Page 176 - 15Diarrhoeaandvomiting

Appendix A

Figure A.5 Monte Carlo simulation showing the incremental costs of IVT against the cost of ORT

would be associated with a longer period of symptoms and morbidity. On the other hand, the
review also presented evidence suggesting that IVT was associated with a statistically significant
increase in length of hospital stay, which might partly reflect increased morbidity and could have
a negative impact itself in terms of cross-infection. Furthermore, it was stated in the Cochrane
review that ‘IVT is a traumatic experience for most children’ and therefore this may be another
difference, albeit small, between the treatments in terms of their impact on quality of life.
To allow for the possibility that the treatments are not equally efficacious, the results in
Tables A.12 and A.13 are presented with a threshold for QALY gain if IVT is to be considered
cost-effective. If, taking into account all other factors, including those mentioned above, ORT
also gives the greatest QALY gain then this simply strengthens the cost-effectiveness implied by
the cost-minimisation analysis. However, if IVT were judged to be the better clinical alternative
then the results of the threshold analysis suggests that, in the base case, IVT could be considered
cost-effective if it delivered a gain of at least 0.032 QALYs over and above that which would
be obtained using ORT. Similarly, for the ‘worst case’ analysis the QALY gain needed for IVT to
be considered cost-effective would need to be 0.018. This is based on a willingness to pay of
£20,000 per QALY, which is a threshold for cost-effectiveness set by NICE.
How likely is it that that such a QALY gain would be attained? An intervention which added a year
of life lived in perfect health would give an incremental gain of 1 QALY. Hence, an intervention
that gave an additional day of life lived in perfect health therefore would yield an incremental
gain of 0.003 QALYs. Therefore, it seems unlikely for the cost differences in these analyses that
IVT would be considered cost-effective. The success of rehydration therapy is usually measured
in hours not days and the incremental QALY weight attached to a state of rehydration compared
with dehydration is likely to be much less than 1.

The sensitivity analysis illustrated in Figures A.2 and A.3 shows that the results are not very
sensitive to changes in the probability of phlebitis with IVT or the probability of paralytic ileus
with ORT. An important driver of this in the model is the relatively low cost assumed to be
associated with such events. If the costs associated with such complications were much higher
than that implied by the model then changes to these probabilities would have a bigger effect on
the final cost-effectiveness conclusions.

151