Selection of comparators.
The authors did not explicitly justify their choice of metoclopramide-based regimens as the comparators to ondansetron-based regimens, which reflected current practice in their hospital. Careful consideration should be given as to whether this alternative is appropriate to current practice in other hospitals and settings.
Validity of estimate of effectiveness.
The design of the study and the selection of the comparators may have limited the internal validity of the study's findings. Patients were not randomised to either group; instead the clinician or the patient chose which regimen they would receive. In addition, the study did not look at the costs and effectiveness of ondansetron and metoclopramide alone, but explored combination regimens, which may have included different proportions of other antiemetic agents administered throughout the 5-day period. A predefined protocol of antiemetic agent was not used in this study. The nature of the 'ideal' antiemetic regimen, in terms of the agents and dosages required to produce the reported ICER, was therefore unclear. In particular, the role of dexamethasone in augmenting the effectiveness of either agent was uncertain.
'Control of emesis' seemed an appropriate effectiveness measure. The study used self-reporting over a 5-day period to record effectiveness, but it made no assessment on how well the patients completed their diaries. The authors cautioned that the sample size might have been insufficient to detect a statistical difference in the quality of life scores.
Validity of estimate of benefit.
The number of cases of emesis avoided was appropriate for this technology, but might be limited in comparisons with other technologies exhibiting different effects. The overall quality of life or preferences would be a more generalisable estimate of benefit.
Validity of estimate of costs.
Only the costs of administering antiemetic therapy, relevant to the hospital perspective, were included in this study. The costs associated with patients who discontinued their chemotherapy due to emesis were not calculated. The inclusion of these costs would have added to the value of the study in terms of informing clinical practice.
The sensitivity analysis was limited and focused only on the acquisition cost of antiemetics. A more extensive sensitivity analysis would have highlighted the impact of the effectiveness of each antiemetic regimen on the baseline ICER. In the sensitivity analysis, the authors also presented average cost-effectiveness ratios. These are misleading since they do not give information about the extra cost per additional unit of outcome, in this instance the number of cases of emesis avoided. They could have been omitted.
Other issues.
The authors considered the issue of generalisability in the sensitivity analysis, and acknowledged that it could have been more extensive. They compared their study with others, highlighting that their study considered effectiveness as opposed to efficacy. However, as the authors pointed out, the use of various combinations of emetics made it difficult to extrapolate the results to different combinations of antiemetics, different patient populations, or other settings.