John Eisenberg takes the helm as AHCPR's new Administrator

John M. Eisenberg, M.D., has been appointed as the new Administrator of the Agency for Health Care Policy and Research. Dr. Eisenberg was a founding Commissioner of the Congressional Physician Payment Review Commission from 1986 through 1995 and served as its Chairman from 1993 to 1995.

Dr. Eisenberg also will serve as the Senior Advisor to the Secretary on Quality, with AHCPR designated as the Department of Health and Human Services' (HHS) lead agency for health care quality improvement issues. One of his early priorities will be to coordinate HHS work on behalf of the Secretary regarding the National Advisory Commission on Consumer Protection and Quality in the Health Care Industry and to chair an interagency committee on quality.

A clinician and researcher, Dr. Eisenberg has held a number of key positions in both academic and clinical medicine. Most recently, he was Chairman of the Department of Medicine, Physician-in-Chief, and Anton and Margaret Fuisz Professor of Medicine at Georgetown University Medical Center. Previously, he served as Chief of the Division of General Internal Medicine and was Sol Katz Professor of General Internal Medicine at the University of Pennsylvania.

Dr. Eisenberg was the first physician to be elected President of the Association for Health Services Research, serving in that capacity in 1991 and 1992, and also was President of the Foundation for Health Services Research. He has been President of the Society for General Internal Medicine, and Vice President of the Society for Medical Decision Making. Dr. Eisenberg is a member of the Institute of Medicine of the National Academy of Sciences. He currently serves on the Board of Regents of the American College of Physicians and has been elected a Master of the College.

Dr. Eisenberg has published over 200 articles and book chapters on topics such as physicians' practices, test use and efficacy, medical education, and clinical economics. His book, Doctor's Decisions and the Cost of Medical Care, was published in 1986. He was co-author of Paying Physicians, published in 1992, and co-editor of The Physicians Practice, published in 1980.

He graduated magna cum laude from Princeton University in 1968 and from the Washington University School of Medicine in St. Louis in 1972. After his residency in internal medicine at the University of Pennsylvania, he was a Robert Wood Johnson Foundation Clinical Scholar and attended the Wharton School where he earned a Master of Business Administration degree with distinction.

Dr. Eisenberg replaces Clifton R. Gaus, Sc.D., who recently resigned as AHCPR Administrator.

Return to Contents

Hospitalization

Adverse drug events cost an average teaching hospital $5.6 million a year

Dangerous and often life-threatening reactions to medications affect 6.5 percent of hospitalized patients, and 28 percent of these reactions are preventable. These adverse drug events (ADEs) often lead to prolonged hospital stays and higher charges; each ADE is associated with $2,595 in additional costs to a hospital, and preventable ADEs account for nearly twice that amount ($4,685). This translates into $5.6 million ($2.8 million for preventable ADEs alone) per year for an average 700-bed teaching hospital, according to a study supported in part by the Agency for Health Care Policy and Research (HS07107). The substantial costs attributable to ADEs, not including the costs of injuries to patients and malpractice costs, justify an investment in efforts to prevent them, concludes David W. Bates, M.D., M.Sc., of Brigham and Women's Hospital, the study's lead author.

Dr. Bates and his colleagues assessed the additional medical resources used by patients harmed by drug treatment (ADEs) by observing 4,108 patients admitted to two tertiary-care hospitals over a 6-month period. They compared the resources used by patients with an ADE with similar non-ADE patients being treated on the same unit. Patients who suffered an ADE stayed in the hospital 2.2 days longer, and patients who suffered preventable ADEs stayed 4.6 days longer than the non-ADE patients. Improving the systems by which drugs are ordered and administered could prevent many of these ADEs, suggest the investigators.

In a previous study, they identified several factors associated with drug prescribing errors. These include failure to notice or account for a decline in the patient's kidney or liver function that would require alteration of drug therapy; failure to notice or account for a history of allergy to the same medication class; using the wrong drug name, dosage form, or abbreviation; incorrect dosage calculations; and atypical or unusual dosage frequency considerations.

More details are in "The costs of adverse drug events in hospitalized patients," by Dr. Bates, Nathan Spell, M.D., David J. Cullen, M.D., M.Sc., and others, in the January 22/29 Journal of the American Medical Association 277(4), pp. 307-311.

Slightly more than half of patients who want life-sustaining treatment receive it

A patient's desire to receive costly life-sustaining treatment has little influence over whether or not the patient actually receives it, according to a study supported by the Agency for Health Care Policy and Research (HS06655). The majority of patients want to receive a life-sustaining treatment—such as cardiopulmonary resuscitation (CPR), ventilatory support, or nasogastric tube feeding—if it will prolong life for any length of time. And, many patients want to have such treatment even if it has little chance of success, explains Marion Danis, M.D., of the University of North Carolina.

Dr. Danis and colleagues interviewed 244 hospitalized patients (50 years of age and older) with a 6- to 12-month life expectancy due to end-stage heart, lung, or liver disease, metastatic cancer, or lymphoma. They followed the patients for 6 months to correlate patient wishes for and subsequent use of life-sustaining treatment during hospitalization. About 58 percent of patients wanted life-sustaining treatment to prolong life for even 1 week. But analysis of 245 subsequent hospitalizations showed that patients desiring such treatment were not significantly more likely to get it than those who did not want it (55 percent vs. 51 percent).

This study demonstrated that patient preferences about the use of life-sustaining treatment did not determine the use of these treatments. While the use of such treatment, particularly intensive care, led to increased hospital costs, these higher costs were not the consequence of patient preferences. One explanation is that doctors were the predominant decisionmakers, and doctors often are reluctant to order procedures that will not prolong life but will increase suffering. Also, factors such as availability of resources (e.g. ICU beds) and provider attitudes influence use of these treatments.

If the amount of patient-physician disagreement about life-sustaining treatment is minimal, as in this study, and care is satisfactory to patients and their families, it may be that patient and family concerns are not necessarily being ignored, note the researchers. They caution, however, that if the use and cost of life-sustaining treatments are not determined by patient preferences, then medical futility policies which override patient wishes may yield little savings and much divisiveness.

See "A prospective study of the impact of patient preferences on life-sustaining treatment and hospital cost," by Dr. Danis, Elizabeth Mutran, Ph.D., Joanne M. Garrett, Ph.D., and others, in Critical Care Medicine 24(11), pp. 1811-1817, 1996.

New classification system for rehabilitation patients provides basis for prospective payment

A new classification system that characterizes particular medical rehabilitation problems shows promise as the basis for another prospective payment system (PPS) similar to the diagnosis-related groups (DRGs) system used by the Health Care Financing Administration (HCFA).

Hospitalized medical rehabilitation patients are currently excluded from HCFA's DRG-based PPS, principally because they differ clinically from and tend to have longer lengths of stay (LOS) than other acute care inpatients. Instead, a different system reimburses medical rehabilitation hospitals and units based on "reasonable" costs per patient, which are subject to a maximum amount regardless of the severity of the case. Because this payment system fails to adjust for a facility's case-mix, it encourages providers to admit the least complex and least disabled patients, potentially denying access to those most in need.

The new system, called the Functional Independent Measure-Function Related Group (FIM-FRG), classifies medical rehabilitation inpatients by complexity, within categories of impairment. The impairment categories distinguish among such conditions as stroke, traumatic brain dysfunction, nontraumatic spinal cord dysfunction, rheumatoid arthritis, amputation, and major multiple trauma.

The FIM-FRG system explains more than 30 percent of variations in length of stay (LOS) for inpatient rehabilitation, according to the final report of a study supported by the Agency for Health Care Policy and Research (HS07595). Use of the system would more adequately scale the level of resources to the clinical needs of patients. It also would encourage the timely discharge of patients and thus provide incentives for cost containment, according to Margaret G. Stineman, M.D., of the University of Pennsylvania, the study's principal investigator. HCFA is currently evaluating the FIM-FRG system for use in the development of a PPS for medical rehabilitation.

The system was developed using discharge data on 85,447 medical rehabilitation patients. The FIM-FRGs first classify patients into one of 20 impairment categories based on the principal medical reason for rehabilitation and then into 65 groups based on functional severity and age at admission to rehabilitation. Like the DRGs, the system would not necessarily reinforce high quality of care. However, Dr. Stineman and her colleagues are working on an outcomes-based FRG system that could be used in conjunction with the LOS-based FIM-FRGs to encourage both high quality and cost-effective rehabilitation.

Copies of the report, Function-Based Rehabilitation Classification, by Dr. Stineman, are available from the National Technical Information Service. Abstract, executive summary, and final report, are available from the National Technical Information Service (NTIS accession no. PB97-131841; 73 pp., $21.50 paper, $10.00 microfiche).

Return to Contents

Patient Outcomes/Effectiveness Research

Patients over 80 benefit most from stroke prevention drug but are least likely to get it

Irregular and often rapid heart beat—or atrial fibrillation—raises the incidence of stroke five-fold, a risk that can be almost entirely removed with long-term use of the anticoagulant warfarin. Despite warfarin's increased use and proven efficacy in preventing strokes in patients with atrial fibrillation, less than 40 percent of such patients were taking warfarin in the early 1990s, according to a study supported in part by the Agency for Health Care Policy and Research (HS07892).

The study found a substantial overall increase in the use of warfarin from 1980 to 1993 (from 7 to 32 percent). But in more recent years, patients 80 years or older with atrial fibrillation were significantly less likely to be prescribed warfarin than younger patients (19 percent vs. 36 percent), even though older patients have the most to gain from its use.

For old and young alike, warfarin remains a complicated therapy whose management imposes burdens on both patients and physicians. This may explain the mere 32 percent of atrial fibrillation patients who were prescribed warfarin in 1992 and 1993. This rate is probably suboptimal given the benefit of anticoagulation in preventing strokes, note Randall S. Stafford, M.D., Ph.D., and Daniel E. Singer, M.D., of Harvard Medical School. They analyzed 1,062 visits by patients with atrial fibrillation to randomly selected office-based physicians included in the National Ambulatory Medical Care Surveys in 1980, 1981, 1985, and 1989 through 1993 to explore national patterns of warfarin and aspirin use.

In 1992 and 1993, anticoagulation therapy was much more likely to be reported in visits to cardiologists (32 percent) and general internists (40 percent) compared with general and family practitioners (15 percent), but reports of its use were similar in women (34 percent) and men (30 percent). Aspirin use increased from 3 percent to 10 percent and showed little overlap with warfarin use. Regional use varied dramatically, with the Midwest (46 percent), West (30 percent), and Northeast (35 percent) having higher rates than the South (16 percent). A more recent year, residence outside the South, patient age 65 to 74 years, and visits to cardiologists and internists increased the independent likelihood of warfarin use.

See "National patterns of warfarin use in atrial fibrillation," by Drs. Stafford and Singer, in the December 9, 1996 Archives of Internal Medicine 156, pp. 2537-2541.

Initial symptom severity predicts outcomes of most men with BPH

About 40 percent of men in their 70s develop benign prostatic hyperplasia (BPH), an enlarged prostate that causes a range of urinary tract symptoms. Prostatectomy (surgical removal of the prostate) is commonly performed on these men. However, recent practice guidelines issued in the United States and internationally suggest "watchful waiting" can be considered along with surgery for men with BPH, even those who have moderate to severe symptoms. According to a recent study supported by the Agency for Health Care Policy and Research (HS06336, HS06540, and HS08397), outcomes for most men who defer surgery depend on the severity of their initial symptoms.

Of the 60 men who began the study with mild symptoms and had 4 full years of followup, 10 percent had undergone prostatectomy within 4 years, 27 percent were taking medication for BPH, and 63 percent remained off active treatment, compared with 24 percent, 31 percent, and 45 percent, respectively, of the 245 men who began with moderate symptoms, and 40 percent, 27 percent, and 33 percent, respectively, of the 66 men who began the study with severe symptoms.

These findings are based on research conducted by the Prostate Patient Outcomes Research Team led by Michael J. Barry, M.D., of Massachusetts General Hospital and his colleagues at the University of Massachusetts, Boston; Colorado Permanente Medical Group, Denver; and the Group Health Cooperative of Puget Sound, Tacoma, WA. For 4 years, they followed the symptom progression and outcomes of 500 men with BPH who were candidates for prostatectomy but who were initially treated nonsurgically in five North American urology practices.

Details are in "The natural history of patients with benign prostatic hyperplasia as diagnosed by North American urologists," by Dr. Barry, Floyd J. Fowler, Jr., Ph.D., Lin Bin, Ph.D., and others in the January 1997 issue of The Journal of Urology 157, pp. 10-15.

Magnesium sulfate appears promising for preventing cerebral palsy in preterm infants

Preterm births account for most infant deaths and many long-term neurologic problems, including cerebral palsy and mental retardation. Infants born weighing less than 1500 g (3 pounds) are most at risk for these problems. About 5 percent of infants weighing less than 1,000 g (about 2 pounds) and as many as 25 percent of the smallest survivors (500 to 600 g, about 1 pound) will have cerebral palsy.

Several retrospective and case control studies have found an association between predelivery administration of magnesium sulfate to mothers and reduced incidence of cerebral palsy in their newborns. Proof of efficacy, however, is lacking. In the interim, magnesium sulfate should not be used to prevent cerebral palsy, warns the Low Birthweight Patient Outcomes Research Team (PORT), which is supported by the Agency for Health Care Policy and Research (PORT contract 290-92-0055).

In a recent article exploring this issue, PORT researchers Robert L. Goldenberg, M.D., and Dwight J. Rouse, M.D., of the University of Alabama at Birmingham, point out that mothers have been placed in serious jeopardy or even died because of excess magnesium administration, and many newborns are thought to have experienced depressed respiration for the same reason. The researchers recommend that clinicians wait for the results of randomized controlled trials on the efficacy of magnesium sulfate to prevent cerebral palsy that are planned or already underway.

The probable benefit of magnesium sulfate may be its action on three major pathways that have been implicated in brain damage and cerebral palsy in very preterm infants. In the first, hemorrhage occurs in the immature vasculature of the periventricular area of the brain during or soon after birth, presumably as a result of the poorly regulated cerebral blood flow of the preterm brain. In the second pathway, perinatal hypoxia combined with altered cerebral blood flow results in damage to periventricular neurons due to asphyxia. Finally, in utero infection stimulates cytokine production, which may adversely affect periventricular neurons causing hemorrhage or other problems. Magnesium may interrupt these pathways by stabilizing vascular tone, reducing reperfusion injury following restricted blood flow, or reducing cytokine and bacterial toxin synthesis.

For more information, see "Preterm birth, cerebral palsy, and magnesium, by Drs. Goldenberg and Rouse, in the February 1997 issue of Nature Medicine 3(2), pp. 146-147.

Researchers examine diagnostic techniques and therapies for acute low back pain

Low back pain affects up to 80 percent of adults at some time during their lives and is a leading cause of disability and lost productivity. The Low Back Pain Patient Outcomes Research Team (PORT), which was supported by the Agency for Health Care Policy and Research (HS06344) and led by Richard A. Deyo, M.D., M.P.H., of the University of Washington, focused its efforts on identifying better ways to diagnose and manage this condition. The PORT project was completed in 1995.

Dr. Deyo also is principal investigator for a followup study, Effectiveness of Treatment Strategies for Low Back Pain (HS08194), which is building on and extending the work of the Low Back Pain PORT. Dr. Deyo and his colleagues recently published the articles summarized below.

Carey, T.S., Garrett, J., and the North Carolina Back Pain Project. (1996). "Patterns of ordering diagnostic tests for patients with acute low back pain." Annals of Internal Medicine 125(10), pp. 807-814.

Nearly half of patients suffering from acute low back pain receive x-rays—usually within 2 weeks of first seeking care for the back pain—to diagnose the source of the problem. About 9 percent receive computed tomography (CT) or magnetic resonance imaging (MRI); slightly less than one-third of these patients receive CTs or MRIs within the first 2 weeks. Many of these tests are unnecessary, and they are done too early and too frequently, according to this study. Clinical guidelines and research studies advise physicians to use x-rays and other imaging technologies conservatively when diagnosing patients with acute low back pain because it often disappears on its own in about 4 weeks.

Patients with relatively long-standing low back pain plus sciatica (pain at the level of the knee or below) are more reasonable candidates for x-rays. Yet this study shows that 20 percent of patients with neither characteristic received x-rays from primary care doctors in North Carolina. Also, 62 percent of patients seen by chiropractors and 70 percent of those seen by orthopedic surgeons in that State received x-rays to diagnose the source of acute low back pain, regardless of the patient's clinical status.

Patients who had pain that began more than 2 weeks before seeing a health care provider and no previous episodes of low back pain were more apt to have an x-ray. Patients who were white; had neurologic deficit, sciatica, or poor functional status (for example, found it difficult to sit or walk); or were treated in a small group practice were more apt to have CT or MRI. Even though patients with sciatica have a more prolonged recovery, little is lost by delaying diagnostic testing for several weeks to determine whether spontaneous recovery will occur, as it does for 55 percent of these patients, according to Timothy S. Carey, M.D., M.P.H., of the University of North Carolina. Dr. Carey and his colleagues studied factors associated with use of x-rays, CTs, or MRIs in 1,580 patients with acute low back pain at community-based, solo, and group practices in North Carolina.

Taylor, V.M., Deyo, R.A., Ciol, M., and Kreuter, W. (1996). "Surgical treatment of patients with back problems covered by workers' compensation versus those with other sources of payment." Spine 21(19), pp. 2255-2259.

Up to 2 percent of the industrial workforce suffers a back injury qualifying for workers compensation each year. These back injuries have been estimated to exceed $50 billion in total annual costs for direct health care and indirect costs for work loss and disability payments. Earlier studies have suggested that patients receiving workers compensation may be overdiagnosed more often than other back pain patients, resulting in excessive surgical intervention. This study revealed similar findings. It shows that in Washington State, patients insured through workers compensation were almost 40 percent more likely to undergo spinal fusion compared with other low back pain patients and almost twice as likely to have subsequent reoperation within 3 years of the initial surgery (18 percent vs. 10 percent). This was true even after adjusting for other factors affecting the likelihood of fusion surgery, such as patient age, sex, coexisting medical conditions, diagnosis, and specialty of the surgeon. These findings reinforce the need for a standardized system of surgical selection criteria, conclude the researchers. They used data from Washington State's automated hospital discharge system for 1988 through 1991. The study group included 1,502 back surgery patients receiving workers' compensation and 2,674 patients with other sources of payment for their care.

Deyo, R.A. (1996, November). "Acute low back pain: A new paradigm for management." British Medical Journal 313, pp. 1343-1344.

In this commentary, Dr. Deyo discusses acute low back pain guidelines recently issued by Britain's Royal College of General Practitioners, which are similar to the AHCPR-sponsored clinical practice guideline on acute low back problems released in late 1994. Gone are the days when strong pain relievers and bed rest were prescribed, notes Dr. Deyo. The new approach to managing acute back pain for otherwise well patients aged 20 to 55 years with pain that does not radiate below the knee is as follows: x-rays, other imaging, and specialist referral are unnecessary; bed rest is not recommended; patients should stay as active as possible and continue normal daily activities; drugs should be prescribed at regular intervals, not as required, and should begin with paracetamol (similar to acetaminophen, but with a slightly different formula) or nonsteroidal antiinflammatory drugs, avoiding narcotics if possible; and finally, spinal manipulation may be considered for relief of symptoms within 6 weeks of onset. Patients who have not returned to ordinary activities and work within 6 weeks should be referred for an exercise program.

Pneumonia PORT publishes recent findings

About 4 million cases of community-acquired pneumonia (CAP) occur in the United States each year, with about one-third of patients hospitalized. The Pneumonia Patient Outcomes Research Team (PORT) was supported by the Agency for Health Care Policy and Research (HS06468) to evaluate geographic variations in hospitalization rates and to determine whether hospitalizations, intensive care unit admissions, diagnostic tests, and certain treatments and costs could be reduced without adversely affecting the health of pneumonia patients.

Led by Wishwa N. Kapoor, M.D., M.P.H., of the University of Pittsburgh, the Pneumonia PORT recently published three studies, discussed here, to explore factors that physicians consider when deciding whether to hospitalize patients with CAP and when to discharge them from the hospital and to compare presenting symptoms, radiographic manifestations, an clinical outcomes of patients with CAP of "atypical" and undetermined origin.

Fine, M.J., Hough, L.L., Medsger, A.R., and others. (1997, January). "The hospital admission decision for patients with community-acquired pneumonia." Archives of Internal Medicine 157, pp. 36-44.

Until recently, reasons for deciding whether patients with CAP should be hospitalized have not been well-defined, resulting in large variations in hospital admission rates. In this study, 292 medical practitioners at four geographically separate hospitals completed questionnaires to identify the factors the clinicians considered when deciding whether or not to hospitalize a CAP patient at low risk for short-term mortality. Low-risk CAP patients are identified as those with a predicted probability of death of less than 4 percent. Results showed that practitioners relied heavily on five clinical factors to make this decision: the patient's respiratory status, coexisting illness, clinical appearance, lung involvement of more than one lobe, and oral intake. Three patient factors were almost always associated with hospitalization: hypoxemia; inability to maintain oral intake of foods, liquids, and antimicrobial medicines; and lack of patient home care support. Patients without these risk factors, but whom practitioners estimated still had more than a 5 percent risk of death, also were more likely to be hospitalized.

However, availability of home intravenous antibiotic therapy and home nursing services would have allowed outpatient treatment of more than half (68 percent and 59 percent, respectively) of CAP patients initially hospitalized for treatment.

Although practitioners also considered psychosocial factors—such as patient reliability, availability of home support, patient's ability to pay, patient and/or family apprehension, and other factors—they rarely considered patient preferences in making the decision about hospitalization. According to the researchers, these findings suggest that improvements in practitioners' ability to identify low-risk patients, greater reliance on patient preferences, and increased availability of ambulatory medical services could facilitate outpatient treatment for a large proportion of low-risk CAP patients.

Fine, M.J., Medsger, A.R., Stone, R.A., and others. (1997, January). "The hospital discharge decision for patients with community-acquired pneumonia." Archives of Internal Medicine 157, pp. 47-56.

Although less than half of all patients with CAP are hospitalized, length of hospital stay is a principal determinant of total medical care costs for all patients with this illness. In this study, the Pneumonia PORT researchers examined the factors considered by physicians in making the hospital discharge decision. Physicians were surveyed to determine the reasons for extending the hospital stay of patients beyond clinical stability and the medical services that could have allowed earlier hospital discharge. The 168 physicians participating in this survey identified the following factors as "very important" in delaying discharge of clinically stable patients: diagnostic evaluation or treatment of comorbid illness (56 percent), completion of a standard course of antimicrobials (15 percent), and delays with arrangements for long-term care (14 percent).

Information on both length of hospital stay and stability at discharge was available for 302 patients. The median length of stay was 7 days for the 29 low-risk patients who remained in the hospital beyond reaching clinical stability and 5 days for the remaining low-risk patients. Median length of stay was 12.5 days for the 42 medium- and high-risk patients who were hospitalized beyond clinical stability and 8 days for the remaining 113 medium- and high-risk patients.

Clinicians frequently cited availability of home intravenous antimicrobial infusion and home visits by nurses as factors that would have allowed earlier discharge of clinically stable patients.

Marrie, T.J., Peeling, R.W., Fine, M.J., and others. (1996, November). "Ambulatory patients with community-acquired pneumonia: The frequency of atypical agents and clinical course." American Journal of Medicine 101, pp. 508-515.

Many physicians believe that atypical pneumonia—that is, pneumonia caused by known agents, such as Mycoplasma pneumoniae, Chlamydia pneumoniae, Legionella species, and respiratory tract viruses—has different symptoms and a different clinical course than typical bacterial pneumonias. However, a recently published study by the Pneumonia PORT shows that both types of CAP have similar symptoms and outcomes and, in fact, have no distinctive clinical or x-ray features that reliably distinguish one from the other.

The PORT researchers examined the etiology and outcomes of 149 patients with acute pneumonia (confirmed by x-ray), who visited the emergency department and outpatient facilities of Victoria General Hospital in Halifax, Nova Scotia, and offices of participating family doctors based in Halifax. Despite a higher proportion of atypical pneumonia patients reporting initial sweats, chills, and headache, their symptom severity was similar to that of patients with pneumonia of undetermined etiology. Thus it was not possible for clinicians to reliably distinguish between these groups based on initial clinical features. Both groups suffered deterioration of physical functioning, with a marked but incomplete recovery at 30 days. Both groups had similar resolution of symptoms, functional status, return to work, and return to usual activities.

Return to Contents
Proceed to Next Section