Page 215

C—
Literature Summary

This Appendix summarizes the literature described in Chapter 2. The references cited are at the end of Chapter 2.



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 215
C Literature Summary This Appendix summarizes the literature described in Chapter 2. The references cited are at the end of Chapter 2. 215

OCR for page 215
216 TO ERR IS HUMAN TABLE C-1 Literature Summary Reference Sample Description Data Source General studies of errors and adverse events Thomas et al., Randomly sampled 15,000 Chart review by trained nurses forthcoming 2000 nonpsychiatric 1992 and board-certified family discharges from a practitioners and internists. representative sample of hospitals in Utah and Colorado. Bhasale et al., 1998 A non-random sample of General practitioner-reported Analysing potential harm 324 general practitioners free-text descriptions of in Australian general reporting incidents incidents and answered practice between October 1993 fixed-response questions. and June 1995.

OCR for page 215
217 APPENDIX C Results Definition(s) Causes/Types of Error Adverse events occurred in Adverse event—“an injury 46.1% of adverse events 2.9% ± 0.2 of caused by medical (22.3% negligent) were hospitalizations in each management (rather than attributable to surgeons and state. 32.6% ± 4 of the disease process) that 23.2% (44.9% negligent) adverse events were due resulted in either a were attributable to to negligence in Utah and prolonged hospital stay internists. 27.4 ± 2.4 were due to or disability at discharge.” negligence in Colorado. Negligence was defined as Death occurred in 6.6% “care that fell below the ± 1.2 of adverse events and standard expected of 8.8% ± 2.5 of negligent physicians in their adverse events. The leading community.” cause of nonoperative adverse events were adverse drug events (19.3% of all adverse events; 35.1% were negligent). Operative events comprised 44.9% of all adverse events and 16.9% were negligent. 805 incidents were reported. Incident—“an unintended Pharmacological management 76% were preventable and event, no matter how related to 51 per 100 27% had potential for severe seemingly trivial or incidents. Poor harm. commonplace, that could communication between have harmed or did harm patients and healthcare a patient.” professionals and actions of others contributed to 23 per 100 incidents each. Errors in judgment contributed to 22 per 100 incidents. Continued

OCR for page 215
218 TO ERR IS HUMAN TABLE C-1 Continued Reference Sample Description Data Source General studies of errors and adverse events (continued) Leape et al., 1993 Record review of 1,133 Harvard Medical Practice Preventing medical injury patients who suffered from Study. an adverse event (AE). McGuire et al., 1992 44,603 consecutive major Resident reports giving name Measuring and managing operations performed at and procedure of each quality of surgery a large medical center patient who suffered any from 1977 to 1990. complication. In a monthly conference, representatives of all specialties determined by consensus the category of each complication (inevitable, inherent risk, error, hospital deficit, coincidence, unknown).

OCR for page 215
219 APPENDIX C Results Definition(s) Causes/Types of Error 70% of adverse events were AE—per Leape (1991), The most common types of found to be preventable, AE is defined as “an preventable errors were 24% unpreventable, and unintended injury that technical errors (44%), 6% potentially preventable. was caused by medical errors in diagnosis (17%), management and that failures to prevent injury resulted in measurable (12%), and errors in the use disability.” of a drug (10%). Preventable AE—an AE Approximately 20% of resulting from an error. technical errors, 71% of Unpreventable AE—an AE diagnostic errors, 50% of resulting from a preventative errors, and complication that cannot 37% of errors in the use of a be prevented at the drug were judged to be current state of knowledge. negligent. Potentially preventable AE—an AE where no error was identified but it is widely recognized that a high incidence of this type of complication reflects low standards of care or technical expertise. 2,428 patients (5.4%) suffered 2,797 complications (6.3%). 49% of these complications were attributable to error. 749 patients (1.7%) died during the same hospitalization. 7.5% of these deaths were attributable to error. Continued

OCR for page 215
220 TO ERR IS HUMAN TABLE C-1 Continued Reference Sample Description Data Source General studies of errors and adverse events (continued) Bedell et al., 1991 203 patients who suffered At least one of the authors Incidence and characteristics from cardiac arrest at a evaluated patients who of preventable iatrogenic teaching hospital during underwent CPR within 24 cardiac arrests 1981. hours of arrest. Information from the medical record was also used. Leape et al., 1991 30,195 randomly selected Hospital records. The nature of adverse events records in 51 hospitals in in hospitalized patients New York state (1984). DuBois et al., 1988 182 deaths from 12 Investigators prepared a Preventable Deaths hospitals for 3 conditions dictated summary of each (cerebrovascular accident, patient’s hospital course. pneumonia, or myocardial Panels of 3 physicians for infarction) each condition then independently reviewed each summary and independently judged whether the death was preventable.

OCR for page 215
221 APPENDIX C Results Definition(s) Causes/Types of Error 28 (14%) of arrests followed Iatrogenic cardiac arrest— The most common causes of an iatrogenic complication. “an arrest that resulted potentially preventable 17 (61%) of the 28 patients from a therapy or arrest were medication died. All 4 reviewers procedure or from a errors and toxic effects considered 18 (64%) of the clearly identified error of (44%), and suboptimal iatrogenic arrests to have omission.” response by physicians to been preventable. clinical signs and symptoms (28%). 1,133 adverse events (AEs) AE—“an unintended injury Drug complications were the occurred in 30,195 patients. that was caused by medical most common type of management and that adverse event (19%), resulted in measurable followed by wound disability.” infections (14%) and technical complications (13%). 58% of the adverse advents were errors in management, among which nearly half were attributable to negligence. The physicians unanimously Preventable deaths from agreed that 14% of the myocardial infarction deaths could have been reflected errors in prevented. 2 out of the 3 management, from physicians found that 27% cerebrovascular accident might have been prevented. reflected errors in diagnosis, and from pneumonia reflected errors in management and diagnosis. Continued

OCR for page 215
222 TO ERR IS HUMAN TABLE C-1 Continued Reference Sample Description Data Source General studies of errors and adverse events (continued) Steel et al., 1981 815 consecutive patients on Record review, clinical Iatrogenic illness on a a university hospital’s personnel interviews, and general medical service general medical service information from utilization- at a university hospital during a 5-month period review coordinators. in 1979. Cooper et al., 1978 47 interviews regarding Interviewees selected at Preventable anesthesia preventable mishaps random from a list of mishaps between September 1975 departmental members. and April 1977 including staff and resident anesthesiologists from a large urban teaching hospital. Dripps et al., 1961 Records of 33,224 patients Patient records The role of anesthesia in anesthetized in a surgical mortality 10-year period.

OCR for page 215
223 APPENDIX C Results Definition(s) Causes/Types of Error 36% of patients had an Iatrogenic illness—“any iatrogenic illness. 9% of illness that resulted from a the patients had an diagnostic procedure or iatrogenic illness that from any form of therapy.” threatened life or produced In addition, the authors considerable disability while, included harmful in another 2%, the illness occurrences (e.g., injuries was believed to contribute from a fall or decubitus to the death of the patient. ulcers) that were not natural consequences of the patient’s disease. 359 preventable critical Critical incident—a mishap 82% of the preventable incidents were identified that “was clearly an incidents reported involved and coded. occurrence that could have human error and 14% led (if not discovered or involved equipment error. corrected in time) or did lead to an undesirable outcome, ranging from increased length of hospital stay to death or permanent disability.” 12 of the 18,737 patients who received spinal anesthesia died from causes definitely related to the anesthetic (1:1,560). 27 of the 14,487 patients who received general anesthesia supplemented with a muscle relaxant died from causes directly related to the anesthetic (1:536). Continued

OCR for page 215
224 TO ERR IS HUMAN TABLE C-1 Continued Reference Sample Description Data Source General studies of errors and adverse events (continued) Beecher and Todd, 1954 All deaths from January 1 team, consisting of an A study of the deaths 1, 1948, through December anesthesiologist, a surgeon, associated with 31, 1952, occurring on and a secretary, worked in anesthesia and surgery the surgical services of each of the 10 hospitals and based on a study of 10 university hospitals. appraised the causes of all 599,548 anesthesias in deaths on the surgical services. ten institutions Medication-related studies Knox, 1999 Analysis of medication Prescription errors tied to errors by 51 lack of advice Massachusetts Globe article pharmacists. Leape, 1999 75 patients randomly Review of medical records and Pharmacist participation on selected from each of 3 pharmacist recommendations. physician rounds and groups: all admissions to adverse drug events in the study unit (2 medical the intensive care unit ICUs at Massachusetts General Hospital) from February 1, 1993, through July 31, 1993 (baseline), and all admissions to the study unit (postintervention) and control unit from October 1, 1994, through July 7, 1995. 50 patients were also selected at random from the control unit during the baseline period.

OCR for page 215
225 APPENDIX C Results Definition(s) Causes/Types of Error 7,977 of the 599,548 patients who received anesthesia died. Gross errors in anesthetic management occurred in 29 of the 384 (7.6%) deaths caused by anesthesia. 88% of medication errors Pharmacists cited factors involved the wrong drug or that led to mistakes. 62% the wrong dose and 63% cited “too many telephone involved first-time calls,” 59% “unusually busy prescriptions rather than day,” 53% “too many refills. customers,” 41% “lack of concentration,” and 32% “staff shortage.” The rate of preventable ADE—per Bates (1993), adverse drug events (ADEs) ADE is defined as “an injury due to ordering decreased resulting from the by 66% from 10.4 per administration of a drug.” 1,000 patient days before the intervention to 3.5 per 1,000 patient days after the intervention. The rate was essentially unchanged during the same time periods in the control unit: 10.9 and 12.4 per 1,000 patient days. Continued

OCR for page 215
243 APPENDIX C Results Definition(s) Causes/Types of Error 315 iatrogenic medication Medication error—“a dose of 60.3% of the 315 errors were errors were reported among medication that deviates attributable to nurses and the 2,147 neonatal and from the physicians’ order 29.6% to pharmacists. Only pediatric care admissions, as written in the medical 2.9% were attributable to an error rate of 1 per 6.8 record. . . . Except for error physicians (because admissions (14.7%). The of omission, the medication prescription errors detected frequency of iatrogenic dose must actually reach before drug administration injury of any sort due to a the patient . . . a wrong were not counted). medication error was 3.1%, dose (or other type of error) or 1 for each 33 intensive that is detected and care admissions. 66 errors corrected before resulted in injury, 33 were administration will not potentially serious, 32 constitute a medication caused mild injuries, and error. . . . Prescription 1 patient suffered acute errors (not dispensed and aminophylline poisoning. administered to the patient) . . . are excluded from this definition . . .” 123,367 medication orders Order with a potential were written. Riley Hospital medication error—“if any for Children had 1,277 aspect of the order was errors out of the 48,034 not in accordance with (2.7%) orders written and information in standard University Hospital had reference text, an approved 1,012 errors out of 75,333 protocol, or dosing (1.3%) orders written. guidelines approved by the 90.4% of the overall orders pharmacy and therapeutics questioned by pharmacists committee of the hospitals.” were confirmed by the physician as being in error. 0.2% of the 2289 errors were classified as potentially lethal, 13.7% were serious, 34.2% were significant, and 51.9% were minor. The number of errors that pharmacists prevent each year approaches 9,000. Continued

OCR for page 215
244 TO ERR IS HUMAN TABLE C-1 Continued Reference Sample Description Data Source Medication-related studies (continued) Nolan and O’Malley, 1988 21 hospital inpatient studies Review of published studies on Prescribing for the elderly, conducted in the United adverse drug reactions (ADRs). part I States, United Kingdom, Israel, New Zealand, Switzerland, Canada, and India published between 1964 and 1981. Folli et al., 1987 101,022 medication orders Copies of errant chart orders Medication error prevention prescribed in two children’s reviewed by a member of by clinical pharmacists teaching hospitals (Miller the pediatric faculty or in two children’s Children’s Hospital of attending physician and by hospitals Memorial Medical Center two pediatric clinical [MMC] and Stanford pharmacist practitioners. University Medical Center [SUMC]) during a six-month period (February through July 1985). Perlstein et al., 1979 43 nursing, pharmacy, and Errors in drug computations medical personnel tested during newborn for accuracy in calculating intensive care drug doses to be administered to newborn infants. (27 registered nurses, 5 registered pharmacists, and 11 pediatricians.)

OCR for page 215
245 APPENDIX C Results Definition(s) Causes/Types of Error Rates of patients experiencing ADRs ranged from 1.5% to 43.5%. A majority of the studies documented ADR rates between 10% and 25%. A combined total of 479 Errant medication order— The most common type of errant medication orders “An order was considered error was incorrect dosage. were identified at the two to be potentially in error if The most prevalent type of institutions. MMC and it was not in accordance error was overdosage. SUMC had similar with standard pediatric frequency of error, 4.9 and references, current 4.5 errors per 1,000 published literature, or medication orders, or 1.37 dosing guidelines approved and 1.79 per 100-patient by the pharmacy and days, respectively. Involving therapeutics committees pharmacists in the reviewing of each hospital.” of drug orders reduced the potential harm resulting from errant medication orders significantly. The mean test score for nurses was 75.6%. 56% of the errors would have resulted in administered doses ten times greater or less than the ordered dose. The mean test score was 96% for pharmacists and none of the errors would have resulted in the administration of doses over 1% greater or less than the dose ordered. Pediatricians averaged a score of 89.1%. 38.5% of the errors would have resulted in the administration of doses ten times higher or lower than the dose ordered. Continued

OCR for page 215
246 TO ERR IS HUMAN TABLE C-1 Continued Reference Sample Description Data Source Medication-related studies (continued) Miller, 1977 Boston Collaborative Drug Interpretation of studies on Surveillance Program adverse drug reactions Burnum, 1976 1,000 adult medical patients Physician observation. Preventability of adverse drawn from a community, drug reactions office-based practice of general internal medicine. Jick, 1974 19,000 inpatients admitted Boston Collaborative Drug Drugs: remarkably to medical wards. Surveillance Program nontoxic

OCR for page 215
247 APPENDIX C Results Definition(s) Causes/Types of Error Adverse drug reactions (ADRs) occur in approximately 30% of hospitalized patients and after about 5% of drug exposures. The rate per patient of life-threatening ADRs in 3% and the rate per course of drug therapy is 0.4%. Adverse drug reactions (ADRs) 23% of the 42 ADRs were occurred in 42 of the attributable to physician individual patients. 23 error (10 out of 42; 6 (55%) were judged because of giving a drug unnecessary and potentially that was not indicated and 4 preventable. because of improper drug administration), 17% to patient or pharmacist error, and 14% to errors shared by the physician, patient and pharmacist. 30% of hospitalized medical patients have at least 1 adverse drug reaction (ADR) while hospitalized. An estimated 3 million hospital patients have an ADR in medical units each year. Continued

OCR for page 215
248 TO ERR IS HUMAN TABLE C-1 Continued Reference Sample Description Data Source Medication-related studies (continued) Phillips et al., 1974 All United States death Increase in U.S. Medication- certificates between 1983 error deaths between and 1993. 1983 and 1993 Talley and Laventurier, 1974 Boston Collaborative Drug Drug-induced illness Surveillance Program and an Israeli study. Cost Thomas et al., 1999 Medical records of 14,732 Two-stage chart review by randomly selected 1992 trained nurses and discharges from 28 board-certified family hospitals in Utah and practitioners and internists. Colorado

OCR for page 215
249 APPENDIX C Results Definition(s) Causes/Types of Error In 1983, 2,876 people died Medication errors— from medication errors. By “‘accidental poisoning by 1993, this number had drugs, medicaments, and risen to 7,391, a 2.57-fold biologicals’ and have increase. Between 1983 and resulted from acknowledged 1993, outpatient errors, by patients or medication error deaths medical personnel.” rose 8.48-fold (from 172 to 1,459) and inpatient medical error deaths rose 2.37-fold (504 to 1,195). An estimated incidence of lethal adverse drug reactions ranges from a low of 60,000 (.18% incidence) to a high of 140,000 (.44% incidence) for hospitalized patients in the U.S. 459 adverse events were Adverse event—“an injury detected, of which 265 were caused by medical preventable. Death occurred management (rather than in 6.6% of adverse events the disease process) that and 6.9% of preventable resulted in either prolonged adverse events. The total hospital stay or disability costs were $661,889,000 at time of discharge.” for adverse events and $308,382,000 for preventable adverse events. Health care costs were $348,081,000 for all adverse events and $159,245,000 for preventable adverse events. 57% of the adverse event health care costs and 46% of the preventable adverse event costs were attributable to outpatient medical care. Continued

OCR for page 215
250 TO ERR IS HUMAN TABLE C-1 Continued Reference Sample Description Data Source Cost (continued) Bates et al., 1997 4,108 admissions to a Stimulated self-reports by The costs of adverse drug stratified random sample nurses and pharmacists and events in hospitalized of 11 medical and daily chart review. 2 patients surgical units in Brigham independent reviewers and Women’s Hospital classified the incidents. (726 beds) and Massachusetts General Hospital (846 beds) in Boston over a 6-month period between February and July 1993. Cases were patients with an adverse drug event (ADE), and the control for each case was a patient on the same unit as the case with the most similar pre-event length of stay. Bootman et al., 1997 To estimate the cost of Survey of an expert panel The health care cost of drug-related problems consisting of consultant drug-related morbidity (DRPs) within nursing pharmacists and physicians and mortality in nursing facilities, a decision with practice experience in facilities analysis technique was nursing facilities and used to develop a geriatric care. probability pathway model.

OCR for page 215
251 APPENDIX C Results Definition(s) Causes/Types of Error 247 ADEs occurred among ADE—“an injury resulting 207 admissions and 60 from medical intervention were preventable. The related to a drug.” additional length of stay Potential ADE—“incidents in was 2.2 days with an ADE which an error was made and 4.6 days with a but no harm occurred.” preventable ADE. The estimated post-event costs attributable to an ADE were $2,595 for all ADEs and $4,685 for preventable ADEs. The estimated annual costs for a 700-bed teaching hospital attributable to all ADEs are $5.6 million and to preventable ADEs are $2.8 million. The national hospital costs of ADEs was estimated at $4 billion; preventable ADEs alone would cost $2 billion. The cost of drug-related DRPs—“an event of morbidity and mortality circumstance involving a with the services of patient’s drug treatment consultant pharmacists that actually or potentially was $4 billion compared interferes with the with $7.6 billion without achievement of an optimal services of consultant outcome.” pharmacists. For every dollar spent on drugs in nursing facilities, $1.33 is consumed in the treatment of DRPs. Continued

OCR for page 215
252 TO ERR IS HUMAN TABLE C-1 Continued Reference Sample Description Data Source Cost (continued) Johnson and Bootman, 1995 A probability pathway model Telephone survey of 15 Drug-related morbidity and was developed for expert practicing mortality drug-related morbidity and pharmacists. mortality based primarily on drug-related problems (DRPs). A panel of experts gave estimates on the numbers of patients affected by DRPs and monetary value data were taken from published reports and statistical reports. Schneider et al., 1995 109 patients at a university- Retrospective chart review. Cost of medication-related affiliated medical center problems at a university hospital who were known hospital to have had clinical consequences from an adverse drug reaction (ADR) or medication error. Bloom, 1988 Retrospective analysis of all Medicaid Management Cost of treating arthritis and direct costs related to the Information System of NSAID-related care of 527 Medicaid Washington, D.C. gastrointestinal recipients treated for side-effects arthritis with non-steroidal anti-inflammatory drugs (NSAIDs) between December 1, 1981 and November 30, 1983.

OCR for page 215
253 APPENDIX C Results Definition(s) Causes/Types of Error Drug-related morbidity and Drug-related problem— mortality costs an estimated “an event or circumstance $76.6 billion in the that involves a patient’s ambulatory setting in the drug treatment that United States. The panel actually, or potentially, members estimated that interferes with the 40% of patients who achievement of an receive drug therapy would optimal outcome.” have some form of DRP. 349 clinical outcomes associated with medical related problems (MRPs) (average of approximately 3 outcomes per patient) were detected. For the 1,911 ADRs and medication errors reported through the voluntary reporting system in 1994, the estimated annual cost was just under $1.5 million. In 1983, an estimated $3.9 Gastrointestinal adverse million was spent on drug reaction—“any treating preventable claim for payment gastrointestinal adverse accompanied by a drug reactions to NSAIDs. diagnosis of peptic ulcer, gastritis/duodenitis, other disorders of the stomach or duodenum, gastrointestinal symptoms, or a pharmacy claim for an H2-recepter antagonist, sucralfate or antacid, which occurred during the arthritis treatment study period.