Recorded presentation

William N. Evans and BeomSoo Kim. department of Economics, University of Maryland, College Park, 3105 Tydings Hall, College Park, MD 20742, 3014053530, kim@econ.umd.edu

In 1999, California passed a new law setting the minimum nurse to patient ratio and other states are preparing similar regulations. This legislation was prompted by research showing that hospitals with lower stafflevels have higher rates of adverse patient outcomes. There is however concern that these results do not reflect a causal relationship, but rather, reflect some other factors that are simply correlated with hospital staff levels. For example, the bulk of studies compare outcome and staff level across hospitals. But hospitals differ along many dimensions and a hospital with higher staff levels may have other characteristics, such as better surgeons, new technology and more diagnostic equipment, that may help explain why these hospitals have higher quality care. Similarly, studies that use within-hospital variation such as those that compare patients admitted to hospitals on weekend versus weekday may suffer from a selection bias. Hospitals may admit a very different type of patient on the weekend compared to the weekday which may explain in part the difference rates of adverse events across the two groups. In this study, we improve on the previous work by examining patient outcomes using an exogenous source of variation in effective staff levels. In particular, we examine whether outcomes are correlated with the number of hospitals admissions over the next two days. Because hospital staff levels are determined in advance, a large influx of patients to a hospital over the next two days will reduce the effective staff levels for patients admitted previously. We can then compare outcomes of two patients admitted on different Tuesdays who experienced different effective staffing ratios due to this exogenous variation. This paper exploits the variation over time in admissions within a hospital to control for permanent differences across hospitals in the quality of care. Unlike studies that rely on within hospital variation by the day of the week, we control for the fact that admissions on a Saturday may be fundamentally different from admissions on Tuesdays. Data for this project is a census of hospital discharges in California over 1996-2000. The data set contains over 18 million discharges and our sample is a special restricted use version of the data that contains the exact day of admission for patients.We find that hospitals discharge patients early when they face lower effective staff level but conflicting evidence about the impact on patient mortality.

Learning Objectives:

• At the conclusion of the session the participant in this session will be able to identify the impact of staffing ratio on patient outcome

Presenting author's disclosure statement:
I do not have any significant financial interest/arrangement or affiliation with any organization/institution whose products or services are being discussed in this session.

Recorded presentation