Australasian Emergency Nursing Journal (2014) 17, 167—175

Available online at www.sciencedirect.com

ScienceDirect journal homepage: www.elsevier.com/locate/aenj

RESEARCH PAPER

Medication errors in ED: Do patient characteristics and the environment influence the nature and frequency of medication errors? Belinda Mitchell Scott, RN, MN a,∗ Julie Considine, RN, PhD, FACN, FFCENA b,c,d Mari Botti, PhD, GDCAP, BA, RN c,d,e a

Northern Health — The Northern Hospital, Emergency Department, 185 Cooper St Epping, Victoria 3076, Australia b Eastern Health — Deakin University Nursing & Midwifery Research Centre, Australia c School of Nursing and Midwifery, Deakin University, Australia d Centre for Quality and Patient Safety Research, Australia e Epworth Deakin Centre for Clinical Nursing Research, Australia Received 9 May 2014; received in revised form 26 July 2014; accepted 27 July 2014

KEYWORDS Emergency medicine; Emergency nursing; Patient safety; Risk management; Medication errors



Summary Background: Medication safety is of increasing importance and understanding the nature and frequency of medication errors in the Emergency Department (ED) will assist in tailoring interventions which will make patient care safer. The challenge with the literature to date is the wide variability in the frequency of errors reported and the reliance on incident reporting practices of busy ED staff. Methods: A prospective, exploratory descriptive design using point prevalence surveys was used to establish the frequency of observed medication errors in the ED. In addition, data related to contextual factors such as ED patients, staffing and workload were also collected during the point prevalence surveys to enable the analysis of relationships between the frequency and nature of specific error types and patient and ED characteristics at the time of data collection. Results: A total of 172 patients were included in the study: 125 of whom patients had a medication chart. The prevalence of medication errors in the ED studied was 41.2% for failure to apply patient ID bands, 12.2% for failure to document allergy status and 38.4% for errors of omission. The proportion of older patients in the ED did not affect the frequency of medication errors. There was a relationship between high numbers of ATS 1, 2 and 3 patients (indicating high levels

Corresponding author. Tel.: +61 421122268. E-mail address: [email protected] (B. Mitchell Scott).

http://dx.doi.org/10.1016/j.aenj.2014.07.004 1574-6267/© 2014 College of Emergency Nursing Australasia Ltd. Published by Elsevier Ltd. All rights reserved.

168

B. Mitchell Scott et al. of clinical urgency) and increased rates of failure to document allergy status. Medication errors were affected by ED occupancy, when cubicles in the ED were over 50% occupied, medication errors occurred more frequently. ED staffing affects the frequency of medication errors, there was an increase in failure to apply ID bands and errors of omission when there were unfilled nursing deficits and lower levels of senior medical staff were associated with increased errors of omission. Conclusions: Medication errors related to patient identification, allergy status and medication omissions occur more frequently in the ED when the ED is busy, has sicker patients and when the staffing is not at the minimum required staffing levels. © 2014 College of Emergency Nursing Australasia Ltd. Published by Elsevier Ltd. All rights reserved.

What is known • Observations of clinical practice suggest that medication administration in the Emergency Department (ED) can be challenging with multiple unknown patients requiring time critical medications at once. • It has been identified that delays in medication administration in the ED lead to increased in-patient mortality and increased patient and family dissatisfaction. • It was identified that 60% of patients had one or more medication errors (178 errors in 192 patients) and 37% of medication errors reached the patient. • It is estimated that 2.3% of medication errors in the ED are medication prescription errors and that 2.1—36.0% of medication errors in ED are administration errors.

What this paper adds • The conditions in the Emergency department when medication errors are more likely to occur. • The frequency of medication errors when incident reporting systems are not used for data collection. • The use of observation as means for data collection in medication errors.

Introduction Observations of clinical practice suggest that medication administration in the Emergency Department (ED) can be challenging with multiple unknown patients requiring time critical medications at once. Commonly, ED patients are administered analgesics for pain and antibiotics for treatment of infection.1,2 It has been identified in previous studies that ED patients are particularly prone to delays in medication administration, particularly related to analgesia.3—5 A number of studies have identified between 68% and 80% of patients experience a delay to analgesia of greater than 1 h.3,4 A further study by Kanjia and Dumaresque5 identified that only 34% of patients with septic shock received antibiotics within the recommended 1 h timeframe. It has been identified that delays in medication administration in the ED lead to increased

in-patient mortality5,6 and increased patient and family dissatisfaction.7 The ED provides an environment where the opportunity for medication errors to occur is high for a variety of reasons. The most compelling reason is the volume of medications administered and the time critical nature of medication administration. First, each ED patient receives an average of 2.5 medications per episode of care.8 Second, high numbers of patients requiring emergency care increases the risk of errors as multiple patients require treatment concurrently, often by the same clinicians. As the burden of chronic disease increases,9 ED patients are becoming more complex. They often have complicated past medical histories and require medications to manage chronic conditions in addition to the acute medical condition that led to the ED presentation that also often requires treatment with medications.8 Third, the need for ED clinicians to administer a wide variety of medication to patients of all ages increases the risk of error. A lack of familiarity with a medication requires additional time to review the medication literature to determine how to prepare the medication for administration. The opportunity for error arises for busy clinicians who may seek alternate advice on the preparation of the medication from colleagues or prepare the drug as they would any other intravenous medication.8 Most other units of the hospital will treat clusters of patients with ‘like’ illnesses, such as neurology, cardiology or orthopaedics, thus providing clinicians with some familiarity with the medications administered. Medication errors were included for further exploration in this study based on the highly variable findings of previous studies that show medication errors contribute to between 1.35% and 60% of all errors in emergency care.8,10—18 There was one prospective study identified through the literature review, this paper described the frequency of medication administration and prescription errors in ED.13 It was identified that 60% of patients had one or more medication errors (178 errors in 192 patients) and 37% of medication errors reached the patient. Of the errors identified, 38.4% were classified as medication administration errors and 53.9% were classified as prescribing errors. It is estimated that 2.3% of medication errors in the ED are medication prescription errors19 and that 2.1—36.0% of medication errors in ED are administration errors.14,15 In order to improve medication administration safety, Blank et al.8 used a pre-test post-test study to determine if the number of medication errors at one study site could be reduced by re-focusing nursing clinicians on

Medication errors in ED Table 1 Timing

169

Data collection schedule. No of occurrences

PPS No.

Total number of patients Emergency Department (n)

Weekday Morning Evening Night Weekend Morning Evening Night

Emergency Department cubicle (n)

1 3 1

1 5, 8, 9 2

34 104 27

21 54 19

2 1 1

4, 6 3 7

118 39 20

43 22 13

342

172

Total

the fundamentals of medication administration. The study involved examining clinicians’ knowledge of medication administration procedures; observing clinical practice for behaviours reflective of recommended medication practices and identification of medication administration errors, via chart review and reported errors. There was a 10% reduction (44% vs. 34%) in the number of medication errors detected as a result of the educational intervention. All of the studies above provide some understanding of the frequency of medication errors in ED however only one study was undertaken in an Australian context and it was a retrospective study reliant on ED clinicians reporting errors. The aim of this study was to determine the frequency and nature of medication errors in emergency care. For the purpose of this study, a medication error was defined as any failure of one or more of the following ‘rights’ of medication safety: right patient, right drug, right time, right reason or right documentation. A secondary aim of the study was to explore whether there were relationships between ED patient characteristics (age, clinical urgency) and ED characteristics (ED occupancy, ED staffing) and the frequency and nature of medication errors in the ED.

Methods A prospective, exploratory descriptive design using point prevalence surveys was used to establish the frequency of observed medication errors in the ED. Point prevalence survey methods derive from the discipline of epidemiology and rely on the researcher collecting ‘snapshots’ (cross sectional data) of the affected population in order to establish the frequency in which an event or phenomenon occurs.20 Point prevalence surveys were used to collect data related to observed errors involving medication management. In addition, data related to contextual factors such as ED patients, staffing and workload were also collected during the point prevalence surveys to enable the analysis of relationships between the frequency and nature of specific error types and patient and ED characteristics at the time of data collection. Approval for the study was granted from the Human Research & Ethics Committees at the study site and Deakin University.

Setting The study was conducted in a government funded health service with five campuses and 613 beds located in the northern suburbs of Melbourne, Australia. The ED in which the study was conducted is located in the only acute care campus of the health service. When the study was conducted, the ED was managing approximately 61,700 attendances per year and 20% of attendances were children aged 16 years or less.21 There were 30 treatment areas including 2 resuscitation cubicles, 11 monitored cubicles, 11 general adult non-monitored cubicles and 6 paediatric cubicles and the admission rate was 25%.21 At the time of the study, the ED used the National Inpatient Medication Chart for prescription of medications, however as a result of historical practices, medication prescriptions could also be written on the Emergency Department Record.

Sample ED patients were included in the study if they were in an ED cubicle during the point prevalence surveys: patients in the waiting room were excluded from the study.

Data collection Data were collected using point prevalence surveys: a total of nine point prevalence surveys were conducted from May, 2009 to July 2009 and occurred over various days of the week and times of the day (Table 1). All ED cubicles were included in each point prevalence survey and the following data were collected for each patient present in an ED cubicle during data collection: (i) Patient characteristics: age, gender (ii) Medication administration variables: presence of patient identification band, documentation of allergy status, presence of allergy band, medications charted but not given, medications given but not charted, duplicate medication orders. In addition, the following data were collected at the conclusion of each point prevalence survey:

170 Table 2

B. Mitchell Scott et al. Frequency of patient ID band use, documentation of patient allergies and application of allergy bands.

Point prevalence surveys

1 2 3 4 5 6 7 8 9 Total

Total number of patients in ED cubicles

ID Band not applied

Allergy status not documented

Allergy band not applied in patients with a documented allergy

n

%

%

N

%

21 19 22 18 14 25 13 23 17

11 4 5 8 10 13 8 8 7

52.3 21.0 22.7 44.4 71.4 52.0 61.5 32.0 41.1

2 5 3 4 2 1 0 2 2

2 0 0 2 3 3 3 3 2

9.5 0.0 0.0 11.1 21.4 12.0 23.1 13.0 11.7

9.5 26.3 13.6 22.2 14.2 4.0 0.0 8.7 11.7

172

71

41.2

21

18

10.4

12.2

(i) ED patient characteristics: age and ATS category of all patients present in the ED (cubicles and waiting room) (ii) ED characteristics: occupancy of resuscitation, monitored, general adult and paediatric cubicles, unfilled nursing and medical deficits, seniority of nursing and medical staff. Patients were subcategorised according to age (aged less than 65 years and patients aged over 65 years) and whether the proportion of older patients was less than or equal to 50% or greater than 50% of the total ED population (cubicle and waiting room patients). Patients were also subcategorised according to clinical urgency measured using triage categories (ATS 1, 2 and 3 and ATS 4 and 5) and whether the proportion of ATS 1, 2 and 3 patients was less than or equal to 50% or greater than 50% of the total ED population (cubicle and waiting room patients). The decision to use 50% as the cut off was arbitrary; there were no other similar studies found during the literature search that could inform the sampling decision, a mathematical approach was taken where over 50% is a majority, under 50% is a minority. The occupancy of resuscitation, monitored, general adult and paediatric cubicles was subcategories into less than or equal to 50% and greater than 50%. Unfilled nursing and medical deficits were subcategorised as yes or no. The seniority of nursing staff was based on the organisational requirement to have at least 5 Registered Nurses with postgraduate qualifications in emergency nursing per shift to safely staff the in-charge, triage and resuscitation positions and was therefore subcategorised as yes or no. Seniority of medical staffing was based on whether at least 50% of medical staff per shift were at Emergency Physician or ED Registrar level and subcategorised as yes or no.

Data analysis Data analysis was performed using IBM SPSS Statistics 18. Descriptive statistics were used to summarise the study data: where data were not normally distributed, medians and inter-quartile ranges (IQR) are presented. Chi-square test was used to compare between group differences.

n

Results A total of 172 patients were included in the study: 125 of whom patients had a medication chart. The median age was not collected however 49.1% (n = 181) were males. The triage category distribution was as follows: 1.3% (n = 5) ATS 1, 15.7% (n = 58) ATS 2, 38.5% (n = 142) ATS 3, 43.4% (n = 160) ATS 4 and 0.8% (n = 3) ATS 5. Of the patients included in the study, 305 patients had been seen by a medical officer and 63 were still waiting to be seen: all patients had undergone ED nursing assessment.

Medication errors ID bands were observed to be absent in 41.2% of patients (n = 71) and 12.2% of patients (n = 21) did not have their allergy status documented. A further 10.8% of patients (n = 18) did not have an allergy band applied despite having an allergy documented on the ED nursing chart (Table 2). Of the 125 patients with a medication chart, 38.4% of patients (n = 48) had errors of omission: that is medications ordered but not given. Errors of omission occurred across all medication point prevalence surveys (Table 3) and overall was 38.4%. There were five patients who had duplicate medication orders. Of the duplicate medication orders, 12 drugs were ordered allowing for multiple opportunities for medication error. The number of duplicate orders per patient ranged from one to five orders. There were also three occasions where nurses had administered medications without a medication order. On one occasion, nurse initiated paracetamol was administered which is permissible under the organisation’s nurse initiated medication schedule (Northern Health Clinical Manual, 2011, Medication Administration Policy) however the nurse initiated administration was not documented on the medication chart by the nurse.

Effect of patient characteristics on medication errors The profile of all patients present in the ED (cubicles and waiting room) was used to examine whether patient

Medication errors in ED Table 3

171

Frequency of errors of omission.

Point prevalence survey

1 2 3 4 5 6 7 8 9 Total

Total number of patients in ED cubicle

Total number of patients with Medication chart

Errors of omission

n

n

n

21 19 22 18 14 25 13 23 17

12 18 19 13 10 14 10 17 12

4 4 5 7 9 5 7 5 2

36.3 22.2 26.3 50.0 90.0 33.3 70.0 29.4 16.6

172

125

48

38.4%

characteristics (age and clinical urgency) had an effect on the frequency and nature of medication errors. The age of patients present in the ED did not have a statistically significant effect on the frequency of errors related to patient identification or documentation of allergy status. Clinical urgency of patients present in the ED had a statistically significant effect on documentation of allergy status: when proportion of ATS 1, 2 and 3 patients was less than or equal to 50% of the total ED population, the failure to document allergy status was 3.5% (n = 6) compared with 8.7% (n = 15) when the proportion of these patients was greater than 50% of the total ED population (p = 0.023) (Table 4). There were no statistically significant relationships between frequency of medication errors of omission and the proportion of older patients in the ED or clinical urgency of ED patients. The frequency of errors of omission was 13.5% (n = 17) when the proportion of patients aged 65 years or over was less than or equal to 50% of the ED population compared with 24.6% (n = 31) when the proportion of patients aged 65 years or over was greater than 50% of the ED population (p = 0.927). The frequency of errors of omission was 30.2% (n = 38) when the proportion of ATS 1, 2 and 3 patients was less than or equal to 50% of the total ED population compared with 31.8% (n = 40) when the proportion of these patients was greater than 50% of the total ED population (p = 0.578).

Table 4

%

Effect of ED characteristics on medication errors There was no significant relationship identified between failure to apply ID bands or failure to document allergy status and the occupancy of the resuscitation, monitored or paediatric cubicles. There was a statistically significant increase in failure to apply ID bands when the general cubicles were over 50% occupied (p = 0.001) (Table 5). Skill mix of nursing and medical had no significant effect on the frequency of failure to apply ID bands or document of allergy status. Unfilled nursing deficits resulted in an 18.1% increase failure to apply ID bands whilst unfilled medical deficits had no significant effect on either ID band application or documentation of allergy status. Increased ED occupancy had a statistically significant effect on errors of omission. When the occupancy of the resuscitation was over 50% of the total cubicles occupied, the frequency of medication omissions increased by 6.3% (p < 0.001) and in the general cubicles, occupancy over 50% increased errors of omission by 23.0% (p = 0.001) (Table 6). Nursing skill mix had no significant effect on the frequency of errors of omission but there was a statistically significant 15.9% decrease in errors of omission when medical staffing per shift had higher numbers of senior medical staff (p = 0.007). There was a statistically significant 19.1% increase in errors of omission when there were unfilled ED

Effect of patient characteristics medication errors. ID Band not applied

Age ≥65 years ≤50% of all ED patients >50% of all ED patients ATS 1, 2 and 3 patients ≤50% of all ED patients >50% of all ED patients

Allergy status not documented

Errors of omission

n

%

p

n

%

p

n

%

p

26 48

15.1 27.9

0.952

4 17

2.3 9.9

0.104

17 31

13.5 24.6

0.927

38 36

42.7 43.6

0.929

6 15

3.5 8.7

0.023

38 40

30.2 31.7

0.587

172

B. Mitchell Scott et al.

Table 5

Effect of ED characteristics on frequency of medication errors. ID Band not applied n

%

p

Allergy status not documented

Errors of omission

n

n

%

p

20 28

15.9 22.2

Medication errors in ED: Do patient characteristics and the environment influence the nature and frequency of medication errors?

Medication safety is of increasing importance and understanding the nature and frequency of medication errors in the Emergency Department (ED) will as...
592KB Sizes 0 Downloads 4 Views