Harvard study shows direct health impact
In the most comprehensive look at the subject to date, a study by the Harvard School of Public Health in Boston, and Vanderbilt University’s School of Nursing in Nashville, TN, has shown that the size and mix of nurse staffing at U.S. hospitals has a direct impact on patient health outcomes.
The import of these findings was underscored by the recent release of a major report, Health Care at the Crossroads: Strategies for Addressing the Evolving Nursing Crisis, by the Joint Commission on Accreditation of Healthcare Organizations (JCAHO), based in Oakbrook Terrace, IL. In it, JCAHO warned that the nursing shortage is "putting patient lives in danger and requires immediate attention."
The Harvard/Vanderbilt study, published in the May 30, 2002, issue of the New England Journal of Medicine, represents a re-examination and refinement of the co-authors’ previous analysis released last year by the Health Resources and Services Administration (HRSA) in cooperation with the Department of Health and Human Services, the Agency for Healthcare Research and Quality, the Centers for Medicare and Medicare Services, and the National Institute for Nursing Research.
"In some subtle ways, this article is different, but the fundamental conclusions hold up," says Jack Needleman, PhD, assistant professor at the Harvard School of Public Health, a co-author of the article with Peter Buerhaus, PhD, RN, FAAN, of Vanderbilt.
The main difference, says Needleman, is that in the original HRSA presentation "we were somewhat equivocal — there was no consistent evidence of an association between LPNs and aide staffing and these outcomes, but some evidence of a total effect. When looked more and harder, we determined that all we really see on these data is a registered nurse effect."
Specifically, the researchers confirmed their initial findings that low levels of RNs among a hospital’s nurses were associated with rates of serious complications such as pneumonia, upper gastrointestinal bleeding, shock, and cardiac arrest, including deaths among patients with these three complications, as well as sepsis or deep vein thrombosis. These complications occurred 3% to 9% more often than in hospitals with higher RN staffing.1 (For a more complete summary of the findings, click here.)
Studies such as these, Needleman says, face a number of challenges, which accounts at least in part for the fact that much of the earlier research on this subject has been inconclusive and perhaps even contradictory.
"Several things account for difficulties when one does research," he says. "Sample size makes a difference in the power of your ability to detect associations that really exist. (His own study used data from 799 hospitals in 11 states and 6 million patients discharged, clearly a large sample.) "You’re also dealing with some data that has a lot of noise’ in it. Take staffing, for example: I’d really like to know how many nurses you have relative to the number of nurses that are needed. But what we deal with in most research are measures of staff in units that don’t take into effect as much as a researcher would like in the differences in patient acuity, plant layout that tells you how many nurses you need, and so on. So when you compare staffing across different hospitals, [you could take into account] other things that influence one hospital in a given eight hours that may not affect another in the same eight hours."
Needleman notes that his team worked very hard to develop a model of risk for anticipating the likelihood of a patient having the complications to be studied, and took the best available adjustments for nurse staffing that were available. "This made our data cleaner," he asserts.
The bottom line, he says, is that "I feel pretty good about what we found; if anything, we were likely to be conservative in the way we drew inferences. He thinks the door "may still be open" on mortality, although he notes that a recent JCAHO Sentinel Event analysis found an association. "I would be shocked to know that, given the work nurses do, being short-staffed has no impact," Needleman says.
Despite problems with data noise and idiosyncrasies at specific institutions, Needleman is unwavering in his belief in the linkage between a shortage of RNs and a drop in quality. "The data at this point, and not just in our study, are very clear; there are some hospitals that have low levels of RN staffing, and patients are being injured by that — some seriously," he asserts.
Solving the problem may be a lot more difficult than identifying it, he concedes. "Fixing that problem is going to require bringing more nurses into the hospital," he says. "There may be ways to restructure work, but efforts in the 90s [to do that] do not seem to have been very successful."
To some degree, he says, the idea that using fewer nurses will save a hospital money can be misleading. "There is some reason to believe that the cost of bringing in more nurses will be associated with savings, in terms of shorter lengths of stay, a reduction in the need to treat complications, and so forth," he notes. "I can’t tell you whether there will be a full offset; but to the extent it isn’t in this climate, it seems to me that the patient may not be prepared to play roulette with his health. Given that, patients will be willing to pay some of the increased costs through higher premiums and Medicare."
His study has a special take-home message for quality managers, Needleman says. "A lot of work on QI in hospitals has focused on very clear events — often those that can be found in a chart — such as a patient not getting a prophylactic antibiotic before surgery and then getting an infection," he observes. "But a lot of the work nurses do, and the lapses they may have, are not well-documented in charts, but these are clearly important issues. The quality managers need to find ways to build QI activities around solving the problems of nursing systems where they may not have the clearest indicators in charts as to what went wrong."
- Associated complications include pneumonia, upper gastrointestinal bleeding, shock, and cardiac arrest.
- Large sample helps bolster validity of findings.
- Quality managers should look beyond charts for issues to address.
For more information, contact:
- Jack Needleman, PhD, assistant professor, The Harvard School of Public Health, 677 Huntington Ave., Boston, MA 02115. Telephone: (617) 432-1318.
1. Needleman J, Buerhaus P, Mattke S, et al. Nurse-staffing levels and the quality of care in hospitals. N Eng J Med 2002; 346:1715-22.