The relationship between nurse staffing and patient outcomes: A case study

BackgroundResearch has shown a direct relationship between staffing levels and patient outcomes for specific nurse-sensitive indicators, with lower patient to nurse ratios (i.e. less patients per nurse) associated with better outcomes.ObjectivesTo explore the relationship between nurse staffing characteristics (the nursing hours worked by permanent and temporary staff and nurse hours per patient day) and patient outcomes: pressure sores, patient falls, upper gastrointestinal bleed, pneumonia, sepsis, shock and deep vein thrombosis.DesignA case study using retrospective hospital data, at ward level.SettingA tertiary cardio-respiratory NHS Trust in England, comprising two hospitals.ParticipantsAll patients, including day cases, who were admitted to either hospital as an in-patient over 12 months.MethodsData were extracted from corporate hospital systems. The clinical areas were categorised as lower dependency, i.e. wards, or critical care which included ICU and high dependency units. The relationship between nurse staffing characteristics and patient outcomes was assessed using either a Poisson or negative binomial regression model as appropriate. We sought to establish whether the outcomes were affected by the nurse hours per patient day, the permanent nurse hours worked as a percentage of the total hours, and the permanent nurse hours worked as a percentage of the permanent and bank hours combined.ResultsIn the lower dependency category wards there was only a weak association demonstrated between nurse staffing and the majority of the outcomes. The results from the high dependency critical care areas showed few significant results with only the rate of sepsis being significantly reduced as the ratio of permanent staff hours increased.ConclusionsThe study demonstrated the possibility of using existing hospital data to examine the relationship between nurse staffing and patient outcomes, however the associations found were weak and did not replicate reliably the findings from previous work.