Morbidity and Mortality from Medication

I find the medicine worse than the malady.

Beaumont and Fletcher, 1647

A fundamental objective of professional practice is to help individual patients or clients solve problems. Specifically, professionals apply general knowledge, e.g., scientific knowledge, to specific circumstances, governed by the principle of beneficence.1 Further, the objectives of beneficent action are, in order of priority:

2. To prevent harm

3. To remove harm

4. To promote good

The first principle is to do no harm. However, this is seldom a satisfactory goal. A professional can avoid committing errors or doing harm by doing nothing. To be worth his fee, so to speak, the professional must act to prevent harm, to remove harm, or even to promote good. To promote good is both a professional aspiration and a motivation.

Drug therapy would appear near the top of most people's lists of how a health professional might remove the harm of disease or promote good. More than half of all physician office visits result in one or more prescriptions. The ideal objective of drug therapy is to improve the quality of a patient's life.2 In part because of legal requirements for the licensing of new drugs, drug products have more rigorous scientific evidence regarding safety, basic efficacy, and often, physiological effects than any other mode of therapy. It seems that drug therapy would exemplify the idea of applying scientific knowledge to improve people's lives. This very often occurs. One need only cite antibiotics to establish this.

Furthermore, an elaborate procedure to regulate drug marketing has developed over the years in response to various disasters, such as the use of a toxic vehicle for sulfanilamide and the birth of thalidomide babies. Almost every nation requires rigorous clinical trials to establish safety and efficacy, and limits claims regarding safety and effectiveness. So, on the one hand, the dangers of drug products are widely recognized. On the other hand, effectiveness and safety exist in a balance. The drug products marketed in the United States and other developed nations are arguably as safe as they can be without sacrificing access to effective drugs.

Premarketing clinical trials are carried out according to strict procedures (research protocols) that were approved by the Food and Drug Administration (FDA). These protocols define the population, especially the diseases to be treated; comorbidities (concomitant diseases) to be excluded; the manner of drug use; and the required clinical testing and reporting. Some populations (e.g., elderly, children, pregnant women) tend to be excluded from drug trials unless it is absolutely necessary to include them. This is for their protection, but it means that scientific studies of drug use in those populations may be scant or slow to appear.

After marketing, drugs may be used for many more indications and for more types of people than those that were included in the clinical trials. This so-called "off-label" use is possible because the states control the practice of medicine. The FDA is specifically prohibited from interfering with the practice of medicine. Doctors often need the flexibility to use their judgment in treating a patient. Also, some populations and indications would probably never be included in labeling. (See the discussion in Chapter 5.)

If medicines could be used as rigorously in daily practice as they are in clinical trials, perhaps they would be as safe and effective as manufacturers and regulators say they are. However, this would be difficult or at best impractical within the normal arrangements of community and hospital practice. It is increasingly obvious that drug product safety is not equivalent to drug therapy safety. This distinction is important and far reaching, and a specialized vocabulary is needed to describe it.

Furthermore, once such a distinction is clear, it may become apparent that the risks of drug therapy are not as widely appreciated as the risks of drug products. Few governments, for example, regulate drug therapy, despite strict regulation of drug products. Furthermore, the medication use policies that do exist in hospitals and managed care organizations, and the priorities of many health care professionals and patients, seem inconsistent with a full appreciation either of the problem or of its possible solutions.

Review of Research Data on Adverse Outcomes of Drug Therapy

An expanding literature documents a widespread problem: the legitimate use of governmentally approved drug products often results in adverse effects and treatment failures, and in turn, costly emergency visits, hospital admissions, transfers to intensive care, and deaths. Another related, but less appreciated problem is treatment failure caused by nonuse of needed medicines, because the doctor did not prescribe it, the pharmacist did not dispense it, or the patient did not take it.

Adverse Drug Reactions

According to the World Health Organization (WHO), an adverse drug reaction (ADR) is a response to a drug which is noxious and unintended and which occurs in man at doses normally used for prophylaxis, diagnosis or therapy of disease, or for modification of physiological function.

There are literally hundreds of published studies on the prevalence of ADRs. Lazarou et al. reviewed 39 studies, conducted among 62,480 patients over a 32-year period, estimating the risks of adverse drug reactions in hospitals. The overall incidence of serious ADRs was 6.7% of hospitalized patients. The incidence of fatal ADRs was 0.32%. They estimated that in 1994, 2,216,000 hospitalized patients had serious ADRs, including 106,000 fatal ADRs, making these reactions between the fourth and sixth leading cause of death in the United States.3

Lazarou et al., among others, assert that ADRs are not preventable by definition. They argue that ADRs (by definition) are unintended and occur at normal doses.3-5 However, other investigators have found, on review of specific cases, that some ADRs (or at least their consequences) may be preventable. This issue is discussed further below.

Preventable Adverse Outcomes

An expanding literature documents preventable illness, hospital admissions, transfers to intensive care, and deaths caused by the misuse of drug products that had been approved as safe and effective. Two basic approaches have been used to estimate prevalence: medical record review and application of preventable drug-related morbidity (PDRM) indicators.

Medical Record Review

Medical record review, or medical record audit, is so called because the data sources are the detailed records of the admission and stay. Medical record review by qualified experts is generally considered to be the gold standard for evaluations of processes requiring judgment. Some studies reviewed a data summary abstracted from medical records, while others interviewed patients and added the interview results to the data available from the patients' official medical records.

The investigators were usually among the reviewers, and presumably chose additional qualified reviewers when necessary. Some reports described formal means to increase reliability, e.g., having more than one reviewer for each case and using criteria for interreviewer agreement.

In a typical study of drug-related admissions, the sample would comprise patients admitted to one or more hospital units. The clinician-investigators reviewed those records or an abstract (summary) to evaluate the reason for hospital admission. In the studies reviewed below, the reviewers used more or less implicit definitions of drug relatedness and preventability. The investigators then counted the number of drug-related admissions (DRAs) and further subdivided them according to whether they had been preventable. The prevalence would then be calculated as the ratio of DRAs (or preventable DRAs) to the total number of admissions reviewed.

Winterstein et al.6 carried out a systematic review of preventable drug-related hospital admissions (PDRAs). They selected 15 studies from 8 nations. These studies are summarized in Table 2.1. Details of sampling and study methods varied widely.

The 15 studies report a median DRA prevalence of 7.1 per 100 hospital admissions (range, 2.5 to 25%). The median prevalence of preventable DRAs was 4.3 per 100 hospital admissions (range, 1.4 to 15%). Overall, the median preventability rate (PDRAs/DRAs) was 58.9% (range, 32 to 86%).

The range of PDRAs in studies from the United States was 2.3 to 15.2%, with a calculated median of about 7.9%. Two studies from the U.K. showed


Studies of Preventable Drug-Related Hospital Admissions


Studies of Preventable Drug-Related Hospital Admissions

Author, Year, Country (reference no.)

Sample Size

DRAs PDRAs as % of as % of Preventability Admissions Admissions Rate (%)

Bero et al., 1991, U.S. (4)





Bigby et al., 1987, U.S. (7)





Courtman and Stallings, 1995, Canada (8)





Cunningham et al., 1997, U.K. (9)





Darchy et al., 1999, France (10)





Dartnell et al., 1996, Australia (11)





Hallas et al., 1992, Denmark (12)





Lakshmanan et al., 1986, U.S. (13)





Lindley et al., 1992, U.K. (14)





Nelson and Talbert, 1996, U.S. (15)





Ng, et al., 1999, Australia (16)





Nikolaus et al., 1992, Germany (17)





Raschetti et al., 1997, Italy (18)





Trunet et al., 1980, France (19)





Trunet et al., 1986, France (20)




















Source: Winterstein et al., Ann. Pharmacother., 36, 1238, 2002.

Source: Winterstein et al., Ann. Pharmacother., 36, 1238, 2002.

that PDRAs account for 3.1 and 4.3% of admissions, with preventability rates of 50 and 80%, about at the median found in studies from other countries.

Table 2.2 summarizes six studies of PDRM among hospitalized inpatients. In these studies, patients already hospitalized would be followed, and patients with possible DRM typically would be identified using screening criteria or voluntary reports from pharmacists or nurses. These patients' medical records would then be evaluated for drug-related morbidity by medical record review. The incidence would then typically be calculated as the ratio of DRM or PDRM to the total number of hospital stays (admissions) during the study.

The incidence of PDRM among inpatients ranges from 0.32 to 3.9%, with a median of about 1.5%. The preventability rate ranges from 20 to 56%, with a calculated median of approximately 41%.

Studies of preventable death caused by drug therapy (contrasted to death from an adverse drug reaction) are difficult to review systematically. Most focus on specific diseases, e.g., asthma. Preventable deaths are rather rare events on a population basis. All studies oversampled for death. That is, in effect, they searched for patients who had died, rather than counting the deaths in a sample drawn sequentially or at random from a general population at risk.27-34 Examples are described below.

Selected Examples Hospitalization

Bero et al.,4 at the University of California, followed 706 elderly patients discharged from a California hospital. Within 6 months of discharge, 247 (35%) reentered the hospital. About one fifth (45) of the readmissions were drug related. The most frequently identified drug-related problems were unexpected adverse drug reactions (10), patient noncompliance (10), overdose (8), lack of a necessary drug therapy (6), and underdose (5). Drug-related factors were a major reason for readmission in half the cases. The majority (76%) of the problems identified were potentially preventable. The authors concluded that specific drug-related problems could become targets for preventive interventions.

Bigby et al.7 studied 686 emergency admissions of patients from their own hospital-based primary care practice. In their judgment, 59 (9%) of the admissions were potentially preventable. Medical care, including inadequate follow-up and adverse drug reactions, caused 40 admissions; lack of patient compliance caused 12; and both medical care and noncompliance caused 7. Adverse drug reactions were the most common cause of treatment problems, and warfarin was the most common cause of an adverse drug reaction. Inadequate follow-up of abnormal physical findings, symptoms, and laboratory test results was also important.

Wayne Ray and colleagues49 at Vanderbilt studied 1021 patients with hip fractures in a study design that matched injured patients to normal controls.


Preventable Drug-Related Morbidity in Inpatients e a e S

Author, Year,

Country (reference no.)

Sampling Type, Sample Size, Setting

Sample Descriptionh

Prevalence of DRM

Prevalence of PDRM

Preventability Rate

Outcome: Significant, Serious, Life-Threatening, or Fatal Adverse Drug Events

Prospective n = 4,031 HAs (21,412

IPDs) 2 tertiary hospitals Prospective n = 420 HAs (2967 IPDs)

Prospective n = 379 HAs

SRS of (+) all adults admitted to 11 units of 2 hospitals over 6 months, February to July 1993;ab (-) obstetric Pts.

(+) All adults admitted to 7 units (2 medical, 2 surgical, 2 obstetric general care, 1 coronary IC) during 37 days in August and September 1990a

(+) All adults admitted over 51 days during October and November 1992 to 3 medical units: 2 general medical, 1 ICUe

Outcome: Disability, Death, or Prolonged Hospital Stay

Wilson et al., 1995, Australia (24)

Retrospective n = 14,179 PRs SRS from 28 hospitals with > 3000 admissions in 2 states

Retrospective n = 30,195 PRs

(+) RS of at least 520 HAs from each hospital (-) Hospitals with less than 3000 eligible admissions per annum, day-only admissions, admissions to psychiatric wards mean age = 43.8f

Outcome: Cardiac Arrest r

Bedell et al., 1991, Prospective (+) All inpatients receiving CPR and 15/203 (7.4%) 8/203 (3.9%) 8/15 (53%) t

U.S. (26) n = 203 discharged patients with cardiac arrest a

Teaching hospital within 24 h after dischargeg S

Note: HA = hospital admission; MRA = medical record audit, or review; Pt. = patient; PR = patient medical record; IPD = inpatient day; RS = random ^

sample; SRS = stratified random sample. o t aPotential cases were assessed by MRAs using two independent reviewers, and a third when necessary to break a tie. sS

bOversampled specific patient groups. Definite and probable ADEs. cNumber of admissions estimated from patient days.

dPotential cases were assessed by medical record review using one reviewer (90%) or two independent reviewers (10%). eRecords assessed by medical record review using two independent reviewers, and a third when necessary to break a tie.

fTwo independent reviewers. e gInternist reviewed hospital medical charts within 24 h after CPR and interviewed staff for clarification; assessment by three internists. h(+) = inclusion criteria; (-) = exclusion criteria.

Was this article helpful?

0 0
Dealing With Asthma Naturally

Dealing With Asthma Naturally

Do You Suffer From ASTHMA Chronic asthma is a paralyzing, suffocating and socially isolating condition that can cause anxiety that can trigger even more attacks. Before you know it you are caught in a vicious cycle Put an end to the dependence on inhalers, buying expensive prescription drugs and avoidance of allergenic situations and animals. Get control of your life again and Deal With Asthma Naturally

Get My Free Ebook

Post a comment