A digital therapeutic for people with borderline personality disorder in Germany (EPADIP-BPD): a pragmatic, assessor-blind, parallel-group, randomised controlled trial

Abstract

Background

Fewer than 25% of patients with borderline personality disorder (BPD) receive guideline-adherent psychotherapy. Digital therapeutics can help reduce this gap. Therefore, we tested the effectiveness and safety of priovi, a digital therapeutic for BPD.

Methods

We conducted a pragmatic, assessor-blind, parallel-group, randomised controlled trial, recruiting patients with BPD, according to DSM-5 classification, of at least moderate severity on the 23-item Borderline Symptom List (BSL-23) via online advertisements in Germany. Patients were aged 18 years or older, agreed on an emergency plan for suicidal crises, and gave informed consent to participate. Patients were excluded if they had a primary diagnosis of a substance use disorder or a diagnosis of a psychotic disorder. Simple computerised coin-toss randomisation (1:1) without stratification was used to assign patients to either the unguided digital therapeutic based on schema therapy together with treatment as usual (TAU; intervention group) or TAU alone (control group) and study investigators were masked to group assignment until after the analysis of the primary outcome and main safety outcome. All patients could use any form of usual care. The control group received freely available BPD-specific self-help material. The primary outcome was the time by treatment interaction in the intention-to-treat sample at 3 months measured by BSL-23 score. Safety was established via the number of serious adverse events (ie, suicide attempts, other life-threatening events, and hospitalisation), assessed via online questionnaire. Individuals with BPD were involved in the development of the digital therapeutic, but not in the design of this study. The trial was registered on German Clinical Trials Register (DRKS00028888) and is completed.

Findings

Between May 3 and Oct 20, 2022, 1766 patients were screened for eligibility and 580 patients (520 [90%] women, 47 [8%] men, and 13 [2%] gender diverse) were randomly assigned to the intervention group (n=302) and control group (n=278). Ethnicity data were not recorded. At the primary timepoint of 3 months, 35 (12%) patients dropped out of the intervention group and 15 (5%) of the control group. The median age of patients was 29 years (IQR 24–37). Intention-to-treat analysis with linear mixed models showed a significant time by treatment interaction (p=0·0005) at 3 months in favour of the intervention group (d=0·24 [95% CI 0·07–0·42]). Regarding safety, there were significantly fewer suicide attempts in the intervention group (n=7) than in the control group (n=21; incidence rate ratio 0·34 [95% CI 0·14–0·79]; p=0·0081) and there were no differences regarding other serious adverse events.

Interpretation

This trial provides tentative evidence of the effectiveness and absence of safety concerns of the digital therapeutic, priovi, in the treatment of individuals with BPD.

Funding

GAIA.