Testing a machine-learning algorithm to predict the persistence and severity of major depressive disorder from baseline self-reports

Abstract

Heterogeneity of major depressive disorder (MDD) illness course complicates clinical decision-making. While efforts to use symptom profiles or biomarkers to develop clinically useful prognostic subtypes have had limited success, a recent report showed that machine learning (ML) models developed from self-reports about incident episode characteristics and comorbidities among respondents with lifetime MDD in the World Health Organization World Mental Health (WMH) Surveys predicted MDD persistence, chronicity, and severity with good accuracy. We report results of model validation in an independent prospective national household sample of 1,056 respondents with lifetime MDD at baseline. The WMH ML models were applied to these baseline data to generate predicted outcome scores that were compared to observed scores assessed 10–12 years after baseline. ML model prediction accuracy was also compared to that of conventional logistic regression models. Area under the receiver operating characteristic curve (AUC) based on ML (.63 for high chronicity and .71–.76 for the other prospective outcomes) was consistently higher than for the logistic models (.62–.70) despite the latter models including more predictors. 34.6–38.1% of respondents with subsequent high persistence-chronicity and 40.8–55.8% with the severity indicators were in the top 20% of the baseline ML predicted risk distribution, while only 0.9% of respondents with subsequent hospitalizations and 1.5% with suicide attempts were in the lowest 20% of the ML predicted risk distribution. These results confirm that clinically useful MDD risk stratification models can be generated from baseline patient self-reports and that ML methods improve on conventional methods in developing such models.

Using patient self-reports to study heterogeneity of treatment effects in major depressive disorder

“…Associations of these predicted values with outcomes over the intervening 10-12 years were then examined using reports obtained in the NCS-2 follow-up survey. These prospective associations were comparable to the retrospective associations found in WMH (Kessler et al, In Press). Importantly, meaningful discrimination was found both at the upper and lower ends of the predicted outcome distributions.…”

Realising stratified psychiatry using multidimensional signatures and trajectories

“…explanatory models) and this has been lacking in over three decades of published studies (see [116] for a review). Validation on truly independent samples is challenging because patient-level data is either not suitably collected, or made available; of the existing studies similar to our proposed framework (reviewed in “Learning models of signatures” section), only two [64, 71] make use of independent samples and all rely on cross-validation for model validation and selection to mitigate against over-optimistic results due to over-fitting e.g. the bias-variance trade-off [57] and inductive bias [117].…”

“…When neuroimaging data were used as the signature, machine learning classifiers were used to predict categorical diagnosis [66], transition from an at-risk state to a dichotomised ‘psychosis versus health’ outcome [67] and dichotomised clozapine response [68, 69], or univariate aggregate predicted univariate global assessment of function (GAF) [70]. Only one study [71] used machine learning to predict multiple outcomes in major depressive disorder although again, these were dichotomised and did not model trajectories as multidimensional constructs. We argue that stratified psychiatry should avoid categorical diagnoses and univariate treatment outcomes.…”

A Preliminary Study of Genetic Variation in the Dopaminergic and Serotonergic Systems and Genome-Wide Additive Genetic Effects on Depression Severity and Treatment Response

“…Last, but not least, it should also be noted that these results are preliminary and require external validation to confirm whether the observed additive genetic effects and genetic correlations are specific to STAR*D or might also exist in other independent samples. Recent machine learning studies of depression (Chekroud et al, 2016; Kessler et al, 2016) underscore the need for validation because findings in a single study may be biased. Specifically, a mathematical model may fit data well in one population, but perform poorly in another.…”