Medication History Taking in Emergency Department Triage Is Inaccurate and Incomplete

Authors


  • Presented at the Society for Academic Emergency Medicine annual meeting, New Orleans, LA, May 2009.

  • The authors have no disclosures or conflicts of interest to report.

  • Supervising Editor: Mark Mycyck, MD.

Address for correspondence and reprints: Maryann Mazer, PharmD, MD; e-mail: maryannmazer@gmail.com.

Abstract

ACADEMIC EMERGENCY MEDICINE 2011; 18:102–104 © 2011 by the Society for Academic Emergency Medicine

Abstract

Objectives:  Medication error prevention has become a priority in health care. The Joint Commission recommends that a list of medications, dosages, and allergies be obtained from all patients. The authors sought to determine the accuracy of medication history taking in emergency department (ED) triage. The hypothesis was that there would be significant discrepancies between medications listed in triage and those the patient was actually taking.

Methods:  This was a prospective, cross-sectional survey of adult patients presenting to the ED. As a part of regular care, nurses recorded a medication list during triage in the electronic medical record (EMR). For this study, the triage medication list was rechecked during an independent patient interview.

Results:  Of 1,797 patients approached, 1,657 completed the survey (92%). The mean age was 39 years (standard deviation [SD] ±16 years). Discrepancies in medication lists obtained during triage were documented in 626 (37%) patients. Discontinued medications (163, 9.8%) were included, additional medications (463, 27.9%) were omitted, and 632 patients (38%) reported taking a nonprescription medication not listed in the EMR.

Conclusions:  Medication histories performed in ED triage are inaccurate and incomplete.

Medication errors are a major cause of morbidity and mortality. It is estimated that 9.7% of patients involved in an adverse drug event (ADE) have a resulting disability.1 Patients who experience an ADE have twice the risk of death as those who do not.2 The Institute of Medicine (IOM) estimates 1.5 million preventable ADEs occur annually, costing $3.5 billion per year.3 ADEs translate into increased emergency department (ED) utilization. Approximately 177,000 ED visits by elders are attributed to drug misadventures annually.4 These numbers are considered an underestimate, as many ADEs go unrecognized or unreported.

The IOM report and the Joint Commission have made medication safety a priority in health care. Medication errors and adverse effects can occur any time in the patient encounter,5 and the majority of ADEs can be prevented.6 The Joint Commission recommends a list of medications, route, frequency, dosage, allergies, and adverse reactions be recorded, a process called medication reconciliation. It is hypothesized that accurate reconciliation prevents therapeutic duplications, drug omissions, drug–drug and drug–disease interactions, dosing errors, allergic reactions, and adverse effects. Medication lists are often inaccurate, and the limited available data suggest that medication reconciliation is often not complete in the ED.7

This study attempts to evaluate the accuracy of medication histories performed during ED triage. We hypothesized that there would be significant discrepancies between medication lists obtained at triage and those medications patients were actually taking. We sought to determine the frequency of omitted medications, discontinued medications, and use of over-the-counter medications by ED patients.

Methods

Study Design and Population

This was a prospective, cross-sectional survey of adult patients presenting to the ED at a tertiary care hospital. Research associates enrolled a convenience sample of patients and administered a questionnaire regarding the medications they were currently taking. The study protocol was approved by the institutional review board, and informed consent was obtained.

The study site is an urban university teaching hospital with 59,000 ED visits annually. Patients over 18 years old presenting to the ED were considered eligible. Patients were excluded if they required emergent treatment or were unable or unwilling to participate.

Survey Content and Administration

As part of usual care, an ED nurse blinded to the study goals generated a medication list at triage from the patient and entered it into the electronic medical record (EMR). Nurses did not receive any additional training in obtaining medication histories. For those patients unable to provide a medication list and with a previous ED encounter, the prior medication list was used as a default. Our EMR automatically performs this function even outside of this study. Research associates administered the survey (Figure 1) in a confidential manner after triage in the treatment area. The associates received standardized training in administering the survey. Subjects were asked to verify all prescription and nonprescription medications they were taking, and these were compared to the list obtained during the triage interview. Omitted and discontinued medications were noted. The patient’s self-reported list was used as the criterion standard. Outcomes included number of medications, discontinued medications, omitted medications, nonprescription medications, and overall medication discrepancies.

Figure 1.

 Survey questionnaire.

Data Analysis

Data were normally distributed and are presented as frequencies with 95% confidence intervals (CIs). Percentages were used for categorical data and means with standard deviations (±SDs) are reported for continuous data. All analyses were performed using SAS statistical software (Version 9.1.3, SAS Institute, Cary, NC).

Results

Of 1,797 patients approached, 1,657 competed the survey (92%). The mean age of patients was 39 years (SD ± 16 years). This was consistent with our general ED population. The mean number of medications taken was 2.6 per patient (range = 0–21). Thirty-one percent of patients (517) were not taking any medications.

Omitted and discontinued medications were noted in 626 patients, yielding a 37% discrepancy rate (95% CI = 35% to 40%). This included 163 (9.8%, 95% CI = 8.5% to 11.4%) patients with discontinued medications and 463 (27.9%, 95% CI = 25.8 to 30.2) patients with omitted medications. Thirty-eight percent (632) of patients reported taking over-the-counter medications. When the type of over-the-counter medication was analyzed, we found 19% of patients were taking nutritional supplements, 21% herbal products, 35% vitamins, and 38% analgesics.

Discussion

Medication reconciliation is composed of five steps: 1) obtain a list of current medications, 2) determine a list of medications to be prescribed, 3) compare the two lists, 4) make clinical decisions, and 5) communicate the new list to appropriate caregivers and the patient. It is also recommended that drug allergies and adverse drug reactions are obtained. These steps are often not completed in ED triage. Medication errors typically occur at “junctions of care,” when a patient is admitted to, transferred within, or discharged from a health care facility.2,3 Of patients presenting to the ED for evaluation, a high percentage will receive pharmacotherapy, so accurate knowledge of current medications is an important part of preventing ADEs.

Medication reconciliation systems have reduced ADEs, and techniques being considered to improve medication reconciliation include using pharmacy personnel to interview patients, reviewing outpatient records and using software to maintain medication lists.8 This process will continue to be refined as the 2010 Joint Commission National Patient Safety Goals are less rigid, focusing on patients who are being admitted or transferred, and not necessarily all patients, as previously recommended. These performance measures are currently not in effect, due to difficulties in implementation.9

Our study raised questions about the medication history taking process, especially how it pertains to ED patients. In our sample, histories were obtained at triage, but it is uncertain when the optimal time for this to occur is. It might be more effective to perform this task during treatment or at admission. It is also unclear who is the best care provider to perform this task: physicians, nurses, pharmacists, or pharmacy technicians. One study demonstrated increased accuracy in ED triage by using pharmacy technicians to review medications.8

The criterion standard for medication reconciliation should also be questioned: should it be the patient’s self-report, outpatient provider, pharmacy record, or a combination of the three? This study used an EMR, which might potentiate the carryover of formerly prescribed medications. This ability to “cut” and “paste” within an EMR has been highlighted as a pitfall, because care providers can copy large amounts of information without prompting critical thinking. Factors such as ED volume may contribute to the phenomenon. The optimal EMR format has yet to be established. Ideally, integration of medical and pharmacy records into one platform may improve prescribing accuracy, while encouraging clinical evaluation of patient data.10

Limitations

Our site is a single, urban teaching hospital, and results may not be applicable to other settings. There was a smaller representation of elders, who are on more medications and at increased risk for ADEs. Only one type of EMR, which prepopulated medications from previous encounters, was studied. Patients who required immediate attention, which might include those with ADEs, were excluded. Patients were not queried if their presenting complaint was related to an ADE. Nurses rotated assignments throughout the ED and not exclusively triage. We did not explore the effect of ED volume or privacy on triage accuracy. The time between triage and interaction with the research associate varied due to staff, number of patients, and interventions which patients required. Our study did not address discrepancies in dosage, frequency, or allergies, which underestimates the true prevalence. We did not examine the clinical relevance of medication discrepancies. The survey was not previously validated. The patient’s response was used as the standard for medications taken, and outpatient records were not verified.

Conclusions

Our study showed significant discrepancies between medication lists generated in ED triage and those later verified by the patient. This has significant implications, as medication history taking is intended to prevent adverse drug events, which are associated with morbidity and mortality. Cautious use of the electronic medical record is needed to prevent perpetuation of inaccurate medication lists. Further research is needed to determine the optimal approach to obtaining medication histories in the ED.

Ancillary