Medication Decisions - Right and Wrong



Bruce Stuart, PhD
Becky Briesacher, MA

The Peter Lamy Center on Drug Therapy and Aging

University of Maryland School of Pharmacy

A background report for the Department of Health and Human Services conference on

Pharmaceutical Costs and Pricing Practices


August 9, 2000

Leavey Conference Center, Georgetown University

Washington, DC

Final Version

Any prescribed medication may be right or wrong given the circumstances of its intended use and the unique condition of the patient being treated. However, there are some drugs, some ways of taking medicines, and some prescription benefit practices that hinder the process of making good decisions about medications, and ultimately the effectiveness of the clinical intervention. These are the gaps in the health care system whereby patients receive less than effective levels of drug therapy—or worse—are exposed to unnecessary hazards due to the medicines they take.

Appropriate medication use requires that drugs be safe and meet the standards of good medical care. This is not simply a process of avoiding medication errors. Medication errors comprise an important, but a relatively small portion of all drug use problems. The landmark Institute of Medicine publication, To Err is Human: Building a Safer Health System (IOM, 1999) focused on preventable failures in the system including the delivery of drug therapies that actually or potentially compromise patients’ safety. The IOM report did not detail the opportunities lost through failure to prescribe useful and necessary medications. As Beers (1999) succinctly puts it, inappropriate prescription therapy encompasses overuse (meaning that medications are prescribed when non-pharmacological therapy is warranted), misuse (a patient needs a medication but is given the wrong one) and underuse (medication need is overlooked or ignored).

The sine qua non of clinical pharmacotherapy is to minimize inappropriate drug use and to maximize appropriate use. Yet in order to realize the full value that pharmaceutical therapy has to offer requires that medication costs be included in the equation. There are cases where “appropriate” may be considered “extravagant” and “cost-ineffective” in light of alternative therapies. The adjectives that one employs frequently depend on where one sits around the health care table. While clinicians, benefit managers, and policy makers may use different language, they face a common set of concerns. These include first, the need for a conceptual framework to assess appropriate and inappropriate drug use; second, an appraisal of what we know and don’t know about the causes and consequences of inappropriate drug use; and third, an agenda for future research.

The organizing theme of this paper is the familiar mantra of best prescribing practices: the right drug, for the right patient, taken in the right way. (Brodie 1967) We have added a fourth dictum to the list, “at the right price.” The payment part has been added to capture the importance of drug prices to individual consumers (particularly those without drug coverage) and the growing role that payers and pharmacy benefit managers play in medicine decision-making. The four-part conceptual framework focuses on the knowledge and behavior of the principal actors in the medication decision making process: manufacturers, regulators, patients, prescribers, and dispensers, as well as purchasers, insurers, and their agents (see Table 1).

Our empirical literature review concentrates on drug use by the elderly. This population group is at much greater risk of drug-related problems than the average younger patient, and much of the published research on prescribing and compliance issues is directed toward them and their special considerations. Our search strategy relied on the traditional biomedical bibliographic indexes augmented with a detailed web search of relevant government publications. The citations are restricted to materials published in English after 1990 from studies conducted in North America. Most of the research on drug appropriateness comes from this part of the world, albeit there is clearly a need for better coordination and sharing of findings from international sources considering the globalization of the drug manufacturing industry. Lessons from the literature are discussed in the final section of the paper.

Table 1. Responsible Parties and Their Roles in Appropriate Use of Drugs

Right Drug

For the Right Patient

Taken the Right Way

At the Right Price

Pharmaceutical Manufacturers

· Reports risk and benefit profiles

· Names & packages Rx



· Sets safety and efficacy standards

· Sets marketing and information standards



· Evaluates Patient

· Selects Rx



· Reviews selection



· Discloses medical history and other treatments


· Sets dosing regimen

· Explains how to use

· Monitors use

· Evaluates outcomes



· Distributes Rx

· Explains how to use

· Monitors use



· Takes Rx as directed

· Reports positive and negative outcomes

Pharmaceutical Manufacturers

· Sets price


Purchasers, insurers, and their agents

· Designs benefit

· Creates formulary

· Negotiates price

· Sets cost-sharing levels



· Considers price when prescribing Rx



· Considers price when filling Rx

The Right Drug

Few prescription drugs are completely problem free, but some present especially high risks of significant adverse consequences—even death—if prescribed. The question addressed in this section is what makes one medication preferable or less risky to take than another? This is a sensitive issue, as all prescription drugs introduced in the United States since 1964 have met efficacy and safety standards set by the Food and Drug Administration (FDA). Drugs are approved by the FDA for one or more specific indications, but except for a very small number of products, physicians have full clinical discretion in terms of to whom they prescribe them and for what reason. Customizable therapy is an important principle in drug selection. Studies of pharmacokinetics (how the body processes a drug) and pharmacogenetics (how genetics influence the drug effect) confirm the wide variability in the way individual patients respond to medications. Sound medical reasons exist for giving certain patients specific drugs despite prevailing wisdom about more appropriate choices. An oft-heard example in this regard is the reluctance of many physicians to switch patients to newer medications when older therapies have been used for years without apparent problems. Nevertheless, research has identified at least three properties of medications that either raise the potential for harm or lower the potential for effective care in most patients. If efficacy is generally equivalent among medications, a drug with any of the following characteristics should be passed over for a better drug.

The first property of an undesirable drug is having a “dirty” side-effect profile relative to its competitors. This means that the medication is not the safest alternative among those that have the same therapeutic effects. All medications have the potential to cause harmful secondary effects, but some carry exceptional risks because of common and noxious side effects that regularly create new medical problems in patients who take them. A well-documented example is of an older patient who complains of insomnia being given a long-acting benzodiazepine (which commonly causes confusion and vertigo), then falls out of bed one morning, and breaks a hip. (Ray 1991) This type of adverse drugs reaction is predictable, preventable, and distinctly unlike the rarer allergic reactions in patients with unknown physiologic sensitivities. The potential danger of giving patients medications with “dirty” side-effects has stimulated the development of newer generation agents that offer similar benefits but much better safety profiles. The problem is that the older, more risky drugs are still available, are widely used, and generally cost a fraction of the newer product.

The extent to which problematic drugs are prescribed is best documented in the population of patients aged 65 or older. There is general consensus among geriatric experts that forty or so particularly unsafe or ineffective medications should not be prescribed to elderly patients under any circumstances or, if used, should be limited to very special situations. (Beers 1991, Stuck 1994, Beer 1997, Beers 1999a, Beers 1999b) For each named drug, the prescribing concerns are documented, and safer or more effective substitutes are available for treating most patients. An example is ticlopidine, for the prevention of stroke. This medication is generally no better than aspirin in preventing clotting and considerably more toxic: over half of all patients taking this drug report adverse effects, including life-threatening neutropenia, agranulocytosis and other hematologic abnormalities. (Beers 1999a) Despite such safety warnings, which have been in place for nearly ten years, research shows that these high-risk drugs are still commonly prescribed. Researchers estimate that from 14 percent to 24 percent of all noninstitutionalized elderly persons take at least one of these potentially toxic or useless medications. (Stuck 1994, Willcox 1994, USGAO 1995) The elderly poor have even higher rates of exposure. Researchers in Kentucky and Maryland found that nearly one in every four Medicaid recipients aged 65 and older in these states had taken at least one of these drugs in the last year. (Zuckerman 1997, Piecoro 2000) Residents of nursing homes and homebound older adults may have the highest rates of all. A study in twelve skilled nursing facilities in Los Angeles found that 40 percent of residents were taking inappropriate drugs. (Beers 1992) A pharmacy review of the drugs used by over 2,000 homebound adults aged 60 or older found a similar 40 percent were prescribed at least one inappropriate medication. (Golden 1999) These and similar research findings prompted the Health Care Financing Administration(now known as Centers for Medicare and Medicaid Services(CMS)) in 1999 to implement drug therapy guidelines based on the Beers criteria as part of the agency’s periodic surveys of nursing facilities.(Beers 1999a)

Inadequate clinical trial information on use in special patient populations is a second contributing property of potentially harmful medications. One difficulty in selecting appropriate drugs for the elderly has been the lack of randomized study data establishing minimum effectiveness guidelines. Since the average participant of a clinical trial (a middle aged male with no comorbidities) looks quite different from the average geriatric patient (an aged female with multiple concomitant conditions) the generalizability of the findings is suspect. Reviews of randomized trials for major diseases affecting the elderly usually find no or rare participation by older patients. A study conducted on 214 acute myocardial infarction trials found that fewer than two in five allowed the inclusion of even one patient over the age of 75. (Gurwitz 1994) Only 20 percent of the patients participating in the five major trials on the use of warfarin for stroke prevention were aged 75 years or older. (Gottlieb 1994) There are efforts underway at NIH to expand the participation of seniors in clinical trials, and President Clinton recently authorized Medicare payment for treatment costs of Medicare beneficiaries involved in clinical trials. These efforts may well lead to better information about geriatric prescribing of new drugs yet to reach market, but they offer little help regarding currently available medications.

Without age-adjusted efficacy information, physicians are left to guess how best to use medicines in older patients, if they are to be used at all. The situation is further complicated as clinical trial data rarely show the comparison of greatest utility---how does the new drug stack up against competitors products. Studies of the consequences of inadequate efficacy and safety information paint a consistent picture of physicians tending to treat elderly patients with doses below established efficacy levels rather than risk adverse effects. Over 86 percent of 10,000 myocardial infarction survivors (mostly geriatric) in Ontario received beta-blocker therapy at doses well below those tested in clinical trials. (Rochon 1999) Similarly, the Cardiovascular Health Study, a prospective study of 5200 older adults followed 5 years, found the average doses of ACE inhibitors used to treat congestive heart failure were below levels generally regarded to be clinically effective. (Smith 1998) When clinical trials do examine drug use specific to older participants, research shows improvement in geriatric care although the change is sometimes slow. Within a year after three major clinical trials in older adults presented evidence that diuretics are more effective in preventing strokes that the newer antihypertensive medications, practice patterns shifted toward the appropriate therapy and away from the inappropriate drugs. (Psaty 1995) In 1997, the FDA established rules for the inclusion of geriatric precautions in medication labeling as an expansion to the 1989 guidelines recommending that clinical trial enrollment reflect the make-up of typical clinical practice patients. (USFDA 1998, USFDA 1989) According to Rochon (2000) the impact of the 1989 guidelines has been negligible, and the influence of 1997 rule is yet unknown.

In cases where prescribers must be guided by experience rather than data, research shows that the more formal training physicians undergo the better. For example, cardiologists are more likely than generalists to use thrombolytic agents to treat acute myocardial infarction and to prescribe beta-blockers and aspirin, all recommended interventions. (Ayanian 1994, Ayanian 1997) Similar findings have been published for other specialty groups. But seeing a specialist is by no means a guarantee of a correct medication regimen. In a study of 214 patients treated by cardiologists or generalists, approximately 75 percent were given the appropriate medication but fewer than 40 percent were receiving dosage levels known to be efficacious, a problem that occurred regardless of physician type. (Chin 1997)

A third property of risky medications is having nondescript packaging or a confusing name. Having a brand name that looks or sounds like another drug’s name is thought to cause approximately one in every four medication errors reported in the United States. (Lambert 1999) The problem is so common that the Institute of Medicine’s report on medical errors called for stricter FDA oversight of drug naming procedures to prevent future products from sounding as homophonic as these three currently available drugs—Celebrex (used for arthritis), Celexa (used for depression) and Cerbyx (used for seizures). (IOM 1999) Stories about medication errors due to product packaging are also common in hospitals. One involved mivacurium, an intravenous muscle relaxant that was inadvertently given to four patients and caused death or serious injuries in all cases. Mivacurium and eleven different products were all packaged in identical foil sleeves with clear windows for label inserts that slid behind frames obscuring the medication name. No research has yet quantified the extent that packaging problems contribute to national medication error rates.

For the Right Patient

This section addresses why patient-level characteristics can make a medication decision appropriate for one population but not for another. We summarize here prescribing issues common in the elderly, noting that individual responses to the aging process vary substantially depending on physiological, genetic, socioeconomic, and environmental factors. The over-60 population is remarkably heterogeneous with respect to physical health and cognitive functioning. (IOM 1997) Nevertheless, elderly patients tend to experience more drug use problems than younger patients for two clear reasons, and possibly a third: 1) the aged use more drugs on average, and drug problems increase with frequency of use, 2) increasing age increases the risk of illness, and any compromises in bodily functions can affect the drug effect, and 3 ) there is some evidence still under debate that as the body ages, it becomes more sensitive to drugs.

Research is still evolving on the physiological mechanisms of aging and the interaction of age-related changes with the actions of drugs. Older patients seem to be more sensitive to the effects of drugs, yet it is difficult to attribute this to the normal aging process as opposed to the presence of coexisting illnesses. At the cellular level, changes due to aging appear to be distinct (e.g., decreased elasticity and DNA degradation) while at the whole organism level the disease process seems to dominate most observed changes. (IOM 1997) The absorption of drugs (which affects the duration and intensity of the drug effect) does not appear to be significantly impaired by aging. However, aging may affect how the drug is distributed throughout the body. (IOM 1997) Changes in the way drugs are delivered to the heart, brain, liver, and other body organs do not necessarily result in higher drug concentrations, but the accompanying decreases in lean body mass and total body water may heighten the effects of certain medications. Aging research is more definitive on the way an older body eliminates drugs through the liver or kidney. Liver mass clearly decreases with age, and renal function also declines. As the elimination process slows, a drug can more readily accumulate in the body, and side effects and toxicity are more likely. The main message from studies on the effects of aging on drug therapies is that the issue is complex and yet unresolved. (IOM 1997, Solomon 1997)

We have clearer evidence that the aged are heavy users of medications. The 13 percent of the U.S. population who are over the age of 65 consume over one-third of all prescription drugs. (Avorn 1995) Nearly 86 percent of all Medicare beneficiaries take at least one prescription drug and their average use is about 20 prescription fills a year. (Stuart 2000, Davis 1999, AARP 1996) The typical nursing home resident takes between five and nine medications concurrently. (Williams 1999, Bernabei 1999, Tobias 1997, Beers 1992)

Research substantiating the relationship between high medication use and adverse drug events is limited to estimates from single institutions, but the results are consistent. A chart review on 301 elderly inpatients over a 4-month period found 37.5 percent experienced an adverse drug reaction and, after statistically controlling for gender and renal function, polypharmacy was the main explanatory factor. (Bowman 1996) A four-year prospective observational study on 332 nursing home residents found over 15 percent admitted for a drug-related hospitalization. (Cooper 1999) The main reasons for admission were complications from nonsteroidal anti-inflammatory use, a psychotropic-related fall with fracture, digoxin toxicity, and insulin hypoglycemia. The strongest predictor of a drug-related hospitalization was the number of medications taken (7.9 vs. 3.3). (Cooper 1999)

A related problem with high medication use is a phenomenon known as the “prescribing cascade,” which is the misdiagnosis of an adverse drug reaction as a new medical condition that is then treated by adding a new drug. (Rochon 1997) This problem occurs because adverse drug reactions often mimic clinical conditions commonly found in geriatric patients. (Avorn 1997) An example is high blood pressure treatment following the use of non-steroidal anti-inflammatory drugs (NSAIDs), which are among the most commonly prescribed drugs for elderly patients. (Ray 1993) An adverse property of these drugs is the repression of a critical enzyme that helps in regulating blood pressure. A case-control study of over 9,000 older Medicaid recipients newly started on blood pressure lowering drugs found that 54 to 80 percent were also using non-steroidal anti-inflammatory drugs. (Gurwitz 1994) Prescribing cascades have been found with metoclopramide and Parkinson’s treatments, and with diuretics and gout therapy. (Rochon 1997) The full extent of medications inappropriately prescribed to older patients to treat overlooked side effects is unknown.

The final factor placing elderly patients at risk for inappropriate drug therapy relates to the burden of chronic conditions. Among U.S. adults aged 65 years or older, 49 percent have arthritis, 40 percent have hypertension, and 29 percent have heart disease. (Desai 1999) Each of these diseases are commonly and successfully treated with medications, but research shows that the likelihood of receiving appropriate care declines as individuals develop multiple diseases. A study of over 1 million elderly residents of Ontario found that patients with chronic conditions were highly likely to be under- treated for common secondary conditions. (Redelmeier 1998) Patients were 60 percent less likely to receive estrogen replacement therapy if they had diabetes, 31 percent less likely to receive lipid-lowering medications if they had pulmonary emphysema, and 41 percent less likely to receive treatments for arthritis if they had psychotic syndromes. These results are supported by other studies. Initiation rates of cholesterol-lowering agents drop substantially with advancing age despite increasing risk of the underlying disease. Lematire (1998) followed a cohort of 5,200 older adults for 7 years and found that persons younger than 70 years were four times more likely to start therapy than those 80 years or older. The typical physician’s focus on “primary” medical problems may help explain the consistent underuse of preventive therapies in this population. Another possibility is that physicians establish different (and lower) treatment goals for older patients compared to younger patients.

Taken the Right Way

Counting just patients who receive a “wrong” drug as a measure of inappropriate medication use surely underestimates the true extent of the problem. (Gurwitz 1994, USGAO 1995) The portion of patients who are given the “right” medication but take it the wrong way is believed to be equally as large. This section reviews the literature on the potential of properly selected medications to interact with other medications, to perform poorly due to under-exposure, or to cause toxicity due to over-exposure. Definitive research here is lacking due to the complex features of the individual prescription (e.g., dose, daily schedule and duration) and the unique clinical presentation of each patient. Another contributing factor is the pivotal behavioral role of the patient, whose willingness and ability to follow the recommended regimen are rarely assessed and documented. These factors combine to make valid measurement of appropriate prescription-taking difficult to gauge.

One common way to evaluate prescribing problems is to screen medication use records against accepted clinical standards of appropriateness, as do pharmacy benefit managers and Medicaid programs when conducting drug utilization reviews. These systems provide reliable estimates of the types of prescribing problems that can be defined within explicit parameters of acceptable care. In a study of five types of drug-related problems (dose, duration of therapy, duplicative therapy, drug-drug interactions, and contraindications) for eight drug categories commonly used to treat the elderly, researchers found that the incidence of inappropriate medications ranged from less than 1 percent to 23 percent depending on the drug and the problem category. (Briesacher 1999) A hopeful finding was that the rates improved over time. The study sample included all community-dwelling elderly in three Medicaid programs (Maryland, Washington and Iowa) between 1989 and 1996. Within each state, fewer prescriptions were written outside of the appropriateness criteria in 1996 compared to 1989. This improvement was not replicated among the younger Medicaid populations in these same states suggesting something unique was at work in the older patient population. One explanation is the abundant physician awareness campaigns waged during this period exhorting clinicians to recognize side-effects in the elderly and to improve geriatric pharmacotherapy. (Briesacher 1999)

Another way to identify medication-taking problems is to examine drug use histories of individual patients. This method has typically been applied only to severely ill patients and captures just the worst-case sequela of drug use problems, but it provides some important evidence nonetheless. A recent example is a review of drug regimens of some 1000 elderly hospitalized patients conducted by Doucet. (1996) Over half were taking at least two drugs with the potential for interactions, and at least 130 patients were experiencing side effects from the interactions that were observable during the hospital stay. The most frequently involved drugs were cardiovascular and psychotropic agents, and the most common problems were neuropsychological impairment, arterial hypotension, and acute renal failure. (Doucet 1996) Another hospital-based study found that more than 53 percent of patients had preventable medication problems. (Lipton 1992) These included wrong dosing schedule, wrong dosage, lack of indication, drug-drug interaction, and therapeutic duplication. However, fewer than 20 percent were considered serious threats to the patient’s health. A similar study of 330 nursing home residents found that over 67 percent had adverse drug reactions and 14 percent were subsequently hospitalized as a result.(Cooper 1999)

How closely patients follow their medication regimen is the other half of the prescription-taking equation. Estimates suggest that 90 percent of all non-institutionalized patients make mistakes in taking their medicines, and that poor compliance with accounts for 30 to 50 percent of the cases where drugs fail to achieve therapeutic goals. (National Council, 1995) Compliance rates for the elderly are usually much better than total population averages, although some drug classes are more problematic for the elderly. In a survey of persons aged 50 or older, 43 percent admitted to either stopping a prescription before it ran out, taking less than prescribed, or not filling the script at all within the last two years. (AARP 1996) Another study found that 77 percent of the elderly Medicaid recipients in New Jersey could not maintain at least an 80 percent level of compliance with their antihypertensive medication during a one-year period. (Monane 1996) Several studies of drug-related hospitalizations by the elderly have shown that noncompliance and inaccurate directions are the primary preventable problems. (Bero 1991, Col 1990, Schmader 1996)

The reasons why patients have difficulties in taking their medicines vary from case to case but one reason—side effects—comes up with regularlity. When persons aged 50 or older were asked to describe their medication-taking problems, 80 percent of those with a problem said the side-effects were bad enough to make them stop the drug all together. (AARP 1996) These results support findings elsewhere that patient noncompliance may be considered a rational response to problematic drug therapy. (Col 1990, Lipton 1992, Bero 1991) Other predictors of medication noncompliance in older adults include low socioeconomic status, large number of medications taken concurrently, high psychological distress, and high frequency of the daily dose. (Coons 1994, Eisen 1990)

At the Right Price

In this section we examine how source of payment and the level of patients’ out-of-pocket obligation can encourage or discourage appropriate use. Given that these are such highly charged topics today, it is surprising to find so little research evidence on the association of coverage and cost to the appropriateness of prescription drug use. One difficulty is in distinguishing higher use from better use. We know that prescription drug coverage and higher incomes are both predictors of getting more medicines and more costly medicines (Stuart 2000), but we do not know if the relationships extend to getting more appropriate drug therapy. (Blustein 2000) There are reasons for believing that improved financial access to prescriptions leads to better care, and other reasons for taking the opposite view. The belief that the risk for inappropriate drug use increases with greater insurance coverage comes from research showing that Medicare beneficiaries with drug benefits fill more prescriptions than those without (Poisal 1999, Stuart 2000), and that the risk of adverse drug events increases as more medications are taken. (USGAO 2000) On the other hand, persons with generous drug coverage are more likely to be in health plans that conduct drug use reviews and disease management programs. (Monane 1998) Such programs should theoretically reduce the risk of adverse drug events although this is largely unproven.

The popular press regularly features stories of elderly people going without medicines because of costs, but the research to date suggests this may be a relatively rare problem limited to certain groups of vulnerable patients. A recent study by the Assistant Secretary of Planning and Evaluation in DHHS showed that only 4 percent of Medicare beneficiaries reported needing a prescription at some point during the year but being unable to afford it. (USDHHS, 2000) Only 7 percent of adult aged 70 or older said they ended up taking less medication than prescribed because of cost. (Sarver 2000) These numbers are low but may not accurately reflect the full spectrum of cost-avoidance behaviors such as skipping prescriptions or postponing refills. A survey of 600 high prescription users in an Arizona Medicare-HMO plan with a capped prescription benefit found that 36 percent obtained free prescription samples from their physicians, 21 percent took less than prescribed, and 19 percent substituted an over-the-counter product for their prescription medicine. (Jernigan 1999)

A closer look at the people who admit to having prescription-filling problems shows that they come from all socio-economic strata, but are more likely to report low income or no insurance coverage. This suggests that financial burden rather than personal preference may be the culprit. Cohen (1997) found that elderly persons living below the poverty level were six times more likely to go without a necessary prescription than those with higher incomes, and elderly persons without any supplemental coverage were three times more likely to go without a necessary prescription than those with private insurance. There are almost no data available on the types of medications unfilled for financial reasons, on the persistence of missed prescriptions, or on the impact of missing prescriptions on patients’ health. One of the few studies to examine these issues is the well-known Soumerai (1991) study of downstream consequences of New Hampshire’s decision to institute a three prescription per month limit in its Medicaid program. One consequence was that frail elderly Medicaid recipients cut their use of essential drugs by 35 percent and subsequently faced an increased risk of institutionalization. (Soumerai 1991)

Research is still accumulating on the association between managed care and appropriate drug use. A review of 2300 Medicare patients with acute myocardial infarctions treated at 20 hospitals showed that the use of lifesaving drug therapies (aspirin, beta-blockers and thrombolytics) was the same if not a bit better for HMO enrollees compared to fee-for-service (FFS) patients. (Soumerai 1999) A comparison of elderly diabetic patients in eight HMOs to fee-for-service plans found use of insulin and oral hypoglycemic agents was nearly identical, though HMO enrollees were slightly more likely to have medication changes. (Retchin 1992) Another HMOs and FFS comparison of care for aged patients with hypertension showed that medication treatment and follow-up were again similar, except FFS physicians adjusted the regimen more often and ordered electrolytes routinely. (Preston 1991)

Reviews of specific drug use policies employed by third-party payers have generally not found any strong correlations to appropriateness of medication use, although this may reflect data problems as much as any lack of culpability. There are theoretical reasons to believe that the financial incentives of managed care organizations (MCOs) and pharmacy benefit managers (PMBs) can come in conflict with the health care needs of patients, but the extent that economic interests trump clinical concerns is largely unknown. Part of the problem is that researchers have difficulty gaining access to proprietary information on the drug use policies of payers and PBMs. The decisions of corporate pharmacy and therapeutics committees (bodies of staff clinicians and administrators who select medications for the formulary) are nearly always made behind closed doors. We do not know, for instance, whether there is any systematic relationship between formulary design and patient compliance. Nor, do we know how formulary changes affect the drug use patterns and outcomes for older patients. In just one Medicare MCO during a 6-month period, formulary coverage changed in 20 different drug classes and affected the medication use of 9,000 of 40,000 beneficiary members. (USGAO 1999)

A formulary practice of particular interest today is the “three tier” copayment for prescription drugs. As of 1999, almost 70 percent of MCOs and PBMs surveyed by Scott-Levin offered the three-tier co-pay option to their employer groups. (Scott-Levin, 1999) The actually number of employer groups using this practice is much lower, but is expected to grow rapidly in response to the double-digit growth in drug costs expected over the next few recent years. Under the three-tier system, patients pay the lowest price for generic drugs, a higher price for “preferred” brand-name drugs on a plan's formulary, and the highest price (sometimes up to $55 a script) for drugs not on formulary. The key element in three tier copay systems is the drug formulary--the more restrictive the formulary, the more products shift to non-formulary status. The potential for abuse here is staggering as plans that use this cost sharing approach are under no legal obligation to include any “preferred” products.

Another controversial policy is the practice of pharmacy benefit managers to favor heavily rebated products on their formularies. A comparison of access to the 200 top-selling prescription drugs on Medicaid formularies showed that the majority of products were available on all lists despite varying degrees of restriction. (Walser 1996) However, Federal law places restrictions on Medicaid formularies that do not apply in the private sector. A related policy issue concerns the practice of “therapeutic interchange,” where physicians are encouraged to switch their patients to preferred products. There is clearly the potential for harm in such practices, but little solid evidence one way or the other. In a survey of 1400 physicians in New York State, nearly 15 percent said that the drugs they wanted to prescribe were not on the formulary and 10 percent said that their patients suffered medication-related problems after their prescriptions had been switched. (Gross 1998)

Righting the Wrongs

The research described in this paper leaves no doubt that the elderly—and by implication other risk groups within the population—are at significant risk for inappropriate medication use. The true dimensions of the problem are murkier. We do not even know if overuse, misuse or underuse is the more problematic. Notwithstanding this lack of evidence, in principle we know that righting the “wrong” aspects of medication decisions has enormous potential for improving health and achieving major cost savings. How that value is realized is another matter.

Take, for instance, the humble bottle of aspirin. Studies have shown that aspirin substantially reduces the risk of death and/or non-fatal heart attacks in patients with a previous MI or unstable angina pectoris (which often occurs before a heart attack). But even low doses of aspirin are by no means harmless. Gastrointestinal distress (nausea, heartburn, pain) is a well-recognized adverse effect. A small portion of the population is hypersensitive to aspirin and cannot tolerate even small amounts. For them, internal bleeding or worse outcomes are possible. In determining the value of aspirin, we must consider two scenarios. In scenario A, aspirin is taken by only people who are at significantly increased risk for myocardial infarction, who lack contraindications, and who are under the care of a doctor to monitor possible side effects. In scenario B, all types of people start taking aspirin. Some are at such low risk for a cardiovascular event that the therapy offers virtually no benefits. Some are allergic to aspirin or take high enough doses to cause intestinal damage that may lead to hospitalization. Some do not take the aspirin at all and suffer heart attacks.

Scenario A might appear ideal, but scenario B typifies the real world for most medications in use today. The problem is moving from B to A in a reasoned and cost-effective manner. In circumstances where side effects, contraindications, and compliance problems are known, this process involves a tradeoff—evaluating the benefits of reaching scenario B against the cost of getting there. Since risks are probabilistic but costs are certain, it is almost never prudent (or possible) to completely execute the transition. However, if the problem profile itself is incomplete and misunderstood—as it is with many drugs—then the tradeoff equation cannot be estimated, let alone implemented.

Given these unknowns, there is a clear need for further research. First, is the need for a systematic evaluation of the standards by which potentially inappropriate drug use is defined. The oft-mentioned Beers criteria for medication use by the elderly are the best known, but there are other contenders incorporated in various drug utilization review program and described in the pharmacological literature.(Beers 1991, Beers 1999a, Knapp 1992) Explicit, evidence-based criteria for good medication use are critical to efforts to reduce prescribing problems as they represent the yardstick for measuring improvement.

Second, is the need for national estimates of the incidence, prevalence, and cost of “wrong” medications. Policy-makers today lack the hard and consistent evidence necessary to build a strong case for widespread and serious interventions to alleviate the problem. In the absence of hard data, researchers have used assumption-based models to estimate the national cost of drug-related complications and mortality. In one oft-quoted study, Johnson and Bootman (1995) estimate that drug-related morbidity and mortality add $76.6 billion in annual health care expenditures. Others come up with more conservative estimates based on different assumptions. Low or high, such estimates engender skepticism because it is nearly impossible to evaluate the validity of the underlying conjectures. When every expert comes up with different numbers, it is easy for policy makers to put off needed action.

Third, is the need to identify the characteristics of individuals most at risk for inappropriate medication use, including those who takes such medications (or should be taking necessary medications if they aren’t), under what circumstances, and for how long. There is a notable lack of research with timeframes long enough to establish whether these problems are episodic or persistence. It stands to reason that if individuals are exposed once or only sporadically to harmful medicines, then the concern is quite different from routine, long-standing exposure. This type of analysis is especially needed for the chronically ill elderly who are at greatest risk for drug-drug interactions and drug-disease interactions.

Lastly is the need for studies on the costs and consequences of treating the sequelae of inappropriate drug use. These studies are among the most difficult to conduct, but represent the best hope for true headway in reaching the goal of appropriate medication use for all. These are the studies that will influence the training of future clinicians and establish the standards by which progress toward that goal is measured. But resources for research are limited as they are for direct patient care. The key is to target research in such a way that drug use problems with significant and severe consequences can be clearly distinguished from those with inconsequential or mild effects on patients’ health. The best approach to improving drug use is to prevent the problems that cause the greatest harm or overlook the greatest benefit.

The authors wish to thank Ilene Verovsky, PharmD, CGP for her helpful comments and suggestions on an earlier draft.


American Association for Retired Persons. Survey on prescription drug issues and usage among americans aged 50 and older. May 1996. (

Avorn J. Medication use and the elderly: current status and opportunities. Health Affairs 1995;14(1):276-86.

Avorn J, Gurwitz JH. Principles of pharmacology. Geriatric Medicine. Eds. Cassel CK, Cohen HJ, Larson EG, et al. Third Edition. Splinger: New York, 1997.

Ayanian J, Hauptman P, Guadagnoli E, Antman E, Pashos C, McNeil B. Knowledge and practices of generalist and specialist physicians regarding drug therapy for acute myocardial infarction. N Engl J Med. 1994;331:1136-1142.

Ayanian JZ. Guadagnoli E. McNeil BJ. Cleary PD. Treatment and outcomes of acute myocardial infarction among patients of cardiologists and generalist physicians. [Journal Article] Archives of Internal Medicine. 1997;157(22):2570-6

Beers MH, Ouslander JG, Rollingher I, et al. Explicit criteria for determining inappropriate medication use in nursing homes. Arch Intern Med 1991;151:1825-1832.

Beers MH, Ouslander JG, Fingold SF, et al. Inappropriate medication prescribing in skilled-nursing facilities. An Inter Med 1992;117:684-689.

Beers MH. Explicit criteria for determining potentially inappropriate medication use by the elderly. Arch Intern Med 1997;157:1531-1536.

Beers MH. Understanding the HCFA(now known as CMS) Guidelines for Potentially Inappropriate Medications in the Elderly---Optimizing Prescribing in Long-Term Care. Merck & Co., Inc, West Point, PA, 1999(a).

Beers MH. Aging as a risk factor for medication-related problems. Consultant Pharmacist 1999(b);14(12):1337-1340.

Bernabei R, Gambassi G, Lapane K, et al. Characteristics of the SAGE database: a new resource for research on outcomes in long-term care. SAGE (Systematic Assessment of Geriatric drug use via Epidemiology) Study Group. J Gerontology. Series A, Biological Sciences & Medical Sciences 1999;54(1):M25-33.

Bero LA, Lipton HL, Bird JA. Characterization of geriatric drug-related hospital readmissions. Med Care 1991;29(10):989-1003.

Blustein J. Drug coverage and drug purchases by medicare beneficiaries with hypertension. Health Affairs 2000;19(2):219-230.

Bowman L, Carlstedt BC, Hancock E, et al. Adverse drug reaction (ADR) occurrence and evaluation in elderly inpatients. Pharmacoepidemiology & Drug Safety 1996;5(1): 9-18.

Briesacher BA, Stuart B, Peluso R. Drug use and prescribing problems in the community-dwelling elderly: a study of three state medicaid programs. Clinical Therapeutics 1999;21(12):2156-2172.

Brodie DC. Drug use control: keystone to pharmaceutical care. Drug Intell Clin Pharm 1967;1:65.

Chin MH, Wang JC, Zhang JX, et al. Utilization and dosing of angiotensin-converting enzyme inhibitors for heart failure: effect of physician specialty and patient characteristics. J Gen Intern Med. 1997;12:563-566

Cohen RA, Bloom B, Simpson G, et al. Access to Health Care. Part 3: Older Adults. National Center for Health Statistics. Vital Health Stat 19(198), 1997.

Col N, Fanale JE, Kronholm P. The role of medication noncompliance and adverse drug reactions in hospitalizations of the elderly. Arch Inter Med 1990;150:841-845.

Coons SJ, Sheahan SL, Martin SS, et al. Predictors of medication noncompliance in a sample of older adults. Clinical Therapeutics 1994;16(1):100-117.

Cooper JW. Adverse drug reaction-related hospitalizations of nursing facility patients: a 4-year study. Southern Medical Journal 1999; 92(5):485-90.

Davis M, Poisal J, Chulis G, et al. Prescription drug coverage, utilization, and spending among medicare beneficiaries. Health Affairs 1999;18(1):231-243.

Desai MM, Zhang P, Hennessy CH. Surveillance for morbidity and mortality among older adults--United States, 1995-1996. Morbidity & Mortality Weekly Report. CDC Surveillance Summaries 1999;48(8):7-25.

Doucet J, Chassagne P, Trivalle CA, et al. Drug-drug interactions related to hospital admissions in older adults: a prospective study of 1000 patients. J Am Geriatr Soc 1996;44:944-948.

Eisen SA. Miller DK. Woodward RS. Spitznagel E. Przybeck TR. The effect of prescribed daily dose frequency on patient medication compliance Archives of Internal Medicine 1990;150(9):1881-4.

Golden AG, Preston RZ, Barnett SD, et al. Inappropriate medication prescribing in homebound older adults. J Am Geriatrics Soc 1999;47(8):948-953.

Gottlieb LK. Salem-Schatz S. Anticoagulation in atrial fibrillation. Does efficacy in clinical trials translate into effectiveness in practice? Archives of Internal Medicine 1994. 154(17):1945-1953.

Gross DJ. Prescription drug formularies in managed care: concerns for the elderly population. Clinical Therapeutics 1998;20(6):1277-1291

Gurwitz J. Suboptimal medication use in the elderly: the tip of the iceberg. JAMA 1994;272(4):316-317.

Gurwitz JH, Avorn J, Bohn R, et al. Initiation of antihypertensive treatment during nonsteroidal anti-inflammatory drug therapy. JAMA 1994;272:781-786.

Furberg CD, Psaty BM, Meyer JV. Nifedipine: dose-related increase in mortality in patients with coronary heart disease. Circulation. 1995;92:1326-1331.

Institute of Medicine. Pharmacokinetics and Drug Interactions in the Elderly and Special Issues in Elderly African-American Populations: Workshop Summary. National Academy Press: Washington, DC, 1997

Institute of Medicine. To Err is Human: Building a Safer Health System. National Academy Press: Washington, DC, 1997

Jernigan CL, Cox ER, Coons SJ, et al. Medicare beneficiaries management of capped prescription benefits. APHA Annual meeting, November 1999.

Johnson JA, Bootman JL. Drug-related morbidity and mortality: a cost-of-illness model. Archives of Internal Medicine 1995;155(18):1949-1956.

Knapp D, Erwin G. Screening Criteria for Outpatient Drug Use Review. Final Report to HCFA(now known as CMS). Baltimore: University of Maryland School of Pharmacy; 1992.

Lambert BL, Lin SJ, Chang KY et al. Similarity as a risk factor in drug-name confusion errors: the look-alike (orthographic) and sound-alike (phonetic) model. Medical Care 1999;37(12):1214-25.

Lemaitre RN, Furberg CD, Newman AB et al. Time trends in the use of cholesterol-lowering agents in older adults. Arch Intern Med 1998;158:1761-1768.

Lipton HL, Bero LA, Bird JA. The impact of clinical pharmacists’ consultations on physicians’ geriatric drug prescribing: a randomized trial. Med Care 1992;30:646-658.

Monane M, Bohn RL, Gurwitz JH, et al. Compliance with antihypertensive therapy among elderly medicaid enrollees: the role of age, gender and race. Am J Public H 1996;86(12):1805-1808.

Monane M, Matthias DM, Nagle BA, et al. Improving Prescribing patterns for the elderly through an online drug utilization review intervention. JAMA 1998;280(14):1249-1252.

National Council on Patient Information and Education. Prescription Medicine Compliance: A Review of the Baseline of Knowledge. Washington, DC, 1995.

Piecoro LT, Browning SR, Prince TS, et al. A database analysis of potentially inappropriate drug use in an elderly medicaid population. Pharmacotherapy 2000;20(2):221-228.

Poisal J, Chulis G. Medicare beneficiaries and drug coverage. Health Affairs 2000;19:248-256.

Preston JA, Retchin SM. The management of geriatric hypertension in health maintenance organizations. J American Geriatrics Society 1991;39(7):683-90.

Psaty BM, Koepsell TD, Yanez ND, et al. Temporal patterns of antihypertensive medication use among older adults, 1989 through 1992. JAMA 1995;273(18):1436-1438.

Ray WA. Psychotropic drugs and injuries among the elderly: a review. J Clinical Psychopharmacology1992;12(6):386-96.

Ray WA, Griffin MR, Avorn J. Evaluating drugs after their approval for clinical use. N Engl J Med 1993;329:2029-2032.

Redelmeier DA. Tan SH. Booth GL. The treatment of unrelated disorders in patients with chronic medical diseases. New England Journal of Medicine 1998;338(21):1516-20.

Retchin SM. Clement DG. Rossiter LF. Brown B. Brown R. Nelson L. How the elderly fare in HMOs: outcomes from the Medicare competition demonstrations. Health Services Research 1992; 27(5):651-69.

Rochon PA, Gurwitz JH. Optimising drug treatment for elderly people: the prescribing cascade. BMJ 1997;315(7115):1096-1099.

Rochon PA, Anderson GM, Tu JV, et al. Age and gender-related use of low dose drug therapy: the need to manufacture “senior” doses and to evaluate the minimal effective dosage. J Am Geriatr Soc 1999;47:954-959.

Rochon PA, Gurwitz JH. Prescribing for elderly persons. JAMA 2000;283(3):339-340.

Sarver JH, Sudano JJ, Baker DW. Financial barriers to medication use among the elderly. Presented at AHSR Annual Meeting, June 2000.

Schmader K, Hanlon JT, Weinberge M, et al. Appropriateness of medication prescribing in ambulatory elderly patients. J Am Geriatr Soc 1994;42:1241-1247.

Scott-Levin, Managed Care Formulary Drug Audit, Fall 1999.

Smith NL, Psaty BM, Pitt B, et al. Temporal patterns in the medical treatment of congestive heart failure with angiotensin-converting enzyme inhibitors in older adults, 1989 through 1995. Arch Inter Med 1998;158:1074-1080.

Solomon DH, Gurwitz JH. Toxicity of nonsteroidal anti-inflammatory drugs in the elderly: is advanced age a risk factor? Amer J Med 1997;102(2):208-215.

Soumerai SB, Ross-Degnan D, Avorn J, et al. Effects of medicaid drug-payment limits on admission to hospitals and nursing homes. N Engl J Med 1991;325:1072-1077.

Soumerai SB, McLaughlin TJ, Gurwitz JH, et al. Timeliness and quality of care for elderly patients with acute myocardial infarction under health maintenance organization vs fee-for-service insurance. Archives Internal Med 1999;159(17):2013-2020.

Stuart B, Shea D, Briesacher BA. Prescription drug costs for medicare beneficiaries: coverage and health status matter. Issue Brief. The Commonwealth Fund, January 2000.

Stuck AE, Beer MH, Steiner A, et al. Inappropriate medication use in community-residing older persons. Arch Intern Med 1994;154:2195-2200.

Tobias DE, Pulliam CC. General and psychotherapeutic medication use in 878 nursing facilities: a 1997 national survey. Consult Pharm 1997;12:1401-1408.

U.S. Department of Health and Human Services. Prescription Drug Coverage, Spending, Utilization, and Prices: Report to the President. US DHHS, April 2000.

U.S. Food and Drug Administration. Guidance for Industry Content and Format for Geriatric Labeling DRAFT GUIDANCE, December 1998.

U.S. Food and Drug Administration. Guideline for the Study of Drugs Likely to Be Used in the Elderly. Rockville, Md; FDA, Center for Drug Evaluation and Research, 1989.

U.S. General Accounting Office. Prescription Drugs and the Elderly: Many Still Receive Potentially Harmful Drugs Despite Recent Improvements. GAO/HHEH-95-152, July 1995.

U.S. General Accounting Office. Prescription Drug Benefits: Implications for Beneficiaries of Medicare HMO Use of Formularies. GAO/HEHS-99-166, July 1999.

U.S. General Accounting Office. Adverse Drug Events: Substantial Problem but Magnitude Uncertain. Statement of Janet Heinrich before Committee on Health Education, Labor, and Pensions. GAO/T-HEH-00-53, February1, 2000.

Walser BL, Ross-Degnan D, Soumerai SB. Do open formularies increase access to clinically useful drugs? Health Affairs 1996;15(3):95-109.

Willcox SM, Himmelstein DU, Woolhandler S. Inappropriate drug prescribing for the community-dwelling elderly. JAMA 1994;272:292-296.

Williams BR, Nichol MB, Lowe B, et al. Medication use in residential care facilities for the elderly. Annals of Pharmacotherapy 1999;33(2):149-155.

Zuckerman IH, Gilner MP, Kreyenbuhl J, et al. Prevalence of potentially inappropriate drug use in the elderly maryland medicaid population. J Am Pharm Assoc 1997;NS37(2):254-255.