[ME1] Assessing generalisability of cost-effectiveness estimates in multinational studies: application to a trial of moxifloxacin in community-acquired pneumonia (CAP)

[ME1] Assessing generalisability of cost-effectiveness estimates in multinational studies: application to a trial of moxifloxacin in community-acquired pneumonia (CAP)

2003 Value in health

Aballea, S. | Hux, M. | Quilici, S. | Drummond, MF. | Volume: 6, Issue: 6, Pages: 616, methods, France, Germany, Spain,

OBJECTIVES: Multi-country trial-based costeffectiveness
analyses often assume that resource utilisation
and clinical efficacy are not country-specific, and
apply country-specific unit costs. We applied econometric
methods to estimate country-specific costeffectiveness,
adjusting for differences in incremental
resource utilisation and case mix across countries. Results
with and without adjustment are compared and methods
described. METHODS: The TARGET multinational trial
compared cure within 21 days for patients with CAP
between sequential IV/PO moxifloxacin monotherapy
and standard comparators. Unit costs were available for
4 countries (France, Germany, Spain, UK) among 10. A
previously published framework, based on a system of
regression equations, was used to determine treatment
impact on resource use and outcome by country, controlling
for baseline characteristics. Clinical efficacy was
held constant across countries, but the impact of cure on
resource utilisation was allowed to vary. Bootstrapping
was also used to estimate uncertainty around countryspecific
cost-effectiveness results. RESULTS: No signifi-
cant inter-country variation in clinical efficacy was
observed (p = 0.9843). Treatment increased the probability
of cure by 5% and the impact of cure on resource
use varied significantly across countries (p < 0.0001). However between-country differences in incremental resource utilisation were not detected statistically (p = 0.7759) so that unadjusted analysis was also a possible approach. Using country-specific unit prices, average incremental costs per patient non-adjusted and adjusted were €-266 and €-436 for Germany, €-381 and €-543 for France, €-281 and €+126 for Spain, €-360 and €-1192 for UK. The probability that moxifloxacin is costsaving was 97% for Germany, 95% for France, 90% for Spain and 87% for the UK in the non-adjusted analyses compared to 99%, 66%, 41% and near 100%. CONCLUSIONS: Where study treatments impact resource use differently across countries or country-specific CEA is desired, adjusted results can differ substantially. Although improved country-specificity is associated with increased variation in cost, country-specific costeffectiveness measures may be more informative.

https://www.doi.org/10.1016/S1098-3015(10)61583-5