Interventions:
The interventions were reported with enough detail and they appear to have been appropriate.
Effectiveness/benefits:
The effectiveness data were well reported. Many patients were lost to follow-up over three months (30 of 135). No assessment of the characteristics of patients with missing data, compared with those with complete data, was reported. Imputation using the last observation carried forward could introduce bias, but the magnitude and direction any bias was unclear. The benefit measure was described. The authors noted that a less specific measure, such as quality-adjusted life-years, would have made the results more generalisable.
Costs:
The perspective was clearly stated, and the cost categories appear to have been appropriate for this perspective. The methods used to estimate the resource use were clearly described and seem to have been appropriate. The use of participant diaries could make the resource use estimates country specific and reduce the generalisability of the results. The costs were from appropriate Dutch sources. Due to the significant loss to follow-up, there was a lot of missing resource use data. These were estimated by mean imputation, which could introduce bias, as the average patient-carer data might differ from the missing patient-carer data. No analysis was reported to identify whether the data were missing at random, or due to the characteristics of the missing patient-carer pairs. The costs were not discounted, which was appropriate, but it was unclear if they were adjusted for inflation, as no price year was stated.
Analysis and results:
The economic evaluation was sufficiently described. Appropriate techniques were used to adjust for initial differences. The results were presented appropriately. Uncertainty was evaluated and the variance statistics were reported, but the bootstrapping methods were not described. The benefit measure was specific, making it difficult to interpret and compare this study's results with those of other studies; this limitation was acknowledged by the authors. The authors compared their results with other published results, and reported the strengths and limitations of their study. They acknowledged that their study was short, and their patients may not have been representative of the population. The intervention required well-trained therapists; the availability of this resource should be considered.
Concluding remarks:
The study seems to have been generally well conducted and reported. The authors' conclusions appear to be appropriate, but may not be generalisable to other settings.