Revisiting the relationship between nurse staffing and quality of care in nursing homes: an instrumental variables approach.

This paper revisits the relationship between nurse staffing and quality of care in nursing homes using an instrumental variables approach. Most prior studies rely on cross-sectional evidence, which renders causal inference problematic and policy recommendations inappropriate. We exploit legislation changes regarding minimum staffing requirements in eight states between 2000 and 2001 as exogenous shocks to nurse staffing levels. We find that registered nurse staffing has a large and significant impact on quality of care, and that there is no evidence of a significant association between nurse aide staffing and quality of care. A comparison of the IV estimation to the OLS estimation of the first-difference model suggests that ignoring endogeneity would lead to an underestimation of how nurse staffing affects quality of care in nursing homes.