Exploring predictors of treatment outcome in cognitive behavior therapy for sleep disturbance following acquired brain injury

Abstract

Purpose: To identify predictors of treatment response to cognitive behavior therapy (CBT) for sleep disturbance following acquired brain injury (ABI).

Methods: Classification and regression tree (CART) analysis was conducted on individual patient data from two pilot randomized controlled trials (RCTs): one in traumatic brain injury (TBI), the other in stroke. The combined sample comprised 32 participants; 15 receiving CBT and 17 allocated to treatment as usual (TAU). The outcome was reliable improvement on the Pittsburgh Sleep Quality Index (PSQI).

Results: Study group was a statistically significant predictor of outcome, with CBT participants more likely to achieve reliable improvements than TAU (OR = 4.88, p = 0.042). Study group (CBT vs. TAU) exhibited an area under the ROC curve (AUROC) of 69%. In separate CART analyzes, verbal memory (CVLT-II >45.5), age (<47.5) and baseline depression (HADS-D > 6) predicted positive outcomes in CBT recipients. Each of these variables added a small (∼5%) but not statistically significant amount to AUROC over study group.

Conclusions: In this ABI sample, better memory, younger age, and higher baseline depression were associated with positive treatment response to CBT although individually these variables were not better than group alone in predicting outcomes. The present findings generate hypotheses for further investigation in future studies.

Implications for rehabilitation

• Cognitive behavior therapy improves sleep quality over treatment as usual in persons with acquired brain injury.

• Individuals who are younger in age with better memory and co-morbid symptoms of depression are more likely to respond to the treatment.

• These findings are based on a small sample and can be considered hypothesis generating for future clinical studies.

Keywords:

• Classification and regression trees
• cognitive behavior therapy
• sleep
• predictors
• treatment outcome
• acquired brain injury

Acknowledgements

The authors wish to thank Dr. Kate Frencham, who was one of the three clinicians delivering the intervention (along with coauthors Adam McKay and Dana Wong), Dr. Moira Junge and Dr. Kerrie Haines for providing clinical supervision and rating treatment fidelity, Dr. Lisa Johnston for managing the randomization process, Ms. Jacqueline Owens for completing the follow-up interviews and Ms. Jacqueline Waite for conducting exercise assessments. We are also thankful to all the participants involved in the study.

Disclosure statement

The authors report no conflicts of interest.

Additional information

Funding

This work was supported by the NHMRC Centre of Research Excellence under grant 1023043, Epworth Research Institute under grant 80969.