The trusted source for
healthcare information and
Patients with certain conditions have better odds for survival if they go to hospitals that treat a high volume of those conditions. It’s unclear why that is, but experts have some hunches.
Some researchers question the validity of the volume-outcome studies in general. Even as the debate proceeds, however, health care purchasers are paying closer attention to such studies. One group of high-profile purchasers is even discussing the controversial practice of selective referrals to high-volume facilities.
While all this might portend unwelcome changes for low-volume hospitals, the astute ones can maintain their appeal by demonstrating quality with solid metrics, adjusted by case mix. Businesses and biostatisticians trust those measures most — when they can lay their hands on them.
While performing a high volume of certain procedures is important to low mortality, it’s unclear why, says R. Adams Dudley, MD, MBA, lead author of a recent volume-outcome study.1 "We don’t know whether practice makes perfect, or whether the good doctors and hospitals get more referrals." He adds that volume-mortality outcomes are an indirect measure of quality. Based at the University of California-San Francisco, Dudley is assistant professor of medicine, health policy, epidemiology, and biostatistics in the division of pulmonary and critical care and the Institute for Health Policy Studies.
Dudley’s study reviewed the 1997 discharge records of California hospitals. Results show that low case volumes were associated with 602 potentially avoidable deaths from 11 procedures and conditions:
For highly specialized procedures, the volume-outcome connection might seem implicitly clear. However, Dudley points out that even in more routine therapies for breast cancer, there’s evidence that long-term outcomes are better when the procedures are done in a high-volume hospital.
At least one research scientist contends the uncertainties surrounding volume and outcomes go much deeper than causation. In some of the studies, "the data have not been analyzed in a way I would like to see. For example, what patients are being analyzed — high-risk or low-risk?" asks Timothy Hofer, MD, MS. As a research investigator at the Veterans Affairs Center for Practice Management and Outcomes Research in Ann Arbor, MI, Hofer has studied the volume-outcome relationship. One of his reports2 concludes that "under virtually all realistic assumptions for model parameter values, sensitivity was less than 20% and predictive error was greater than 50%."
He also speculates whether there is a volume threshold at which outcomes plateau or deteriorate. Once the medical community understands why the differences exist, it will be in a better position to make policy decisions aimed at giving patients the best possible chance for positive outcomes, Hofer says.
One seemingly obvious solution to the disparity is regionalization. That’s the practice of selective referral by payers and primary care providers to the nearest high-volume facility. Most of the patients in the California study could have reached one by going an extra 10 miles or less. However, those living in rural areas would have had to travel greater distances.
Hofer notes that regionalization is a highly contentious solution to the volume-outcomes issue. Some patients might accept the burden of extra travel for potentially better results. For others, it might be out of the question. But the complexities of regionalization go deeper than acceptability to patients, and they’re not widely discussed, Hofer insists. Of course, certain providers would sustain income losses. "And it’s not obvious that the outcomes will be better in all cases. The dangers of transporting patients might reduce the outcomes," he asserts.
"For patients who need angioplasty immediately, there’s no way we can transport all of them to regional centers in time. Even if the results were as beneficial as some believe, would they outweigh the cost of transportation?" Hofer asks. For some patients, transportation costs could extend throughout follow-up and rehabilitation.
Dudley and colleagues1 suggest that large-scale referrals of patients to high-volume hospitals could disrupt clinical processes or cause treatment delays, both of which could worsen the odds for good outcomes.
All told, Hofer says, it’s a fallacy to jump on regionalization as the only solution to the volume issue. "[But] so far, nobody has come up with an alternative." Dudley would rather see quality measures that drill down to the process and system level because, he says, "There are good low-volume hospitals and bad high-volume hospitals."
Dudley chastises providers for what he sees as a general disregard for practice metrics, volume-related or otherwise. "Clinicians don’t take heed of the real measures of quality, either — those adjusted for case mix. Unfortunately, when there is not an economic incentive, they don’t take heed." Even if providers aren’t paying attention, some of their big customers are.
One group, called Leapfrog, based in Washington, DC, says its mission is "to trigger giant leaps forward in patient safety, as well as the overall quality, customer service, and affordability of health care." Its founders include the Buyers Healthcare Action Group in Minneapolis; General Electric in Fairfield, CT; General Motors in Detroit; the Pacific Business Group on Health in San Francisco; the U.S. Office of Personnel Management in Washington, DC; and the Health Care Financing Ad-ministration in Baltimore, as well as some state Medicaid agencies.
For two years, it operated informally, according to Suzanne Delbanco, PhD, the group’s executive director. "We’ve been talking with large purchasers about using their leverage to improve quality. We finally focused on patient safety because it was a theme that could rally businesses with different cultures."
Late last year, Leapfrog formally invited public and private purchasers to adopt purchasing principles based on quality at the provider level. Subscribers to the principles will agree to do business with providers who meet three patient safety criteria:
Leapfrog has placed a number of qualifiers on evidence-based hospital referrals, Delbanco explains. "We would focus on elective, nonemergency procedures, and the practice would take place in areas of high population," she says. "Quality is not due to the numbers; the numbers are a proxy. The characteristics of high-volume hospitals are the key. Purchasers actually pay more attention to what the evidence shows about good clinical practice. Those are the things that purchasers will reward."
While purchasers won’t start a wholesale shift of patients to high-volume hospitals anytime soon, astute low-volume organizations should start considering their options. "As a low-volume hospital, if you’re confident that your outcomes are good, you can go to the trouble to prove it," explains Mike Killian, director of marketing and public affairs for William Beaumont hospitals in Royal Oak and Troy, MI. This would involve data collection and presentation of your results. "I would also look at the credentials of the nurses and key people in the organization," he advises. "Or you can benchmark against the high-volume hospitals." Review their processes and systems, and bring yours up to their norms. This could improve your outcomes, he notes.
It’s especially challenging, though, for clinicians in isolated facilities to keep their skills sharp for complicated procedures such as surgery for esophageal cancer. Killian suggests that in those cases, the service should be discontinued. "At Beaumont, for instance, we stopped heart transplants because our volumes were not high enough," he explains.
Regardless of its volume, Killian suggests that an organization use quality measurements as in-house tools instead of public relations material. "Don’t spend a lot of time telling people about the quality of your services," he says. "Spend time doing the improvements to achieve good outcomes. Good public relations is in what you do, not in what you say."
While it’s important to understand whether people survive hospitalization, some might carry the question further. From a purchaser’s perspective, it would be ideal to know the functional levels with which patients survive their hospitalization, notes Delbanco. "More research will go in that direction."
But we’re not there yet, Dudley notes. "While functional level is an important measure of quality, the statistical methodology and databases are not refined enough to measure it." And, since each state requires hospitals to report whether patients leave the hospital alive, mortality data are more accessible to researchers, he adds.
Dudley’s research1 bores a bit deeper than mortality. He concedes that in-hospital mortality reductions may not translate into long-term survival benefits. Hospital mortality can be affected by discharge policies and availability of transition care, such as nursing home beds. His report reviews 30-day mortality rates for esophageal cancer, pancreatic cancer, and carotid endarterectomy, as well as five-year mortality for breast cancer. Once again, though, the data suggest that patients treated at higher-volume hospitals have better survival rates.
[For more information on volume-clinical outcomes relationships, contact:
• Mike Killian, Director of Marketing and Public Affairs, Corporate Administration, William Beaumont Hospital, 3601 W. 13 Mile Road, Royal Oak, MI 48073. E-mail: firstname.lastname@example.org.
• Timothy Hofer, MD, MS, Research Investigator, Center for Practice Management and Outcomes Research, Department of Veterans Affairs, Ann Arbor, MI. E-mail: email@example.com.
For more information on the purchaser’s response to the volume-outcomes data, contact:
• Suzanne Delbanco, PhD, Executive Director, The Leapfrog Group, c/o Porter Novelli, 1909 K St., N.W., Washington, DC 20006. Telephone: (202) 973-2953. E-mail: firstname.lastname@example.org.]
1. Dudley RA, Johansen KL, Brand R, et al. Selective referral to high-volume hospitals: Estimating potentially avoidable deaths. JAMA 2000; 283:1,159-1,166.
2. Thomas JW, Hofer TP. Accuracy of risk-adjusted mortality rate as a measure of hospital quality of care. Med Care 1999; 37:83-92.
• Gordon TA, Burleyson GP, Tielsch JM, et al. The effects of regionalization on cost and outcome for one general high-risk surgical procedure. Ann Surg 1995; 221:43-49.