Dataset size versus homogeneity: A machine learning study on pooling intervention data in e-mental health dropout predictions.

Objective: This study proposes a way of increasing dataset sizes for machine learning tasks in Internet-based Cognitive Behavioral Therapy through pooling interventions. To this end, it (1) examines similarities in user behavior and symptom data among online interventions for patients with depression, social anxiety, and panic disorder and (2) explores whether these similarities suffice to allow for pooling the data together, resulting in more training data when prediction intervention dropout. Methods: A total of 6418 routine care patients from the Internet Psychiatry in Stockholm are analyzed using (1) clustering and (2) dropout prediction models. For the latter, prediction models trained on each individual intervention's data are compared to those trained on all three interventions pooled into one dataset. To investigate if results vary with dataset size, the prediction is repeated using small and medium dataset sizes. Results: The clustering analysis identified three distinct groups that are almost equally spread across interventions and are instead characterized by different activity levels. In eight out of nine settings investigated, pooling the data improves prediction results compared to models trained on a single intervention dataset. It is further confirmed that models trained on small datasets are more likely to overestimate prediction results. Conclusion: The study reveals similar patterns of patients with depression, social anxiety, and panic disorder regarding online activity and intervention dropout. As such, this work offers pooling different interventions' data as a possible approach to counter the problem of small dataset sizes in psychological research.

Primary Care Behavioral Health in Sweden - a protocol of a cluster randomized trial evaluating outcomes related to implementation, organization, and patients (KAIROS).

BACKGROUND: Providing comprehensive and continuous care for patients whose conditions have mental or behavioral components is a central challenge in primary care and an important part of improving universal health coverage. There is a great need for high and routine availability of psychological interventions, but traditional methods for delivering psychotherapy often result in low reach and long wait times. Primary Care Behavioral Health (PCBH) is a method for organizing primary care in which behavioral health staff provide brief, flexible interventions to a large part of the population in active collaboration with other providers. While PCBH holds promise in addressing important challenges, it has not yet been thoroughly evaluated. METHODS: This cluster randomized trial will assess 17 primary care centers (PCCs) that are starting a PCBH implementation process. The PCCs will be divided into two groups, with one starting immediate implementation and the other acting as a control, implementing six months later. The purpose of the study is to strengthen the evidence base for PCBH regarding implementation-, organization-, and patient-level outcomes, taking into consideration that there is a partially dependent relationship between the three levels. Patient outcomes (such as increased daily functioning and reduction of symptoms) may be dependent on organizational changes (such as availability of treatment, waiting times and interprofessional teamwork), which in turn requires change in implementation outcomes (most notably, model fidelity). In addition to the main analysis, five secondary analyses will compare groups based on different combinations of randomization and time periods, specifically before and after each center achieves sufficient PCBH fidelity. DISCUSSION: A randomized comparison of PCBH and traditional primary care has, to our knowledge, not been made before. While the naturalistic setting and the intricacies of implementation pose certain challenges, we have designed this study in an effort to evaluate the causal effects of PCBH despite these complex aspects. The results of this project will be helpful in guiding decisions on how to organize the delivery of behavioral interventions and psychological treatment within the context of primary care in Sweden and elsewhere. TRIAL REGISTRATION: ClinicalTrials.gov: NCT05335382. Retrospectively registered on March 13th, 2022.

Predicting remission after internet-delivered psychotherapy in patients with depression using machine learning and multi-modal data.

This study applied supervised machine learning with multi-modal data to predict remission of major depressive disorder (MDD) after psychotherapy. Genotyped adult patients (n = 894, 65.5% women, age 18-75 years) diagnosed with mild-to-moderate MDD and treated with guided Internet-based Cognitive Behaviour Therapy (ICBT) at the Internet Psychiatry Clinic in Stockholm were included (2008-2016). Predictor types were demographic, clinical, process (e.g., time to complete online questionnaires), and genetic (polygenic risk scores). Outcome was remission status post ICBT (cut-off ≤10 on MADRS-S). Data were split into train (60%) and validation (40%) given ICBT start date. Predictor selection employed human expertise followed by recursive feature elimination. Model derivation was internally validated through cross-validation. The final random forest model was externally validated against a (i) null, (ii) logit, (iii) XGBoost, and (iv) blended meta-ensemble model on the hold-out validation set. Feature selection retained 45 predictors representing all four predictor types. With unseen validation data, the final random forest model proved reasonably accurate at classifying post ICBT remission (Accuracy 0.656 [0.604, 0.705], P vs null model = 0.004; AUC 0.687 [0.631, 0.743]), slightly better vs logit (bootstrap D = 1.730, P = 0.084) but not vs XGBoost (D = 0.463, P = 0.643). Transparency analysis showed model usage of all predictor types at both the group and individual patient level. A new, multi-modal classifier for predicting MDD remission status after ICBT treatment in routine psychiatric care was derived and empirically validated. The multi-modal approach to predicting remission may inform tailored treatment, and deserves further investigation to attain clinical usefulness.

Very long-term outcome of cognitive behavioral therapy for insomnia: one- and ten-year follow-up of a randomized controlled trial.

Insomnia is a common and chronic disorder, and cognitive behavioral therapy (CBT) is the recommended treatment. Very long-term follow-ups of CBT are very rare, and this study aimed to investigate if improvements were stable one and ten years after CBT for insomnia (CBT-i). Based on a three-armed randomized controlled trial of bibliotherapeutic CBT-i, participants received an insomnia-specific self-help book and were randomized to therapist guidance, no guidance, or a waitlist receiving unguided treatment after a delay. Six weeks of treatment was given to 133 participants diagnosed with insomnia disorder. After one and ten years, participants were assessed with self-reports and interviews. Improvements were statistically significant and well maintained at one- and ten-year follow-ups. Average Insomnia Severity Index score [95%CI] was 18.3 [17.7-18.8] at baseline, 10.1 [9.3-10.9] at post-treatment, 9.2 [8.4-10.0] at one- and 10.7 [9.6-11.8] at ten-year follow-up, and 64% and 66% of participants no longer fulfilled criteria for an insomnia diagnosis at one and ten years, respectively. Positive effects of CBT were still present after ten years. Insomnia severity remained low, and two-thirds of participants no longer fulfilled criteria for an insomnia diagnosis. This extends previous findings of CBT, further confirming it as the treatment of choice for insomnia.

An investigation and replication of sleep-related cognitions, acceptance and behaviours as predictors of short- and long-term outcome in cognitive behavioural therapy for insomnia.

The objectives were to investigate the potential for sleep-related behaviours, acceptance and cognitions to predict outcome (insomnia severity) of cognitive behavioural therapy for insomnia (CBT-I). Baseline and outcome data from four randomised controlled trials (n = 276) were used. Predictors were the Dysfunctional Beliefs and Attitudes about Sleep-10 (DBAS-10), Sleep-Related Behaviours Questionnaire (SRBQ), and Sleep Problems Acceptance Questionnaire (SPAQ), and empirically derived factors from a factor analysis combining all items at baseline (n = 835). Baseline values were used to predict post-treatment outcome, and pre-post changes in the predictors were used to predict follow-up outcomes after 3-6 months, 1 year, or 3-10 years, measured both as insomnia severity and as better or worse long-term sleep patterns. A majority (29 of 52) of predictions of insomnia severity were significant, but when controlling for insomnia severity, only two (DBAS-10 at short-term and SRBQ at mid-term follow-up) of the 12 predictions using established scales, and three of the 40 predictions using empirically derived factors, remained significant. The strongest predictor of a long-term, stable sleep pattern was insomnia severity reduction during treatment. Using all available predictors in an overfitted model, 21.2% of short- and 58.9% of long-term outcomes could be predicted. We conclude that although the explored constructs may have important roles in CBT-I, the present study does not support that the DBAS-10, SRBQ, SPAQ, or factors derived from them, would be unique predictors of outcome.

A very brief self-report scale for measuring insomnia severity using two items from the Insomnia Severity Index - development and validation in a clinical population.

OBJECTIVE: To develop a very brief scale with selected items from the Insomnia Severity Index (ISI), and to investigate the psychometric properties of the proposed scale in a psychiatric sample. METHODS: Patient data from seven Cognitive Behavioral Therapy (CBT) for insomnia trials and from regular care were used in psychometric analyses (N = 280-15 653). The samples included patients screening (N = 6936) or receiving treatment (N = 1725) for insomnia and other psychiatric conditions. Six criteria relating to component structure, sensitivity to change and clinical representativeness were used to select items. Psychometric analyses for the proposed very brief scale were performed. RESULTS: One item representing satisfaction/dissatisfaction with current sleep pattern and one item representing interferences with daily functioning, were selected to create the 2-item ISI version. Correlations with the full scale were high at screening, pre and post, and for change (0.82-0.94). Categorical omega was ⍵C = 0.86. With a cut-off of 6 points, the scale could detect Insomnia Disorder with a sensitivity of 84% and a specificity of 76%, which was close to the full ISI showing 86% and 80% respectively. CONCLUSIONS: The systematic psychometric evaluation based on a large sample from different contexts makes the proposed 2-item ISI version (ISI-2) a strong candidate for a very brief scale measuring insomnia, both for detecting cases and for measuring change during CBT with an overall high discriminative validity. ISI-2 is especially useful in clinical settings or population studies where there is a need to measure more than one condition at a time without overburdening patients. CLINICAL TRIALS: Trials used in this analysis: ClinicalTrials.gov identifier: NCT01105052 (https://www.clinicaltrials.gov/ct2/show/NCT01105052) (sample b), ClinicalTrials.gov identifier: NCT01256099 (https://clinicaltrials.gov/ct2/show/NCT01256099) (sample c and d), German clinical trial (DRKS), registration ID: DRKS00008745 (https://www.drks.de/drks_web/navigate.do?navigationId=trial.HTML&TRIAL_ID=DRKS00008745) (sample e), ClinicalTrials.gov identifier: NCT01663844 (https://clinicaltrials.gov/ct2/show/NCT01663844) (sample f and g), ClinicalTrials.gov Identifier: NCT02743338 (https://clinicaltrials.gov/ct2/show/NCT02743338) (sample h).