Page 94 - Effective healthcare cost containment policies Using the Netherlands as a case study - Niek W. Stadhouders

Chapter 4
and that correction for reverse causality may be incomplete may suggest a downward bias of the estimated elasticity (Raftery, 2014). For example, when only mortality-related patient groups are included, a larger marginal effect and lower threshold is obtained. Based on this, our estimates may be interpreted as a conservative threshold.
The findings, as presented in this paper, are important for policy makers in a number of ways. First of all, the results can be used by Dutch policy makers as a reference value to evaluate new technologies. Our estimates are close to the upper bound of the reference value for new pharmaceuticals of €80,000 euro per QALY. According to our estimates, under a constrained budget new technologies may displace care valued at €73,600 per QALY, suggesting that new treatments and medicines should provide value of more than €73,600 per QALY to maximise total health. In resource allocation, policy makers should compare hospital spending to the value of all other spending alternatives, not just in healthcare but also to other public spending categories like education or infrastructure. Our methodology is suited to estimate marginal value in other areas of health, but estimating value of other government spending would require different methodologies.
4.4.3 Policy relevance and future recommendations
Between-country differences of marginal benefits for different disease categories or age groups could reflect clinical differences between countries. While our research provides disease-specific thresholds, several issues require further research before these comparisons may be used to improve allocation of hospital funds. Three additions would be most valuable. Firstly, LYoL costs should be specified to disease category since this is the most important source of transformation uncertainty. Secondly, additional years should be analyzed. This serves three goals: to increase the precision around the estimators; to estimate how thresholds change over time; and to reduce the risk of overfitting. Thirdly, QoL monitoring should be improved, for example by using patient reported outcome measures (PROMs). These three improvements would allow identification of areas that are relatively overfunded or underfunded with more certainty. The relevance of this research for policy making calls for further studies focusing on a single outcome or disease type - requiring fewer assumptions - in order to verify the robustness of our results. Recently, for cardiovascular hospital spending in the Netherlands a cost-effectiveness threshold of €41,000 was estimated for 2010 (van Baal et al., 2018). This indicates that up to 2010, cardiovascular care may have been relatively cost-effective, and shifting additional resources to this patient group could have improved total health. To improve efficient allocation of spending, more research on cost effectiveness of single disease groups should be encouraged.
86