Interventions:
The intervention was a combination of those used by several clinical experts, including general practitioners, specialists, and health practitioners. None of these treatments was described and it was unclear whether they would be relevant for other settings.
Effectiveness/benefits:
The clinical and epidemiological estimates were from a review of the literature and interviews with experts. Little information on the designs of the source studies was provided, making it difficult to judge the validity of this evidence. All inputs were varied in the sensitivity analysis. QALYs were an appropriate measure, as insomnia has an impact on quality of life and can increase mortality due to work or car accidents. The utility weights were obtained using an appropriate instrument, completed by local experts, and these values were combined with published estimates. This appears to have been valid.
Costs:
The authors stated that the perspective of society was taken, but productivity costs were not directly included. The sources for the unit costs were provided and reflected the authors’ setting. The resource use was mainly from interviews with experts and published studies, which were not fully described. The costs were varied in the sensitivity analysis and the price year was reported allowing reflation exercises.
Analysis and results:
The total costs and QALYs for treated versus untreated patients were reported and an incremental analysis was conducted. It was unclear why the net benefit per QALY was calculated as treatment was dominant and the calculation of a ratio was unnecessary. Deterministic and probabilistic sensitivity analyses were conducted and showed that the main results were robust. The authors stated that their results might be conservative against treatment, as the costs of car and work accidents were not included. The findings appear to be specific to the authors' setting and might be difficult to transfer to other countries without making appropriate adjustments.
Concluding remarks:
Most of the data were from interviews with experts and the treatment pathways were not fully described, but the authors’ conclusions appear to be valid and robust.