Nurse staffing and patient outcomes: a longitudinal study on trend and seasonality

In a 2002 national survey of physicians and the public, nurse understaffing was ranked as one of the greatest threats to patient safety in hospitals within the United States (US) [1]. According to the American Nurses Association (ANA), “when health care employers fail to recognize the association between RN staffing and patient outcomes, laws and regulations become necessary” [2]. In 2003 California mandated minimum registered nurse (RN)-to-patient ratios for hospitals to be met by January 1, 2004 [3]. Twelve other states subsequently issued staffing laws requiring hospitals to have either staffing committees responsible for plans and staffing policy or some form of disclosure and/or public reporting of staffing levels [2]. Perhaps due in part to policy changes, including the 2008 Centers for Medicare Medicaid Services (CMS) rule change ending reimbursement for costs of certain hospital-acquired conditions, as well as to ongoing concerns with quality improvement, nurse staffing levels in US hospitals have increased substantially in recent years. From 2004 to 2011, total nursing hours per patient day (HPPD) on general care units of US hospitals increased by 11.5 % and registered nurse hours per patient day (RN HPPD) increased by 22.9 % [4].

It may be intuitive that more nurses can provide better patient care, but research findings about the association between nursing staffing and patient outcome have been inconclusive. The staffing law in California has been in effect for more than 10 years, but researchers did not find evidence of quality improvement associated with the legislation [5, 6]. A meta-analysis based on findings from a systematic review of the literature identified a consistent relationship between higher nurse staffing and lower patient mortality; however, findings regarding the association between nurse staffing and other outcomes such as falls, pressure ulcers, and urinary tract infections varied across studies, and overall results were inconclusive in a pooled analysis [7]. Recent studies provide little empirical evidence to clarify this finding. For pressure ulcers, Park et al. showed that higher RN HPPD was associated with lower unit-acquired pressure ulcer rates in adult care units [8], whereas other researchers reported that staffing and hospital-acquired pressure ulcers were not meaningfully associated [9] or they were associated in the opposite direction, that is, the higher staffing the higher pressure ulcer rate [10–12]. Other studies have also reported associations between higher nurse staffing and higher risk of adverse patient outcomes [13–15]. Most of these studies were based on cross-sectional analyses and some researchers considered the counter-intuitive findings as the result of inadequate risk adjustment [10, 15].

Examining the association of nurse staffing and patient outcomes from a longitudinal perspective may provide new information. According to some longitudinal studies and national surveys, rates of falls and hospital-acquired pressure ulcers in US hospitals decreased significantly in recent years [9, 16–18]. As nurse staffing levels in the US have increased during the same time period [4], it is expected that nurse staffing and rates of these nursing sensitive outcomes were inversely associated at trend level. At seasonal level, the rate of hospital-acquired pressure ulcers was found to be the highest in Quarter 1 of a year; the researchers hypothesized that the seasonality (seasonal pattern) in pressure ulcer rate was related to decreased staffing level in Quarter 1 attributable to patient volume [11]. When patient outcomes are sensitive to nurse staffing, we would expect that they are associated not only at trend level but also at seasonal level.

This study was designed to examine the longitudinal association between nurse staffing and patient outcomes, such as falls and hospital-acquired pressure ulcers, using data from the National Database of Nursing Quality Indicators® (NDNQI®). Our hypothesis was that nurse staffing and rates of these two outcomes were associated inversely at both trend and seasonal levels.