Expected survival: we assume that DES confers no advantage in lifetime survival over bare metal stent | No evidence of survival difference has been detected in short term trials, but this may only appear after follow up of 10 years or more | Cost effectiveness analysis is based on very small utility gains of short duration from impaired quality of life before repeat intervention (equivalent to less than one month of extended survival). Thus, any survival advantage would be of greater effect and invalidate current model results |
Trial protocol contamination: most trial evidence is confounded by protocol investigations that inflate repeat intervention rates. Corrections for “clinically driven” interventions are suspect as a posteriori adjustments | Comparison of trial subgroups with and without trial angiography allow adjustments to be made in some cases. It is not clear whether this process is generalisable | This inherent bias generally favours DES, since it inflates the apparent recurrent risk amenable to improvement by DES. This effect is largely offset by use of audit data for patient risks and use of relative (not absolute) measures of DES efficacy |
Duration of effect: we assume that the benefits of DES in avoiding restenotic recurrence of symptoms is limited to 12 months. | Assumption is consistent with trial and registry evidence, but these are not of sufficient duration to detect longer term differences. | Since the outcome gains in the current cost effectiveness analysis are very small, any additional benefit is likely to change results significantly |
Absence of true outcomes: most trials report a variety of intermediate outcomes and virtually none provide the total number of repeat revascularisations | For low risk patients, use of target vessel revascularisation appears to be a reasonable proxy for true outcomes. It is not clear whether this assumption is valid for high risk groups | Where additional interventions are omitted from both trial arms, the apparent relative benefit of using DES is inflated and this may have a large effect on cost effectiveness results |
Unit costs: the new NHS system of fixed reimbursement costs applies to virtually all the important resources in the model and is due to be extended in future. We therefore assume that from the perspective of a service commissioner, there is no uncertainty in unit costs | NHS financial returns suggest large intersite differences in unit costs (possibly due to local application of accounting rules). It is not clear how the new system will affect hospitals | Sensitivity analysis of resource use suggests that cost effectiveness results are vulnerable to unit cost changes in only one area: the relative cost of coronary artery bypass grafting episodes of care compared with percutaneous transluminal angioplasty episodes |