The Prescription Paradox

By John F. Lauerman / March / April 1999
November 19th, 2007

This story begins almost fifteen years ago in a quiet nursing home somewhere in the United States. Little was extraordinary or garish about it: No filth lay on the floor. No meals sat uneaten in front of staring patients. No sudden outbursts of uncomprehending rage and frustration punctuated the stillness. But were nursing homes like this one places of comfort and happiness - or of sterility and despair?

"You saw television sets in the rooms," Sidney Katz, former director of Brown's Center for Gerontology and Health Care Research, recalls of that time. "There was almost no noise, it was very clean, and I once noticed that almost all the people in their beds were facing in the same direction. It looked almost ideal, but in reality the residents were half-doped. That was how they controlled things."

Until the late 1980s, this Stepford-like tableau constituted day-to-day life at many nursing homes throughout the United States. Residents were out of sight and frequently out of their minds, their personalities dulled by prescription drugs. The care of older people received little attention from most medical institutions and funding agencies.In fact, the National Institute on Aging, the branch of the National Institutes of Health devoted to research into the health of older people, was not founded until 1974. Implicit in this general state of neglect was a prevailing sense that being older was equivalent to being incapacitated, demented, frustrated, or depressed. If nursing home residents were unaware of their condition, what did it matter? Their basic needs were being met, and death was probably not that far away.

It was somnolent nursing homes like this one that prompted Katz to seek more information on the condition of the institutionalized elderly. Katz, working with the National Academy of Science's Institute of Medicine, joined other researchers to push for a national effort to gather data on nursing homes. Over the past decade they have been helped by a seemingly unrelated development: medicine's hearty embrace of the computer. Processes such as diagnosis, treatment, outplacement, and billing have all been streamlined, and although some critics have decried the depersonalization that sometimes accompanies this increased computerization, it has made available rich new veins of data and the processing power to analyze them. The opportunity now exists for researchers to scrutinize all kinds of national trends in patient management that would have been impenetrable only a few years ago.

One of the most exciting of these research efforts is based at Brown. Begun in 1996 jointly by the Center for Gerontology and Health Care Research and the Institute of Internal Medicine and Geriatrics of the Catholic University School of Medicine in Rome, the Systematic Assessment of Geriatric Drug Use via Epidemiology, or SAGE Study Group, has been examining how prescription drugs fit into the picture. The SAGE group is international and interdisciplinary; it includes physicians, pharmacists, epidemiologists, biostatisticians, and a variety of programmers and data managers. The group strives for a comprehensive look at everything from how certain drugs interact with certain diseases to how a drug's use in the real world compares with its recommended use.

The work of the SAGE group could provide insights into a number of medical advances, such as which drugs might be most effective in treating Alzheimer's, for example, or what might be the best course of treatment for diabetes and Parkinson's disease among the elderly. More troubling, however, have been the study group's revelations of the inconsistent and sometimes shocking medical practices of many nursing homes. As a result, the SAGE effort has come under intense scrutiny itself, with some physicians - surprised and even embarrassed by the prescribing trends SAGE has uncovered - questioning its very utility. Despite such controversy, though, this database of remarkably detailed information on half a million individual nursing home residents is a harbinger of a new approach to researching the health care of older people. What it is revealing could lead to better, more humane treatment of a nursing home population whose neglect and abuse has long been a national scandal.

The most disturbing of the SAGE group's findings focus on how doctors prescribe medication to older patients for such conditions as cancer pain and hypertension. In June 1998, Professor of Gerontology Vincent Mor, who now leads the Center for Gerontology and Heath Care Research, and a team from Brown, New Jersey, Boston, London, and Rome published a study in the Journal of the American Medical Association that looked at the use of pain relief medication by 13,625 nursing home residents who were at least sixty-five years old and who had been diagnosed with cancer. The study drew on every patient with cancer pain in every nursing home in Kansas, Maine, Mississippi, New York, and South Dakota. In the past, many studies of this size would have looked at the subjects' age, sex, race, medication use, cancer type, and perhaps a few other variables. The SAGE researchers, on the other hand, included such details as how many medications each resident was taking, whether he or she was fed by tube, whether the resident had an in-dwelling catheter, and whether he or she was in physical restraints or was terminally ill, depressed, married, widowed, religious, or bedridden. Nothing as comprehensive had ever been done on this population.

Leading the study at Brown was Assistant Research Professor of Gerontology Kate Lapane, who scrutinized the cancer data with co-investigator Giovanni Gambassi of the Catholic University medical school in Rome. They found that 4,003 of the residents were in daily pain. Of these, 26 percent received such relatively weak, non-narcotic painkillers as aspirin or Tylenol. Sixteen percent of the 4,003 residents received much stronger analgesics, such as morphine. But most surprising of all was that another 26 percent of these older patients with daily cancer pain received no painkillers whatsoever. In addition, the study found that African-American residents in documented daily pain were 39 percent less likely to receive pain medication than their white counterparts, and older patients of all races fared worse than younger patients.

"When I first saw the data, I assumed I had made a mistake somewhere," recalled Lapane recently while sitting in a small office in the Biomedical Center at the corner of Brown and Meeting streets. "I went back through all my programming line by line, asking myself 'Is this right? Is this right? Is this right?' Once I was satisfied there was no error there, I was very upset. I felt sick to my stomach."

The pain study received wide, if fleeting, coverage in the press. Acknowledging the reliability of such broad-based findings, neurologists and pain specialists admitted in print and on the air to a nationwide failure to give full credence and attention to older patients who claim to be in pain. "We're doing a lousy job of educating pharmacists, nurses, and physicians in the management of pain," Charles Cleeland of the M.D. Anderson Cancer Center in Texas said on National Public Radio. No one disputed the findings; at the same time, no one could explain them.

That was in June. Five months later, the SAGE group published a study in the November 23, 1998, issue of the Archives of Internal Medicine, this time in a totally different area. The group found that of 80,206 nursing home residents with high blood pressure, 30 percent received no treatment. Unlike the results of the cancer-pain study, this finding was not completely unexpected, since there is some debate about whether the side effects on the elderly of blood-pressure medications are more of a problem than the hypertension they are intended to address.

However, after studying which hypertension drugs were used to treat the 80,206 nursing home residents, Gambassi, Lapane, and their SAGE colleagues found that in the vast majority of cases, the individuals were prescribed drug regimens that did not follow the guidelines developed after the drugs' clinical trials. Large randomized trials here and abroad have shown that certain drugs - diuretics, possibly beta-blockers, and some calcium-channel blockers - reduce the rate of strokes, heart attacks, and deaths among older people with high blood pressure. Among the SAGE subjects, however, only one quarter received diuretics, the therapy proven most effective by international studies. Only 8 percent were prescribed beta-blockers. And somewhat surprisingly, 26 percent of the nursing home residents in the study received calcium-channel blockers, while 22 percent were taking something called angiotensin converting enzyme, or ACE, inhibitors, both of which have far less research backing their use.

"The patterns of drug use that we have documented suggest that these recommendations have little, if at all, informed the practice of nursing home physicians," the SAGE study concluded. But why? Are nursing home doctors using these less-studied medications because they know the evidence for all these hypertension drugs is seldom based on research involving old people, who can react very differently to some medications? Or have these doctors simply fallen prey to the marketing push of drug companies that make calcium-channel blockers and ACE inhibitors?

ertainly doctors treating the elderly often find themselves reaching for their prescription pads and scratching their heads over the potential interactions among all the drugs any one patient may be taking. A 1988 study of 718 hospitalized patients who were at least fifty years old found that 13 percent of their admissions were due to adverse drug side effects. Lapane refers to it as "the paradox": Older people are the largest consumers of drugs in the nation, yet older people are seldom included in the clinical trials used to prove the safety and efficacy of these drugs. The average SAGE subject takes seven prescription medications, yet their effects on this population singly and in combination with one another has long been one of the most understudied areas in pharmacology.


In 1989 the U.S. Food and Drug Administration recommended that all new prescription drugs be evaluated for their effects on people over sixty-five. As a guideline, the recommendation has no enforcement punch, however; as a result, serious gaps in knowledge persist, such as the unknowns in the treatment of heart disease among the elderly. In the September 16, 1992, Journal of the American Medical Association, Jerry Gurwitz, an expert in medications and older people, coauthored a paper that looked at 214 studies of medications used on heart-attack patients; the research-ers found that more than 60 percent of the studies excluded people older than seventy-five, and because women tend to have heart attacks at an older age than men, most of the studies also excluded women, who, incidentally, far outnumber men in nursing homes.

"I don't think the situation has changed much since," says Gurwitz, who is executive director of the Meyers Primary Care Insti-tute in Worcester, Massachusetts. "There are few elderly people in clinical trials, and there are even fewer of the oldest old - people over eighty-five. And even when people of these ages are included, they're much healthier than the average person who would be taking the drug in question."

There are many legitimate barriers to including older people in clinical trials. Transportation to and from testing sites may be difficult. Older people may be more sensitive to the unknown side effects of experimental drugs and could more easily be harmed by them. Because so many elderly are already taking a number of drugs, sorting out the effects of the experimental drug can be difficult, if not impossible.

But some of the bias against older persons is built into medicine itself. As Knight Steel, professor of geriatrics at the New Jersey Medical School and a member of the SAGE Study Group, points out, medical students are poorly prepared for the particular problems of the elderly and their medications, even though older people are increasingly a part of every primary care physician's practice. "You could make the argument that there is no reason for [medical students] to do pediatrics," says Steel, "because the vast majority of us will never do that in our practice. But 95 percent of us will have extensive experience in geriatrics, and yet most medical students never set foot inside a nursing home during their training."

Even at Brown's medical school, where there is a rich tradition of geriatric care and scholarship, a geriatrics rotation is not required. According to Director of Curriculum Affairs Richard Dollase, however, four of seven required student clinical rotations involve significant exposure to elderly patients, and geriatricians direct two of these rotations. In the classroom, problem-based learning sessions often focus on such issues as dementia, fainting, and other disorders most common among older people.

In any case, the low profile of geriatrics in the health care scene allowed nursing homes to fly under the radar for years, and it was not until the late 1980s, after the release of an Institute of Medicine report on nursing-home care, and after several widely reported lawsuits were filed, that the sordid reality of many nursing homes began to emerge. Congressional hearings led to adding a mandate in the 1987 Omnibus Budget Reconciliation Act (OBRA) to create a national data collection system.

There is a direct line from the 1987 OBRA to SAGE. Vincent Mor, of Brown's Center for Gerontology and Health Care Research, was one of a team of researchers engaged to design and implement the monitoring system Congress envisioned. Working with University of Michigan mathematician and Professor of Public Health Brant Fries and others, he designed the computerized Resident Assessment Instrument now used at virtually every nursing home in the United States. "We were looking for other uses for the data," recalls Mor. He approached Roberto Bernabei, an Italian geriatrician from the Catholic University medical school in Rome and Mor's frequent squash partner at international meetings, and Bernabei introduced him to Gambassi, a clinician with the analytical capability to examine complicated epidemiological questions. Mor then recruited Lapane for her expertise in drug use and epidemiology - and for her database skills. "We realized that we could do epidemiological research on drug efficacy in a population that uses a lot of drugs but is seldom included in clinical trials," Mor continues. "We identified five states that were already computerizing the Resident Assessment Instrument data in 1992, and we began to collect their data on drug use."

Thus SAGE was born. In addition to the nursing home information, Mor "cross-walked" the database to hospitalization records kept by the Health Care Finance Administration and to nursing-home assessments held in another large national database. By piling on the databases, SAGE researchers now have access to each subject's complete care history as well as the performance rating of the nursing home each is living in.

Not all researchers have been impressed with the SAGE group's work, however. Some charge that the data is of varying quality, having been collected in inconsistent ways by doctors and nurses untrained in filling out research-oriented forms. An editorial in The Gerontologist cautioned that research depending on databases like SAGE should be viewed with "a healthy degree of skepticism."

One reason for skepticism, in the view of critics, is that, although SAGE is extremely detailed, few so-called clinical measures, such as the results of CAT scans or cardiac stress tests, are included. As a result, when the SAGE group looked at residents' hypertension, for example, they had none of the measures doctors in the nursing home would use to determine the severity of the disease. "Some people who reviewed the [hypertension] paper wanted us to go back and look at all the patients' charts to get that data," recalls Lapane. In response to such critics, Lapane says the SAGE database compensates with what geriatricians call "functional measures" - whether and how well a patient can walk, talk, eat, or socialize, for instance. Although rating these abilities is not as clear-cut as measuring a heart rate, they are what most geriatricians are concerned about when dealing with older patients.

Critics have also objected to the uncontrolled nature of the population whose data the SAGE group utilizes. The experts who review scientific articles before publication have been especially uneasy that the SAGE studies are not based on carefully chosen samples. "Sometimes the reviewers want me to apply statistical tests to make sure I have a representative sample," says Michigan's Brant Fries. "I just say, 'Look, I've got the whole state. You can't get any more representative than that.' "

Even after having gained the scientific credibility that comes with publishing in the major medical journals, the SAGE group sometimes hears grumblings that it is composed of what David Gifford, an assistant professor in Brown's department of community health and community medicine and a SAGE collaborator, calls doctor-bashers. "We as a group have wanted to use the data set not necessarily to look at inappropriate prescribing," Gifford says. "We'd prefer to identify the populations that benefit most from certain treatments - to optimize the treatments that are being used."

Unfortunately, it doesn't always work out that way, as Gifford himself found out when he began using the database to detect which patients might benefit most from the anti-Alzheimer's drug tacrine, which can be effective during the early stages of the disease. When Gifford began analyzing the data, the first thing he saw was that the drug was used in very few patients and that almost one-third of those who did receive tacrine had already reached the late stages of Alzheimer's. Gifford concluded that the drug was often used to treat people who were unlikely to benefit from it.

Similarly, Professor of Clinical Neuroscience Joseph Friedman and Teaching Fellow in Clinical Neuroscience Hubert Fernandez recently used the database to answer what Friedman calls "a zillion questions" about the management of Parkinson's disease. "We've identified some 24,000 patients with Parkinson's in the database," Friedman says. "After Alzheimer's disease and stroke, it could be the third most common neurological diagnosis [among this population]." But like Gifford, Friedman and Fernandez, have been getting unexpected results. Their preliminary findings suggest that mismanagement of these patients' drugs may also be widespread.

Despite such repeated findings, the SAGE researchers are eager to shed their gadfly reputation. They hope that their analyses will eventually lead to more effective treatment by suggesting better matches between such diseases as Alzheimer's and Parkinson's and the drugs prescribed to help control them. Clearly, the approach taken by the SAGE Study Group is no substitute clinical trials; but in a population that is often ill-suited for the protocols of such studies, it may be the only hope. "No one is going to try to design the kind of prospective studies for these patients that we see in clinical trials," says Lapane. "They're too weak and frail, and we couldn't ask them to take placebos in place of the drugs they need every day to treat their chronic conditions. We have to use the data that we have to study them analytically, and that's a challenge. My goal is to prevent human suffering and to work towards better treatment of a neglected population. To the extent SAGE research allows us to do that, it presents a fantastic opportunity."

John F. Lauerman is coauthor of Living to 100: Lessons in Living to Your Maximum Potential at Any Age, published in April by Basic Books.

What do you think?
See what other readers are saying about this article and add your voice. 
Related Issue
March / April 1999