rationale, design, and methods. American Journal of Medical Genetics. Part B, Neuropsychiatric Genetics: The Official Publication of the International Society of Psychiatric Genetics, 183(1), 38–50. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. McCarthy S, Das S, Kretzschmar W, Delaneau O, Wood AR, Teumer A, Kang HM, Fuchsberger C, Danecek P, Sharp K, Luo Y, Sidore C, Kwong A, Timpson N, Koskinen S, Vrieze S, Scott LJ, Zhang H, Mahajan A, … Haplotype Reference Consortium. (2016). A reference panel of 64,976 haplotypes for genotype imputation. Nature Genetics, 48(10), 1279–1283. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. McDonald S, Melkonian M, Karin E, Dear BF, Titov N & Wootton BM (2023). Predictors of response to cognitive behavioural therapy (CBT) for individuals with obsessive-compulsive disorder (OCD): a systematic review. Behavioural and Cognitive Psychotherapy, 51(4), 302–319. [DOI] [PubMed] [Google Scholar]
  36. Meier SM, Mattheisen M, Mors O, Schendel DE, Mortensen PB & Plessen KJ (2016). Mortality Among Persons With Obsessive-Compulsive Disorder in Denmark. JAMA Psychiatry , 73(3), 268–274. [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Mullins N, Forstner AJ, O’Connell KS, Coombes B, Coleman JRI, Qiao Z, Als TD, Bigdeli TB, Børte S, Bryois J, Charney AW, Drange OK, Gandal MJ, Hagenaars SP, Ikeda M, Kamitaki N, Kim M, Krebs K, Panagiotaropoulou G, … Andreassen OA (2021). Genome-wide association study of more than 40,000 bipolar disorder cases provides new insights into the underlying biology. Nature Genetics, 53(6), 817–829. [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. National Collaborating Centre for Mental Health (UK). (2006). Obsessive-Compulsive Disorder: Core Interventions in the Treatment of Obsessive-Compulsive Disorder and Body Dysmorphic Disorder. British Psychological Society (UK). [PubMed] [Google Scholar]
  39. Okbay A, Wu Y, Wang N, Jayashankar H, Bennett M, Nehzati SM, Sidorenko J, Kweon H, Goldman G, Gjorgjieva T, Jiang Y, Hicks B, Tian C, Hinds DA, Ahlskog R, Magnusson PKE, Oskarsson S, Hayward C, Campbell A, … Young AI (2022). Polygenic prediction of educational attainment within and between families from genome-wide association analyses in 3 million individuals. Nature Genetics, 54(4), 437–449. [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Olatunji BO, Davis ML, Powers MB & Smits JAJ (2013). Cognitive-behavioral therapy for obsessive-compulsive disorder: a meta-analysis of treatment outcome and moderators. Journal of Psychiatric Research, 47(1), 33–41. [DOI] [PubMed] [Google Scholar]
  41. Öst L-G, Havnen A, Hansen B & Kvale G (2015). Cognitive behavioral treatments of obsessive-compulsive disorder. A systematic review and meta-analysis of studies published 1993–2014. Clinical Psychology Review, 40, 156–169. [DOI] [PubMed] [Google Scholar]
  42. Purcell S, Neale B, Todd-Brown K, Thomas L, Ferreira MAR, Bender D, Maller J, Sklar P, de Bakker PIW, Daly MJ & Sham PC (2007). PLINK: a tool set for whole-genome association and population-based linkage analyses. American Journal of Human Genetics, 81(3), 559–575. [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. R Core Team. (2023). R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing, Vienna, Austria. https://www.R-project.org/ [Google Scholar]
  44. Rayner C, Coleman JRI, Purves KL, Hodsoll J, Goldsmith K, Alpers GW, Andersson E, Arolt V, Boberg J, Bögels S, Creswell C, Cooper P, Curtis C, Deckert J, Domschke K, El Alaoui S, Fehm L, Fydrich T, Gerlach AL, … Eley TC (2019). A genome-wide association meta-analysis of prognostic outcomes following cognitive behavioural therapy in individuals with anxiety and depressive disorders. Translational Psychiatry, 9(1), 150. [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Rosa-Alcázar AI, Sánchez-Meca J, Gómez-Conesa A & Marín-Martínez F (2008). Psychological treatment of obsessive-compulsive disorder: a meta-analysis. Clinical Psychology Review, 28(8), 1310–1325. [DOI] [PubMed] [Google Scholar]
  46. Ruscio AM, Stein DJ, Chiu WT & Kessler RC (2010). The epidemiology of obsessive-compulsive disorder in the National Comorbidity Survey Replication. Molecular Psychiatry, 15(1), 53–63. [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. Savage JE, Jansen PR, Stringer S, Watanabe K, Bryois J, de Leeuw CA, Nagel M, Awasthi S, Barr PB, Coleman JRI, Grasby KL, Hammerschlag AR, Kaminski JA, Karlsson R, Krapohl E, Lam M, Nygaard M, Reynolds CA, Trampush JW, … Posthuma D (2018). Genome-wide association meta-analysis in 269,867 individuals identifies new genetic and functional links to intelligence. Nature Genetics, 50(7), 912–919. [DOI] [PMC free article] [PubMed] [Google Scholar]
  48. Scahill L, Riddle MA, McSwiggin-Hardin M, Ort SI, King RA, Goodman WK, Cicchetti D & Leckman JF (1997). Children’s Yale-Brown Obsessive Compulsive Scale: reliability and validity. Journal of the American Academy of Child and Adolescent Psychiatry, 36(6), 844–852. [DOI] [PubMed] [Google Scholar]
  49. Strom NI, Gerring ZF, Galimberti M, Yu D, Halvorsen MW, Abdellaoui A, Rodriguez-Fontenla C, Sealock JM, Bigdeli T, Coleman JR, Mahjani B, Thorp JG, Bey K, Burton CL, Luykx JJ, Zai G, Alemany S, Andre C, Askland KD, … Mattheisen M (2024). Genome-wide association study identifies 30 obsessive-compulsive disorder associated loci. In medRxiv (p. 2024.03.13.24304161). 10.1101/2024.03.13.24304161 [DOI] [Google Scholar]
  50. Svanborg P & Asberg M (1994). A new self-rating scale for depression and anxiety states based on the Comprehensive Psychopathological Rating Scale. Acta Psychiatrica Scandinavica, 89(1), 21–28. [DOI] [PubMed] [Google Scholar]
  51. Trubetskoy V, Pardiñas AF, Qi T, Panagiotaropoulou G, Awasthi S, Bigdeli TB, Bryois J, Chen C-Y, Dennison CA, Hall LS, Lam M, Watanabe K, Frei O, Ge T, Harwood JC, Koopmans F, Magnusson S, Richards AL, Sidorenko J, … Schizophrenia Working Group of the Psychiatric Genomics Consortium. (2022). Mapping genomic loci implicates genes and synaptic biology in schizophrenia. Nature, 604(7906), 502–508. [DOI] [PMC free article] [PubMed] [Google Scholar]
  52. UK biobank. (2018, 8. January). http://www.nealelab.is/uk-biobank/
  53. van Buuren S & Groothuis-Oudshoorn K (2011). mice: Multivariate Imputation by Chained Equations in R. Journal of Statistical Software, 45, 1–67. [Google Scholar]
  54. Wallert J, Boberg J, Kaldo V, Mataix-Cols D, Flygare O, Crowley JJ, Halvorsen M, Ben Abdesslem F, Boman M, Andersson E, Hentati Isacsson N, Ivanova E & Rück C (2022). Predicting remission after internet-delivered psychotherapy in patients with depression using machine learning and multi-modal data. Translational Psychiatry, 1(12), 357. [DOI] [PMC free article] [PubMed] [Google Scholar]
  55. Wannemüller A, Kumsta R, Jöhren H-P, Eley TC, Teismann T, Moser D, Rayner C, Breen G, Coleman J, Schaumburg S, Blackwell SE & Margraf J (2021). Genes in treatment: Polygenic risk scores for different psychopathologies, neuroticism, educational attainment and IQ and the outcome of two different exposure-based fear treatments. The World Journal of Biological Psychiatry: The Official Journal of the World Federation of Societies of Biological Psychiatry, 22(9), 699–712. [DOI] [PubMed] [Google Scholar]
  56. World Health Organization. (2008). The global burden of disease : 2004 update. World Health Organization. [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supinfo
NIHMS2063255-supplement-Supinfo.docx (928.3KB, docx)

Data Availability Statement

Patient data are deemed sensitive personal information and therefore the dissemination of individual level data is prohibited by Swedish and Norwegian law. Free third-party access is, therefore, not possible. Given obtained ethical approval, data may be used in further analyses.

RESOURCES