Association for Public Policy Analysis & Management

Hospital nurse staffing is a widely discussed issue that has received much research and policy attention.  Over the past few years, proposals to establish minimum nurse-to-patient ratios in hospitals have been debated at both the federal and state levels, with California being the first state to implement such a requirement in 2004.  In addition, there is ongoing concern over nurse shortage and its implications for hospital capacity and the quality of inpatient care.

What is often missing in the discussion about hospital nursing labor is nursing productivity, in particular how return to nurse staffing has changed as the landscape of the health care system evolved over time.  Incentivized by policies and competition, hospitals continue to expand capital inputs, adopting information technology and new medical equipment.  Various payment reforms are being experimented across the country, aiming at improving patient outcomes and controlling cost.  In the meanwhile, there is an increase in severity and medical complexity in inpatients, resulting in demand for more resources and higher intensity care.  These changes likely require hospital nurses to develop new skills and take on new tasks, which could in turn impact their productivity. 

Recent research shows that hospitals productivity has grown considerably in the last decade, which raises the question of whether nursing labor contributed to such a trend – the topic of our study.  Combining data from Medicare claims for a random sample of 20 percent of fee-for-service beneficiaries and the American Hospital Association Annual Survey from 2002 to 2013, we construct measures of patient outcomes and inpatient nurse staffing at the hospital level.  With this set up, we investigate evolution in the relationship between nurse staffing and the delivery of high-quality stays over time.

This analysis builds upon research on hospital production and considers fixed effects models to reduce bias.  Our regressions account for hospital capital inputs, patient severity using the AHRQ Inpatient Quality Indicator risk adjustment algorithm, and a number of hospital- and county-level characteristics such as payer mix and Medicare managed care penetration.  The main outcome of interest is overall 30-day survival rate.

Preliminary results show that nursing labor contributes to improved 30-day survival rate, as expected.   Over time, the return to nursing labor has declined to varying degrees depending on nurse type.  We observe a small, insignificant decrease in productivity among registered nurses and a small but significant decrease in productivity among licensed practical/vocational nurses.