Centre for Health Economics, University of York, July 2018
- This paper considers the detailed breakdown of hospital inpatient expenditures across the period 2007/08 to 2014/15. Decomposition techniques are used to unpick the observed rise in expenditure into a component due to a change in the distribution of characteristics (eg, greater prevalence of morbidity) and a component due to structural changes in the impact of such characteristics on expenditures (eg due to technological change).
Abstract
Understanding the drivers of growth in health care expenditure is crucial for forecasting future health care requirements and to ameliorate inefficient expenditure. This paper considers the detailed breakdown of hospital inpatient expenditures across the period 2007/08 to 2014/15. Decomposition techniques are used to unpick the observed rise in expenditure into a component due to a change in the distribution of characteristics, for example, greater prevalence of morbidity, and a component due to structural changes in the impact of such characteristics on expenditures (coefficient effects, for example, due to technological change). This is undertaken at the mean using standard decomposition techniques, but also across the full distribution of expenditures to gain an understanding of where in the distribution growth and its determinants are most relevant. Decomposition at the mean indicates a larger role for a structural change in characteristics rather than a change in coefficients. A key driver is an increased prevalence of comorbidities. When considering the full distribution we observe a decrease in expenditure at the bottom of the distribution (bottom two quintiles) but increasing expenditure thereafter. The largest increases are observed at the top of the expenditure distribution. Where changes in structural characteristics dominate changes in coefficients in explaining the rise in expenditure. Increases in comorbidities (and the average number of first diagnoses) across the two periods, together with increases in non-elective long stay episodes and non-elective bed days are important drivers of expenditure increases
