Interventions:
The study evaluated the outcomes of infliximab. There appeared to be an implicit comparison of infliximab against no infliximab rather than infliximab compared to another relevant treatment for a specific population. The authors indicated that adalimumab was also found to be effective in the treatment of Crohn's disease but were unable to evaluate their comparative effectiveness given the study design.
Effectiveness/benefits:
The retrospective no-protocol design of the study means that the outcomes may better reflect effectiveness in practice than efficacy in a trial but the non-randomised before-and-after design also means that there is the possibility of confounding of the results. The study had a follow-up of two years but Crohn's disease is a chronic lifelong disease and the study did not capture long-term effects of treatment. The authors indicated that mixed- and random-effects models were used to pool data from different study centres. It would have been useful to know the between-hospital variance.
As the authors acknowledged patient quality of life was not among the study outcomes, which inhibits cost-effectiveness evaluation in the UK study setting.
Costs:
The cost estimates were based on the retrospective design and the outcomes may better reflect effectiveness in practice than efficacy in a trial but the non-randomised before-and-after design also means that there is the possibility of confounding of the results. The authors may not have intended to undertake a full cost-effectiveness analysis but enough information was reported for readers to consider drawing cost-effectiveness conclusions. Costs of the concomitant medications for the two periods were not reported. These would need to be accounted for in a full cost-effectiveness analysis and the net incremental costs of the treatments and outcomes calculated. It appeared as though the cost of concomitant treatments decreased with the provision of infliximab.
Costs were derived from appropriate UK sources and were clearly reported. Costs were reported at aggregate categorical levels without a price year which makes it difficult to transfer the data to other settings. Eighteen UK hospitals was a good sized sample and increased the likelihood that the study included representative costs of treatment for moderate to severe Crohn's disease in the UK.
Analysis and results:
The authors presented results comprehensively with consistent reporting of variances and p-values. The authors thoroughly compared their results to those of other studies, including a study that indicated adalimumab was an effective treatment for moderate to severe Crohn's disease.
The authors considered the limitations of their study appropriately: they indicated that the retrospective design could lead to inconsistent treatment protocols and data recording across centres; that patient consultations were not at standardised times, which necessitated the use of annualised means; and that patients initiating treatment with infliximab may have improved or worsened without infliximab, which the study design could not capture due to the before-and-after study design.
The authors may not have intended to undertake a full cost-effectiveness analysis. For a full cost-effectiveness analysis, net incremental costs of treatments and outcomes needed to be calculated and the value of the incremental benefit needed to be considered.
Concluding remarks:
Although the outcomes may reflect effectiveness in practice, potential for confounding and the lack of a comprehensive incremental cost analysis made it difficult to draw cost-effectiveness conclusions from this study.