Data Collecting Strategies for the Improvement of

Data Collecting Strategies for the Improvement of Healthcare Service Analyzing Dispensing Errors in a Hospital Pharmacy

Yoshiko HABUCHI

Toshitaka TAKENOUCHI

Yoichi MOTOMURA

Center for Service Research National Institute of Advanced Industrial Science and Technology, AIST Tsukuba, JAPAN [email protected]

Department of Pharmacy Services Showa University Hospital Tokyo, JAPAN [email protected]

Center for Service Research National Institute of Advanced Industrial Science and Technology, AIST Tsukuba, JAPAN [email protected]

Abstract— In this report, we first describe the importance of both a quality management system and a risk factor monitoring system for healthcare service. This report also discusses a case study that monitored dispensing errors in a hospital pharmacy. The authors used audit records and then interlinked error data with both location and prescription data to investigate the circumstances and potential causes of these errors. The results suggest that the frequency of prescription of certain drugs is a key factor in these types of errors. Understanding the effects of this high volume of prescriptions in the hospital pharmacy could help prevent errors and increase productivity. This report showed that an effective monitoring system could create concrete improvements and enhance utilization of information. Keywords-component; Quality Management System for Healthcare Service; Evidence Based Improvement; Hospital phamacy; Dispensing errors; Service engineering

I.

BACKGROUND

One of the most significant challenges faced by healthcare organizations today is how to effectively demonstrate accountability for the quality and safety of the medicine provided to patients. Reducing drug-related iatrogenic incidents is a concern in Japan, as well as the rest of the world. Mistakes in the prescription and dispensing of medication are recognized as a major cause of medical errors. These types of mistakes can have serious implications for the health of patients. In the service business, if companies are to compete, survive and prosper in the information age, they must use measurement and management systems derived from their strategies and capabilities [1]. This same situation applies to healthcare service. Today, almost all healthcare organizations use quality initiatives and quality management programs. The issue of measurement is central to these types of programs, so above all healthcare organizations are already familiar with a variety of process quality and safety measurements. Some medical organizations start to construct quality and safety management systems along the lines of a quality management standard system such as the ISO9001 [2]. However, those systems and standards only create guidelines

that set out what needs to be done. Medical organizations that want to follow these plans then have to deal with the problems of how these things can actually be done by themselves. In healthcare service, there are some processes that naturally resist definition and standardization. These processes are in fact often more art than science [3]. For instance, complex operations in changeable environments cannot be performed as if one were following instructions in a manual; these situations require the technique and experience of a talented craftsperson. However, as all aspects of the healthcare field do not involve this level artistic process, there needs to be a way to bring the general environment under control by process standardization. It probably makes sense for those in healthcare service to define and standardize the fundamental issues involved in this field. Today, organizations are gathering a greater volume of data than they ever have before. The difficulty involved in this collection deals primarily with ensuring data quality, as well as integrating and deciding what subsets of data need to be made easily available in data stock. Most healthcare service organizations still face significant difficulties with these basic data issues. Each healthcare organization establishes its own reporting system for medical errors and adverse incidents. This is undertaken with the aim of investigating medical errors, understanding their causes, and establishing preventative measures to reduce their frequency. A comprehensive national reporting of medical nearmiss/adverse incidents by health care providers has created a precedent that may lead to methods to enhance patient safety. In 2004, a multi-institutional, Internet-based reporting system for adverse medical incidents was developed by the Japan Council for Quality Health Care. This system provides information about medical near-miss/adverse incidents and evaluates the circumstances surrounding these incidents in order to maintain public confidence in healthcare services and improve the quality of these services. Although the systems discussed above constitute an important strategy for enhancing patient safety, they have been seen as limited in several ways [4]. First, because the reporting is voluntary, the frequency of incidents reported

does not represent the true scope of these types of errors. All voluntary systems are likely to suffer from selective reporting and thus incomplete ascertainment of data. Second, the strength and comprehensiveness of the local investigations of these errors varied widely, and the reliability with which contributing factors were identified probably varied as well. Caamaño et al. (2002) reviewed methods used to study dispensing quality in community pharmacies. This study concluded that the use of external observers and simulated clients was likely to increase the validity of the reported results. [4]. However, both of these methods are very time-consuming, expensive, and associated with potential ethical problems [5]. Valid meta-analysis regarding errors in the dispensing of medication is extremely difficult to perform because of differences in clinical settings, organization members, patient populations, and drug dispensing methods. The most important factor in preparing for sophisticated analysis is the standardization and availability of a sufficient volume of high-quality data. For this reason, we used the audit reporting system at the final check stage at the hospital pharmacy that participated in this study as an easily administered screening tool. We collected error data from the audit reports and crossreferenced this with location information. Then we determined the types of dispensing errors and the frequency of their occurrence in terms of location in order to explore the reasons why they occurred. This would allow us to suggest preventive measures. Finally we tried to suggest improved strategies for using the collected data. II.

METHODS

The study was conducted in two parts. The first focused on counting and classifying the errors from audit reports that were gathered from the final check stage in the hospital pharmacy. Then we focused on analysis of these errors based on the nature of the prescription and the location in the pharmacy in which these errors occurred. Frequency of errors and prescription medicine Frequency of errors Setting The survey took place in a hospital pharmacy department during the period from 23rd July 2008 to 31st December 2008. The hospital that participated in the study has been certified by the Ministry of Health, Labor and Welfare as an advanced treatment facility. The hospital has 853 beds and 35 diagnosis and treatment departments. The dispensing activities of the pharmacy between 8:30 a.m. and 9:30 a.m. for inpatient items, and between 9:30 a.m. and 5:00 p.m. for outpatient items were surveyed. All members of the pharmacy staff except administrative staff (n=30) are involved in the dispensing of inpatient medicine. Five members of the staff work in the pharmacy throughout the day and cover dispensing services for outpatients. Figure 1 shows the process of medical prescriptions in the hospital and the audit checks by pharmacy inspectors. First, all prescribed medications were subjected to an audit when the prescriptions were checked. Then the dispensing

works were divided into three categories depending on the type of medicine. This study focused on mix-up errors during the preparation of tablets. All dispensing drugs were checked in two steps by two inspectors, who visually checked the drug contents against the prescription before the drugs were delivered to the wards. Prescribed medication

Checking prescription

Powdered medicine

Tablets and External medicine

Liquid medicine

Primary checkup

Audit Report 1

Secondary checkup

Audit Report 2

Ward delivery Figure 1.

Process chart of dispensing work

Definition In this study, errors were classified into fifteen categories (Table I). Particular attention was given to “Incorrect drug strength” errors and dispensing “Incorrect drug” errors, because they had the potential to severely harm patients. Frequency of prescription medicine To estimate the total amount of medicine prescribed in the pharmacy, we examined prescription medicine records through a stratified random sampling method. This was done because the hospital pharmacy in this study did not record prescriptions as data classified by the type of medicine used. The survey involved 31,275 prescriptions dispensed during the period between July 23rd and October 31st in the hospital pharmacy department. The minimum number of samples, n (drawn from a population of N prescriptions), required needed in order to reach a reliability of 95% and a range of error of 5% was 380 according to the finite population sampling formula. For subsequent analysis, 997 prescriptions have been extracted using a stratified random sampling method. All prescriptions were first divided into 28 categories based on the month and a day of the week that the medicine was originally dispensed. The information was then randomly extracted from these categories.

TABLE I.

AUDIT DETAILS IN SURVEYED HOSPITAL PHARMACY

1) Incorrect drug: Dispensing a drug that is different from that which was prescribed. 2) Incorrect strength: Dispensing a dose unit containing the wrong amount of the correct drug, without an appropriate adjustment to the dosing instructions. 3) Dose added or missing doses: Dispensing a larger/smaller quantity of drug than that which was prescribed. 4) Omission of item: Failure to dispense a prescribed item. 5) Omission of attached document: Failure to attach document containing information regarding doses. 6) Omission of modifying instruction on the drug bag or the label: The instructions printed on the drug bag, or the label had not been modified, despite the dosage being changed. 7) Incorrect drug bag: Mixing the bags with those of other patients' or different types of drug bags. 8) Incorrect scale marks for liquid medicine: Incorrect dosage instructions on the label. Forgetting to put a label on the bottle. Forgetting to put in a cup for liquid medicines. 9) Inappropriate doses of diluents: The amount of diluents that are added is not defined. 10) Disagreement in numbers of divided powder: The numbers of divided powders differ from the prescription. 11) Incorrect check colors: The line color painted on the divided powder packages does not follow the operational rule. 12) Unconfirmed dosage: Unconfirmed dosage though there is a question about dosage. 13) Unconfirmed dose regimen: Unconfirmed dose regimen though there is a question about the dose regimen. 14) Drug interactions: The interaction of the medicine described in the prescription is not confirmed. 15) Miscellaneous: Any errors not included in the above categories.

III.

RESULTS AND DISCUSSION

A. Frequency of errors During the study period, a total of 50,194 prescriptions were issued. At the final check stage, a total of 1,540 errors were reported (3%). Table II indicates the classifications of dispensing errors identified at the final check stage during the study period. The top three frequent errors were “Dose added, or missing doses,” “Omission of attached document,” and “Omission of items”. These errors accounted for 63.9% of total errors due to absent-mindedness. Such errors are the easiest to detect, so they may not result in immediate harm to patients, but they do reduce work efficiency. By contrast, although “Incorrect drug strength” and dispensing of “Incorrect drugs” errors were the fourth and fifth most frequent errors, such errors have the potential to severely harm patients. For this reason, it is important to reduce the frequency of these errors and finally prevent them altogether. The other errors occurred less frequently during the study period. B. Relationship between the frequency of errors and the prescription medicine The only data in which drug names were recorded by the pharmacy inspectors were error data that involved mix-ups in the dispensing of drugs. For this reason, the following analyses were limited to errors involving “incorrect strength” and “incorrect drug”. Figure 2 shows the correlation between the frequency of the medicine prescribed and dispensing

errors. “Error (Original drug)” indicates the error, of being “prescribed but not executed,” and “Error (Error drug)” indicates the wrong drug was executed. Both these errors that involved mix-ups were positively correlated with the frequency of sampling drug [Original drug: r=.22, p