Who benefits from electronic communications data in psychotherapy? Predictors of outcome and alliance in a pooled randomized trial database.

OBJECTIVE: Three recent pilot randomized clinical trials found no overall advantage of clinician dashboards containing information from patients' electronic communications over treatment-as-usual (TAU). The current pooled post-hoc study examined patient characteristics and electronic-communication measures as exploratory moderators of depressive-symptom outcomes and predictors of therapeutic alliance, with the goal of generating hypotheses for future trials. METHODS: Outcome measures were the Patient Health Questionnaire-8 (PHQ-8) and Working Alliance Inventory (WAI). Using pooled data across the three studies (N = 289 with PHQ-8 slopes; N = 301 in the full baseline database), regression analyses tested demographic, functioning, electronic-communication activity, and language variables, controlling for study membership. RESULTS: Gender and Black racial identity showed interactions with treatment in predicting PHQ-8 slopes, although the Black racial identity finding was based on only seven intervention patients and requires caution. Number of electronic-communication platforms also interacted with treatment: high-platform users had relatively poor outcomes in TAU but improved comparably to low-platform users in the intervention. Greater first-person singular pronoun use was associated with stronger therapeutic alliance (partial r = .39). CONCLUSION: These exploratory findings identify candidate moderators and process markers for a prospective study. Future trials should directly assess patient preference for incorporating electronic communications, standardize and measure clinician dashboard engagement, and replicate candidate moderators before clinical recommendations are made.Trial registration: ClinicalTrials.gov identifier: NCT04011540.Trial registration: ClinicalTrials.gov identifier: NCT03712267.Trial registration: ClinicalTrials.gov identifier: NCT03925038.