These economic analyses differ widely in their methods and results because not only do they make different assumptions, they also consider different costs and apply different sensitivity analyses (to factors including costs). All of the economic analyses considered hospitalization costs and drug costs, not surprisingly, because these are the two main variables in the equation. Only the Dutch and UK studies, however, gave consideration to general practice costs, only the Dutch study included nursing home and district nursing costs, and only the UK study included initiation costs, monitoring costs and out-patient costs. Only the Dutch study attempted to make some allowance for indirect costs, although it was somewhat dismissive of their effects. The SOLVD study was the only one to attempt to estimate the direct costs of deaths and none of the studies made any attempt to estimate the indirect costs. With respect to sensitivity analysis, only the Dutch, UK, and USA studies examined the effect of varying the proposed efficacy of treatment, and all found the economic analysis to be sensitive to this variable; indeed, the Dutch study found it to be highly sensitive. The UK, USA, as well as SOLVD studies, examined the sensitivity of the analysis to varying hospital costs and all found it to be sensitive. The UK and USA studies found their analyses to be relatively insensitive to the effects of a discount rate. The USA study found greater sensitivity to varying mortality than the Dutch study. Other factors considered included initiation and follow-up strategies in the UK study and dose regime in the SOLVD study.