Can we predict failure in couple therapy early enough to enhance outcome?

•Tested whether weekly therapy progress could detect off-track couples.•An expert derived algorithm was used to predict therapy outcome.•Detected a proportion of couples not benefiting from therapy at Session 3.•Prediction was substantially improved at Session 4.•Eventual outcome was accurately predicted for 70% of couples at Session 4.

Feedback to therapists based on systematic monitoring of individual therapy progress reliably enhances therapy outcome. An implicit assumption of therapy progress feedback is that clients unlikely to benefit from therapy can be detected early enough in the course of therapy for corrective action to be taken. To explore the possibility of using feedback of therapy progress to enhance couple therapy outcome, the current study tested whether weekly therapy progress could detect off-track clients early in couple therapy. In an effectiveness trial of couple therapy, 136 couples were monitored weekly on relationship satisfaction and an expert derived algorithm was used to attempt to predict eventual therapy outcome. As expected, the algorithm detected a significant proportion of couples who did not benefit from couple therapy at Session 3, but prediction was substantially improved at Session 4 so that eventual outcome was accurately predicted for 70% of couples, with little improvement of prediction thereafter. More sophisticated algorithms might enhance prediction accuracy, and a trial of the effects of therapy progress feedback on couple therapy outcome is needed.