Health-Economic Evaluation of Psychological Interventions for Anxiety Prevention: A Systematic Review.

OBJECTIVE: Although evidence supports the effectiveness of psychological interventions for prevention of anxiety, little is known about their cost-effectiveness. The aim of this study was to conduct a systematic review of health-economic evaluations of psychological interventions for anxiety prevention. METHODS: PubMed, PsycInfo, Web of Science, Embase, Cochrane Central Register of Controlled Trials, EconLit, National Health Service (NHS) Economic Evaluations Database, NHS Health Technology Assessment, and OpenGrey databases were searched electronically on December 23, 2022. Included studies focused on economic evaluations based on randomized controlled trials of psychological interventions to prevent anxiety. Study data were extracted, and the quality of the selected studies was assessed by using the Consensus on Health Economic Criteria and the Cochrane risk-of-bias tool. RESULTS: All included studies (N=5) had economic evaluations that were considered to be of good quality. In two studies, the interventions showed favorable cost-effectiveness compared with usual care groups. In one study, the intervention was not cost-effective. Findings from another study cast doubt on the cost-effectiveness of the intervention, and the cost-effectiveness of the intervention in the remaining study could not be established. CONCLUSIONS: Although the findings suggest some preliminary evidence of cost-effectiveness of psychological interventions for preventing anxiety, they were limited by the small number of included studies. Additional research on the cost-effectiveness of psychological interventions for anxiety in different countries and populations is required.

Predicting Undesired Treatment Outcomes With Machine Learning in Mental Health Care: Multisite Study.

Background: Predicting which treatment will work for which patient in mental health care remains a challenge. Objective: The aim of this multisite study was 2-fold: (1) to predict patients' response to treatment in Dutch basic mental health care using commonly available data from routine care and (2) to compare the performance of these machine learning models across three different mental health care organizations in the Netherlands by using clinically interpretable models. Methods: Using anonymized data sets from three different mental health care organizations in the Netherlands (n=6452), we applied a least absolute shrinkage and selection operator regression 3 times to predict the treatment outcome. The algorithms were internally validated with cross-validation within each site and externally validated on the data from the other sites. Results: The performance of the algorithms, measured by the area under the curve of the internal validations as well as the corresponding external validations, ranged from 0.77 to 0.80. Conclusions: Machine learning models provide a robust and generalizable approach in automated risk signaling technology to identify cases at risk of poor treatment outcomes. The results of this study hold substantial implications for clinical practice by demonstrating that the performance of a model derived from one site is similar when applied to another site (ie, good external validation).

A roadmap for applying machine learning when working with privacy-sensitive data: predicting non-response to treatment for eating disorders.

OBJECTIVES: Applying machine-learning methodology to clinical data could present a promising avenue for predicting outcomes in patients receiving treatment for psychiatric disorders. However, preserving privacy when working with patient data remains a critical concern. METHODS: In showcasing how machine-learning can be used to build a clinically relevant prediction model on clinical data, we apply two commonly used machine-learning algorithms (Random Forest and least absolute shrinkage and selection operator) to routine outcome monitoring data collected from 593 patients with eating disorders to predict absence of reliable improvement 12 months after entering outpatient treatment. RESULTS: An RF model trained on data collected at baseline and after three months made 31.3% fewer errors in predicting lack of reliable improvement at 12 months, in comparison with chance. Adding data from a six-month follow-up resulted in only marginal improvements to accuracy. CONCLUSION: We were able to build and validate a model that could aid clinicians and researchers in more accurately predicting treatment response in patients with EDs. We also demonstrated how this could be done without compromising privacy. ML presents a promising approach to developing accurate prediction models for psychiatric disorders such as ED.

Closing the gap between screening and depression prevention: a qualitative study on barriers and facilitators from the perspective of public health professionals in a school-based prevention approach.

BACKGROUND: The prevalence of depression has increased among adolescents in western countries. Prevention is needed to reduce the number of adolescents who experience depression and to avoid negative consequences, including suicide. Several preventive interventions are found to be promising, especially multi-modal approaches, for example combining screening and preventive intervention. However, an important bottleneck arises during the implementation of preventive intervention. Only a small percentage of adolescents who are eligible for participation actually participate in the intervention. To ensure that more adolescents can benefit from prevention, we need to close the gap between detection and preventive intervention. We investigated the barriers and facilitators from the perspective of public health professionals in screening for depressive and suicidal symptoms and depression prevention referral in a school-based setting. METHODS: We conducted 13 semi-structured interviews with public health professionals, who execute screening and depression prevention referral within the Strong Teens and Resilient Minds (STORM) approach. The interviews were recorded, transcribed verbatim, and coded in several cycles using ATLAS.ti Web. RESULTS: Three main themes of barriers and facilitators emerged from the interviews, namely "professional capabilities," "organization and collaboration," and "beliefs about depressive and suicidal symptoms and participation in prevention". The interviews revealed that professionals do not always feel sufficiently equipped in terms of knowledge, skills and supporting networks. Consequently, they do not always feel well able to execute the process of screening and prevention referral. In addition, a lack of knowledge and support in schools and other cooperating organizationorganizations was seen to hinder the process. Last, the beliefs of public health professionals, school staff, adolescents, and parents -especially stigma and taboo-were found to make the screening and prevention referral process more challenging. CONCLUSIONS: To further improve the process of screening and prevention referral in a school-based setting, enhancing professional competence and a holding work environment for professionals, a strong collaboration and a joint approach with schools and other cooperating organizations and society wide education about depressive and suicidal symptoms and preventive intervention are suggested. Future research should determine whether these recommendations actually lead to closing the gap between detection and prevention.

Longitudinal clinical and functional outcome in distinct cognitive subgroups of first-episode psychosis: a cluster analysis.

BACKGROUND: Cognitive deficits may be characteristic for only a subgroup of first-episode psychosis (FEP) and the link with clinical and functional outcomes is less profound than previously thought. This study aimed to identify cognitive subgroups in a large sample of FEP using a clustering approach with healthy controls as a reference group, subsequently linking cognitive subgroups to clinical and functional outcomes. METHODS: 204 FEP patients were included. Hierarchical cluster analysis was performed using baseline brief assessment of cognition in schizophrenia (BACS). Cognitive subgroups were compared to 40 controls and linked to longitudinal clinical and functional outcomes (PANSS, GAF, self-reported WHODAS 2.0) up to 12-month follow-up. RESULTS: Three distinct cognitive clusters emerged: relative to controls, we found one cluster with preserved cognition (n = 76), one moderately impaired cluster (n = 74) and one severely impaired cluster (n = 54). Patients with severely impaired cognition had more severe clinical symptoms at baseline, 6- and 12-month follow-up as compared to patients with preserved cognition. General functioning (GAF) in the severely impaired cluster was significantly lower than in those with preserved cognition at baseline and showed trend-level effects at 6- and 12-month follow-up. No significant differences in self-reported functional outcome (WHODAS 2.0) were present. CONCLUSIONS: Current results demonstrate the existence of three distinct cognitive subgroups, corresponding with clinical outcome at baseline, 6- and 12-month follow-up. Importantly, the cognitively preserved subgroup was larger than the severely impaired group. Early identification of discrete cognitive profiles can offer valuable information about the clinical outcome but may not be relevant in predicting self-reported functional outcomes.

Dealing with confounding in observational studies: A scoping review of methods evaluated in simulation studies with single-point exposure.

The aim of this article was to perform a scoping review of methods available for dealing with confounding when analyzing the effect of health care treatments with single-point exposure in observational data. We aim to provide an overview of methods and their performance assessed by simulation studies indexed in PubMed. We searched PubMed for simulation studies published until January 2021. Our search was restricted to studies evaluating binary treatments and binary and/or continuous outcomes. Information was extracted on the methods' assumptions, performance, and technical properties. Of 28,548 identified references, 127 studies were eligible for inclusion. Of them, 84 assessed 14 different methods (ie, groups of estimators that share assumptions and implementation) for dealing with measured confounding, and 43 assessed 10 different methods for dealing with unmeasured confounding. Results suggest that there are large differences in performance between methods and that the performance of a specific method is highly dependent on the estimator. Furthermore, the methods' assumptions regarding the specific data features also substantially influence the methods' performance. Finally, the methods result in different estimands (ie, target of inference), which can even vary within methods. In conclusion, when choosing a method to adjust for measured or unmeasured confounding it is important to choose the most appropriate estimand, while considering the population of interest, data structure, and whether the plausibility of the methods' required assumptions hold.

Designing and testing of a health-economic Markov model to assess the cost-effectiveness of treatments for Bipolar disorder: TiBipoMod.

Background: Bipolar disorder is an often recurrent mood disorder that is associated with a significant economic and health-related burden. Increasing the availability of health-economic evidence may aid in reducing this burden. The aim of this study is to describe the design of an open-source health-economic Markov model for assessing the cost-effectiveness of interventions in the treatment of Bipolar Disorders type I and II, TiBipoMod. Methods: TiBipoMod is a decision-analytic Markov model that allows for user-defined incorporation of both pharmacological and non-pharmacological interventions for the treatment of BD. TiBipoMod includes the health states remission, depression, (hypo)mania and death. Costs and effects are modeled over a lifetime horizon from a societal and healthcare perspective, and results are presented as the total costs, Quality-Adjusted Life Years (QALY), Life Years (LY), and incremental costs per QALYs and LYs gained. Results: Functionalities of TiBipoMod are demonstrated by performing a cost-utility analysis of mindfulness-based cognitive therapy (MBCT) compared to the standard of care. Treatment with MBCT resulted in an increase of 0.18 QALYs per patient, and a dominant incremental cost-effectiveness ratio per QALY gained for MBCT at a probability of being cost-effective of 71% when assuming a €50,000 willingness-to-pay threshold. Conclusion: TiBipoMod can easily be adapted and used to determine the cost-effectiveness of interventions in the treatment in Bipolar Disorder type I and II, and is freely available for academic purposes upon request at the authors.

(Cost)-effectiveness and implementation of integrated community-based care for patients with severe mental illness: a study protocol.

BACKGROUND: As severe mental illness (SMI) is associated with a high disease burden and persistent nature, patients with SMI are often subjected to long-term mental healthcare and are in need of additional social support services. Community-based care and support services are organized via different providers and institutions, which are often lacking structural communication, resulting in a fragmented approach. To improve the efficiency of care provision and optimize patient wellbeing, an integrated multi-agency approach to community-based mental health and social services has been developed and implemented. AIM: To present a research protocol describing the evaluation of flexible assertive community teams integrated with social services in terms of effectiveness, cost-effectiveness, and implementation. METHODS/DESIGN: A quasi-experimental study will be conducted using prospective and retrospective observational data in patients with severe mental illness. Patients receiving care from three teams, consisting of flexible assertive community treatment and separately provided social support services (care as usual), will be compared to patients receiving care from two teams integrating these mental and social services into a single team. The study will consist of three parts: 1) an effectiveness evaluation, 2) a health-economic evaluation, and 3) a process implementation evaluation. To assess (cost-)effectiveness, both real-world aggregated and individual patient data will be collected using informed consent, and analysed using a longitudinal mixed model. The economic evaluation will consist of a cost-utility analysis and a cost-effectiveness analysis. For the process and implementation evaluation a mixed method design will be used to describe if the integrated teams have been implemented as planned, if its predefined goals are achieved, and what the experiences are of its team members. DISCUSSION: The integration of health and social services is expected to allow for a more holistic and recovery oriented treatment approach, whilst improving the allocation of scarce resources. This study aims to identify and describe these effects using a mixed-method approach, and support decision-making in the structural implementation of integrating mental and social services.

Design and testing of a health economic Markov model for treatment of anorexia nervosa.

OBJECTIVES: Anorexia Nervosa (AN) is a severe psychiatric disorder and knowledge about the cost-effectiveness of potential interventions is limited. The aim of this paper is to introduce the Trimbos Institute health economic cost-effectiveness model for Anorexia Nervosa (AnoMod-TI), a flexible modeling tool for assessing the long-term cost-effectiveness of interventions for AN in late adolescent and adult patients, which could support clinical decision making. METHODS: AnoMod-TI is a state-transition cohort simulation (Markov) model developed from a Dutch societal perspective, which consists of four health states - namely full remission (FR), partial remission (PR), AN and death. Results are expressed as total healthcare costs, QALYs and incremental cost-effectiveness ratio. RESULTS: For the purpose of demonstrating AnoMod-TI and how it could be used to estimate cost-effectiveness over a 20-year time horizon, it was applied to a hypothetical treatment scenario. Results illustrate how a relatively costly intervention with only modest effects can still be cost-effective in the long term. CONCLUSIONS: AnoMod-TI can be used to examine long-term cost-effectiveness of various interventions aimed at either treating AN or preventing relapse from a state of partial or full remission. AnoMod-TI is freely available upon request to the authors.

Patients' and Psychologists' Preferences for Feedback Reports on Expected Mental Health Treatment Outcomes: A Discrete-Choice Experiment.

In recent years, there has been an increasing focus on routine outcome monitoring (ROM) to provide feedback on patient progress during mental health treatment, with some systems also predicting the expected treatment outcome. The aim of this study was to elicit patients' and psychologists' preferences regarding how ROM system-generated feedback reports should display predicted treatment outcomes. In a discrete-choice experiment, participants were asked 12-13 times to choose between two ways of displaying an expected treatment outcome. The choices varied in four different attributes: representation, outcome, predictors, and advice. A conditional logistic regression was used to estimate participants' preferences. A total of 104 participants (68 patients and 36 psychologists) completed the questionnaire. Participants preferred feedback reports on expected treatment outcome that included: (a) both text and images, (b) a continuous outcome or an outcome that is expressed in terms of a probability, (c) specific predictors, and (d) specific advice. For both patients and psychologists, specific predictors appeared to be most important, specific advice was second most important, a continuous outcome or a probability was third most important, and feedback that includes both text and images was fourth in importance. The ranking in importance of both the attributes and the attribute levels was identical for patients and psychologists. This suggests that, as long as the report is understandable to the patient, psychologists and patients can use the same ROM feedback report, eliminating the need for ROM administrators to develop different versions.

Predicting Future Service Use in Dutch Mental Healthcare: A Machine Learning Approach.

A mental healthcare system in which the scarce resources are equitably and efficiently allocated, benefits from a predictive model about expected service use. The skewness in service use is a challenge for such models. In this study, we applied a machine learning approach to forecast expected service use, as a starting point for agreements between financiers and suppliers of mental healthcare. This study used administrative data from a large mental healthcare organization in the Netherlands. A training set was selected using records from 2017 (N = 10,911), and a test set was selected using records from 2018 (N = 10,201). A baseline model and three random forest models were created from different types of input data to predict (the remainder of) numeric individual treatment hours. A visual analysis was performed on the individual predictions. Patients consumed 62 h of mental healthcare on average in 2018. The model that best predicted service use had a mean error of 21 min at the insurance group level and an average absolute error of 28 h at the patient level. There was a systematic under prediction of service use for high service use patients. The application of machine learning techniques on mental healthcare data is useful for predicting expected service on group level. The results indicate that these models could support financiers and suppliers of healthcare in the planning and allocation of resources. Nevertheless, uncertainty in the prediction of high-cost patients remains a challenge.

Health-economic evaluation of psychological interventions for depression prevention: Systematic review.

Psychological interventions have been proven to be effective to prevent depression, however, little is known on the cost-effectiveness of psychological interventions for the prevention of depression in various populations. A systematic review was conducted using PubMed, PsycINFO, Web of Science, Embase, Cochrane Central Register of Controlled Trials, Econlit, NHS Economic Evaluations Database, NHS Health Technology Assessment and OpenGrey up to January 2021. Only health-economic evaluations based on randomized controlled trials of psychological interventions to prevent depression were included. Independent evaluators selected studies, extracted data and assessed the quality using the Consensus on Health Economic Criteria and the Cochrane Risk of Bias Tool. Twelve trial-based economic evaluations including 5929 participants from six different countries met the inclusion criteria. Overall, the quality of most economic evaluations was considered good, but some studies have some risk of bias. Setting the willingness-to-pay upper limit to US$40,000 (2018 prices) for gaining one quality adjusted life year (QALY), eight psychological preventive interventions were likely to be cost-effective compared to care as usual. The likelihood of preventive psychological interventions being more cost-effective than care as usual looks promising, but more economic evaluations are needed to bridge the many gaps that remain in the evidence-base. ETHICS: As this systematic review is based on published data, approval from the local ethics committee was not required.

Mapping of the World Health Organization's Disability Assessment Schedule 2.0 to disability weights using the Multi-Country Survey Study on Health and Responsiveness.

OBJECTIVES: To develop and test an internationally applicable mapping function for converting WHODAS-2.0 scores to disability weights, thereby enabling WHODAS-2.0 to be used in cost-utility analyses and sectoral decision-making. METHODS: Data from 14 countries were used from the WHO Multi-Country Survey Study on Health and Responsiveness, administered among nationally representative samples of respondents aged 18+ years who were non-institutionalized and living in private households. For the combined total of 92,006 respondents, available WHODAS-2.0 items (for both 36-item and 12-item versions) were mapped onto disability weight estimates using a machine learning approach, whereby data were split into separate training and test sets; cross-validation was used to compare the performance of different regression and penalized regression models. Sensitivity analyses considered different imputation strategies and compared overall model performance with that of country-specific models. RESULTS: Mapping functions converted WHODAS-2.0 scores into disability weights; R-squared values of 0.700-0.754 were obtained for the test data set. Penalized regression models reached comparable performance to standard regression models but with fewer predictors. Imputation had little impact on model performance. Model performance of the generic model on country-specific test sets was comparable to model performance of country-specific models. CONCLUSIONS: Disability weights can be generated with good accuracy using WHODAS 2.0 scores, including in national settings where health state valuations are not directly available, which signifies the utility of WHODAS as an outcome measure in evaluative studies that express intervention benefits in terms of QALYs gained.

More than cost-effectiveness? Applying a second-stage filter to improve policy decision making.

BACKGROUND: Apart from cost-effectiveness, considerations like equity and acceptability may affect health-care priority setting. Preferably, priority setting combines evidence evaluation with an appraisal procedure, to elicit and weigh these considerations. OBJECTIVE: To demonstrate a structured approach for eliciting and evaluating a broad range of assessment criteria, including key stakeholders' values, aiming to support decision makers in priority setting. METHODS: For a set of cost-effective substitute interventions for depression care, the appraisal criteria were adopted from the Australian Assessing Cost-Effectiveness initiative. All substitute interventions were assessed in an appraisal, using focus group discussions and semi-structured interviews conducted among key stakeholders. RESULTS: Appraisal of the substitute cost-effective interventions yielded an overview of considerations and an overall recommendation for decision makers. Two out of the thirteen pairs were deemed acceptable and realistic, that is investment in therapist-guided and Internet-based cognitive behavioural therapy instead of cognitive behavioural therapy in mild depression, and investment in combination therapy rather than individual psychotherapy in severe depression. In the remaining substitution pairs, substantive issues affected acceptability. The key issues identified were as follows: workforce capacity, lack of stakeholder support and the need for change in clinicians' attitude. CONCLUSIONS: Systematic identification of stakeholders' considerations allows decision makers to prioritize among cost-effective policy options. Moreover, this approach entails an explicit and transparent priority-setting procedure and provides insights into the intended and unintended consequences of using a certain health technology. PATIENT CONTRIBUTION: Patients were involved in the conduct of the study for instance, by sharing their values regarding considerations relevant for priority setting.

Real-World Treatment Costs and Care Utilization in Patients with Major Depressive Disorder With and Without Psychiatric Comorbidities in Specialist Mental Healthcare.

BACKGROUND: The majority of patients with major depressive disorder (MDD) have comorbid mental conditions. OBJECTIVES: Since most cost-of-illness studies correct for comorbidity, this study focuses on mental healthcare utilization and treatment costs in patients with MDD including psychiatric comorbidities in specialist mental healthcare, particularly patients with a comorbid personality disorder (PD). METHODS: The Psychiatric Case Register North Netherlands contains administrative data of specialist mental healthcare providers. Treatment episodes were identified from uninterrupted healthcare use. Costs were calculated by multiplying care utilization with unit prices (price level year: 2018). Using generalized linear models, cost drivers were investigated for the entire cohort. RESULTS: A total of 34,713 patients had MDD as a primary diagnosis over the period 2000-2012. The number of patients with psychiatric comorbidities was 24,888 (71.7%), including 13,798 with PD. Costs were highly skewed, with an average ± standard deviation cost per treatment episode of €21,186 ± 74,192 (median €2320). Major cost drivers were inpatient days and daycare days (50 and 28% of total costs), occurring in 12.7 and 12.5% of episodes, respectively. Compared with patients with MDD only (€11,612), costs of patients with additional PD and with or without other comorbidities were, respectively, 2.71 (p < .001) and 2.06 (p < .001) times higher and were 1.36 (p < .001) times higher in patients with MDD and comorbidities other than PD. Other cost drivers were age, calendar year, and first episodes. CONCLUSIONS: Psychiatric comorbidities (especially PD) in addition to age and first episodes drive costs in patients with MDD. Knowledge of cost drivers may help in the development of future stratified disease management programs.

Design of a health-economic Markov model to assess cost-effectiveness and budget impact of the prevention and treatment of depressive disorder.

Background/objective: To describe the design of 'DepMod,' a health-economic Markov model for assessing cost-effectiveness and budget impact of user-defined preventive interventions and treatments in depressive disorders.Methods: DepMod has an epidemiological layer describing how a cohort of people can transition between health states (sub-threshold depression, first episode of mild, moderate or severe depression (partial) remission, recurrence, death). Superimposed on the epidemiological layer, DepMod has an intervention layer consisting of a reference scenario and alternative scenario comparing the effectiveness and cost-effectiveness of a user-defined package of preventive interventions and psychological and pharmacological treatments of depression. Results are presented in terms of quality-adjusted life years (QALYs) gained and healthcare expenditure. Costs and effects can be modeled over 5 years and are subjected to probabilistic sensitivity analysis.Results: DepMod was used to assess the cost-effectiveness of scaling up preventive interventions for treating people with subclinical depression, which showed that there is an 82% probability that scaling up prevention is cost-effective given a willingness-to-pay threshold of €20,000 per QALY.Conclusion: DepMod is a Markov model that assesses the cost-utility and budget impact of different healthcare packages aimed at preventing and treating depression and is freely available for academic purposes upon request at the authors.

Applying machine learning on health record data from general practitioners to predict suicidality.

BACKGROUND: Suicidal behaviour is difficult to detect in the general practice. Machine learning (ML) algorithms using routinely collected data might support General Practitioners (GPs) in the detection of suicidal behaviour. In this paper, we applied machine learning techniques to support GPs recognizing suicidal behaviour in primary care patients using routinely collected general practice data. METHODS: This case-control study used data from a national representative primary care database including over 1.5 million patients (Nivel Primary Care Database). Patients with a suicide (attempt) in 2017 were selected as cases (N = 574) and an at risk control group (N = 207,308) was selected from patients with psychological vulnerability but without a suicide attempt in 2017. RandomForest was trained on a small subsample of the data (training set), and evaluated on unseen data (test set). RESULTS: Almost two-third (65%) of the cases visited their GP within the last 30 days before the suicide (attempt). RandomForest showed a positive predictive value (PPV) of 0.05 (0.04-0.06), with a sensitivity of 0.39 (0.32-0.47) and area under the curve (AUC) of 0.85 (0.81-0.88). Almost all controls were accurately labeled as controls (specificity = 0.98 (0.97-0.98)). Among a sample of 650 at-risk primary care patients, the algorithm would label 20 patients as high-risk. Of those, one would be an actual case and additionally, one case would be missed. CONCLUSION: In this study, we applied machine learning to predict suicidal behaviour using general practice data. Our results showed that these techniques can be used as a complementary step in the identification and stratification of patients at risk of suicidal behaviour. The results are encouraging and provide a first step to use automated screening directly in clinical practice. Additional data from different social domains, such as employment and education, might improve accuracy.

Economic evaluations of non-pharmacological interventions and cost-of-illness studies in bipolar disorder: A systematic review.

Background Bipolar disorder (BD) is associated with substantial societal burden. Therefore, economic studies in BD are becoming increasingly important. The goal of the current study is three-fold: (1) summarize the evidence regarding economic evaluations (EEs) of non-pharmacological interventions for BD, (2) summarize cost-of-illness studies (COIs) for BD published 2012 or later and (3) assess the quality of the identified studies. Methods A systematic search was conducted in MedLine, EMBASE and PsycINFO. For both EEs and COIs, quality assessments were conducted and general and methodological characteristics of the studies were extracted. Outcomes included incremental-cost-effectiveness ratios for EEs and direct and indirect costs for COIs. Results Eight EEs and ten COIs were identified. The included studies revealed high heterogeneity in general and methodological characteristics and study quality. All interventions resulted in improved clinical outcomes. Five studies additionally concluded decreased total costs. For COIs, we found a wide range of direct ($881-$27,617) and indirect cost estimates per capita per year ($1,568-$116,062). Limitations High heterogeneity in terms of interventions, study design and outcomes made it difficult to compare results across studies. Conclusions Interventions improved clinical outcomes in all studies and led to cost-savings in five studies. Findings suggest that non-pharmacological intervention for BD might be cost-effective. Studies on the costs of BD revealed that BD has a substantial economic burden. However, we also found that the number of EEs was relatively low and methodology was heterogenous and therefore encourage future research to widen the body of knowledge in this research field and use standardized methodology.

Predicting future suicidal behaviour in young adults, with different machine learning techniques: A population-based longitudinal study.

BACKGROUND: The predictive accuracy of suicidal behaviour has not improved over the last decades. We aimed to explore the potential of machine learning to predict future suicidal behaviour using population-based longitudinal data. METHOD: Baseline risk data assessed within the Scottish wellbeing study, in which 3508 young adults (18-34 years) completed a battery of psychological measures, were used to predict both suicide ideation and suicide attempts at one-year follow-up. The performance of the following algorithms was compared: regular logistic regression, K-nearest neighbors, classification tree, random forests, gradient boosting and support vector machine. RESULTS: At one year follow up, 2428 respondents (71%) finished the second assessment. 336 respondents (14%) reported suicide ideation between baseline and follow up, and 50 (2%) reported a suicide attempt. All performance metrics were highly similar across methods. The random forest algorithm was the best algorithm to predict suicide ideation (AUC 0.83, PPV 0.52, BA 0.74) and the gradient boosting to predict suicide attempt (AUC 0.80, PPV 0.10, BA 0.69). LIMITATIONS: The number of respondents with suicidal behaviour at follow up was small. We only had data on psychological risk factors, limiting the potential of the more complex machine learning algorithms to outperform regular logistical regression. CONCLUSIONS: When applied to population-based longitudinal data containing multiple psychological measurements, machine learning techniques did not significantly improve the predictive accuracy of suicidal behaviour. Adding more detailed data on for example employment, education or previous health care uptake, might result in better performance of machine learning over regular logistical regression.

Cost-Effectiveness of Virtual Reality Cognitive Behavioral Therapy for Psychosis: Health-Economic Evaluation Within a Randomized Controlled Trial.

BACKGROUND: Evidence was found for the effectiveness of virtual reality-based cognitive behavioral therapy (VR-CBT) for treating paranoia in psychosis, but health-economic evaluations are lacking. OBJECTIVE: This study aimed to determine the short-term cost-effectiveness of VR-CBT. METHODS: The health-economic evaluation was embedded in a randomized controlled trial evaluating VR-CBT in 116 patients with a psychotic disorder suffering from paranoid ideation. The control group (n=58) received treatment as usual (TAU) for psychotic disorders in accordance with the clinical guidelines. The experimental group (n=58) received TAU complemented with add-on VR-CBT to reduce paranoid ideation and social avoidance. Data were collected at baseline and at 3 and 6 months postbaseline. Treatment response was defined as a pre-post improvement of symptoms of at least 20% in social participation measures. Change in quality-adjusted life years (QALYs) was estimated by using Sanderson et al's conversion factor to map a change in the standardized mean difference of Green's Paranoid Thoughts Scale score on a corresponding change in utility. The incremental cost-effectiveness ratios were calculated using 5000 bootstraps of seemingly unrelated regression equations of costs and effects. The cost-effectiveness acceptability curves were graphed for the costs per treatment responder gained and per QALY gained. RESULTS: The average mean incremental costs for a treatment responder on social participation ranged between €8079 and €19,525, with 90.74%-99.74% showing improvement. The average incremental cost per QALY was €48,868 over the 6 months of follow-up, with 99.98% showing improved QALYs. Sensitivity analyses show costs to be lower when relevant baseline differences were included in the analysis. Average costs per treatment responder now ranged between €6800 and €16,597, while the average cost per QALY gained was €42,030. CONCLUSIONS: This study demonstrates that offering VR-CBT to patients with paranoid delusions is an economically viable approach toward improving patients' health in a cost-effective manner. Long-term effects need further research. TRIAL REGISTRATION: International Standard Randomised Controlled Trial Number (ISRCTN) 12929657; http://www.isrctn.com/ISRCTN12929657.