Towards Outcome-Driven Patient Subgroups: A Machine Learning Analysis Across Six Depression Treatment Studies.

BACKGROUND: Major depressive disorder (MDD) is a heterogeneous condition; multiple underlying neurobiological and behavioral substrates are associated with treatment response variability. Understanding the sources of this variability and predicting outcomes has been elusive. Machine learning (ML) shows promise in predicting treatment response in MDD, but its application is limited by challenges to the clinical interpretability of ML models, and clinicians often lack confidence in model results. In order to improve the interpretability of ML models in clinical practice, our goal was to demonstrate the derivation of treatment-relevant patient profiles comprised of clinical and demographic information using a novel ML approach. METHODS: We analyzed data from six clinical trials of pharmacological treatment for depression (total n = 5438) using the Differential Prototypes Neural Network (DPNN), a ML model that derives patient prototypes which can be used to derive treatment-relevant patient clusters while learning to generate probabilities for differential treatment response. A model classifying remission and outputting individual remission probabilities for five first-line monotherapies and three combination treatments was trained using clinical and demographic data. Prototypes were evaluated for interpretability by assessing differences in feature distributions (e.g. age, sex, symptom severity) and treatment-specific outcomes. RESULTS: A 3-prototype model achieved an area under the receiver operating curve of 0.66 and an expected absolute improvement in remission rate for those receiving the best predicted treatment of 6.5% (relative improvement of 15.6%) compared to the population remission rate. We identified three treatment-relevant patient clusters. Cluster A patients tended to be younger, to have increased levels of fatigue, and more severe symptoms. Cluster B patients tended to be older, female, have less severe symptoms, and the highest remission rates. Cluster C patients had more severe symptoms, lower remission rates, more psychomotor agitation, more intense suicidal ideation, and more somatic genital symptoms. CONCLUSION: It is possible to produce novel treatment-relevant patient profiles using ML models; doing so may improve interpretability of ML models and the quality of precision medicine treatments for MDD.

Machine-learning based routing of callers in an Israeli mental health hotline.

BACKGROUND: Mental health contact centers (also known as Hotlines) offer crisis intervention and counselling by phone calls and online chats. These mental health helplines have shown great success in improving the mental state of the callers, and are increasingly becoming popular in Israel and worldwide. Unfortunately, our knowledge about how to conduct successful routing of callers to counselling agents has been limited due to lack of large-scale data with labeled outcomes of the interactions. To date, many of these contact centers are overwhelmed by chat requests and operate in a simple first-come-first-serve (FCFS) scheduling policy which, combined, may lead to many callers receiving suboptimal counselling or abandoning the service before being treated. In this work our goal is to improve the efficiency of mental health contact centers by using a novel machine-learning based routing policy. METHODS: We present a large-scale machine learning-based analysis of real-world data from the online contact center of ERAN, the Israeli Association for Emotional First Aid. The data includes over 35,000 conversations over a 2-years period. Based on this analysis, we present a novel call routing method, that integrates advanced AI-techniques including the Monte Carlo tree search algorithm. We conducted an experiment that included various realistic simulations of incoming calls to contact centers, based on data from ERAN. We divided the simulations into two common settings: standard call flow and heavy call flow. In order to establish a baseline, we compared our proposed solution to two baseline methods: (1) The FCFS method; and (2) a greedy solution based on machine learning predictions. Our comparison focuses on two metrics - the number of calls served and the average feedback of the callers (i.e., quality of the chats). RESULTS: In the preliminary analysis, we identify indicative features that significantly contribute to the effectiveness of a conversation and demonstrate high accuracy in predicting the expected duration and the callers' feedback. In the routing methods evaluation, we find that in heavy call flow settings, our proposed method significantly outperforms the other methods in both the quantity of served calls and average feedback. Most notably, we find that in the heavy call flow settings, our method improves the average feedback by 24% compared to FCFS and by 4% compared to the greedy solution. Regarding the standard-flow setting, we find that our proposed method significantly outperforms the FCFS method in the callers' average feedback with a 12% improvement. However, in this setting, we did not find a significant difference between all methods in the quantity of served-calls and no significant difference was found between our proposed method and the greedy solution. CONCLUSION: The proposed routing policy has the potential to significantly improve the performance of mental health contact centers, especially in peak hours. Leveraging artificial intelligence techniques, such as machine learning algorithms, combined with real-world data can bring about a significant and necessary leap forward in the way mental health hotlines operate and consequently reduce the burden of mental illnesses on health systems. However, implementation and evaluation in an operational contact center is necessary in order to verify that the results replicate in practice.

Treatment selection using prototyping in latent-space with application to depression treatment.

Machine-assisted treatment selection commonly follows one of two paradigms: a fully personalized paradigm which ignores any possible clustering of patients; or a sub-grouping paradigm which ignores personal differences within the identified groups. While both paradigms have shown promising results, each of them suffers from important limitations. In this article, we propose a novel deep learning-based treatment selection approach that is shown to strike a balance between the two paradigms using latent-space prototyping. Our approach is specifically tailored for domains in which effective prototypes and sub-groups of patients are assumed to exist, but groupings relevant to the training objective are not observable in the non-latent space. In an extensive evaluation, using both synthetic and Major Depressive Disorder (MDD) real-world clinical data describing 4754 MDD patients from clinical trials for depression treatment, we show that our approach favorably compares with state-of-the-art approaches. Specifically, the model produced an 8% absolute and 23% relative improvement over random treatment allocation. This is potentially clinically significant, given the large number of patients with MDD. Therefore, the model can bring about a much desired leap forward in the way depression is treated today.