Nurse Staffing and Patient Outcomes:
Examining the Evidence in Acute Care and Nursing Homes
December 2006 (download pdf file)

I. Introduction | II. Acute Care Settings | III. Nursing Homes

IV. Conclusion | V. References

II. Acute Care Settings

Nurse Staffing and Patient Outcomes in Acute Care
Three systematic reviews were identified that examine nurse staffing and patient outcomes. The three systematic reviews include the Institute of Medicine’s ¹ report ‘To Err is Human: Building a Safer Health System’ published in 2000, Lang et al. ⁹ 2004 ‘Systematic review on the effects of nurse staffing on patient, nurse employee, and hospital outcomes’, and the latest systematic review authored by Lankshear et al. ¹⁰ 2005 titled ‘Nurse staffing and healthcare outcomes a systematic review of the international research evidence’.

Nurse Staffing Levels
As shown in Table 5, most studies tend to be correlational in nature. Although there is a paucity of evidence that tends to suggest that nurse staffing is negatively associated with unplanned hospital readmission and failure to rescue. 117,119-121 There remains no substantive evidence supporting a cause and effect relationship between these measures based on the correlational nature of the studies. This is also true for evidence that suggests that nurse staffing is negatively associated with increased length of stay, nosocomial infection (urinary tract infection, postoperative infection, and pneumonia), and pressure ulcers. 122-125

Study results are inconsistent as to whether higher nurse staffing levels have a positive effect on patient outcomes. Although six 30,89,118,120,129 of the seventeen studies in Table 5 reported no association between richer nurse staffing and positive patient outcomes, the other 11 that report an association tend to be more recent, with larger samples and more sophisticated methods for accounting for confounders. However, these studies examined a variety of different types and acuities of patients which may not be representative of other patient populations. Within some patient groups such as common surgical patients there appears to be some correlational evidence that nurse staffing is associated with patient outcomes. None of the studies specifically identify the ratios or hours of care that produce the best outcomes for different groups of patients or different nursing units.

Associations between Staffing and Patient Outcomes

Failure to rescue
The incidence of failure to rescue (death within 30 days among patients who experienced complications) was lower among surgical, but not medical, patients at higher levels of RN hours per day and higher total nursing hours per day ^13,29 and at lower patient loads per nurse. ¹² Aiken et al. ^11,12 reporting two different analyses conducted on the same Pennsylvania data identified that lower post surgical patient mortality was associated with a) lower patient-to-registered nurse ratios and b) a higher proportion of BS and MS prepared RNs.

Tourangeau et al. ³⁰ after adjusting for case mix and patient care need found that a richer skill mix of RNs was associated with lower 30-day mortality for surgical patients whereas the total amount of nursing staff was not related. The evidence, although equivocal, supports a potential inverse association between nurse staffing and failure to rescue among surgical patients (Table 6). However, longitudinal evidence is required to support these findings in surgical patients.

In-patient mortality
In a longitudinal study, Mark et al. ³¹ analyzed data collected from 422 hospitals (in 11 US states) and found that an increase in RN staffing levels was associated with reduced rates of mortality. Manheim et al., ³² after adjusting for case mix, found that more RNs per admission and a richer skill mix were each associated with lower mortality rates in 3,796 hospitals in 1992. Hartz et al. ²⁰ also reported that more RNs and a stronger RN skill mix were associated with lower mortality among 3,100 hospitals. Krakauer et al. ³³ compared 2 predictive models constructed from different data sets. Both models supported a relationship between a richer RN skill mix and lower inpatient mortality. Aiken et al. ³⁴ found similar relationships in 22 Magnet hospitals but not in a control group of 314 nonfederal hospitals. Finally, in recent analyses, Aiken et al. ^11,12 found that better RN staffing was associated with improved mortality in surgical patients. (These studies analyze the same Pennsylvania data.) Bond et al. ³⁵ found, in examination of 3,763 hospitals, weak significant relationships between the proportion of RNs per occupied bed and mortality rates among Medicare patients, adjusted for severity. Needleman et al. ²⁹ reported no association in medical or surgical patients, and Robertson and Hassan, ²⁷ analyzing 1989 to 1991 data, found no association between the proportion of RNs, LPNs, or NAs and 30-day post admission mortality from chronic obstructive pulmonary disease. The evidence remains inconclusive. Standardized nurse staffing measures and longitudinal evidence are required to support these findings (Table 6).

Pneumonia
The evidence between skill mix and pneumonia reported by 3 key studies is mixed: the American Nurses Association (ANA) ^17,18,36 study found a relationship for California hospitals in 1992 and 1994 but not for New York hospitals for the same year, and Needleman et al. ²⁹ found a relationship for both medical and surgical units. Mark et al. ³¹ reported an inverse relationship between RN staffing levels and pneumonia. Kovner and Gergen ²² found an inverse relationship between the number of RNs per patient day and pneumonia in patients after surgery but not after invasive vascular procedures. However, three studies by Cho et al., ³⁷ Kovner et al., ³⁸ and Unruh ³⁹ did not find this relationship. Thus, the evidence remains unclear whether a significant inverse relationship exists between nurse staffing and pneumonia rates among medical–surgical patients (Table 6).

Urinary tract infections
The ANA study ¹⁸ found a relationship between nurse staffing and urinary tract infection (UTI) rates in California hospitals for both 1992 and 1994 and for New York hospitals only in 1994. Needleman et al. ²⁹ reported a relationship in medical patients but not in surgical patients. Mark et al. ³¹ found an inverse relationship between RN staffing levels and incidence of urinary tract infections. Sovie et al. ⁴⁰ found that total nursing hours per patient day was associated with a decrease in UTI rates among medical students. This finding was present only in 1998 data, however, not in 1997 data, and the clinical importance of the effect could not be assessed as a result of discrepant data. Kovner and Gergen ²² found that a higher number of RN full-time equivalents (RN FTEs) per patient day was statistically associated with lower rates, but the clinical importance of the lower rates was marginal. Recent studies report no relationship: Cho et al. ³⁷ found no relationship between UTI rates and total nursing staffing, total RN hours, and percent of RN staffing, and Kovner et al. ³⁸ found no association between UTI rates and RN hours per severity adjusted patient day or LPN hours per severity-adjusted patient day. There are mixed findings concerning the relationship between UTI rates and nurse staffing (Table 6).

Pressure ulcers
The 1997 ANA report found that richer skill mixes were associated with lower rates of pressure ulcers in California and New York hospitals in 1992 and 1994. ³⁶ Total nursing hours were associated with lower rates of ulcers in New York in 1992 but not in 1994, and in California in 1994 but not in 1992. Mark et al. ³¹ found inverse relationships between rates of pressure ulcers and RN staffing levels. Blegen et al. ^15,16 also found that a higher skill mix, up to 87.5% RN, was associated with lower rates in forty-two nursing units from one hospital. Needleman et al. ²⁹ reported no association. Most findings from the five studies using this endpoint show no association. Whitman et al. ⁴¹ examined staffing and patient outcomes in 95 patient care units across 10 hospitals in the Eastern US and found no significant relationships between staffing and rate of pressure ulcers. In 2005, Donaldson et al ⁴² reported on the first analysis of the impact of mandated minimum staffing ratios in a convenience sample of 68 acute hospitals in California. The data indicated that assessment of the impacts of the mandated ratios on the prevalence of decubiti did not reveal significant changes. The evidence is inconsistent and is not strong enough to support a relationship between nurse staffing and the incidence of pressure ulcers (Table 6).

Falls

Donaldson et al. ⁴² reporting on the mandated minimum staffing ratios in California found there was no significant impact on the incidence of patient falls. A case-control study of patient, education, and care-related risk factors for inpatient falls took place in one tertiary teaching hospital in St Louis, Missouri in 2002. The study examined 6 predictors of inpatient falls using multivariate analysis and found patient-to-nurse ratio as significantly associated (OR 1.6% CI: 1.2-2.0) with fall rates. However, the significance of effect was not reported. Dunton et al. 2004 ⁴³ using 2002 data from 1,751 hospital units in the National Database of Nursing Quality Indicators found that percent of registered nurse hours had a significant inverse association with fall rates for step-down (p<0.01) and medical units (p<0.05), but not for surgical and combined medical-surgical units. However, it is unclear whether these findings are clinically significant. The paucity of evidence concerning impact of staffing levels on fall rates is equivocal and further research is required.

Other patient outcomes
Other than the single report by Needleman et al. ²⁹ of a significant relationship between lower staffing and shock in medical patients, the evidence indicates that associations between nurse staffing and other patient outcomes studied are unclear (Table 6).