Assessing treatment fidelity and contamination in a cluster randomised controlled trial of motivational interviewing and cognitive behavioural therapy skills in type 2 diabetes

Background: Competencies in psychological techniques delivered by primary care nurses to support diabetes self-management were compared between the intervention and control arms of a cluster randomised controlled trial as part of a process evaluation. The trial was pragmatic and designed to assess effectiveness. This article addresses the question of whether the care that was delivered in the intervention and control trial arms represented high fidelity treatment and attention control, respectively. Methods: Twenty-three primary care nurses were either trained in motivational interviewing (Ml) and cognitive behavioural therapy (CBT) skills or delivered attention control. Nurses' skills in these treatments were evaluated soon after training (treatment arm) and treatment fidelity was assessed after treatment delivery for sessions midway through regimen (both arms) using the Motivational Interviewing Treatment Integrity (MITI) domains and Behaviour Change Counselling Index (BECCI) based on consultations with 151 participants (45% of those who entered the study). The MITI Global Spirit subscale measured demonstration of Ml principles: evocation, collaboration, autonomy/support. Results: After training, median MITI Mi-Adherence was 86.2% (IQR 76.9-100%) and mean MITI Empathy was 4.09 (SD 1.04). During delivery of treatment, in the intervention arm mean MITI Spirit was 4.03 (SD 1.05), mean Empathy was 4.23 (SD 0.89), and median Percentage Complex Reflections was 53.8% (IQR 40.0-71.4%). in the attention control arm mean Empathy was 3.40 (SD 0.98) and median Percentage Complex Reflections was 55.6% (IQR 41.9-71.4%). Conclusions: After Ml and CBT skills training, detailed assessment showed that nurses had basic competencies in some psychological techniques. There appeared to be some delivery of elements of psychological treatment by nurses in the control arm. This model of training and deliver)of Ml and CBT skills integrated into routine nursing care to support diabetes self management in primary care was not associated with high competency levels in all skills.