Abstract WP402: Randomised Controlled Trial Of Motivational Interviewing Early After Acute Stroke - The Effect Of Using Different Cut-Points For Mood

Abstract

Background: Mood problems are detrimental to recovery post-stroke. Of the different approaches to treating mood disorders, Motivational Interviewing (MI) - a patient-centred talk-based therapy - has been shown to have a beneficial effect at 3- and 12-months post-stroke in a single-centre randomized controlled trial (RCT). Mood was assessed by the General Health Questionnaire-28 (GHQ-28) using a cut-point 4/5. The purpose of this study was to explore the association between the effect of MI on mood and different recommended cut-points on the GHQ-28. Methods: In the RCT, consecutive patients on a hospital stroke register who met the inclusion criteria were randomised, balanced on age, sex, function, and location, to either usual stroke care (USC) or USC plus MI. Patients in the MI group received four weekly sessions of MI starting 2- to 4-weeks post-stroke. Logistic regression modelling was used to explore the effects of MI on mood when cut-points 5/6 and 11/12 were used. Results: Four hundred eleven patients were randomized: 207 to USC (mean age 69, female 41%); 204 to MI (mean age 69, female 42%). At 3-months post-stroke, the different cut-points produced varying results. For the trial cut-point (4/5), 81 (39%) USC patients and 100 (49%) MI patients had a good outcome, which was significant [p<0.04; OR=1.60, 95%CI=1.04-2.46]. A cut-point of 5/6 meant that 98 (47%) USC and 114 (56%) MI patients had a good outcome [p=0.08; OR=1.47, 95%CI=0.96-2.27]. A cut-point of 11/12, meant that 151 (73%) USC and 155 (75%) MI patients had a good outcome [p=0.49; OR=1.19, 95%CI=0.73-1.94]. Similarly, at 12-months the cut-point of 4/5 showed a significant benefit of MI over USC, while the cut-points of 5/6 and 11/12 showed a non-significant effect. Conclusion: The RCT showed a significant benefit of MI over USC on mood using the cut-point of 4/5: the power of the original trial was based on this cut-point. Post-hoc analysis showed this effect was not seen using other recommended cut-points. This analysis demonstrates the importance of identifying a valid cut-point to power a trial, the need to be consistent in its use when analysing and interpreting the data, and that caution is needed in performing and interpreting post-hoc analyses.