Page 81 - Effective healthcare cost containment policies Using the Netherlands as a case study - Niek W. Stadhouders

The marginal benefits of healthcare spending in the Netherlands
(Claxton et al., 2015a). Due to availability of rich data, including health questionnaire outcome data, we were able to first transform both the mortality data and the health questionnaire data to obtain estimated total QALYs (sum of death-related QALYs and morbidity-related QALYs), and then estimate the effect of spending on total QALYs. Figure 4.1 shows the transformations we used to arrive at the level of analysis. Light blocks represent individual level data and dark blocks represent group level data. The arrows show the transformation steps. Transformations that introduce uncertainty (dark arrows) were subject to Monte Carlo analyses.
Figure 4.1: Data transformation and estimation strategy
We combined three datasets on patient group level: health questionnaires, mortality statistics and hospital claims. Hospital claims data contained the euro amount of the claim, a patient follow-up code, patient gender and age, and a DBC code (Dutch alternative to the DRG system) for all hospitals in the Netherlands from 2012 to 2014 (Zorginstituut Nederland, 2017). Claims data prior to 2012 were considered insufficiently comparable due to differences in data collection. Data after 2014 were incomplete at the time of analysis and
73