Marten J. Poley - Erasmus Universiteit Rotterdam

28Chapter 2occur in ARM patients. Taking medical consumption as an indication of quality oflife, the quality of life of ARM patients is lower than that of the generalpopulation. Nevertheless, the EQ-5D results show that the impact of stooldifficulties on the overall quality of life is limited. Given the concerns about thequality of life of ARM patients mentioned above, this finding is encouraging. Costeffectiveness(cost per QALY of € 2,482) is good.We were interested to know whether the type of malformation (high or low) andthe mode of treatment affect the long-term consequences of ARM. However, itwas only possible to elicit this information for patients born between 1993 and1996 (who constitute less than one seventh of all responders). To retrieve thisinformation for older patients would require in-depth, time-consuming research,given the long time interval and the lack of computerized data. With theexception of the direct medical costs (Table 2.2), we were, therefore, unable tomake this distinction for all costs and effects. Our knowledge of the impact of thetype of malformation and the mode of treatment is, therefore, restricted to thefirst few years after birth. No conclusions can be made currently on the effect onlong-term costs and effects and on cost-effectiveness.To determine direct medical costs, we adopted a combination of top down andbottom up calculations. It can be disputed whether the efforts required to applytime-consuming bottom up methods for all cost items are in proportion to theinfluence on the comparison between treatment and no treatment. We verified inadvance whether abandoning precise bottom up methods would crucially affectthe outcomes, and, when in doubt, we selected the bottom up technique. We,therefore, expect that the outcomes will not change substantially when all costsare calculated using precise bottom up methods.Economic analyses always are partly based on assumptions. As far as possible wechose solutions that were more likely to overestimate the costs of treatment, sowe consider it to be unlikely that we have presented a too favorable picture ofcost-effectiveness. This applies to all 4 cost categories.The skew of the distribution of the direct medical costs is striking (Table 2.2).This can be explained partly by 3 'outliers' who spent respectively 128, 46, and30 days in the intensive care department. Obviously, a certain spread in the costsis foreseeable because of 'regular' complications. However, when taking this intoaccount, part of this spread could probably be attributed to comorbidity.Therefore, the costs that can be attributed to the treatment of ARM are likely tobe overestimated. This observation is not only valid in these 3 apparent 'outliers'but also in other patients. These questions also apply to the indirect medicalcosts. As mentioned previously, it was our intention to exclude future healthcarecosts of unrelated diseases, but costs of comorbidity inevitably enter into thecalculations. It is particularly difficult to understand the higher number of nightsspent in hospital by patients aged 20 years and older as a consequence of ARM.