Errors in Health Care:
A Leading Cause of Death and Injury
Health care is not as safe as it should be. A substantial body of evidence points to medical errors as a leading cause of death and injury.
• Sizable numbers of Americans are harmed as a result of medical errors. Two studies of large samples of hospital admissions, one in New York using 1984 data and another in Colorado and Utah using 1992 data, found that the proportion of hospital admissions experiencing an adverse event, defined as injuries caused by medical management, were 2.9 and 3.7 percent,1 respectively. The proportion of adverse events attributable to errors (i.e., preventable adverse events) was 58 percent in New York, and 53 percent in Colorado and Utah.2
• Preventable adverse events are a leading cause of death in the United States. When extrapolated to the over 33.6 million admi ssions to U.S. hospitals in 1997, the results of these two studies imply that at least 44,000 and perhaps as many as 98,000 Americans die in hospitals each year as a result of medical errors.3 Even when using the lower estimate, deaths in hospitals due to preventable adverse events exceed the number attributable to the 8th-leading cause of death.4 Deaths due to preventable adverse events exceed the deaths attributable to motor vehicle accidents (43,458), breast cancer (42,297) or AIDS (16,516).5
• Total national costs (lost income, lost household production, disability, health care costs) are estimated to be between $37.6 billion and $50 billion for adverse events and between $17 billion and $29 billion for preventable adverse events.6 Health care costs account for over one-half of the total costs. Even when using the lower estimates, the total national costs associated with adverse events and preventable adverse events represent approximately 4 percent and 2 percent, respectively, of national health expenditures in 1996.7 In 1992, the direct and indirect costs of adverse events were slightly higher than the direct and indirect costs of caring for people with HIV and ADS.8
• In terms of lives lost, patient safety is as important an issue as worker safety. Although more than 6,000 Americans die from workplace injuries every year,9,10 in 1993 medication errors are estimated to have accounted for about 7,000 deaths.11 Medication errors account for one out of 131 outpatient deaths and one out of 854 inpatient deaths.
• Medication-related errors occur frequently in hospitals; not all result in actual harm, but those that do are costly. One recent study conducted at two prestigious teaching hospitals found that almost two percent of admissions experienced a preventable adverse drug event, resulting in average increased hospital costs of $4,700 per admission or about $2.8 million annually for a 700-bed teaching hospital.12 If these findings are generalizable, the increased hospital costs alone of preventable adverse drug events affecting inpatients are about $2 billion for the nation as a whole.
• Hospital patients represent only a fraction of the total population at risk of experiencing a medication-related error. In 1998, nearly 2.5 billion prescriptions were dispensed by U.S. pharmacies at a cost of about $92 billion.13 Numerous studies document errors in prescribing medications,14,15 dispensing by pharmacists,16 and unintentional nonadherence on the part of the patient.17 Medication errors have the potential to increase as a major contributor to avoidable morbidity and mortality as new medications are introduced for a wider range of indications.
This chapter provides a summary of findings in the literature on the frequency and cost of health care errors and the factors that contribute to their occurrence.
Although the literature pertaining to errors in health care has grown steadily over the last decade and some notable studies are particularly strong
methodologically, we do not yet have a complete picture of the epidemiology of errors. Many studies focus on patients experiencing injury and provide valuable insight into the magnitude of harm resulting from errors. Other studies, more limited in number, focus on the occurrence of errors, both those that result in harm and those that do not (sometimes called ''near misses"). More is known about errors that occur in hospitals than in other health care delivery settings.
Synthesizing and interpreting the findings in the literature pertaining to errors in health care is complicated due to the absence of standardized nomenclature. For purposes of this report, the terms error and adverse event are defined as follows:
An error is defined as the failure of a planned action to be completed as intended (i.e., error of execution) or the use of a wrong plan to achieve an aim (i.e., error of planning).18
An adverse event is an injury caused by medical management rather than the underlying condition of the patient. An adverse event attributable to error is a "preventable adverse event."19 Negligent adverse events represent a subset of preventable adverse events that satisfy legal criteria used in determining negligence (i.e., whether the care provided failed to meet the standard of care reasonably expected of an average physician qualified to take care of the patient in question).20
When a study in the literature has used a definition that deviates from the above definitions, it is noted below.
Medication-related error has been studied extensively for several reasons: it is one of the most common types of error, substantial numbers of individuals are affected, and it accounts for a sizable increase in health care costs.21,22,23 There are also methodologic issues: (1) prescription drugs are widely used, so it is easy to identify an adequate sample of patients who experience adverse drug events; (2) the drug prescribing process provides good documentation of medical decisions, and much of this documentation resides in automated, easily accessible databases; and (3) deaths attributable to medication errors are recorded on death certificates. There are probably other areas of health care delivery that have been studied to a lesser degree but may offer equal or greater opportunity for improvement in safety.
Efforts to assess the importance of various types of errors are currently hampered by the lack of a standardized taxonomy for reporting adverse events, errors, and risk factors.24,25 A limited number of studies focus di-
rectly on the causes of adverse events, but attempts to classify adverse events according to "root causes" are complicated by the fact that several interlocking factors often contribute to an error or series of errors that in turn result in an adverse event.26,27 In recent years, some progress toward a more standardized nomenclature and taxonomy has been made in the medication area, but much work remains to be done.28
The following discussion of the literature addresses four questions:
1. How frequently do errors occur?
2. What factors contribute to errors?
3. What are the costs of errors?
4. Are public perceptions of safety in health care consistent with the evidence?
How Frequently do Errors Occur?
For the most part, studies that provide insight into the incidence and prevalence of errors fall into two categories:
1. General studies of patients experiencing adverse events. These are studies of adverse events in general, not studies limited to medication-related events. These studies are limited in number, but some represent large-scale, multi-institutional analyses. Virtually all studies in this category focus on hospitalized patients. With the exception of medication-related events discussed in the second category, little if any research has focused on errors or adverse events occurring outside of hospital settings, for example, in ambulatory care clinics, surgicenters, office practices, home health, or care administered by patients, their family, and friends at home.
2. Studies of patients experiencing medication-related errors. There is an abundance of studies that fall into this category. Although many focus on errors and adverse events associated with ordering and administering medication to hospitalized patients, some studies focus on patients in ambulatory settings.
An adverse event is defined as an injury caused by medical management rather than by the underlying disease or condition of the patient.29 Not all, but a sizable proportion of adverse events are the result of errors. Numerous
studies have looked at the proportion of adverse events attributable to medical error. Due to methodologic challenges, far fewer studies focus on the full range of error—namely, those that result in injury and those that expose the patient to risk but do not result in injury.
The most extensive study of adverse events is the Harvard Medical Practice Study, a study of more than 30,000 randomly selected discharges from 51 randomly selected hospitals in New York State in 1984.30 Adverse events, manifest by prolonged hospitalization or disability at the time of discharge or both, occurred in 3.7 percent of the hospitalizations. The proportion of adverse events attributable to errors (i.e., preventable adverse events) was 58 percent and the proportion of adverse events due to negligence was 27.6 percent. Although most of these adverse events gave rise to disability lasting less than six months, 13.6 percent resulted in death and 2.6 percent caused permanently disabling injuries. Drug complications were the most common type of adverse event (19 percent), followed by wound infections (14 percent) and technical complications (13 percent).31,32
The findings of the Harvard Medical Practice Study in New York have recently been corroborated by a study of adverse events in Colorado and Utah occurring in 1992.33 This study included the review of medical records pertaining to a random sample of 15,000 discharges from a representative sample of hospitals in the two states. Adverse events occurred in 2.9 percent of hospitalizations in each state. Over four out of five of these adverse events occurred in the hospital, the remaining occurred prior to admission in physicians' offices, patients' homes or other non-hospital settings. The proportion of adverse events due to negligence was 29.2 percent, and the proportion of adverse events that were preventable was 53 percent.34 As was the case in the New York study, over 50 percent of adverse events were minor, temporary injuries. But the study in New York found that 13.6 percent of adverse events led to death, as compared with 6.6 percent in Colorado and Utah. In New York, about one in four negligent adverse events led to death, while in Colorado and Utah, death resulted in about 1 out of every 11 negligent adverse events. Factors that might explain the differences between the two studies include: temporal changes in health care, and differences in the states' patient populations and health care systems.35
Both the study in New York and the study in Colorado and Utah identified a subset of preventable adverse events that also satisfied criteria applied by the legal system in determining negligence. It is important to note that although some of these cases may stem from incompetent or impaired providers, the committee believes that many could likely have been avoided had better systems of care been in place.
Extrapolation of the results of the Colorado and Utah study to the over 33.6 million admissions to hospitals in the United States in 1997, implies that at least 44,000 Americans die in hospitals each year as a result of preventable medical errors.36 Based on the results of the New York study, the number of deaths due to medical error may be as high as 98,000.37 By way of comparison, the lower estimate is greater than the number of deaths attributable to the 8th-leading cause of death.38
Some maintain these extrapolations likely underestimate the occurrence of preventable adverse events because these studies: (1) considered only those patients whose injuries resulted in a specified level of harm; (2) imposed a high threshold to determine whether an adverse event was preventable or negligent (concurrence of two reviewers); and (3) included only errors that are documented in patient records.39
Two studies that relied on both medical record abstraction and other information sources, such as provider reports, have found higher rates of adverse events occurring in hospitals. In a study of 815 consecutive patients on a general medical service of a university hospital, it was found that 36 percent had an iatrogenic illness, defined as any illness that resulted from a diagnostic procedure, from any form of therapy, or from a harmful occurrence that was not a natural consequence of the patient's disease.40 Of the 815 patients, nine percent had an iatrogenic illness that threatened life or produced considerable disability, and for another two percent, iatrogenic illness was believed to contribute to the death of the patient.
In a study of 1,047 patients admitted to two intensive care units and one surgical unit at a large teaching hospital, 480 (45.8 percent) were identified as having had an adverse event, where adverse event was defined as "situations in which an inappropriate decision was made when, at the time, an appropriate alternative could have been chosen."41 For 185 patients (17.7 percent), the adverse event was serious, producing disability or death. The likelihood of experiencing an adverse event increased about six percent for each day of hospital stay.
Some information on errors can also be gleaned from studies that focus on inpatients who died or experienced a myocardial infarction or postsurgical complication. In a study of 182 deaths in 12 hospitals from three conditions (cerebrovascular accident, pneumonia, or myocardial infarction), it was found that at least 14 percent and possibly as many as 27 percent of the deaths might have been prevented.42 A 1991 analysis of 203 incidents of cardiac arrest at a teaching hospital,43 found that 14 percent followed an iatrogenic complication and that more than half of these might have been prevented. In a study of 44,603 patients who underwent surgery between
1977 and 1990 at a large medical center, 2,428 patients (5.4 percent) suffered complications and nearly one-half of these complications were attributable to error.44 Another 749 died during the same hospitalization; 7.5 percent of these deaths were attributed to error.
Patients who died during surgery requiring general anesthesia have been the focus of many studies over the last few decades. Anesthesia is an area in which very impressive improvements in safety have been made. As more and more attention has been focused on understanding the factors that contribute to error and on the design of safer systems, preventable mishaps have declined.45,46,47,48 Studies, some conducted in Australia, the United Kingdom and other countries, indicate that, today, anesthesia mortality rates are about one death per 200,000–300,000 anesthetics administered, compared with two deaths per 10,000 anesthetics in the early 1980s.49 The gains in anesthesia are very impressive and were accomplished through a variety of mechanisms, including improved monitoring techniques, the development and widespread adoption of practice guidelines, and other systematic approaches to reducing errors.50
Lastly, some studies have relied on incident reporting systems to identify and analyze errors. For example, in Australia, 324 general practitioners participating voluntarily in an incident reporting system reported a total of 805 incidents during October 1993 through June 1995, of which 76 percent were preventable and 27 percent had the potential for severe harm.51 These studies provide information on the types of errors that occur but are not useful for estimating the incidence of errors, because the population at risk (i.e., the denominator) is generally unknown.
Even though medication errors that result in death or serious injury occur infrequently, sizable and increasing numbers of people are affected because of the extensive use of drugs in both out-of-hospital and in-hospital settings. In 1998, nearly 2.5 billion prescriptions were dispensed in U.S. pharmacies at an estimated cost of about $92 billion.52 An estimated 3.75 billion drug administrations were made to patients in hospitals.53
In a review of U.S. death certificates between 1983 and 1993, it was found that 7,391 people died in 1993 from medication errors (accidental poisoning by drugs, medicaments, and biologicals that resulted from acknowledged errors by patients or medical personnel), compared with 2,876 people in 1983, representing a 2.57-fold increase.54 Outpatient deaths due
to medication errors rose 8.48-fold during the 10-year period, compared with a 2.37-fold increase in inpatient deaths.
Medication Errors in Hospitals
Medication errors occur frequently in hospitals. Numerous studies have assessed the incidence of adverse drug events (ADEs), defined as an injury resulting from medical intervention related to a drug.55 Not all ADEs are attributable to errors. For example, a patient with no history of allergic reactions to drugs, who experiences an allergic reaction to an antibiotic, has suffered an ADE, but this ADE would not be attributable to error. However, an error would have occurred if an antibiotic was prescribed to a patient with a history of documented allergic reactions, because the medical record was unavailable or not consulted. We discuss only those studies of ADEs that identified the subset of ADEs determined to be preventable (i.e., attributable to errors).
In an analysis of 289,411 medication orders written during one year in a tertiary-care teaching hospital, the overall error rate was estimated to be 3.13 errors for each 1,000 orders written and the rate of significant errors to be 1.81 per 1,000 orders.56 In a review of 4,031 adult admissions to 11 medical and surgical units at two tertiary care hospitals, Bates et al. identified 247 ADEs for an extrapolated event rate of 6.5 ADEs per 100 nonobstetrical admissions, and a mean number per hospital per year of approximately 1,900 ADEs.57 Twenty-eight percent were judged preventable.
In a study of patients admitted to coronary intensive care, medical, surgical, and obstetric units in an urban tertiary care hospital over a 37-day period, the rate of drug-related incidents was 73 in 2,967 patient-days: 27 incidents were judged ADEs; 34, potential ADEs; and 12, problem orders.58 Of the 27 ADEs, five were life threatening, nine were serious, and 13 were significant. Of the 27 ADEs, 15(56 percent) were judged definitely or probably preventable. In a study of prescribing errors detected and averted by pharmacists in a 631-bed tertiary care teaching hospital between July 1994 and June 1995, the estimated overall rate of errors was 3.99 per 1,000 medication orders.59
Children are at particular risk of medication errors, and as discussed below, this is attributable primarily to incorrect dosages.60,61 In a study of 101,022 medication orders at two children's teaching hospitals, a total of 479 errant medication orders were identified, of which 27 represented potentially lethal prescribing errors.62 The frequency of errors was similar at
the two institutions, 4.9 and 4.5 errors per 1,000 medication orders. The error rate per 100 patient-days was greater in the pediatric intensive care units (PICUs) than in the pediatric ward or neonatal intensive care units, and the authors attribute this to the greater heterogeneity of patients cared for in PICUs and the broad range of drugs and dosages used. In a four-year prospective quality assurance study, 315 medication errors resulting in injury were reported among the 2,147 neonatal and pediatric intensive care admissions, an error rate of one per 6.8 admissions.63 The frequency of iatrogenic injury of any sort due to a medication error was 3.1 percent—one injury for each 33 intensive care admissions.
Not surprisingly, the potential for medication-related error increases as the average number of drugs administered increases. In a prospective cohort study of 4,031 adult admissions to 11 medical and surgical units in two tertiary care hospitals (including two medical and three surgical ICUs), the rate of preventable ADEs and preventable potential ADEs in ICUs was 19 events per 1,000 patient-days, nearly twice the rate of non-ICUs.64 When adjusted for the number of drugs used in the previous 24 hours or ordered since admission, there were no differences in error rates between ICUs and nonICUs.
Current estimates of the incidence of medication errors are undoubtedly low because many errors go undocumented and unreported.65,66,67,68 For example, in a study of patients admitted to five patient care units at a tertiary care hospital during a six month period in 1993, it was found that incident reports were filed with the hospital's quality assurance program or called into the pharmacy hotline for only three of the 54 people experiencing an adverse drug event.69
Some errors are also difficult to detect in the absence of computerized surveillance systems. In a study of 36,653 hospitalized patients, Classen et al. identified 731 ADEs in 648 patients, but only 92 of these were reported by physicians, pharmacists, and nurses.70 The remaining 631 were detected from automated signals, the most common of which were diphenhydramine hydrochloride and naloxone hydrochloride use, high serum drug levels, leukopenia, and the use of phytonadione and antidiarrheals.
Medication Errors in Ambulatory Settings
There is evidence indicating that ADEs account for a sizable number of admissions to inpatient facilities, but we do not know what proportion of these ADE-related admissions are attributable to errors. One study found
that between three and 11 percent of hospital admissions were attributable to ADEs.71 A review of 14 Australian studies published between 1988 and 1996 reported that 2.4 to 3.6 percent of all hospital admissions were drug related, and between 32 and 69 percent were definitely or possibly preventable. Drug groups most commonly involved were cytotoxics, cardiovascular agents, antihypertensives, anticoagulants, and nonsteroidal anti-inflammatory drugs.72
ADEs also result in increased visits to physician offices and emergency departments. In an analysis of 1,000 patients drawn from a community of-rice-based medical practice who were observed for adverse drug reactions, adverse effects were recorded in 42 (4.2 percent), of which 23 were judged to be unnecessary and potentially avoidable.73 In an analysis of 62,216 visits to an emergency department by patients enrolled in a health maintenance organization (HMO), it was found that 1,074 (1.7 percent) were related to medication noncompliance or inappropriate prescribing.74
There is a sizable body of literature to document the incidence of patient noncompliance with medication regimens, but less is known about the proportion of noncompliance attributable to medical error (defined as accidental or unintentional nonadherence to a therapeutic program) as opposed to intentional noncompliance. In a meta-analysis of seven studies, Sullivan et al. estimate that 5.5 percent of admissions can be attributed to drug therapy noncompliance, amounting to 1.94 million admissions and $8.5 billion in hospital expenditures in 1986.75 Similar results were obtained by Einarson in a meta-analysis of 37 studies published between 1966 and 1989, which found that hospital admissions caused by ADEs, resulting from noncompliance or unintentionally inappropriate drug use, ranged from 0.2 to 21.7 percent with a median of 4.9 percent and a mean of 5.5 percent.76 Patient noncompliance is clearly an important quality issue, but it should be emphasized that we do not know the extent to which noncompliance is related to errors.
Factors that Contribute to Errors
Studies of Adverse Events
Patient safety problems of many kinds occur during the course of providing health care. They include transfusion errors and adverse drug events; wrong-site surgery and surgical injuries; preventable suicides; restraint-related injuries or death; hospital-acquired or other treatment-related infections; and falls, burns, pressure ulcers, and mistaken identity. Leape et al.
have characterized the kinds of errors that resulted in medical injury in the Medical Practice Study as diagnostic, treatment, preventive, or other errors (see Box 2.1).
More than two-thirds (70 percent) of the adverse events found in this study were thought to be preventable, with the most common types of preventable errors being technical errors (44 percent), diagnosis (17 percent), failure to prevent injury (12 percent) and errors in the use of a drug (10 percent). The contributions of complexity and technology to such error rates is highlighted by the higher rates of events that occur in the highly technical surgical specialties of vascular surgery, cardiac surgery, and neurosurgery. In hospitals, high error rates with serious consequences are most likely in intensive care units, operating rooms and emergency departments.
Thomas et al., in their study of admissions to hospitals in Colorado and Utah experiencing adverse events, found that about 30 percent were attributable to negligence.77 The hospital location with the highest proportion of
negligent adverse events (52.6 percent) was the emergency department. The authors note the complexity inherent in emergency medical care and point to the need to improve teamwork and standardize work procedures.
Other studies have made similar attempts to classify errors. Dubois and Brook studied 49 preventable deaths from 12 hospitals, and found that for those who died of a myocardial infarction, preventable deaths reflected errors in management; for cerebrovascular accident, most deaths reflected errors in diagnosis; and for pneumonia, some deaths reflected errors in management and some reflected errors in diagnosis.78 In an analysis of 203 cardiac arrests at a teaching hospital, Bedell et al. found that of the half that might have been prevented, the most common causes of potentially preventable arrest were medication errors and toxic effects, and suboptimal response by physicians to clinical signs and symptoms.79
Studies of Medication Errors
Ensuring appropriate medication use is a complex process involving multiple organizations and professionals from various disciplines; knowledge of drugs; timely access to accurate and complete patient information; and a series of interrelated decisions over a period of time. As shown in Box 2.2, errors can creep into this process at various points. Some errors are errors of commission (e.g., administration of improper drug), while others are errors of omission (e.g., failure to administer a drug that was prescribed).
Medication errors are often preventable, although reducing the error rate significantly will require multiple interventions. In the study of prescribing errors conducted by Lesar et al.,80 the most common factors associated with errors were decline in renal or hepatic function requiring alteration of drug therapy (13.9 percent); patient history of allergy to the same medication class (12.1 percent); using the wrong drug name, dosage form, or abbreviation (11.4 percent for both brand name and generic name orders); incorrect dosage calculations (11.1 percent); and atypical or unusual and critical dosage frequency considerations (10.8 percent). The most common groups of factors associated with errors were those related to knowledge and the application of knowledge regarding drug therapy (30 percent); knowledge and use of knowledge regarding patient factors that affect drug therapy (29.2 percent); use of calculations, decimal points, or unit and rate expression factors (17.5 percent); and nomenclature—for example incorrect drug name, dosage form, or abbreviations (13.4 percent).
Many studies have identified inappropriate prescribing as a particu-
larly important factor in accounting for medication errors. In an analysis of 1987 National Medical Expenditure Survey data, it was found that physicians prescribe potentially inappropriate medications for nearly a quarter of all older people living in the community.81 In a study of 366 consecutive patients admitted to a department of cardiology, ''definite" or "probable" drug events (i.e., adverse drug reactions and dose-related therapeutic failures) accounted for 15 admissions, of which five were judged to be due to error in prescription and another five judged to have been avoidable had appropriate measures been taken by prescribing physicians.82 In an analysis of 682 children admitted to a Congenital Heart Disease Center at a teaching hospital in the United Kingdom, 441 medication errors were reported by
nurses, doctors, and pharmacists, of which prescribing errors accounted for 68 percent, followed by administration errors (25 percent) and supply errors (seven percent).83 In Burnum's84 analysis of 1,000 patients drawn from a community office-based medical practice who experienced adverse drug reactions, 23 patients were judged to have experienced an "unnecessary and potentially avoidable" event, 10 of which were due to physician error (i.e., six due to administration of a drug not indicated and four to improper drug administration).
Physicians do not routinely screen for potential drug interactions, even when medication history information is readily available. In an analysis of 424 randomly selected visits to a hospital emergency department, 47 percent led to added medication, and in 10 percent of the visits in which at least one medication was added, the new medication added a potential adverse interaction.85 In all cases, a medication history was recorded on the patients and available to the physicians.
Errors can occur in the dispensing of drugs by pharmacists. In a recent investigation of pharmacists, the Massachusetts State Board of Registration in Pharmacy estimated that 2.4 million prescriptions are filled improperly each year in Massachusetts.86 Eighty-eight percent of the errors involved giving patients the wrong drug or the wrong strength.
Errors in the ordering and administration of medications are common in hospitals. Bates et al.,87 in an analysis of more than 4,000 admissions to two tertiary care hospitals, found that about 28 percent of 247 adverse drug events were preventable and most of these resulted from errors that occurred at the stages of ordering and administration. Davis and Cohen88 in their review of the literature and other evidence on errors report an error rate of 12 percent to be common in the preparation and administration of medications in hospitals. In a study of medication orders at two children's teaching hospitals, Folli et al.89 found that errors occurred in almost five out of every 1,000 orders and that the most prevalent error was overdose.
Patients make errors too. With greater emphasis on community-based long-term care, increased ambulatory surgery, shorter hospital lengths of stay, and greater reliance on complex drug therapy, patients play an increasingly important role in the administration of drugs. Greenberg et al.90 found that 4.3 percent of the elderly enrolled in Medicare social HMOs required assistance with the administration of medications. The inability to manage complex drug therapies explains why some elderly are in institutional rather than community-based long-term-care settings.91
Automated information and decision support systems are effective in reducing many types of errors. In an analysis of admissions to 11 medical
and surgical units at two tertiary care hospitals, Leape et al.92 identified 334 errors as the causes of 264 preventable ADEs and potential ADEs. About three out of four errors were caused by one of seven types of systems failures (drug knowledge dissemination, dose and identity checking, patient information availability, order transcription, allergy defense, medication order tracking, and interservice communication), and all could have been improved by better information systems that disseminate knowledge about drugs and make drug and patient information readily accessible at the time it is needed.
Computerized drug order entry systems have much potential to reduce errors. In a study of 379 consecutive admissions to three medical units at an urban tertiary care hospital, 10,070 medication orders were written and 530 medication errors were identified (5.3 errors per 100 orders). More than half of the medication errors involved at least one missing dose of a medication.93 Of the 530 medication errors, five (0.9 percent) resulted in adverse drug events that were judged preventable, and another 35 represented potential adverse drug events (i.e., medication errors with the potential for injury but in which no injury occurred). Physician computer order entry could have prevented 84 percent missing dose medication errors, 86 percent of potential adverse drug events, and 60 percent of preventable adverse drug events. However, more sophisticated technology is not the only option; involving pharmacists in reviewing drug orders significantly reduced the potential harm resulting from errant medication orders.94,95
The Cost of Errors
In addition to the unfortunate health consequences suffered by many as a result of medical error, there are direct and indirect costs borne by society as a whole as a result of medical errors. Direct costs refer to higher health care expenditures, while indirect costs include factors such as lost productivity, disability costs, and personal costs of care.
Based on analysis of 459 adverse events identified by reviewing the medical records of 14,732 randomly selected 1992 discharges from 28 hospitals in Colorado and Utah, Thomas et al. estimated the total costs (lost income, lost household production, disability and health care costs) to be nearly $662 million of which health care costs totaled $348 million.96 The total costs associated with the 265 of the 459 adverse events that were found to be preventable were $308 million, of which $159 million represented health care costs. Based on extrapolation to all hospital admissions in the United
States, the authors estimate the national costs of adverse events to be $37.6 billion and of preventable adverse events to be $17 billion. The total national costs associated with adverse events was approximately 4 percent of national health expenditures in 1996. In 1992, the direct and indirect costs of adverse events were slightly higher than the direct and indirect costs of caring for people with HIV and AIDS.
It has been estimated that for every dollar spent on ambulatory medications, another dollar is spent to treat new health problems caused by the medication.97 Studies of the direct costs of medication-related errors fall into three categories; (1) population-based studies of patients in a community or health plan; (2) studies of medication-related errors that occur in hospitals; and (3) studies of medication-related errors that occur in nursing homes.
One estimate places the annual national health care cost of drug-related morbidity and mortality in the ambulatory setting as high as $76.6 billion in 1994.98 Not all drug-related morbidity and mortality is preventable, but numerous studies document errors in prescribing,99,100 dispensing by pharmacists,101 and unintentional nonadherence on the part of the patient.102
Medication-related errors occur frequently, most do not result in actual harm, but those that do are costly. One recent study conducted at two prestigious teaching hospitals found that almost two percent of admissions experienced a preventable ADE, resulting in an average increased length of stay of 4.6 days and an average increased hospital cost of nearly $4,700 per admission.103 This amounts to about $2.8 million annually for a 700-bed teaching hospital, and if these findings are generalizable, the increased hospital costs alone of preventable adverse drug events affecting inpatients are about $2 billion for the nation as a whole.
In a matched case-control study of all patients admitted to a large teaching hospital from January 1990 through December 1993, it was found that adverse drug events complicated 2.43 admissions per 100.104 Controls were matched to cases on primary discharge diagnosis related group (DRG), age, sex, acuity, and year of admission. The occurrence of an ADE was associated with an increased length of stay of 1.91 days and an increased cost of $2,262. The increased risk of death among patients experiencing an adverse drug event was 1.88.
Other studies corroborate the high cost of medication-related errors. One study conducted in a university-affiliated medical center hospital estimated that the annual costs of treating the 1,911 medication-related problems identified through the hospital's voluntary reporting system in 1994
totaled slightly less than $1.5 million.105 Bloom has estimated that $3.9 billion was spent in 1983 to manage the preventable gastrointestinal adverse effects of nonsteroidal anti-inflammatory drugs.106
Medication-related errors also occur in nursing homes. For every dollar spent on drugs in nursing facilities, $1.33 is consumed in the treatment of drug-related morbidity and mortality, amounting to $7.6 billion for the nation as a whole, of which $3.6 billion has been estimated to be avoidable.107
Public Perceptions of Safety
Although the risk of dying as a result of a medical error far surpasses the risk of dying in an airline accident, a good deal more public attention has been focused on improving safety in the airline industry than in the health care industry. The likelihood of dying per domestic jet flight is estimated to be one in eight million.108 Statistically, an average passenger would have to fly around the clock for more than 438 years before being involved in a fatal crash. This compares very favorably with a death risk per domestic flight of one in two million during the decade 1967–1976. Some believe that public concern about airline safety, in response to the impact of news stories, has played an important role in the dramatic improvement in safety in the airline industry.
The American public is aware that health care is less safe than some other environments, but to date, it has made few demands on the health care industry to demonstrate improvement. In a public opinion poll conducted by Louis Harris & Associates for the National Patient Safety Foundation, the health care environment was perceived as "moderately safe" (rated 4.9 on a scale of one through seven where one is not safe at all and seven is very safe).109 Respondents viewed the health care environment as much safer than nuclear power or food handling, but somewhat less safe than airline travel or the work environment.
Americans have a very limited understanding of health care safety issues. When asked, What comes to mind when you think about patient safety issues in the health care environment? 28 percent of respondents did not mention anything, 20 percent mentioned exposure to infection, 13 percent cited the general level of care patients receive, and 11 percent cited qualifications of health professionals.110 When asked about the main cause of medical mistakes, respondents most frequently cited carelessness or negligence (29 percent) of health care professionals, who are overworked, worried, or stressed (27 percent).
Most people learn about medical mistakes through anecdotes. More than four out of five respondents have heard about a situation in which a medical mistake was made.111 When asked how they heard about the most recent medical mistake, 42 percent cited a friend or relative; 39 percent, television, newspaper, or radio; and 12 percent, personal experience.
Most people view medical mistakes as an "individual provider issue" rather than a failure in the process of delivering care in a complex delivery system. When asked about possible solutions to prevent medical mistakes, actions rated very effective by respondents were "keeping health care professionals with bad track records from providing care" (75 percent) and ''better training of health care professionals" (69 percent).112
There are numerous factors that might contribute to the "disconnect" between public perceptions and actual health care error rates. The various accreditation and licensure programs for health care organizations and providers have been promoted as "Good Housekeeping Seals of Approval," yet they fail to provide adequate assurance of a safe environment. Reducing medical errors and improving patient safety are not an explicit focus of these processes. Even licensed and accredited organizations may have implemented only rudimentary systems and processes to ensure patient safety.
For the most part, media coverage has been limited to occasional reporting of anecdotal cases. The impact of anecdotal information on safety may also be less effective in health care than in the nuclear waste or airline industries, where an individual event often impacts dozens or hundreds of people at a time.
Patient safety is also hindered through the liability system and the threat of malpractice, which discourages the disclosure of errors. The discoverability of data under legal proceedings encourages silence about errors committed or observed. Most errors and safety issues go undetected and unreported, both externally and within health care organizations.
1. Brennan, Troyen A.; Leape, Lucian L.; Laird, Nan M., et al. Incidence of adverse events and negligence in hospitalized patients: Results of the Harvard Medical Practice Study I. N Engl J Med. 324:370–376, 1991. See also: Leape, Lucian L.; Brennan, Troyen A.; Laird, Nan M., et al. The Nature of Adverse Events in Hospitalized Patients: Results of the Harvard Medical Practice Study II. N Engl J Med. 324(6):377–384, 1991. Thomas, Eric J.; Studdert, David M.; Burstin, Helen R., et al. Incidence and Types of Adverse Events and Negligent Care in Utah and Colorado. Med Care forthcoming Spring 2000.
2. Thomas, Eric J.; Studdert, David M.; Newhouse, Joseph P., et al. Costs of Medi-
cal Injuries in Utah and Colorado. Inquiry. 36:255–264, 1999. See also: Leape, et al., 1991. Brennan, et al., 1991.
3. American Hospital Association. Hospital Statistics. Chicago. 1999.
4. Centers for Disease Control and Prevention (National Center for Health Statistics). Deaths: Final Data for 1997. National Vital Statistics Reports. 47(19):27, 1999.
5. Centers for Disease Control and Prevention (National Center for Health Statistics). Births and Deaths: Preliminary Data for 1998. National Vital Statistics Reports. 47(25):6, 1999.
6. Thomas, et al., 1999.
7. Thomas, et al., 1999.
8. Thomas, et al., 1999.
9. Occupational Safety and Health Administration. The New OSHA: Reinventing Worker Safety and Health [Web Page]. Dec. 16, 1998. Available at: www.osha.gov/oshinfo/reinvent.html.
10. Bureau of Labor Statistics. National Census of Fatal Occupational Injuries, 1998. U.S. Department of Labor: Washington, D.C. August 1999.
11. Phillips, David P.; Christenfeld, Nicholas; Glynn, Laura M. Increase in US Medication-Error Deaths between 1983 and 1993. Lancet. 351:643–644, 1998.
12. Bates, David W.; Spell, Nathan; Cullen, David J., et al. The Costs of Adverse Drug Events in Hospitalized Patients. JAMA. 277:307–311, 1997.
13. National Wholesale Druggists' Association. Industry Profile and Healthcare Factbook. Reston, VA: 1998.
14. Hallas, Jesper; Haghfelt, Torben; Gram, Lars F., et al. Drug Related Admissions to a Cardiology Department: Frequency and Avoidability. J Intern Med. 228:379–384, 1990.
15. Willcox, Sharon M.; Himmelstein, David U.; Woolhandler, Steffie. Inappropriate Drug Prescribing for the Community Dwelling Elderly. JAMA. 272:292, 1994.
16. Knox, Richard. Prescription Errors Tied to Lack of Advice: Pharmacists Skirting Law, Massachusetts Study Finds. Boston Globe. February 10, 1999:(Metro);B1.
17. Einarson, Thomas R. Drug-Related Hospital Admissions. Ann Pharmacother. 27:832–840, 1993.
18. Reason, James T. Human Error. Cambridge, MA: Cambridge University Press; 1990.
19. Brennan, et al., 1991.
20. Leape, et al., 1991. See also Brennan, et al., 1991.
21. Leape, et al., 1991.
22. Manasse, Henri R. Jr. Medication Use in an Imperfect World: Drug Misadventuring as an Issue of Public Policy, part 1. Am J Hosp Pharm. 46:929–944, 1989.
23. Johnson, Jeffrey A., and Bootman, J. Lyle. Drug-Related Morbidity and Mortality: A Cost-of-Illness Model. Arch Intern Med. 155(18):1949–1956, 1995.
24. Victoroff, Michael S. The Right Intentions: Errors and Accountability. J Fam Pract. 45:38–39, 1997.
25. Andrews, Lori B.; Stocking, Carol; Krizek, Thomas, et al. An Alternative Strategy for Studying Adverse Events in Medical Care. Lancet. 349:309–313, 1997.
26. Vincent, Charles; Taylor-Adams, Sally and Stanhope, Nicole. Framework for Analysing Risk and Safety in Clinical Medicine. BMJ. 316(11): 1154–1157, 1998.
27. Cook, Richard and Woods, David. Operating at the Sharp End: the Complexity
of Human Error. Bogner, Marilyn Sue, Ed. Human Errors in Medicine. Hillsdale, NJ: Lawrence Erlbanm Associates; 1994; pp. 255–310.
28. Nadzam, Deborah M. Development of Medication-Use Indicators by the Joint Commission on Accreditation of Health Care Organizations. AJHP. 48:1925–1930, 1991.
29. Brennan, et al., 1991.
30. Brennan, et al., 1991.
31. Leape, et al., 1991.
32. Brennan, et al., 1991.
33. Thomas, et al., 2000.
34. Thomas, et al., 1999.
35. Thomas, et al., 2000.
36. American Hospital Association. Hospital Statistics. Chicago. 1999. See also: Thomas, Eric J.; Studdert, David M.; Burstin, Helen R., et al. Incidence and Types of Adverse Events and Negligent Care in Utah and Colorado. Med Care forthcoming March 2000. See also: Thomas, Eric J.; Studdert, David M.; Newhouse, Joseph P., et al. Costs of Medical Injuries in Utah and Colorado. Inquiry. 36:255–264, 1999.
37. American Hospital Association. Hospital Statistics. Chicago. 1999. See also: Brennan, Troyen A.; Leape, Lucian L.; Laird, Nan M., et al. Incidence of adverse events and negligence in hospitalized patients: Results of the Harvard Medical Practice Study I. N Engl J Med. 324:370–376, 1991. See also: Leape, Lucian L.; Brennan, Troyen A.; Laird, Nan M., et al. The Nature of Adverse Events in Hospitalized Patients: Results of the Harvard Medical Practice Study II. N Engl J Med. 324(6):377–384, 1991.
38. Centers for Disease Control and Prevention (National Center for Health Statistics). Deaths: Final Data for 1997. National Vital Statistics Reports 47(19):27, 1999. See also: Centers for Disease Control and Prevention (National Center for Health Statistics). Births and Deaths: Preliminary Data for 1998. National Vital Statistics Reports. 47(25):6, 1999.
39. Andrews, et al., 1997.
40. Steel, Knight; Gertman, Paul M.; Crescenzi, Caroline, et al. Iatrogenic Illness on a General Medical Service at a University Hospital. N Engl J Med. 304:638–642, 1981.
41. Andrews, et al., 1997.
42. Dubois, Robert W. and Brook, Robert H. Preventable Deaths: Who, How Often, and Why? Ann Intern Med. 109:582–589, 1988.
43. Bedell, Susanne E.; Deitz, David C.; Leemah, David, et al. Incidence and Characteristics of Preventable Iatrogenic Cardiac Arrests. JAMA. 265(21):2815–2820, 1995.
44. McGuire, Hunter H.; Horsley, J. Shelton; Salter, David R., et al. Measuring and Managing Quality of Surgery: Statistical vs Incidental Approaches. Arch Surg. 127:733–737, 1992.
45. Cooper, Jeffrey B.; Newbower, Ronald S.; Long, Charlene D., et al. Preventable Anesthesia Mishaps: A Study of Human Factors. Anesthesiology. 49:399–406, 1978.
46. Gaba, David M. Human Error in Anesthetic Mishaps. Int Anesthesiol Clin. 27(3):137–147, 1989.
47. Duncan, Peter G., and Cohen, Marsha M. Postoperative Complications: Factors of Significance to Anaesthetic Practice. Can J Anaesth. 34:2–8, 1987.
48. Cohen, Marsha M.; Duncan, Peter G.; Pope, William D. B., et al. A Survey of 112,000 Anaesthetics at One Teaching Hospital (1975–1983). Can Anaesth Soc J. 33:22–31, 1986.
49. Sentinel Events: Approaches to Error Reduction and Prevention. Jt Comm J Qual Improv. 24(4):175–186, 1998.
50. Chassin, Mark R. Is Health Care Ready for Six Sigma Quality? Milbank Q. 764:565–591, 1998.
51. Bhasale, Alice L.; Miller, Graeme C.; Reid, Sharon E., et al. Analysing Potential Harm in Australian General Practice: an Incident-Monitoring Study. MJA. 169:73–76, 1998.
52. National Wholesale Druggists' Association. Industry Profile and Healthcare Factbook, op cit.
53. Manasse, 1989.
54. Phillips, et al., 1998.
55. Bates, David W.; Boyle, Deborah L.; Vander Vilet, Martha, et al. Relationship between Medication Errors and Adverse Drug Events. J Gen Intern Med. 10:199–205, 1995.
56. Lesar, Timothy S.; Briceland, Laurie, and Stein, Daniel S. Factors Related to Errors in Medication Prescribing. JAMA. 277(4):312–317, 1997.
57. Bates, David W.; Cullen, David J.; Laird, Nan M., et al. Incidence of Adverse Drug Events and Potential Adverse Drug Events: Implications for Prevention. JAMA. 274:29–34, 1995.
58. Bates, David W.; Leape, Lucian L., and Petrycki, Stanley. Incidence and Preventability of Adverse Drug Events in Hospitalized Adults. J Gen Intern Med. 8:289–294, 1993.
59. Lesar, Briceland, and Stein, 1997.
60. Koren, Gideon and Haslam, Robert H. Pediatric Medication Errors: Predicting and Preventing Tenfold Disasters. J Clin Pharmacol. 34:1043–1045, 1994.
61. Perlstein, Paul H.; Callison, Cornelia; White, Mary, et al. Errors in Drug Computations During Newborn Intensive Care. Am J Dis Child. 33:376–379, 1979.
62. Folli, Hugo L.; Poole, Robert L.; Benitz, William E., et al. Medication Error Prevention by Clinical Pharmacists in Two Children's Hospitals. Pediatrics. 79:718–722, 1987.
63. Raju, Tonse N. K.; Kecskes, Susan; Thornton, John P., et al. Medication Errors in Neonatal and Paediatric Intensive-Care Units. Lancet. 374–376, 1989.
64. Cullen, David J.; Sweitzer, Bobbie Jean; Bates, David W., et al. Preventable Adverse Drug Events in Hospitalized Patients: A Comparative Study of Intensive Care and General Care Units. Crit Care Med. 25(8): 1289–1297, 1997.
65. Manasse, 1989.
66. Davies, D.M., ed. Textbook on Adverse Drug Reactions, 3rd. ed. Oxford: Oxford University Press; 1985.
67. Griffin, J.P., and Weber, J. C. P. Voluntary Systems of Adverse Reaction Reporting—Part I. Adverse Drug React Acute Poisoning Rev. 4:213–230, 1985.
68. Griffin, J.P., and Weber, J. C. P. Voluntary Systems of Adverse Reaction Reporting—Part II. Adverse Drug React Acute Poisoning Rev. 5:23–25, 1986.
69. Cullen, David J.; Bates, David W.; Small, Stephen D., et al. The Incident Reporting System Does Not Detect Adverse Drug Events. Jt Comm J Qual Improv. 21(10):541–548, 1995.
70. Classen, David C.; Pestonik, Stanley, L.; Evans, Scott; Burke, John P., et al. Com-
puterized Surveillance of Adverse Drug Events in Hospital Patients. JAMA. 266(20):2847–2851.
71. Beard, Keith. Adverse Reactions as a Cause of Hospital Admissions in the Aged. Drug Aging. 2:356–361, 1992.
72. Roughead, Elizabeth E.; Gilbert, Andrew L.; Primrose, J. G., et al. Drug-Related Hospital Admissions: A Review of Australian Studies Published 1998–1996. Med J. Aust. 168:405–408, 1998.
73. Burnum, John F. Preventability of Adverse Drug Reactions. Ann Intern Med. 85:80, 1976.
74. Schneitman-McIntire, Orinda.; Farnen, Tracy A.; Gordon, Nancy, et al. Medication Misadventures Resulting in Emergency Department Visits at an HMO Medical Center. Am J Health Syst Pharm. 3:1416–1422, 1996.
75. Sullivan, Sean D.; Kreling, David H.; Hazlet, Thomas K., et al. Noncompliance with Medication Regimens and Subsequent Hospitalizations: A Literature Analysis and Cost of Hospitalization Estimate. J Res Pharm Econom. 2(2):19–33, 1990.
76. Einarson, 1993.
77. Thomas, et al., 2000.
78. Dubois, Robert W. and Brook, Robert H. Preventable Deaths: Who, How Often, and Why? Ann Int Med. 109:582–589, 1988.
79. Bedell, Susanna E.; Deitz, David C.; Leeman, David; Delbanco, Thomas, L. Incidence and Characteristics of Preventable Iatrogenic Cardiac Arrests. JAMA. 265:2815–2820, 1991.
80. Lesar, Briceland, and Stein, 1997.
81. Willcox, Sharon M.; Himmelstein, David U.; Woolhandler, Steffie, et al. Inappropriate Drug Prescribing for the Community Dwelling Elderly, op cit.
82. Hallas, et al., 1997.
83. Wilson, Dirk G.; McArtney, R. G.; Newcombe, Robert G., et al. Medication Errors in Paediatric Practice: Insights from a Continuous Quality Improvement Approach. Eur J Pediatr. 157:769–774, 1998.
84. Burnum, 1976.
85. Beers, Mark H.; Storrie, Michele; and Lee, Genell. Potential Adverse Drug Interactions in the Emergency Room. Ann Intern Med. 112:61–64, 1990.
86. Knox, 1999.
87. Bates, et al., 1995.
88. Davis, Neil M. and Cohen, Michael R. Medication Errors: Causes and Prevention. Philadelphia: George F. Stickley Co., 1981.
89. Folli, et al., 1987.
90. Greenberg, Jay; Leutz, Walter; Greenlick, Merwyn, et al. The Social HMO Demonstration: Early Experience. Health Affairs. 7:66, 1988.
91. Strandberg, Lee R. Drugs as a Reason for Nursing Home Admissions. Am Health Care Assoc J. 10:20–23, 1984.
92. Leape, Lucian L.; Bates, David W., and Cullen, David J. Systems Analysis of Adverse Drug Events. JAMA. 274:35–43, 1995.
93. Bates, et al., 1995.
94. Folli, et al., 1987.
95. Blum, Keith V.; Abel, S.R.; Urbanski, Chris J., et al. Medication Error Prevention by Pharmacists. Am J Hosp Pharm. 45:1902–1903, 1988
96. Thomas, et al., 1999.
97. Alliance for Aging Research. When Medicine Hurts Instead of Helps. Washington, DC: The Alliance for Aging Research; 1998.
98. Johnson, et al., 1995.
99. Hallas, et al., 1997.
100. Willcox, et al., 1994.
101. Knox, 1999.
102. Einarson, 1993.
103. Bates, David W., et al. The Costs of Adverse Drug Events in Hospitalized Patients, op cit, 1995.
104. Classen, David C.; Pestotnik, Stanley L.; Evans, R. Scott, et al. Adverse Drug Events in Hospitalized Patients: Excess Length of Stay, Extra Costs, and Attributable Mortality. JAMA. 277:301–306, 1997.
105. Schneider, Philip J.; Gift, Maja G.; Lee, Yu-Ping, et al. Cost of Medication-Related Problems at a University Hospital. Am J Health Syst Pharm. 52:2415–2418, 1995.
106. Bloom, Bernard S. Cost of Treating Arthritis and NSAID-Related Gastrointestinal Side-Effects. Aliment Pharmacol Ther. l(Suppl 2):131–138, 1998.
107. Bootman, J. Lyle; Harrison, LTC Donald L., and Cox, Emily. The Health Care Cost of Drug-Related Morbidity and Mortality in Nursing Facilities. Arch Intern Med. 157(18):2089–2096, 1997.
108. Federal Aviation Administration, Office of System Safety. Aviation Safety Reporting System (ASRS) Database [Web Page]. 1999. Available at: http://nasdac.faa.gov/safety data.
109. National Patient Safety Foundation. Diverse Groups Come Together to Improve Health Care Safety Through the National Patient Safety Foundation [Web Page]. 1997 Aug 29. Available at: http://www.ama-assn.org/med-sci/npsf/pr897.htm. Note: press release.
110. National Patient Safety Foundation, 1997.
111. National Patient Safety Foundation, 1997.
112. National Patient Safety Foundation, 1997.