Psychosocial rehabilitation (PSR) is at the core of psychiatric recovery. There is a paucity of evidence regarding how the needs and characteristics of patients guide clinical decisions to refer to PSR interventions. Here, we used explainable machine learning methods to determine how socio-demographic and clinical characteristics contribute to initial referrals to PSR interventions in patients with serious mental illness.

Data were extracted from the French network of rehabilitation centres, REHABase, collected between years 2016 and 2022 and analysed between February and September 2022. Participants presented with serious mental illnesses, including schizophrenia spectrum disorders, bipolar disorders, autism spectrum disorders, depressive disorders, anxiety disorders and personality disorders. Information from 37 socio-demographic and clinical variables was extracted at baseline and used as potential predictors. Several machine learning models were tested to predict initial referrals to four PSR interventions: cognitive behavioural therapy (CBT), cognitive remediation (CR), psychoeducation (PE) and vocational training (VT). Explanatory power of predictors was determined using the artificial intelligence-based SHAP (SHapley Additive exPlanations) method from the best performing algorithm.

Data from a total of 1146 patients were included (mean age, 33.2 years [range, 16–72 years]; 366 [39.2%] women). A random forest algorithm demonstrated the best predictive performance, with a moderate or average predictive accuracy [micro-averaged area under the receiver operating curve from ‘external’ cross-validation: 0.672]. SHAP dependence plots demonstrated insightful associations between socio-demographic and clinical predictors and referrals to PSR programmes. For instance, patients with psychotic disorders were more likely to be referred to PE and CR, while those with non-psychotic disorders were more likely to be referred to CBT and VT. Likewise, patients with social dysfunctions and lack of educational attainment were more likely to be referred to CR and VT, while those with better functioning and education were more likely to be referred to CBT and PE.

A combination of socio-demographic and clinical features was not sufficient to accurately predict initial referrals to four PSR programmes among a French network of rehabilitation centres. Referrals to PSR interventions may also involve service- and clinician-level factors. Considering socio-demographic and clinical predictors revealed disparities in referrals with respect to diagnoses, current clinical and psychological issues, functioning and education.

Children and adolescents with a history of adverse childhood experiences (ACEs) are more likely than their peers to develop mental health difficulties, but not enough is known about their help-seeking behaviours and preferences. We aimed to determine whether ACEs are associated with access to and perceived unmet need for mental health services and support amongst secondary school students.

We used multi-level logistic regression with data from the 2020 OxWell Student Survey to assess whether ACEs were associated with (1) prior access to mental health support and (2) perceived unmet need for mental health services in a community sample of English secondary school students. We assessed ACEs as a cumulative score from the Center for Youth Wellness Adverse Childhood Experiences Questionnaire: Teen Self-Report version and accounted for current mental health difficulties as measured by the 25-item Revised Children’s Anxiety and Depression Scale (RCADS).

Our analysis included 2018 students across 64 schools, of whom 29.9% (598/2002) reported prior access to mental health support. Of those not reporting prior access, 34.1% (469/1377) reported a perceived unmet need for services. In the unadjusted models, cumulative ACE scores were significantly positively associated with both prior access to mental health support (odds ratio (OR) = 1.36; 95% confidence interval (CI): 1.29–1.43) and perceived unmet need for mental health services (OR = 1.47; 95% CI: 1.37–1.59), meaning that students who had experienced adversity had a greater chance of having previously accessed support as well as perceiving an unmet need for services. After adjusting for mental health difficulties and other sociodemographic variables, cumulative ACE scores were positively associated with prior access (adjusted OR (aOR) = 1.25; 95% CI: 1.17–1.34 with a significant interaction between RCADS and ACE scores, aOR = 0.88; 95% CI: 0.84–0.93) as well as perceived unmet need (aOR = 1.32; 95% CI: 1.21–1.43 with a significant interaction between RCADS and ACE scores, aOR = 0.85; 95% CI: 0.78–0.91).

Although it is encouraging that adolescents with experience of adversity are more likely than their peers with similar levels of depression and anxiety symptoms to have accessed mental health support, there remains a concern that those who have not accessed support are more likely to perceive an as-yet unmet need for it. Mental health support must be available, accessible and acceptable to all who need it, especially for those groups that traditionally have not accessed services, including the more marginalised and vulnerable populations.