Interventions:
The interventions were sufficiently reported.
Effectiveness/benefits:
The model assumed 10 lost QALYs from the results of another published study. The methods of utility measurement were not reported from the referenced published study. Discounting was appropriate for time horizons longer than one year, so should have been applied in this study, although it was not clear if it was possible. The screening was assumed to be 100% sensitive and specific when it was not, but this strengthened the general conclusion that screening was not cost-effective as the assumption made favoured screening.
The model assumed that all patients with an intracranial aneurysm of greater than 7mm would have a corrective intervention. The authors justified this cut-off point by quoting publications and the very low five-year risk of rupture for intracranial aneurysms smaller than 7mm. However, the authors might have done a sensitivity analysis to further increase the size of an aneurysm referral for treatment, as the risk of rupture needed to be sufficiently high to outweigh the risk of death from treatment. In addition, because the risk of rupture was much greater for posterior aneurysms than for anterior aneurysms (assuming these could be distinguished), a sensitivity analysis may have been justified where only patients with a posterior aneurysm of 7mm or larger, or an anterior aneurysm of 12mm or larger, would be referred for treatment.
Costs:
It was unclear what was included in the cost of disability; this cost was based on a 1999 study.
The study perspective was not reported. The price year was not reported for the cost data, and it was not clear whether reflation was undertaken. The source of UK intervention and hospital costs was reported as been taken from national tariffs, but their source was not included in the text or bibliography. These limitations made it difficult to assess the validity and generalisability of the study results. No discounting of future costs was reported; this was appropriate and should have been included in the model.
Analysis and results:
No incremental cost effectiveness ratios were reported in the body of the text. The graphs used to display the sensitivity analysis did not display the base case. The combinations of risk of rupture and the prevalence required to make screening produce a positive health gain and be cost-effective was adequately reported.
A thorough discussion of model limitations was given. The authors' conclusions appear reasonable in spite of the limitations of the reporting and model structure.
Concluding remarks:
While the model structure was simple and there were limitations in the reporting and in some data, the discussion appropriately identifies some limitations of the model and the authors' conclusions appear appropriate.