A Novel Approach for Constructing Personalized Networks from Longitudinal Perceived Causal Relations

IntroductionThe network approach to psychopathology aims to solve problems with heterogeneity and comorbidity, as well as advance psychiatric precision medicine. Statistical relations in networks can be estimated from intensive longitudinal data, but causal interpretations of such networks are limited by strong statistical assumptions and the frequency of assessments heavily influences which relations can be discovered.ObjectiveAn alternative is to create networks from patient perceptions, which comes with other limitations such as retrospective bias. As a synthesis, we introduce the Longitudinal Perceived Causal Relations (L-PCR) approach.Methods20 participants screening positive for depression completed up to 28 days of brief assessments of experienced symptoms and perceived symptom-symptom influences. Quality criteria of this new method are introduced via a bootstrapping algorithm, answering questions such as “Which symptoms should be included in networks?”, “How many datapoints need to be collected to achieve stable networks?”, and “Does the network change over time?”.ResultsTo achieve stability, networks had to be created using only a core of frequently experienced symptoms, i.e. symptoms experienced during at least a third of days. About 40% of respondents achieved stable networks and only a few respondents exhibited changes in network structure across time. The method was time efficient (on average 7.4 minutes per day) and well received by participants.ConclusionsFuture directions of L-PCR include individualized core symptoms in a controlled clinical population, as well as additional assessment types such as counterfactual questions. Overall, L-PCR addresses several of the prevailing issues found in statistical networks and therefore provides a clinically-meaningful method for personalized analysis.