Productivity Change in Health Care

Abstract

Research on productivity change in health care has surged in recent years. This interest reflects both policy interest in the value of health care and improving data capabilities and methods for productivity research. Because of the central importance of quality change in health care, this research has directly or indirectly considered not only changes in the costs of producing health services (e.g., the cost of a hospital day), but also changes in the benefits of health services for patient health. Some studies have compared overall changes in population health to changes in aggregate medical expenditures. For example, recent studies by Kevin Murphy and Robert Topel (1999) and William Nordhaus (1999) suggest that the value to current and future generations of Americans of improvements in life expectancy in recent decades has exceeded $2 trillion per year. Accounting for the improvements in ageadjusted functional health that also appear to have occurred in recent decades (e.g., Kenneth Manton et al., 1997) makes the improvement in health even greater. Cutler and Elizabeth Richardson (1999) estimate that, even if only 25 percent of the overall improvement in health is attributable to medical care, then health-care productivity has risen. Yet translating this into health-care productivity calculations leaves many issues unresolved. It is not immediately clear how to determine the share of health improvements that result from medical care. Further, even if overall productivity improvements have been high, it is possible that many changes in health-care productivity have been less valuable. Identifying areas of high and low past and potential future productivity improvements would be helpful for guiding policymakers. For all of these reasons, much of the recent research on health-care productivity has focused on explicit analysis of costs and outcomes for certain common, serious health problems, where other factors can be controlled for and relevant inputs and health outcomes can be measured. In this paper, we review the state of the art of the evidence on health-care productivity. We first summarize a set of recent productivity studies of common conditions that account for a substantial fraction of overall medical spending. These studies also illustrate the range of methods that have been used in disease-level productivity studies. In general, the studies show rather substantial productivity gains in care. We then present new evidence on productivity of treatment for breast cancer, a disease that, at least in its most common forms in adults, many experts believe has seen little improvement in benefits of care over time. Considering cancer allows us to focus on a condition where there is no presumption that medical care has been worthwhile, and where there are a host of complex issues related to case-finding, the timing of diagnoses, and chronic care. We find that the treatment of cancer has had at best small productivity improvements. Outcomes have improved more on a per-case basis than when considering the population as a whole.