Skip Navigation U.S. Department of Health and Human Services
Agency for Healthcare Research Quality
Archive print banner

Feature Story

This information is for reference purposes only. It was current when produced and may now be outdated. Archive material is no longer maintained, and some links may not work. Persons with disabilities having difficulty accessing this information should contact us at: Let us know the nature of the problem, the Web address of what you want, and your contact information.

Please go to for current information.

Beta blockers substantially improve survival in elderly heart attack patients, but few receive them

Elderly patients 65 years of age and older who receive beta blockers following a heart attack are 43 percent less likely to die within 2 years and 22 percent less likely to be rehospitalized than elderly patients who don't receive this therapy. These beta blocker benefits are similar to those experienced by younger patients. Yet only one in five (21 percent) elderly patients who are eligible to receive beta blockers are prescribed them, according to a recent study supported by the Agency for Health Care Policy and Research (HS07631 and HS06311). The study also found that many physicians substitute calcium channel blockers for the more effective beta blockers, a substitution associated with a doubled risk of death in elderly patients.

This underuse of beta blockers among surviving heart attack patients contradicts clinical guidelines and substantial research findings recommending beta blockers as routine preventive therapy for surviving AMI patients, notes Stephen B. Soumerai, Sc.D., of Harvard Medical School, the study's principal investigator. Dr. Soumerai and his colleagues examined hospital discharge, drug use, and mortality data for 5,332 Medicare beneficiaries in New Jersey, who were 30-day survivors of an acute myocardial infarction (AMI or heart attack). Of these patients, all had insurance coverage for drugs, and 3,737 were eligible for beta blocker therapy. Analysis showed that eligible patients were nearly three times as likely to receive a new prescription for a calcium channel blocker than a beta blocker after their AMI. Advanced age (over 75 years), heart failure, and use of calcium channel blockers often meant that the patient would not receive beta blocker therapy.

This is the first study to demonstrate the survival benefits of long-term beta blocker treatment in patients over age 75, a group that has been excluded from clinical trials. The National Committee for Quality Assurance used the findings from this study as the basis for changing the performance measurement for beta blocker use after AMI to include patients over 75 years of age in the most recent version of the Health Plan Employer Data and Information Set (HEDIS 3). HEDIS is being used by the Health Care Financing Administration to assess the quality of care provided by Medicare HMOs.

More details are in "Adverse outcomes of under use of beta-blockers in elderly survivors of acute myocardial infarction," by Dr. Soumerai, Thomas J. McLaughlin, Sc.D., Donna Spiegelman, Sc.D., and others, in the January 8, 1997, Journal of the American Medical Association 277(2), pp. 115-121.

Return to Contents

Patient Outcomes/Effectiveness Research

New prediction method could cut hospital admissions and costs for pneumonia patients

The Agency for Health Care Policy and Research recently announced a simple and accurate method to predict which patients with pneumonia may be treated at home rather than in a hospital. The prediction method—a clinical model used to help doctors assess the need for hospitalization—also could help reduce the over $4 billion spent annually for inpatient care.

About 600,000, or 15 percent, of the 4 million Americans who develop pneumonia each year are hospitalized. Because of the lack of evidence-based admission criteria and the tendency to overestimate the risk of death, many low-risk patients who could be safely treated outside the hospital are admitted for inpatient care. Hospitalization costs for this disease are estimated to be 10 to 15 times higher than outpatient therapy.

The prediction method is intended to help doctors identify more accurately and easily which pneumonia patients do not require intensive treatment. Projections made from a prospective study by the researchers of roughly 2,300 individuals in Pittsburgh, PA, Boston, MA, and Halifax, Nova Scotia, who were treated for pneumonia at home or in the hospital, suggest that 26 to 31 percent of those hospitalized could have been treated on an outpatient basis, had the model been used. Another 13 to 19 percent of the inpatients could have been hospitalized only briefly for observation.

While the researchers caution that there may be other factors which should be considered before assigning a patient to outpatient care—including patient preferences, the ability to drink fluids and take medication by mouth, presence of a caregiver, or certain severe medical conditions—they conclude that there is enough preliminary evidence to show their prediction method could reduce hospitalizations without jeopardizing patient health and quality of care. In addition, it may assure that high-risk patients are appropriately admitted to the hospital.

The findings, which were reported by Michael J. Fine, M.D., M.Sc., Associate Professor of Medicine at the University of Pittsburgh School of Medicine, and his colleagues from the University of Pittsburgh, Harvard Medical School, and Dalhousie University, are from the AHCPR-funded study on the Assessment of Variations and Outcomes for Community-Acquired Pneumonia (AHCPR grant HS06468), which was directed by Wishwa N. Kapoor, M.D., M.P.H., also of the University of Pittsburgh.

See "A prediction rule to identify low-risk patients with community-acquired pneumonia," by Dr. Fine, Thomas E. Auble, PH.D., Donald M. Yealy, M.D., and others, in the January 23, 1997, issue of the New England Journal of Medicine, pp. 243-250.

AHCPR's pneumonia study is part of a major effort to reduce physician uncertainty and improve treatment effectiveness. The findings are summarized in two fact sheets—one for clinicians and one for consumers—that are available online. Print copies of What's New for Clinicians: Pneumonia, New Prediction Model Proves Promising (AHCPR Publication No. 97-R031) and What's New for Consumers: Pneumonia, More Patients May Be Treated at Home (AHCPR Publication No. 97-R030) are available from the AHCPR Publications Clearinghouse.

A limited number of copies of this Patient Outcomes Research Team's final report, Community-Acquired Pneumonia, are available from the AHCPR Publications Clearinghouse (AHCPR Publication No. 97-N009). The report also can be purchased from the National Technical Information Service (NTIS accession number PB97-117808; 109 pp, $28.00 paper, $14.00 microfiche).

Return to Contents

Hospital Use/Quality of Care

Areas with higher hospitalization rates do not necessarily have more inappropriate hospital admissions

Many cost-containment strategies are based on the presumption that hospital service areas (and the hospitals within them) with higher admission rates for the same clinical conditions than lower rate areas have more inappropriate admissions (that is, admissions that could be treated adequately on an outpatient basis) and are good targets for cost-containment. However, a recent study supported by the Agency for Health Care Policy and Research (HS06048) suggests otherwise. It shows that areas with high hospitalization rates for elderly persons (aged 65 and older) did not have a higher proportion of inappropriate admissions than did low-rate areas. This finding suggests that small-area variation studies may be of limited use in targeting areas for cost-containment strategies, notes Joseph Restuccia, Dr.P.H., of Boston University, the study's principal investigator.

Dr. Restuccia and his colleagues tested this hypothesis on elderly patients admitted to hospitals in 70 small geographic areas in Massachusetts from 1990 to 1992 for eight conditions. They calculated the relative hospitalization rate, that is, observed versus expected number of admissions for each condition, given the patient age and sex distribution in the area and the Statewide age- and sex-specific hospitalization rates for these conditions.

Based on the Appropriateness Evaluation Protocol (AEP) criteria applied by the Massachusetts Peer Review Organization and the clinical judgment of MassPRO physician reviewers, the rate of inappropriate admissions among patients admitted for a medical diagnosis averaged 13 percent and ranged among the 70 small areas from 3.8 percent to 21.3 percent. The AEP-based and clinician-based inappropriateness rates showed no relationship to the relative hospitalization rate, even after adjusting for factors such as supply and demand that may have masked the relationship between inappropriateness and hospital admission rates.

The study also involved development and application of protocols to a sample of elderly patients in the 10 highest and 10 lowest rate areas to judge whether hospital admission was medically necessary to treat certain conditions and the appropriateness of performing several diagnostic procedures. For cardiac catheterization, 17.7 percent of admissions failed admission criteria in high rate areas, and 15.8 percent failed in low rate areas, an insignificant difference; corresponding percentages for revascularization were 7.1 percent and 9.9 percent, again insignificant.

More details related to the study of medical admissions are in the journal article, "High hospital admission rates and inappropriate care: Does inappropriate use of hospital care explain much of the small-area variation in admission rates?" by Dr. Restuccia, Michael Shwartz, Ph.D., Arlene Ash, Ph.D., and Susan Payne, Ph.D., in the Winter 1996 issue of Health Affairs 15(4), pp. 156-163.

A complete report of this project, including findings related to specific medical conditions and cardiac procedures, can be found in the grant final report, Inappropriateness and Variations in Hospital Use, which is available from the National Technical Information Service (NTIS accession no. PB96-189899, abstract, executive summary, final report, and appendixes A-E, 220 pp., $41.00 paper; $19.50 microfiche).

Severity of illness measures and length of stay found to be poor indicators of hospital quality of care

Hospitals frequently are judged on their quality of care based on length of hospital stay for a particular condition and on mortality data, that is, how many patients die versus how many would be expected to die in their condition. Some hospitals rightfully argue that they have much sicker patients and that their higher mortality rates and longer hospital stays reflect sicker patients, not poorer quality of care. As a result, many hospitals, insurance companies, and other payers use severity-of-illness measures to adjust hospital outcomes and length of stay.

Two recent studies, supported by the Agency for Health Care Policy and Research (HS06742), show that different measures produce different expected-versus-actual hospital mortality rates and lengths of stay, so that hospital quality of care determinations will depend on the measure used. Nor are hospital readmission rates reliable indicators of hospital quality of care, according to a third study, supported in part by AHCPR (HS07345). The three studies are summarized here.

Iezzoni, L.I., Ash, A.S., Schwartz, M., and others (1996, October). "Judging hospitals by severity-adjusted mortality rates: The influence of the severity-adjustment method," American Journal of Public Health 86(10), pp. 1379-1387.

This study of 100 acute care hospitals nationwide used 10 different severity measures to assess the severity of illness of 11,880 adults admitted in 1991 for acute myocardial infarction (AMI) and to calculate expected versus observed AMI death rates at each hospital. Unadjusted mortality rates ranged from 5 percent to 26 percent for the 100 hospitals. Severity measure agreement ranged from only fair to good on which hospitals had the highest (more deaths than expected) or lowest (fewer deaths than expected) mortality rates. Severity measures based on medical records frequently disagreed with measures based on discharge abstracts. Thirty-two hospitals had significantly better or worse death rates than expected using one or more, but not all 10, severity measures.

Thus, whether individual hospitals were identified as especially good or poor frequently depended on the particular severity measure used. This is a significant problem because hospitals and third-party payers often purchase a software package of one severity measure and use it to unilaterally judge hospital quality of care, explains Lisa I. Iezzoni, M.D., M.Sc., of Harvard Medical School.

These findings raise important questions for report card efforts to judge hospital performance by using severity-adjusted death rates. Dr. Iezzoni and colleagues conclude that using such information to guide punitive actions against providers is not indicated until its meaning is clearer.

Shwartz, M., Iezzoni, L.I., Ash, A.S., and Mackiernan, Y.D. (1996, October). "Do severity measures explain differences in length of hospital stay?" HSR: Health Services Research 31(4), pp. 365-385.

The researchers used 14 severity-of-illness measures to evaluate hospitals expected versus observed length of stay for elderly hip fracture patients. Hospitals are reimbursed a set amount by Medicare based on a patients diagnosis, called the diagnosis-related group (DRG). Hospitals have argued that they should be reimbursed more for more severely ill patients in the same DRG. In this study, the researchers used different severity measures to evaluate whether the measures had any impact on one cost factor, hospital length of stay (LOS). They used each measure to evaluate expected LOS for 5,664 elderly patients admitted to 80 hospitals nationwide in 1992 for surgical repair of a hip fracture.

They found that the patients had a mean LOS of 12 days. Observed average LOS for hospitals ranged from 8 to 24 days. Although the 14 severity measures showed excellent agreement in ranking hospitals based on z-scores (observed average LOS minus expected LOS), no severity measure explained the differences between hospitals with the shortest and longest LOS. The researchers conclude that severity measures did little to explain LOS differences among hospitals that could justify higher reimbursements for sicker hip fracture patients and are of little value in identifying the best and worst hospitals in terms of average LOS for these patients.

Thomas, J.W. (1996, Fall). "Does risk-adjusted readmission rate provide valid information on hospital quality?" Inquiry 28, pp. 258-270.

The presumption is that poorer quality of care during hospitalization is more apt to lead to rehospitalization than better quality of care. However, this study did not demonstrate a link between quality of care and hospital readmission rates. J. William Thomas, Ph.D., of the University of Michigan, used Medicare patient discharge data from Michigan hospitals for 1989 to 1991 to develop condition-specific models of 12 clinical conditions. He used the models to calculate readmission risk probabilities for patients whose medical records had been reviewed for quality of care by a Michigan peer review organization. Results showed that readmission rates were essentially the same for patients with each of the 12 clinical conditions studied who were judged to have received poor quality care and those whose care was judged acceptable. This was true for unadjusted readmission rates and rates that were adjusted for patients demographic and clinical characteristics. Dr. Thomas concludes that readmission rates do not provide valid information on hospitals' quality of care performance.

Return to Contents

Patient Care/Clinical Decisionmaking

Except for Medicare enrollees, private insurance does not seriously affect patients' choice of physician type

With the exception of Medicare beneficiaries, those with health insurance in 1987 were no more apt than the uninsured to seek care by specialists instead of generalists. However, people insured by health maintenance organizations (HMOs) were less likely than others to seek specialist care. This suggests that HMOs were using generalist physicians in gatekeeping roles as early as 1987, and that HMO enrollees received somewhat different styles of care than prevailed in fee-for-service medicine at that time, according to researchers at the Agency for Health Care Policy and Research.

Mechanisms may have been in place for the uninsured in 1987—through a combination of self-payment, charity care, and cross-subsidized care—that enabled them to obtain the same mix of specialist/generalist physician services as the insured received, explain Philip F. Cooper, Ph.D., and Amy K. Taylor, Ph.D., of AHCPR, and Len M. Nichols of the Urban Institute. They point out that insurance coverage per se is not the key variable in determining access to specialist services, or at least this was not the case in 1987. Instead, practice style could be much more important in determining the type of physician services that patients receive. The researchers studied a nationally representative sample of outpatient visits to physicians, using data from the 1987 National Medical Expenditure Survey and the American Medical Association Masterfile.

Their analysis also showed that well-to-do women under 65 years of age were more apt to choose specialists over generalists than poor women, but these differences did not exist for women over age 65, supporting the argument that Medicare, regardless of supplemental coverage, tends to equalize access to specialists across income classes. However, Medicare enrollees with private supplemental insurance were more likely to gain access to specialist care. Physician characteristics, such as age and board certification, also affected the probability of visiting specialists. The researchers were unable to determine whether these differences were due to patient preferences, physician supply, or a combination of the two.

See "Patient choice of physician: Do health insurance and physician characteristics matter?" by Drs. Cooper, Nichols, and Taylor, in the Fall 1996 issue of Inquiry 33, pp. 237-246. Reprints (AHCPR Publication No.97-R032) are available from the AHCPR Publications Clearinghouse.

Certain findings on lung x-rays raise the probability of short-term death from pneumonia

Lung x-rays are commonly used to diagnose the severity of pneumonia. When these x-rays show accumulation of fluid in the lining of both lungs (bilateral pleural effusions), the likelihood of dying from pneumonia within 1 month increases substantially. This is the conclusion of a study by the Pneumonia Patient Outcomes Research Team (PORT), which is led by Wishwa N. Kapoor, M.D., and supported by the Agency for Health Care Policy and Research (HS06468). This x-ray finding can help physicians assess the prognosis of patients with pneumonia and decide the initial site and intensity of care for patients who are hospitalized for the condition, notes Peggy B. Hasley, M.D., of the University of Pittsburgh, the paper's lead author.

The researchers used a panel of radiologists to evaluate the x-ray patterns of 1,906 outpatients and inpatients with community-acquired pneumonia, who were enrolled in a multicenter, prospective study. Of these patients, 9 percent had x-ray evidence of fluid in the lung—unilateral for two-thirds of patients, and bilateral for one-third of patients. In univariate analyses, x-ray evidence of fluid in both the right and left lung increased the risk of death within 1 month by seven-fold (risk ratio, or RR of 6.5; 1 is equal risk). Univariate analyses also revealed that a higher risk of death was associated with x-ray evidence of moderate to massive pleural effusion (RR of 3.4), radiographic infiltrate involving accumulation of fluid in two lobes (RR of 2.2) or three or more lobes (RR of 3.1), bilateral infiltrates (RR of 2.8), and bronchopneumonia (RR of 1.6).

After controlling for patient demographic and other clinical factors known to be associated with death in patients with community-acquired pneumonia, only x-ray evidence of bilateral pleural effusions remained independently associated with death within 30 days.

For more information, see "Do pulmonary radiographic findings at presentation predict mortality in patients with community-acquired pneumonia?" by Dr. Hasley, Michael N. Albaum, M.D., Yi-Hwei Li, Ph.D., and others, in the October 28, 1996, Archives of Internal Medicine 156, pp. 2206-2212.

Interactive videos help heart disease patients choose the treatment that's right for them

Interactive video programs can help undecided patients with heart disease to select a treatment and increase confidence in their treatment decision, according to a pilot study by the Ischemic Heart Disease Patient Outcomes Research Team (PORT) supported by the Agency for Health Care Policy and Research (HS06503). Patients who are more active in decisions affecting their health generally fare better than more passive patients, explain the Duke University Medical Center researchers, who are led by Elizabeth R. DeLong, Ph.D. In fact, patient participation is receiving increasing attention by clinicians as a means to improve patient outcomes. One approach is use of an interactive video such as the Shared Decision-Making Program (SDP) for Ischemic Heart Disease (IHD).

The SDP-IHD compares three treatments: medication, angioplasty, and bypass surgery, through a physician narrator, patient testimonials, and empirically based, patient-specific outcome estimates of short-term complications and long-term survival. The video's interactive format uses a videodisk player, microcomputer, touchscreen video monitor, and printer, which can be wheeled to the patient's bedside.

On the basis of each patient's severity of illness data, the SDP estimates the risk of in-hospital mortality for bypass surgery and angioplasty and 5-year survival with all three treatments. The program also provides an overview of ischemic heart disease, treatment descriptions, and explanations of possible complications.

The SDP helped 7 of 16 (44 percent) initially undecided patients to select a treatment. In addition, 10 of the 60 patients in the study (16 percent) changed their initial treatment choice after viewing the SDP. Although the SDP increased viewers confidence about their treatment decisions, it also increased anxiety among some patients. Nevertheless, patients found the SDP to be more helpful than all other decision aids except for their physician's advice.

The SDP increased patient involvement, but it did not decrease patient preference for revascularization procedures such as angioplasty or bypass surgery over medication. In fact, most patients who initially preferred revascularization did not change their preference after viewing the program. This may be because the SDP's models demonstrate long-term survival advantages of revascularization over medication for many individuals. These findings were based on 60 patients diagnosed with significant coronary artery disease at two hospitals in 1994 and 1995 who watched and rated the video.

Details are in "Impact of an interactive video on decision making of patients with ischemic heart disease," by Lawrence Liao, B.A., James G. Jollis, M.D., Dr. DeLong, and others, in the Journal of General Internal Medicine 11, pp. 373-376, 1996.

Return to Contents
Proceed to Next Article

The information on this page is archived and provided for reference purposes only.


AHRQ Advancing Excellence in Health Care