Background: The core symptoms of psychotic and affective illnesses are important, yet challenging to measure, indicators of disease severity and progression. In our longitudinal study, we monitored individual changes in illness symptoms using daily self-reported ecological assessments, and analyzed their predictive potential for mania, psychosis, and depression as captured in clinical ratings of monthly encounters.
Methods: From 55 participants diagnosed with a primary psychotic or affective disorder followed over 20 months in average, we analyzed 455 monthly clinical interviews and daily ecological assessments completed within 14 days prior to the interview. Ecological assessments consisted of 32 items on mood, psychotic symptoms, and physical well-being, answered via the smartphone application Beiwe. In each interview, a trained rater scored the Positive and Negative Syndrome Scale (PANSS), Montgomery-Asberg Depression Rating Scale (MADRS), and Young Mania Rating Scale (YMRS). For each clinical total score, we independently tested associations with 32 ecological items using linear mixed-effect modeling.
Results: Our selected ecological items were most strongly associated with MADRS total (Bs ranging from 0.07-5.10 with a mean of 2.10). Specifically, the average score of negative affect ecological items was significantly associated with MADRS (B = 3.05, 95% CI [2.64, 3.47]). Independent associations of each ecological item with YMRS and PANSS subscales were generally modest.
Conclusions: Our preliminary analysis showed promise that ecological assessments could be used for tracking symptom fluctuations in patients with severe mental illness. In future work, we will use machine learning approaches to estimate specific symptom dimension scores based on ecological assessments.
