1138 — Nurse Staffing Levels and Escalation to Intensive Care or Mortality: A National Study
Lead/Presenter: Laura Petersen,
COIN - Houston
All Authors: Petersen LA (VA HSR&D Center for Innovations in Quality, Effectiveness, and Safety, Houston, TX; Michael E. DeBakey VA Medical Center; Baylor College of Medicine), Knox, MK (VA HSR&D Center for Innovations in Quality, Effectiveness, and Safety, Houston, TX; Michael E. DeBakey VA Medical Center; Baylor College of Medicine) Eck, CS (VA HSR&D Center for Innovations in Quality, Effectiveness, and Safety, Houston, TX; Michael E. DeBakey VA Medical Center; Baylor College of Medicine) Yang, C (VA HSR&D Center for Innovations in Quality, Effectiveness, and Safety, Houston, TX; Michael E. DeBakey VA Medical Center; Baylor College of Medicine) Dorsey, LE (Michael E. DeBakey VA Medical Center, Houston, TX) Mehta, PD (University of Houston, Houston, TX)
Studies of registered nurse (RN) staffing have shown that when staffing is low, nurses' surveillance of patients is impaired, and the risk of adverse events such as mortality increases. Less is known about the relationship between staffing and other outcomes such as transfer to the intensive care unit (ICU). We studied the relationship between patient exposure to lower RN staffing and ICU transfer or mortality, controlling for other sources of increased risk for mortality or ICU transfer.
We studied all Veterans admitted to one of 234 medical, surgical, or medical-surgical units in 94 VA Medical Centers in FY2019 who spent 2-10 days on the admitted unit and were either discharged or died. RN hours per patient day (RNHPPD) were calculated using estimated daily RN hours and unit-level prorated patient days. A day with low RNHPPD was defined as having less than 85 percent of the unit median RNHPPD. Using logistic regression, we examined the relationship between the number of days with low RNHPPD on the patientâ€™s admitting unit and two outcomes: the likelihood of inpatient mortality and likelihood of transfer to ICU at any point during the patient stay. We controlled for patient age and relative risk score (RRS). We also controlled for unit-level mean RRS and unit-level across-year proportion of low-staffing days.
The mean age of the sample (n = 127,726 patients) was 69.5 years, the mean RRS was 5.4. The number of deaths was 1,241 and number of patients transferred to ICU was 4,604. Adjusting for unit-level covariates, mean patient RRS (OR = 1.8), proportion of RNHPPD days less than 85 percent of the unit median (OR = 22.8) were significantly associated with mortality. After adjusting for the unit-level proportion of low staffing days, deviation from the unit-norm of patientâ€™s exposure to low-staffing days was not significantly associated with mortality. At the unit level, mean RRS (OR = 1.0) and proportion of low RNHPPD days (OR = 0.1) were significantly associated with transfer to ICU. After adjusting for unit-level predictors, number of low staffing days during a patientâ€™s unit-stay was not significantly related to likelihood of ICU transfer.
Patients admitted to units with a high number of days where RN staffing levels are less than 85 percent of the unit median have a significantly increased likelihood of mortality and significantly decreased likelihood of transfer to ICU during their hospitalization, controlling for severity of illness and other confounders. The results suggest that the effect on mortality of patientsâ€™ exposure to low staffing is driven by the inpatient unit environment.
Although low RN staffing has been associated with inpatient mortality in prior studies, less is understood about how RN staffing affects escalation to ICU care. To our knowledge, this is the first study to show that low RN staffing levels reduce ICU transfers, accounting for patient risk. This result suggests that low RN staffing may increase mortality through a reduction in appropriate transfers to the ICU and gives insight into a possible mechanism through which low RN staffing may increase the risk of mortality.