Interventions:
The comparators were not well described. They were simply stated to be home care and institutional care. This makes it impossible to assess if the interventions can be generalised to other settings.
Effectiveness/benefits:
The study methods were not well reported; it appears to have been a retrospective cohort study (home or institutional care were determined retrospectively). It was unclear if the patient characteristics differed between groups before receiving the intervention. This design has a high risk of bias due to confounding, and no efforts to control for this were reported. The health outcomes were not fully assessed as only the utility scores at one point in time were reported. A thorough evaluation would assess the differences in health change over a sufficient period of time. These points make it impossible to determine if any differences in quality of life and costs were due to patient characteristics or the intervention. It was unclear whether the quality of life measurements were retrospective, or measured the patients' health states at that time. The quality of life of caregivers was not included in the analysis.
Costs:
The estimates of resource use and costs were from semi-structured interviews with caregivers. It was not clear to what extent this retrospective approach was affected by recall issues. The resource use and costs were from local sources, which was appropriate, but the sources for the unit costs were not reported. The perspective was not stated, but appears to have been societal. The price year was not reported, making it difficult to reproduce the results, but it was likely to have been 2008.
Analysis and results:
The analysis and results were adequately reported. The authors discussed the limitations of their analysis, the main ones being the retrospective nature of the study, the small sample, and the lack of evaluation of the care burden on caregivers.
Concluding remarks:
The study was generally poorly reported. Its retrospective design introduced a high risk of bias in the results due to population and intervention factors. It is unclear if the authors' conclusions were appropriate.