Wristband Errors in Small Hospitals

Transcription

1 PHLEBOTOMY J a n e C. Dale, MD Stephen W. Renner, MD Wristband Errors in Small Hospitals A College of American Pathologists' Q-Probes Study of Quality ssues in Patient dentification Although methods of identification have improved from fingerprinting, first introduced in the scientific literature in 1880,1 to the DNAbased techniques in use today, patient identification in the hospital setting has remained largely unchanged. n this scenario, a health care worker asks the patient to state his or her name and applies an identification band to his or her wrist. The wristband then is used to ensure the accurate identification of that patient throughout the hospital stay. Although patient identification is a crucial first step in quality laboratory testing, the laboratory must rely on other hospital services for the timely placement of accurate identification bands. The College of American Pathologists' (CAP) Q-Probes program is a voluntary quality improvement program that develops and uses standardized tools to gather data, determine performance, and present opportunities for improvement in a wide range of laboratory activities. n 1991, a study of wristband identification errors was conducted as a part of the Q-Probes program.2 The average bed size of hospitals participating in that study was 265. n 1993, in response to the need for studies designed for smaller hospitals, the CAP introduced a separate Q-Probes program for hospitals with a bed size of less than 200. As a part of this program, we studied wristband errors in participating small hospitals. The purpose of the study was to determine the frequency of different types of errors and to identify policies and procedures that influenced wristband error rates. Methods and Materials Subscribers to the CAP Small Hospital Q-Probe program for the first quarter of 1993 participated in this study. The distribution of the 204 participating institutions according to bed size was as follows: ABSTRACT We compared wristband errors for 204 small hospitals. Phlebotomists examined wristbands on 451,436 occasions and identified 25,800 errors (total error rate, 5.7%). The absence of a wristband accounted for 64.6% of all errors reported; wristbands with missing, 12.4%; multiple wristbands with different, 12.1%; wristbands with erroneous, 6.7%; illegible wristbands, 3.5%; and patients wearing another patient's wristband, 0.7%. Factors found to correlate with lower error rates were the practice of sending written correspondence to the nursing service involved for each error detected, the practice of having nursing staff monitor wristbands on patient transfer, and laboratory accreditation from the College of American Pathologists (CAP). Factors found to correlate with higher error rates were the practice of allowing wristbands to be placed on objects that may become separated from the patient (eg, chart, beds, wall) and the practice of o o o having nurses responsible for initial wristband placement..161 (80.5%) institutions in the 1- to 150-bed.37 (18.5%) institutions in the 151- to 300-bed. 2 (0.1%) institutions in the 301- to 450-bed 4 institutions did not provide about bed size. Written instructions and data collection forms were provided. Phlebotomists monitored patient identification during all venipunctures and skin punctures performed during a 3-month period. Patients treated in the emergency department (ED) were included only if that patient population was routinely wristbanded. From the Department of Laboratory Medicine and Pathology, Mayo Clinic, Rochester, Minn (Dr Dale), and Pathology and Laboratory Medicine and Research Services, Department of Veterans Affairs, Medical Center, West Los Angeles, Calif (Dr Renner). Reprint requests to Dr Dale, Mayo Medical Laboratories, Hilton 378, Mayo Clinic, 200 First Street SW, Rochester, MN LABORATORY MEDCNE i

2 <1 1-<2 2-<4 4-<6 6-<8 8-<10 i0-<15 15-<20 Total Error Rate (%) Fig 1. Distribution of total wristband error rates (%) among participating institutions (N - 204). All patient encounters and wristband errors were prospectively tallied. Wristband errors were categorized as the following: Absent wristband Patient wearing another patient's wristband Patient wearing more than one wristband with conflicting Wristbands with missing Wristbands with partially erroneous Wristbands with illegible Participants also answered the following questions pertaining to the patient identification and wristband policies and procedures at their institutions: Fig 2. Proportions of different types of wristband errors, as percentages of total errors (based on aggregate data). Who is primarily responsible for performing phlebotomy at your institution? Does your institution have a written protocol for identifying patients at the time of phlebotomy? What is included on the identification band? Are identification bands allowed to be placed somewhere else other than on the patient's wrist? r- More than one band (12.11%) m^^- Wrong (6.74%) ^ X Other patient's band (0.68%) Missing (12.43%) wr- llegible (3.46%) Absent band (64.59%) LABORATORY MEDCNE Who is primarily in charge of initially placing an identification band on a patient? s a physician's written order required for wristband removal? 8.29 Are wristband errors continuously monitored by the phlebotomy staff? By the nursing staff? f wristband errors are discovered by phlebotomists, are nursing service and the ward clerk or 20-<60 secretary notified in person immediately? s phlebotomy refused until the wristband error is corrected? f a patient needing phlebotomy is encountered without a wristband, what do you do? f you routinely refuse to draw a specimen until a wristband is replaced, how is the laboratory usually notified that the wristband has been placed? s written correspondence generated for each wristband error identified? To whom is that correspondence sent? Do you have a regularly scheduled interdisciplinary forum during which patient identification problems are routinely discussed? Are wristband errors a permanent quality assurance indicator at your institution? The total error rate was calculated by dividing the total number of errors detected by the total number of wristbands monitored. The error rate for each type of error was similarly calculated. Participating laboratories then were ranked by percentiles for total error rate and the error rates for each of the type of errors in order to compare institutions. nstitutions with lower error rates were given higher percentile rankings while those with higher error rates were given lower percentile rankings. The influence of various policies and procedures on total error rates was assessed. Results Data were collected from 204 small hospitals during a 3-month study period in Of 451,436 patients examined for wristband errors, 25,800 errors were detected, for a total error rate of 5.7%. For individual institutions, the number of patients monitored ranged from 82 to 24,475; total errors detected ranged from 2 to 1,080. The mean and median institutional total error rates were 8.4% and 6.5%, respectively. The distribution of total Wristband error rates is shown in Fig 1.

3 ~ 1 (%) The proportions of different types of wristband errors, as a percentage of total errors, are shown in Fig 2. Rates for each type of error, as a percentage of total patients monitored, are shown in Fig 3. The most common error was the absence of a wristband, accounting for 64.6% of all errors detected and 3.7% of patients examined. A number of practice factors were shown to be correlated with error rates (at a significance level of 0.10) through a step-wise regression model. The relative importance of factors that correlated with error rates is shown in Fig 4. CAP accreditation also was shown to correlate with lower total error rates. The majority of facilities (55%) used phlebotomists, aides, or assistants to perform most phlebotomies. Nineteen percent used medical laboratory technicians and/or medical technologists primarily. The remainder of facilities used a combination of these. Error detection rates were found to be independent of the classification of employee who performed phlebotomy. Almost all participating facilities included the patient's name and hospital number on wristbands (99.5% and 93.2% of respondents, respectively). A majority also included date of birth and sex (65.9%), patient's location (58%), and date of admission (50.2%). Many also included the patient's doctor's name (44.9%). Fewer included age (14.1%), race (13.2%), address (8.3%), religion (6.8%), and other identifying numbers. The majority of facilities used supplementary wristbands for emergency use (58.9%). Some facilities (24%) included a check-sum to ensure correctness of the wristband number. (A check-sum is a mathematical manipulation of a string of numbers such that transposition of one of the numbers in the series would be detected.) Error rates were independent of the type of included on the wristband. None of the participants in this study used bar-coded wristbands. The majority of institutions (69.8%) allowed wristbands to be placed around the ankle. One fourth of institutions also allowed wristbands to be attached to the patient's chart, bed, or walls. The practice of allowing an identification band to be placed on objects from which the patient might become separated was associated with an increased total error rate. The admissions department (56%) or nursing staff (51%) most commonly placed wristbands. The policy of nursing service being primarily responsible for placing wristbands was associated with an increased error rate.,1 1 ~ ummm Absent band Missing More than one band Wrong llegible Other patient's band Wristband Error Type n the majority of facilities (60%), phlebotomy continuously monitored for wristband errors. Nursing service performed continuous monitoring at 33% of facilities. The frequency of monitoring by nursing service was variable 14% before transfusion, 13% daily, 13% per shift, and 10% on admission. n 8% of facilities, the nursing service monitored for wristband errors when patients were transferred. This practice was associated with a lower total error rate. Almost all facilities had written protocols for patient identification. Only 1.5% required a physician's order to remove a wristband. Neither of these two policies affected total error rates. Sixty-nine percent of facilities routinely wristbanded patients in the emergency department. This practice was associated with a lower total error rate. Almost all facilities (97%) notified nursing service immediately when wristband errors were detected. Three fourths notified the ward clerk/secretary immediately; one fourth refused to collect the specimen until the error was corrected; Fig 3. Error rates of different types of wristband errors as percentages of total patients monitored. Fig 4. Participant practices influencing total error. WB indicates wristband; ED, emergency department. The length of the bar reflects the strength with which each practice is correlated with wristband error rates (the longer the bar in either direction, the stronger the correlation). Fewer Errors More Errors Nursing primarily places WB WB allowed on chart/bed/walls ED patients routinely banded WB monitored at transfer Correspondence sent to nursing LABORATORY MEDCNE 20S

4 of processes and outcomes are interdisciplinary and collaborative."4 All factors found to be associated with wristband errors in our study involved nursing and/or administrative policies and procedures and represent opportunities for crossdepartmental efforts to improve patient care. Discussion The placement, use, and misuse (ie, removal) Accurate patient identification is a cardinal rule of wristbands can occur in a complex series of of all patient care activities. The Clinical procedures involving a variety of hospital services Laboratory mprovement Amendments of 1988 and employees (Fig 5). Disruption of any step in require that laboratories have written policies and the process may lead to inappropriate or delayed procedures for specimen collection and labeling delivery of patient care services. Although ranthat ensure positive specimen identification. 3 dom errors may occur anywhere in the process, Patient identification wristbands (D band) play systematic problems may be uncovered by an integral role in maintaining accurate identifi- reviewing the entire patient identification process cation of hospitalized patients. n order for the and determining the type of wristband errors freidentification process to work, wristbands must quently detected. This type of continuous quality contain complete, accurate, and legible informa- improvement activity requires cooperation tion, and must be promptly placed on the correct among a wide range of hospital departments. individual. The laboratory must rely on other Because three of the factors shown to correlate services for the timely placement of accurate with wristband error rates in our study directly wristbands. The phlebotomist must confirm the involve nursing service (nursing service responsipatient's identification, assure that the name and ble for placing wristbands, nursing service responother demographic data on the D band and test sible for monitoring wristbands when patients are requisitions agree before collection, and accu- transferred, phlebotomy corresponds with nursing rately label all specimen tubes. The testing labo- service when wristband errors are identified), ratory relies on the phlebotomist to guarantee cooperative activities between the phlebotomy and that the specimens received have been correctly nursing services may reduce patient identification labeled. errors. Our study showed that the practice of sendthe 1996 Joint Commission on Accreditation of ing written reports to the nursing service involved Healthcare Organizations fjcaho) Accreditation for each identification error discovered had the Manual stresses the need for careful coordination strongest correlation with lower error rates. and collaboration among departments in quality mplementation or improvement in this type of improvement activities.4 JCAHO performance practice should be considered. improvement standard (P) 1.1 addresses whether nstitutions in which nursing service was pri"designing, measuring, assessing, and improving marily responsible for placing D bands had one fifth generated a written report with each incident. Many laboratories sent reports to the hospital QA committee and nursing service. Sending correspondence to nursing service was associated with a lower total error rate. Fig 5. Hypothetical wristband f l o w diagram. ED indicates emergency department; WB indicates w r i s t b a n d ; V intravenous. Patient presents to ED T Preliminary paperwork Surgery Dietary Pharmacy Phlebotomy Nursing Blood bank Other ancillary services Medical staff Patient sees physician Physician decides to admit patient X Patient sent to floor WB applied by nursing WB generated by admissions department Positive patient identification Proper/appropriate patient care activities WB sent to floor WB removed from patient to wall, chart, or bed, or discarded Request for new WB Patient Surgery Nursing LABORATORY MEDCNE V team Physical therapy Other ancillary services

5 higher error rates, in general, than did those institutions that delegated that responsibility to other departments. Given the many patient care duties of nursing service personnel, consideration should be given to delegating responsibilities for D band placement to another department that typically encounters the patient early in the admitting process (eg, admissions personnel), when feasible. Ten percent of respondents indicated that nursing service monitored for D band errors on admission, 13% daily, 13% every shift, and 8% upon patient transfer. The practice of monitoring D bands on patient transfer was associated with lower error rates. n our study, only 14% of respondents indicated that nursing service monitored for D band errors before transfusions. Clearly, transfusion policies mandate positive patient identification, which requires checking the D band for accuracy and completeness.5 The response to this question may indicate that laboratory personnel are unfamiliar with nursing practices. Because the study did not require input from nursing, however, we are unable to confirm this theory. n a 1994 QProbes study of blood transfusion practices in 497 hospitals, 70% of participating institutions required two individuals to read aloud to each other the identifying on the patient's wristband and the blood bag to verify patient and blood unit identification before infusion.6 When transfusions were monitored in the study, this identification procedure was not observed in 11% of the transfusions. The practice of checking the patient's wristband for identification and matching with the intended blood product must be in place and should be emphasized at regular intervals with all hospital staff responsible for the preparation and administration of blood products. Two other nonlaboratory practices were associated with D error rates. nstitutions that routinely banded patients seen in the emergency department had lower error rates than those that did not. Because many patients are admitted through the emergency department, consideration should be given to routinely applying D bands on arrival. The practice of allowing D bands to be placed on objects from which the patient may become separated was associated with higher error rates. This practice should be prohibited. The final factor found to be associated with lower wristband error rates was CAP accreditation of the participating laboratory. This finding reinforces the fact that sound laboratory practices and policies can have a tangible impact on the quality of patient care beyond the confines of the laboratory. Our small hospital study (median bed size 99) showed a higher total error rate (5.7%) than did a previous CAP Q-Probes study conducted in larger hospitals (median bed size 265), which showed a 2.7% total error rate.2 Because of differences in study design, including number of participants, length of study, exclusion of pediatric patients as well as those admitted to the emergency department, and different patient tally procedures, a direct comparison of both studies cannot be made. t should be noted, however, that in the 1991 study, 67% of participating laboratories indicated that they refused to perform phlebotomies until any wristband error was corrected a practice that was associated with lower error rates. n this study, only 26% of participants adhered to that policy. Consideration should be given to implementation and enforcement of the policy of delaying specimen collections until after wristband errors are corrected. n addition, nursing service was more frequently responsible for the initial placement of wristbands in the small hospital study (51% of participating institutions) than in the large hospital study (41%). Also, the practice of giving primary responsibility to nursing service was associated with a higher error rate. The 5.7% error rate reported by this study suggests that there is opportunity for further reducing wristband errors, especially because the errors detected may represent only a fraction of the errors that actually occur. While efforts to detect errors must continue, greater emphasis should be placed on the prevention of errors. mprovement in the accuracy and timely placement of wristbands requires the cooperative efforts of a number of hospital departments, including administration, medical staff, nursing staff, nursing service, and the laboratory. References 1. Faulds H. On the skin furrows of the hand. Nature. Oct 28, 1880; Renner SW, Howanitz PJ, Bachner P. Wristband identification error reporting in 712 hospitals. Arch Path Lab Med. 1993;117: Clinical Laboratory mprovement Amendments of 1988, final rule. Federal Register. February 1992; 57: Joint Commission on Accreditation of Healthcare Organizations Comprehensive Accreditation Manual for Hospitals. Oakbrook, 111: Joint Commission on Accreditation of Healthcare Organizations. 1996: American Association of Blood Banks. Technical Manual. 12th ed. Bethesda, Md: American Association of Blood Banks; 1996: Renner SW, Howanitz PJ. Transfusion errors: data analysis and critique. Northfield, 111; College of American Pathologists; 1994; Q-Probes. LABORATO