Therapy after therapy: Entry rates into subsequent psychological treatment among patients receiving therapist-guided internet-delivered or face-to-face psychotherapy.

OBJECTIVES

Despite numerous psychotherapy trials, knowledge on service-system efficiency in terms of return to treatment is limited, especially regarding internet-delivered cognitive behavioural therapy (iCBT). We estimated the hazard ratio (HR) of subsequent psychological treatment over several years following the initial psychological intervention.

METHODS

This naturalistic register follow-up study in Finland included patients receiving therapist-guided iCBT (2013-2021, n = 30,934) or ≤20-session psychotherapy (2018-2021, n = 3348), and matched population controls (n = 92,846). Their long-term psychotherapy data (≤200 sessions/3 years, requiring prior treatment) were obtained from the Social Insurance Institution. We used Cox proportional hazard regression, adjusting for age, sex, first purchase of psychotropic drugs and onset of the first psychiatric diagnosis.

RESULTS

Given the adjustments, the hazard of subsequent long-term psychotherapy was fourfold after iCBT (HR = 4.08; 95% CI 3.81-4.37) and nearly ninefold after ≤20-session psychotherapy (HR = 8.94; CI 7.79-10.26), compared to those without these prior treatments. Prior ≤20-session psychotherapy was associated with reduced hazard of entering subsequent iCBT (HR .12, CI .09-.16), while prior long-term psychotherapy was not (p = .087). Prior iCBT was associated with a lower hazard of subsequent ≤20-session psychotherapy (HR = .41, CI = .35-.47), while prior long-term psychotherapy was not (p = .332).

CONCLUSIONS

Mapping the succession of psychological treatments added knowledge and revealed surprises. For example, patients receiving therapist-guided iCBT were less likely to access subsequent face-to-face psychotherapy than those initially treated face-to-face. While past services are used as a convenience indicator for future services, future research on successive psychological treatments should continue to disentangle clinical need from service systems effects.