Researchers at odds on agency’s nurse staffing report

Those cited in HPC document disagree with some of the findings

THE HEALTH POLICY COMMISSION’S REPORT early last month on the potential effects of Question 1 was portrayed as an independent, objective analysis, but many of the researchers whose work was cited in the report disagree with at least some of the findings.

The commission, an independent watchdog agency set up to monitor and find ways to rein in health care costs, commissioned the report during the summer but did not disclose it was under development until just before it was released. The focus of the report was Question 1 on the ballot which, if approved by voters on Tuesday, would set mandatory nurse-to-patient staffing ratios throughout medical facilities.

The report, written by a staff researcher at the commission and a faculty expert at a California university, estimated the ballot question could increase health care costs in Massachusetts by up to $1 billion without an appreciable improvement in quality of care. The findings, coming from a respected state agency, have played a powerful role in the debate over the question. They are regularly cited by news outlets and by opponents of Question 1, while the Massachusetts Nurses Association, the chief sponsor of the question, has claimed the report was biased.

In interviews with the researchers who wrote and were cited in the report, it is apparent that experts can review the same data and come to very different conclusions. The researchers interviewed were Joanne Spetz, a University of California, San Francisco, professor who coauthored the commission’s report; Linda Aiken from the University of Pennsylvania; Jack Needleman from the University of California, Los Angeles; and Judith Shindul-Rothschild from Boston College. All are familiar with each other’s work; Spetz and Aiken have even worked on papers together.

All of the researchers agreed the goal of the Massachusetts ballot measure, to improve staffing and patient outcomes in lower-staffed hospitals, is important. Needleman, who used to live in Massachusetts when he taught at Harvard, said, “There are a lot of well-staffed hospitals there. But just as is the case in other states, there are some hospitals that are so low-staffed in Massachusetts I would not want to be a patient in them.”

The researchers also agree that, in general, the research shows that lowering the number of patients per nurse improves a wide range of patient outcomes. “This is something everyone pretty much agrees on – you add more nurses and you get better results,” said Needleman.

According to her research, Shindul-Rothschild said, the best-staffed hospitals in the state, and those meeting the standards of Question 1, received $500,000 on average in bonuses from Medicare last year for good patient outcomes, while those with lower nurse staffing levels were penalized $500,000 due to poor patient outcomes.

One major point of disagreement between the researchers is whether laws that mandate nurse-to-patient ratios work. Most research on nurse staffing mandates focuses on California, the only state that has a law in place limiting the number of patients per nurse.

When the Health Policy Commission’s report was first unveiled, Spetz said studies on the effect of the California law were “a little bit of a wash” – mixed results that indicated there had been “no systematic improvement” in patient outcomes.

“Based on the generic research [of nurse staffing], we would expect to see improved outcomes in California,” Spetz said in an interview. “We were surprised when we did not see it. You have to ask, what else is going on? Maybe the hospitals in California laid off aides. Maybe they took money out of other things, like the infection control nurse. The Massachusetts ballot question would not keep you from laying off the infection control nurse. Or maybe the ratios just weren’t strong enough. Patient care is complex, and nursing is an important factor in improving patient care, but it’s just one factor.”

Aiken, however, said a series of studies she has done over the years show that nurse staffing ratios have played a critical role in lowering mortality rates in California hospitals. “For every patient you add to a nurse’s workload, mortality rates increase by 7 percent,” she said.

Spetz and the Health Policy Commission report’s co-author, agency research director David Auerbach, cited Aiken’s studies, but Spetz is skeptical about the broader implications of Aiken’s findings. She said the results do not prove the ratios actually improve outcomes because the studies are cross-sectional, not longitudinal, meaning they look at a group of hospitals at only one point in time.

“The limitations section [of Aiken’s 2006 study] clearly says the results were a correlation and not causal,” Spetz said. “But there is a difference between the limitations section of the paper and how the press release was written, which said the study proves ratios work.”

Spetz said groups advocating for laws mandating nursing staffing ratios have utilized the language in the press release as proof that ratios work.

Aiken responded that her most recent study, published last month in Medical Care, looks at hospitals in California and other states over time. “This new longitudinal data,” she said, “shows nearly the same improvements in mortality, further strengthening the argument that mandated nurse-to-patient ratios work. I stand by the results of my research.”

The Health Policy Commission report calculated how much it would cost for Massachusetts hospitals to add the 2,200 to 3,100 nurses needed to staff up the state’s hospitals to meet the requirements spelled out in Question 1. Their high estimate, just shy of $1 billion, includes an estimated $34 million to $47 million from savings due to shorter patient stays in hospitals.

In a 2006 Health Affairs article, Needleman estimated the cost impact of lowering nurses’ workloads in nearly 800 hospitals around the country. He agreed with the general methodology behind the Health Policy Commission’s cost estimates, but thinks the report underestimated the savings.

While the commission estimated savings would come in at about 10 percent of the total costs, Needleman said he would expect savings somewhere between 15 to 30 percent. According to Needleman, studies show that lighter nurse workloads reduce the length of time patients stay in the hospital. Needleman said a lot of hospital costs are fixed, so the savings per day of a shorter stay are roughly half of the average daily costs of care.

Shindul-Rothschild, who developed much lower cost estimates of the ballot question for the proponents of Question 1, said the Health Policy Commission inflated estimates of wage growth because the researchers didn’t account for salary increases locked into contracts negotiated by union nurses, who make up the majority of hospital nurses. She also disputed the commission’s estimate of additional staff needed to ensure ratios are maintained during meal and bathroom breaks.

“The HPC added 20 percent more RNs to cover ‘breaks,’” Schindul-Rothschild said. “Just do the math, if the HPC high estimate is for 7 percent more RNs to take care of patients, why do you need to add another 20 percent to cover breaks?” She said the HPC calculations suggest that most Massachusetts hospitals do not currently have enough staff coverage to allow their nurses to take breaks, which, if true, would be a violation of federal labor laws.

Spetz said Shindul-Rothschild’s wage calculations were off mainly because she didn’t include fringe benefits. “That made her estimate severely low,” Spetz said. Spetz also said it is not surprising that Shindul-Rothschild’s calculations varied so dramatically from those of the group opposing Question 1. “These are stakeholder estimates,” she said, an apparent reference to the fact that Shindul-Rothschild prepared estimates for the Massachusetts Nurses Association, an organization she previously headed.

Based on the cost estimates contained in the Health Policy Commission report, members of the commission and opponents of the ballot question have raised the possibility that passage of the measure could prompt hospitals to close wards to cut costs.

Aiken said similar predictions were made in California before the law there took effect in 2004, but they never came to pass. “There is no evidence that hospitals closed in California due to ratios legislation,” she said. Indeed, lawsuits claiming the law forced hospitals to close were ultimately thrown out, due to press reports and other evidence the hospitals were in financial trouble prior to the law’s implementation.

Meet the Author
David Schildmeier, a spokesman for the Massachusetts Nurses Association, said the way the Health Policy Commission developed its report was improper. With past reports, he said, the commission would meet with all sides and conduct its own research. “In this case, the union didn’t know about the report until a week before its release,” he said. “Meanwhile, the HPC was using the hospital’s data, communicating with them, but not us. That’s not right. The HPC is supposed to be objective.”

In an email responding to the union’s claims, commission spokesman Matt Kitsos said the agency was objective. “In publishing its research findings, the HPC did not promote or oppose the pending ballot question, but rather added its independent analysis to the policy discussions on the issue,” he said.