Nurse staffing levels and patient outcomes: A systematic review of longitudinal studies

Abstract

Background: The contribution of registered nurses towards safe patient care has been demonstrated in many studies. However, most of the evidence linking staffing levels to outcomes is cross-sectional with intrinsic limitations including an inability to establish that presumed cause (staffing) precedes the effect. No reviews have summarised longitudinal studies considering nurse staffing and patient outcomes.

Objectives: To identify and assess the evidence for an association between nurse staffing levels, including the composition of the nursing team, and patient outcomes in acute care settings from longitudinal studies.

Methods: We undertook a systematic review of studies where the association between nurse staffing with patient outcomes was assessed in a longitudinal design. Studies with repeated cross-sectional analyses were excluded unless a difference-in-difference design was used. We searched Medline, CINAHL, Embase and the Cochrane Library up to February 2022. We used the ROBINS-I tool to assess risk of bias. We synthesised results in a tabular form and a narrative grouped by outcome.

Results: 27 papers were included. Studies were conducted in a variety of settings and populations, including adult general medical/surgical wards and adult and neonatal intensive care units. Staffing measures were operationalised in a variety of different ways, making direct comparisons between studies difficult and pooled estimates impossible. Most studies were either at serious (n = 12) or critical (n = 5) risk of bias, with only 3 studies at low risk of bias. Studies with the most risk of bias were judged as likely to underestimate the effect of higher registered nurse staffing. Findings are consistent with an overall picture of a beneficial effect from higher registered nurse staffing on preventing patient death. The evidence is less clear for other patient outcomes with a higher risk of bias, but in general the proposition that higher registered nurse staffing is likely to lead to better patient outcomes is supported. Evidence about the contribution of other nursing staff groups is unclear.

Conclusion: The causal relationship between low registered nurse staffing and mortality is plausible and these estimates of relationships from longitudinal studies provide further support. To address residual uncertainties, future studies should be conducted in more than one hospital and using standardised measures when reporting staffing levels.

Tweetable abstract: Having more registered nurses on hospital wards is causally linked to reduced mortality - new review shows there is little room for doubt @ora_dall @workforcesoton @turnel.

Keywords: Inpatients; Longitudinal studies; Nursing staff; Personnel staffing and scheduling; Workforce.