Hospitals’ increased reliance on agency nurses and overtime shifts for in-house nurses appears to have implications for patient safety, a study published Wednesday in JAMA Network Open suggests.
Among a sample of 70 hospitals, researchers from George Washington University (GWU) and Premier Inc. found breakpoint thresholds at which increasing these nurse staffing approaches was associated with higher risk of pressure ulcers—an Agency for Health Research and Quality (AHRQ) patient safety indicator long tied to nursing care and staffing.
The analysis reviewed hospital data from January 2019 through December 2022, a period in which many providers faced staffing challenges due to pandemic constraints.
During that window, it found the sampled hospitals had exceeded those safe thresholds by 140% for agency staffing and by 63.6% for overtime use, as measured by average hours per patient day. This translated to a 6.4% increase in pressure ulcer incidence tied to excess agency staffing and a 2.1% increase tied to excess overtime.
“Our study shows that when hospitals over rely on travel nurses or overtime for the regular nursing staff, patient safety care may be compromised,” Patricia Pittman, lead author of the study and director of the Fitzhugh Mullan Institute for Health Workforce Equity at the GWU Milken Institute School of Public Health, said in a release. “Hospitals that want to improve safety should closely track their reliance on overtime and especially nurse staffing agencies.”
There are about 2.5 million pressure ulcer cases recorded in the U.S. healthcare system annually that result in as many as 60,000 deaths, according to data cited in the study. Each hospital stay associated with pressure ulcers add $72,000 to all-payer costs, and $43,000 for Medicare specifically.
The researchers conducted their analysis across 10 different AHRQ patient safety indicators. Beyond pressure ulcers, only postsurgery hemorrhage/hematoma rates were significantly associated with an increase in agency nurse hours per patient day, though this indicator did not show a statistically significant threshold under the team’s structural breakpoint model.
The team noted that their 70-hospital pool reflected “a reasonable national distribution” but “remains a convenience sample” of hospitals affiliated with Premier, which could introduce bias.
And, while it may have provided “natural experiment” of contingency staffing tactics, the researchers acknowledged the unusual circumstances COVID-19 presented for hospitals.
“We did attempt to control for external factors that could have been associated with the outcome and COVID-19 itself in order to mitigate any influence they may have had,” they wrote.
How much should nurse staffing analyses influence patient safety policy?
John Martin, another coauthor and vice president of data science for Premier, said the team’s findings contribute to a “growing body of data that highlights the importance of using contingent labor judiciously.” He added that the takeaways are also in line with data Premier has seen across participants in its Workforce Innovation Collaborative, “which shows that use of contingent labor among CMS 4- and 5-Star Hospitals in our workforce database more rapidly declined in 2023 and 2024 when compared to the rest of the country.”
More practically, the researchers wrote that differentiated reporting of nurse staffing type could be a useful transparency measure for policymakers, payers and patient safety advocacy groups to explore.
And, while the specific pressure ulcer breakpoints they identified—0.090 agency nurse hours per patient day for agency nurses and 0.055 overtime hours per patient day—may not be generalizable to the entire sector, individual hospitals could use a similar approach to “create their own analyses and establish breakpoints for overtime and agency nurse hours to maximize patient safety,” the researchers wrote.
Others advised more caution and continued evaluation before applying new patient safety policy based on these or similar findings.
In an invited commentary published alongside the study, Sean Clarke, Ph.D., of the NYU Rory Meyers College of Nursing, and Christopher DePesa, Ph.D., an emergency department nurse specialist and researcher at Massachusetts General Hospital, wrote that the “findings are intriguing, but raise a number of questions and have a number of methodological loose ends typical of similar studies.”
They warned of the “considerable” methodological challenges faced by similar studies, which over nearly 30 years have yielded “frustratingly inconsistent” results around staffing parameter correlations. The pair also highlighted that the sample’s use of overtime and agency labor was “well below that generally reported for U.S. acute care hospitals.”
Further, while institutional-level policy changes addressing overtime and agency use “may make sense for reasons including cost considerations and suspected decreases in quality of care,” they warned that management doesn’t make such decisions in a vacuum. Other organizational conditions and considerations, such as the availability of labor, and a need to better understand hospital-to-hospital variation should give policymakers considering broader changes based on these types of studies pause.
“With any luck, in the coming years, we will move beyond the data sources and variables that researchers currently have routine access to, such as the data Pittman and colleagues have leveraged here, to data perhaps better suited for examining changes in the nurse workforce and their consequences for quality of care, such as more detailed staffing and patient outcomes measures from electronic health records and hospitals’ information systems,” the pair wrote in their commentary.
“In the meantime, we acknowledge the contributions of this article and the work of others in the broader field while recommending continued cautious weighing of existing and new research findings in this area, particularly in terms of guidance for policy change.”