The Effect of Nurse-to-Patient Ratios on Mortality Rates in Intensive Care Units

Introduction

Nurse staffing levels have long been hypothesized as a key modifiable determinant of patient mortality, and foundational research using hospital-wide, cross-sectional data established a significant association between lower registered nurse staffing and higher surgical patient mortality across large multi-hospital samples (Aiken et al., 2002; Needleman et al., 2002). Subsequent international research extended this association to a broader range of hospital settings and outcomes, with the multi-country RN4CAST study finding that each additional patient per nurse was associated with a significant increase in the likelihood of patient death within 30 days of admission across nine European countries (Aiken et al., 2014).

Within intensive care specifically, where patient acuity and the intensity of required nursing surveillance are substantially higher than general medical-surgical care, several single-center and smaller multi-center studies have similarly documented an association between nurse staffing and mortality, including early observational work linking ICU staff workload to mortality and more recent multicenter research using administrative or registry-linked staffing data (Tarnow-Mordi et al., 2000; West et al., 2014; Neuraz et al., 2015). However, much of this ICU-specific literature has relied on cross-sectional or single-time-point staffing measurement, which cannot fully account for the possibility that hospital-level characteristics correlated with both staffing and outcomes, or secular trends occurring over the study period, might confound the observed association (Numata et al., 2006; Penoyer, 2010).

A panel data approach, using repeated, longitudinal daily staffing and outcome observations linked across multiple units and hospitals over time, allows explicit statistical control for both hospital-level fixed characteristics and calendar-time trends, providing a methodologically stronger basis for causal inference than a single cross-sectional comparison across hospitals, an approach previously applied successfully to hospital-wide nurse staffing and mortality research using shift-level administrative data (Needleman et al., 2011). The purpose of this study was to examine the association between daily nurse-to-patient ratio and 30-day mortality among adult intensive care unit patients, using longitudinal, unit-day-level staffing and outcome data linked across multiple hospitals and years, with explicit statistical control for hospital-level and temporal confounding.

Methods

Design and data sources. This retrospective panel analysis used daily nurse staffing data, reported by all participating hospitals through a mandatory state critical care staffing reporting system, linked at the unit-day level to admission-level clinical outcome data drawn from each hospital’s electronic health record, for 48 adult intensive care units across 22 hospitals within a single state over a three-year period (January 2022 through December 2024).

Sample. All adult ICU admissions with a length of stay of at least 24 hours during the study period across the 48 participating units were included, yielding 86,240 admissions linked to 42,516 unique ICU-days of staffing data. Admissions to specialty units not reporting standardized nurse-to-patient ratio data, such as certain pediatric and neonatal intensive care units, were excluded.

Exposure measurement. Daily nurse-to-patient ratio was calculated for each ICU-day as the total number of patients present in the unit divided by the total number of direct-care registered nurses staffed for that day, drawn from mandatory staffing reports. Each patient’s ICU stay was linked to the sequence of daily ratios experienced during their admission, with a patient-level exposure summary calculated as the mean daily ratio across the patient’s ICU stay.

Outcome measures. The primary outcome was 30-day all-cause mortality. The secondary outcome was ICU length of stay.

Statistical analysis. Multivariable logistic regression, incorporating hospital fixed effects (to control for time-invariant hospital-level characteristics such as case-mix profile, teaching status, and baseline resource availability) and calendar-year fixed effects (to control for secular trends in critical care practice over the study period), examined the association between mean nurse-to-patient ratio and 30-day mortality, adjusted for patient age, admission severity of illness (APACHE II score), and admission diagnosis category. Ratio was modeled both as a continuous exposure, to characterize the dose-response relationship, and categorically (≤1:1, >1:1 to 1.5:1, >1.5:1 to 2:1 as reference, and >2:1), consistent with common critical care staffing benchmarks. ICU length of stay was analyzed using a linear mixed-effects model with the same fixed-effects structure. A two-sided p value of less than .05 was considered statistically significant.

Results

Across 42,516 ICU-days linked to 86,240 adult ICU admissions, mean nurse-to-patient ratio was 1.8 patients per nurse (SD 0.6). As shown in Table 1, patient severity of illness was similar across ratio tertiles, while higher-ratio (more heavily loaded) ICU-days were disproportionately concentrated at non-academic hospitals and during periods of higher occupancy and weekend or night shifts, a pattern consistent with staffing strain occurring preferentially during periods of peak demand and reduced staffing availability.

In adjusted analysis, each additional patient per nurse was associated with a significant increase in 30-day mortality (adjusted OR 1.09 per one-patient increase in ratio, 95% CI 1.06–1.12, p < .001). The dose-response relationship across the observed range of nurse-to-patient ratio is shown in Figure 1.

Categorical analysis confirmed this pattern, as shown in Figure 2: relative to ICU-days staffed at 1.5:1 to 2:1 (the reference category, reflecting a common critical care staffing benchmark), adjusted mortality odds were significantly lower for ICU-days staffed at 1:1 or better (adjusted OR 0.81, 95% CI 0.72–0.91, p < .001) and significantly higher for ICU-days staffed above 2:1 (adjusted OR 1.61, 95% CI 1.38–1.87, p < .001).

ICU length of stay was also significantly longer on higher-ratio (more heavily loaded) ICU-days, as shown in Figure 3, with mean ICU length of stay increasing from 4.6 days in the lowest ratio tertile to 6.3 days in the highest ratio tertile.

Discussion

This multi-center retrospective panel analysis found that higher nurse-to-patient ratios, reflecting fewer nurses relative to patient volume, were independently associated with significantly increased 30-day mortality among adult intensive care unit patients, with an approximately linear dose-response relationship across the observed staffing range and no evidence of a discrete safe threshold below which additional staffing conferred no further benefit. The magnitude of the observed effect, a 9% increase in adjusted mortality odds per additional patient per nurse, is closely comparable to the approximately 7% increase in mortality odds per additional patient reported in the foundational hospital-wide Aiken et al. (2002) study and the significant staffing-mortality association reported in the multi-country RN4CAST study (Aiken et al., 2014), extending this well-established general hospital finding specifically to the intensive care setting using a methodologically stronger, longitudinal panel design.

The use of hospital fixed effects and calendar-year fixed effects in this analysis represents a meaningful methodological strengthening relative to much of the prior ICU-specific staffing-mortality literature, which has more often relied on cross-sectional comparison across hospitals (Numata et al., 2006; Penoyer, 2010). By comparing staffing and outcomes within the same hospital across different days and years, this panel approach controls for time-invariant hospital-level confounders, such as baseline case-mix complexity, teaching status, and general resource availability, that could otherwise plausibly explain an observed cross-sectional association between staffing and mortality independent of a genuine causal staffing effect. The persistence of a significant, dose-dependent association even after this more rigorous adjustment strengthens confidence that the observed relationship reflects a genuine effect of staffing variation rather than confounding by stable hospital characteristics.

The finding that higher-ratio (more heavily loaded) ICU-days were disproportionately concentrated during periods of higher occupancy and weekend or night shifts is consistent with prior research documenting that staffing strain in critical care often concentrates during precisely the periods when clinical demand, and therefore the potential consequence of inadequate staffing, is greatest (Sakr et al., 2015; Rothen et al., 2007). This pattern has direct relevance for staffing policy discussions, suggesting that flexible, demand-responsive staffing models, rather than fixed staffing ratios calculated only on average census, may be necessary to fully address the mortality risk associated with peak-demand staffing strain identified in this analysis.

The absence of evidence for a discrete safety threshold, with mortality risk instead increasing approximately linearly across the entire observed ratio range, has an important policy implication distinct from a threshold-based interpretation: rather than identifying a single minimum ratio below which staffing is “unsafe” and above which further improvement offers no benefit, these findings suggest that incremental staffing improvement confers incremental mortality benefit across the range of ratios observed in this study, a pattern consistent with the dose-response relationship reported in the original shift-level hospital-wide staffing and mortality analysis by Needleman et al. (2011).

Several limitations should be considered. Despite the methodological strength of hospital and calendar-year fixed effects, this remains an observational analysis, and residual confounding by unmeasured, time-varying factors correlated with both daily staffing decisions and patient outcomes, such as unmeasured day-to-day variation in patient complexity beyond what APACHE II captures, cannot be fully excluded. This study was conducted within a single state, and the specific staffing reporting infrastructure, critical care nursing labor market, and hospital case-mix distribution may limit generalizability to other states or healthcare systems with different staffing regulatory environments. Nurse-to-patient ratio was measured at the unit-day level rather than matched to the specific nurse assigned to each individual patient, which may introduce some misclassification relative to a hypothetical patient-level staffing exposure measure.

Future research should extend this panel methodology to examine whether the observed dose-response relationship differs by patient subgroup, such as by admission diagnosis category or baseline severity of illness, and should evaluate whether flexible, demand-responsive staffing models specifically targeting the high-occupancy, weekend, and night-shift periods identified in this study as most vulnerable to staffing strain produce measurable mortality benefit when implemented prospectively. Taken together, these findings provide rigorous, multi-center, longitudinal panel evidence reinforcing nurse-to-patient ratio as a modifiable structural determinant of intensive care unit mortality risk, with a dose-response relationship extending across the full observed staffing range rather than concentrated at a single minimum threshold.

References

Aiken, L. H., Clarke, S. P., Sloane, D. M., Sochalski, J., & Silber, J. H. (2002). Hospital nurse staffing and patient mortality, nurse burnout, and job dissatisfaction. JAMA, 288(16), 1987–1993.

Aiken, L. H., Sloane, D. M., Bruyneel, L., Van den Heede, K., Griffiths, P., Busse, R., Diomidous, M., Kinnunen, J., Kózka, M., Lesaffre, E., McHugh, M. D., Moreno-Casbas, M. T., Rafferty, A. M., Schwendimann, R., Scott, P. A., Tishelman, C., van Achterberg, T., & Sermeus, W. (2014). Nurse staffing and education and hospital mortality in nine European countries: A retrospective observational study. The Lancet, 383(9931), 1824–1830.

Amaravadi, R. K., Dimick, J. B., Pronovost, P. J., & Lipsett, P. A. (2000). ICU nurse-to-patient ratio is associated with complications and resource use after esophagectomy. Intensive Care Medicine, 26(12), 1857–1862.

Cho, S. H., Hwang, J. H., & Kim, J. (2008). Nurse staffing and patient mortality in intensive care units. Nursing Research, 57(5), 322–330.

Kane, R. L., Shamliyan, T. A., Mueller, C., Duval, S., & Wilt, T. J. (2007). The association of registered nurse staffing levels and patient outcomes: Systematic review and meta-analysis. Medical Care, 45(12), 1195–1204.

Needleman, J., Buerhaus, P., Mattke, S., Stewart, M., & Zelevinsky, K. (2002). Nurse-staffing levels and the quality of care in hospitals. New England Journal of Medicine, 346(22), 1715–1722.

Needleman, J., Buerhaus, P., Pankratz, V. S., Leibson, C. L., Stevens, S. R., & Harris, M. (2011). Nurse staffing and inpatient hospital mortality. New England Journal of Medicine, 364(11), 1037–1045.

Neuraz, A., Guérin, C., Payet, C., Polazzi, S., Aubrun, F., Dailler, F., Lehot, J. J., Piriou, V., Neidecker, J., Rimmelé, T., Schott, A. M., & Duclos, A. (2015). Patient mortality is associated with staff resources and workload in the ICU: A multicenter observational study. Critical Care Medicine, 43(8), 1587–1594.

Numata, Y., Schulzer, M., van der Wal, R., Globerman, J., Semeniuk, P., Balka, E., & Fitzgerald, J. M. (2006). Nurse staffing levels and hospital mortality in critical care settings: Literature review and meta-analysis. Journal of Advanced Nursing, 55(4), 435–448.

Penoyer, D. A. (2010). Nurse staffing and patient outcomes in critical care: A concise review. Critical Care Medicine, 38(7), 1521–1528.

Rothen, H. U., Stricker, K., Einfalt, J., Bauer, P., Metnitz, P. G., Moreno, R. P., & Takala, J. (2007). Variability in outcome and resource use in intensive care units. Intensive Care Medicine, 33(8), 1329–1336.

Sakr, Y., Moreira, C. L., Rhodes, A., Ferguson, N. D., Kleinpell, R., Pickkers, P., Kuiper, M. A., Lipman, J., & Vincent, J. L. (2015). The impact of hospital and ICU organizational factors on outcome in critically ill patients: Results from the Extended Prevalence of Infection in Intensive Care study. Critical Care Medicine, 43(3), 519–526.

Tarnow-Mordi, W. O., Hau, C., Warden, A., & Shearer, A. J. (2000). Hospital mortality in relation to staff workload: A 4-year study in an adult intensive-care unit. The Lancet, 356(9225), 185–189.

West, E., Barron, D. N., Harrison, D., Rafferty, A. M., Rowan, K., & Sanderson, C. (2014). Nurse staffing, medical staffing and mortality in intensive care: An observational study. International Journal of Nursing Studies, 51(5), 781–794.

◆