Multldlsciplinary program for detecting and evaluating

Medication-dispensing errors Raporta

References 1. Betz RP, Levy HB. An interdisciplinary method of classifying and monitoring medication errors. Am J Hasp Pharm. 1985; 42:1724-32.

2. Cobb MD. Evaluating medication errors. Hofp Pharm. 1986; 21:925-9. 3. Barker KN, Harris JA, Webster DB et al. Consultant evalua­ tion of a hospital medication system; implementation and evaluation of the new svstem. Am I Hcsp Pharm. 1984; 41:2022-9.

MATTHEW R. KEITH, RENEE A. BELLANGER-MCCLEERY, AND JOHN E. FUCHS, JR.

Abstract: An adverse drug reaction (ADR)-reporting program involv­ ing detection of charted ADRs by quality assurance nurses and data collection and causality assessment by staff pharmacists is described. The voluntary ADR-reporting mechanism used in a 900-bed, uni­ versity-based hospital complex pro­ duced less than one ADR report per month. The newly implemented system depends on nurses to detect and report documented ADRs through concurrent chart review. Staff pharmacists are then responsi­ ble for follow-up chart review, data collection, and causality assign­ ment based on two published algo­ rithms. An inservice education pro­ gram designed to increase the

awareness and understanding of ADRs was provided to the depart­ ment of pharmacy and the quality assurance nurses. The clinical staff provides quality assurance through weekly ADR committee meetings. Drug information center personnel complete the causality algorithms by using the data collected by the staff pharmacists. The ADR com­ mittee then compares the algorithm results of the two assessors. Dis­ crepancies in scoring are evaluated to determine whether a change in the system is necessary. An FDA re­ port is generated if the staff phar­ macist assessor and the drug in­ formation center assessor obtain results of "probable" for both algo­ rithms.

An ADR-reporting program that relies on quality assurance nurses to detect charted ADRs and on staff pharmacists to evaluate reported ADRs increased the average num­ ber of ADRs reported from 0.4 to 20 per month.

Index terms: Administration; Com­ mittees; Documentation; Drug in­ formation centers; Drugs, adverse reactions; Food and Drug Adminis­ tration (U.S.); Hospitals; Nurses; Pharmacists, hospital; Pharmacy, institutional, hospital; Protocols; Quality assurance; Reports;Toxicity Am J Hosp Pharm. 1989, 46:180912

Epidemiologic studies show that 17-36% of pa­ tients admitted to hospitals will experience an ad­ verse drug reaction (ADR).'"5 Furthermore, 3-5% of hospital admissions may be due to ADRs.'-^-^-^ In 1986, the FDA determined that of the 26,753 spon­ taneously reported ADRs, 24% were serious; of these, 18% resulted in hospitalization and 6% were a factor leading to death." Numerous methods of detecting and reporting ADRs have been described in the literature. Pro­ grams that rely on voluntary physician reporting of ADRs or audits of discharge summaries by medi­ cal records department personnel have met with only limited success.®"" In their survey of ADRreporting programs in selected hospitals. Case and

Guzzetti'^ found that although 58 of 76 hospitals had concurrent programs in place, the average time required to detect and report 10 ADRs was 4.9 months. Lack of familiarity with reporting mecha­ nisms, failure to recognize ADRs, and fear of possi­ ble legal repercussions all have been cited as rea­ sons for this poor showing. Programs that involve hospital drug information services have fared bet­ ter,'® but drug information personnel frequently spend a great deal of time collecting, assessing, and reporting data. The University of Texas Medical Branch is a 900bed, tertiary-care, teaching hospital complex. The department of pharmaceutical services employs 145 pharmacists and technicians, including two

MATTHBV R. KEITH is Associate Coordinator of Drug Infor­ mation, Pharmaceutical Services, Universit\' of Texas Medical Branch, Galveston (UTMB), and Clinical Instructor, College of Pharmacy,'Universitj- of Texas, Austin (UTA). RENEE A. BEL­ LANGER-MCCLEERY, PHARM. D., is Nutrition Coordinator, UTMB, and Clinical Assistant Professor, UTA. JOHN E. FCJCHS, JR., PHARM. D., is Coordinator of Clinical Services, UTMB, and

Clinical Assistant Professor, UTA. Address reprint requests to Mr. Keith at the Department of Pharmacy, University of Texas Medical Branch, 301 University Boulevard, Galveston, T,X 77550. Copyright © 1989, American Society of Hospital Pharmacists, Inc. Al'l rights reserved. 0002-9289/89/0901-1809501.00.

Vol 46 Sep 1989 American Journal of Hospital Pharmacy 1809

Downloaded from https://academic.oup.com/ajhp/article-abstract/46/9/1809/5174858 by guest on 21 February 2019

Multldlsciplinary program for detecting and evaluating adverse drug reactions

Reports Adverse drug reactions

Description of the Program To circumvent the detection problem in the re­ porting mechanism, we solicited the services of the institution's newly formed Division of Quality Uti­ lization Management. This division comprises 11 nurses who concurrently audit all patient charts at least once every three days during hospitalization and then again after discharge. The nurses monitor conformity to diagnosis-related groups, drug-use evaluations, and other quality-of-care criteria. They also monitor patient charts for written docu­ mentation of an ADR. DIG personnel explained the reporting mechanism to the nurses, provided them with the hospital's definition of an ADR,^ and in­ structed them to anticipate ADR documentation in nurses' and physicians' daily notes, notes on pa­ tient histories and physical examinations, written consultation reports, and medication orders. The nurses were also instructed to report toxicity caused by excessive doses (both intentional and accidental) of drugs. Auditors report suspected ADRs to the quality

utilization management office daily. DIG person­ nel retrieve and review the audit report sheets each morning. ADRs that are assessed by DIG personnel to be unique or rare or to have some abnormal characteristic are flagged for chart review and data collection. After the system had been in place and operated solely by DIG personnel for three months, chart review and data collection became part of the staff pharmacists' daily responsibilities. An ADR data collection sheet (Figures 1 and 2) was designed for standardizing the collection of pertinent informa­ tion from patient charts. A 12-part inservice education program was con­ ducted by the clinical staff and pharmacy residents before the new ADR scheme was implemented in order to familiarize staff pharmacists with the data collection sheet, algorithms for assessing the prob­ ability of ADRs, and the procedure for data collec­ tion. The topics covered in the 30-minute presenta­ tions are listed in Table 1. The first three lectures were presented before staff pharmacists became involved in ADR data collection and assessment. All 12 presentations were a required part of staff development and training. Handout materials and audio cassettes were made available to individuals not able to attend a given lecture. When a serious or rare ADR (potentially report­ able to the FDA) is discovered, DIG personnel sup­ ply the patient's name and location and the nature of the suspected ADR, along with a data collection form, to the staff pharmacist in charge of the ser­ vice to which the patient was admitted. The phar­ macist is then responsible for the data collection and probability assessment. The probability rating is based on the algorithms of Kramer et al.'^ and Naranjo et al.,'' which have been found to be equivalent predictors of ADR causality.'® We have chosen to use both algorithms because each has its own advantages. The Kramer algorithm is more detailed and is useful in teaching thought process­ es that are important when one is clinically assess­ ing an ADR. The Naranjo algorithm is less time consuming, allows less room for differences in in­ terpretation, and provides a rating for comparison with that based on the Kramer algorithm. When an ADR data collection sheet is returned by a staff pharmacist, DIG personnel review it for completeness. The coordinator or associate coordi­ nator of the DIG then completes both algorithms by drawing on the data collection sheet. This dou­ ble-check system has provided valuable assistance in locating weaknesses in the ADR data collection sheet. Scorings by DIG personnel and the staff pharmacists are compared. Discrepancies are eval­ uated to determine whether the ADR sheet did not provide sufficient information (and therefore should be revised) or whether interpretation and understanding of the algorithms was at fault. Occasionally, ADR outcome is not resolved at the

1810 American Journal of Hospital Pharmacy Vol 46 Sep 1989

Downloaded from https://academic.oup.com/ajhp/article-abstract/46/9/1809/5174858 by guest on 21 February 2019

pharmacy residents and four clinical pharmacists who do not have drug distribution responsibilities. Comprehensive 24-hour unit dose drug distribu­ tion and i.v. admixture services are provided, and the pharmacists participate in pediatric cardiopul­ monary resuscitation attempts and are members of several medical teams. Before December 1987, we relied on our institution's health-care profession­ als to voluntarily report «serious, rare, or novel ADRs to the associate director of pharmacy or the drug information center (DIG). The reporting phy­ sician in charge of the patient in question was re­ sponsible for completing and mailing form 1639 to FDA. The drug information service compiled data about reported ADRs and presented these data to the pharmacy and therapeutics (P&T) committee on a quarterly basis. This type of voluntary, unsolicited reporting was not successful in our institution; fewer than five ADRs were reported each year. Our attempts to increase awareness of ADR-reporting mecha­ nisms through newsletters and bulletins were inef­ fective. We identified the weak link in ADR report­ ing as being the detection and timely reporting of ADRs to the pharmacy department. Miwa and RandalP'^ achieved some success with a voluntary ADR-monitoring program that was based on par­ ticipation by the pharmacy and nursing staff. Kimelblatt et al.'^ found that an ADR-reporting sys­ tem based on reporting by staff pharmacists was effective in increasing the number of reported re­ actions. In this paper, we describe an ADR-reporting pro­ gram in which quality assurance nurses identify ADRs and staff pharmacists concurrently review and assess the causality of ADRs.

Adverse drug reactions Raport* Figure 1. Adverse drug reaction data collection sheet, page 1.

Figure 2. Adverse drug reaction data collection sheet, page 2. Cutccae cf AT* (include descTiptico arnl datesi

rr Csncurreit R«»xe*

^ Fetrcepective Fewrw Was there a *echaUen