Electronic Health Record Assignment

O R I G I N A L A R T I C L E

Effect of an electronic medication administration record application on patient safety

Noelia Vicente Oliveros PharmD1 | Teresa Gramage Caro PharmD PhD1 |

Covadonga Pérez Menendez‐Conde PharmD PhD1 | Ana María Álvarez‐Díaz PharmD1 |

Sagrario Martín‐Aragón Álvarez PhD3 | Teresa Bermejo Vicedo PharmD PhD2 |

Eva Delgado Silveira PharmD PhD1

1 Hospital Pharmacist, Hospital Universitario

Ramón y Cajal, Department of Pharmacy,

Madrid, Spain

2 Chief of Pharmacy, Hospital Universitario

Ramón y Cajal, Department of Pharmacy,

Madrid, Spain

3 Professor, Universidad Complutense de

Madrid, School of Pharmacy, Department of

Pharmacology, Madrid, Spain

Correspondence

Noelia Vicente Oliveros, Hospital Universitario

Ramón y Cajal (Department of Pharmacy),

Carretera de Colmenar Viejo km 9,1; 28034

Madrid, Spain.

Email: noelia.vicente@salud.madrid.org

Abstract

Rationale, aims, and objectives: To evaluate the effect of an electronic medication admin-

istration record (eMAR) application on the rate of medication errors in medication administration

recording (ME‐MAR).

Methods: A before‐and‐after, quasiexperimental study was conducted in a university hospital

that implemented the eMAR application in March 2014. Data collection was conducted in April

2012 (pre‐) and June 2014 (post‐) by two pharmacists. The ME‐MARs were analysed by the staff

involved to identify their cause. The two pharmacists independently classified the ME‐MARs. In

the case of disagreement, a research team examined the ME‐MARs and categorized them by

consensus. Three classifications were used: A classic medication error taxonomy and 2

technology‐induced error taxonomies.

Results: The pharmacists analysed 2835 (pre‐) and 2621 (post‐) medication administration

records (MAR), respectively. Overall, the ME‐MAR rate decreased from 48.0% (pre‐) to 36.9%

(post‐) (P < .05). The same types of ME‐MAR were observed in both phases except for “MAR with

incomplete information,” which was not observed in the postimplementation phase. In both

phases, the most frequent ME‐MAR was “MAR at the wrong time” (MAR before or after medica-

tion administration) (31.6% vs 30.2%). The main cause of ME‐MARs in both phases was the fail-

ure to follow work procedures. The potential future risk of ME‐MARs significantly decreased

after the eMAR implementation (P < .05). All ME‐MARs were “use errors” because of human

factors. New ME‐MARs (1.24%; n = 12) were observed in the postimplementation phase.

Conclusion: Use of the eMAR application significantly reduces the rate of ME‐MAR and their

potential risk. The main cause of ME‐MAR was the failure to follow work procedures.

KEYWORDS

clinical safety, evaluation, medical error, medical informatics

1 | INTRODUCTION

More than 15 years have passed since the “To Err Is Human” report

was published and considerable controversy remains on how much

improvement in patient safety has actually been achieved.1 Clearly,

some progress has been made, but improvement is still proceeding at

a glacial pace. Nevertheless, the implementation of healthcare informa-

tion technology (HIT) has provided an opportunity for continuing

improvement.2 A great deal of clinical care involves gathering and syn-

thesizing information. In healthcare systems with increasing patient

complexity and distribution of care, high standards of patient care

can no longer be supported by traditional paper‐based information

management.3

Particular emphasis has been placed on the use of HIT to reduce

medication errors.4,5 Advocates of HIT contend that the widespread

use of systems such as Computerized physician order entry (CPOE)

Received: 10 November 2016 Revised: 7 March 2017 Accepted: 8 March 2017

DOI: 10.1111/jep.12753

888 © 2017 John Wiley & Sons, Ltd. J Eval Clin Pract. 2017;23:888–894.wileyonlinelibrary.com/journal/jep

 

 

and electronic medication administration records (eMAR) will improve

the efficacy of care delivery and help meet the challenges of medica-

tion management.6,7 It is now well recognized that HIT innovations

offer many benefits through the improved management of health

information, but it should be taken into account that any new develop-

ments have the potential to introduce new errors and risks in

healthcare delivery.3,8,9 Thus, the unanticipated negative conse-

quences of such systems should be identified. Unfortunately, the

extent of HIT‐associated patient harm is difficult to quantify due to

the lack of empirical data.2

Safety is an emergent system property that needs to be addressed

throughout the lifecycle of HIT systems, including their design, con-

struction, implementation, and use.2,3 In our hospital, an eMAR appli-

cation was developed using continuous usability evaluation. Even so,

it was not possible to predict all possible interactions between the sys-

tem components during the design stage. Safety problems or hazards

tend to emerge from unexpected interactions between system compo-

nents and human users. There is a potential for unsafe interactions

when HIT systems are integrated with local clinical workflows, includ-

ing other technologies and the organizational structure itself. There-

fore patient safety should also be addressed during and after the

implementation of systems, and problems and hazards should be con-

tinuously evaluated and promptly mitigated.2,3

The aim of this study was to evaluate the impact of the eMAR

application on patient safety. A before‐and‐after study was conducted

to measure the impact of this application on the medication error rate

in medication administration recording (ME‐MAR) after the implemen-

tation of the eMAR application.

2 | METHODOLOGY

2.1 | Study design

A before‐and‐after, quasiexperimental study was conducted between

2012 and 2014 in a 947‐bed teaching hospital that implemented the

eMAR application. The primary outcome measure was the ME‐MAR

rate before and after the implementation of the eMAR application.

An ME‐MAR was defined as the omission of the medication adminis-

tration record (MAR), the wrong MAR, or a MAR lacking sufficient

information on medication administration.10

2.2 | Setting

A medical and a surgical hospitalization unit was chosen for the

study. Both hospitalization units worked with CPOE and automated

dispensing cabinets. The CPOE software Prescriwin® (Baxter®)

was provided with basic clinical decision support systems (CDSS),

such as drug allergy and drug interaction alerts and drug information

resources, and was integrated with ancillary applications in

pharmacy. Electronic Health Record Assignment

Nurse records in the preimplementation phase:

All nurse records were paper‐based. In the case of MAR, once the

electronically‐assisted prescriptions had been made, the physicians

printed the medical records in which the nurses subsequently docu-

mented the medication administration.

Nurse records in the postimplementation phase:

The nurse records were created using the electronic system

(eMAR) as well as paper records (the remaining nurse records). In the

case of MAR, once the prescriptions had been made, the nurses

directly documented subsequent medication administration in the

eMAR application.

The eMAR application was integrated within the CPOE‐CDSS and

pharmacy validation process, which allowed nurses to acknowledge

orders, document the medications administered to the patient, and to

communicate online with physicians and pharmacists. Moreover, the

eMAR application reminded nurses about medications that were due

for each patient and made the MAR visible to every team member. A

vendor (Baxter®) designed the eMAR application, which was based

on the CPOE‐CDSS application and current paper MARs and installed

on desktop computers. Thus, after the medication administration ward

round, nurses had to return to the centralized nursing station to sign

the medication administration.

The implementation of the eMAR application entailed changes in

hospital procedures and workflow. Among other aspects, the eMAR

application included justifying an omission or change of medication

administration dose, working in real time, and standardizing adminis-

tration times. Before the eMAR was implemented, and once drugs

had been prescribed, a nurse scheduled the doses to specific drug

round times and indicated the drug round at which the first dose had

to be given. After implementation, administration times were

established at the moment of the prescription and the nurses followed

the new schedule.

2.3 | Data collection

Data collection was conducted in April 2012 (pre‐) and June 2014

(post‐). The postimplementation phase started 3 months after imple-

mentation (March 2014).

Two pharmacists directly observed MAR for 14 hours per day

(8:00 am to 10:00 pm) from Monday to Friday, for 4 weeks before

eMAR implementation and afterwards. Before beginning the data col-

lection, two researchers examined a small training set (100 MAR) to

measure their interrater reliability for classifying observations as med-

ication errors (k = 0.75 (95% CI 0.59‐0.901)).11

One of the pharmacists collected data during the morning shift

and the other during the afternoon shift. The pharmacists reviewed

MAR after the medication rounds, 9 am, 12 pm and 1 pm in the morn-

ing shift and 4 pm, 6 pm, and 8 pm in the afternoon shift. Whenever an

ME‐MAR was found, the researchers asked the healthcare staff

involved to discover the cause of the error. Other data included the

hospital unit, characteristics of the patients (sex and age), date, shift,

medication, active substance, Anatomical Therapeutic Chemical (ATC)

group, dose, route, time of administration, and a detailed description

of how the error occurred and its impact on the patient. Electronic Health Record Assignment

2.4 | Classification of errors

Each ME‐MAR was classified according to 3 taxonomies: a classic ME

taxonomy in both phases and 2 technology‐induced error taxonomies

VICENTE OLIVEROS ET AL. 889

 

 

for classifying the errors after the implementation of the eMAR appli-

cation (appendix 1).

1. Classic ME taxonomy: ME‐MARs were classified according to the

Ruiz‐Jarabo Group classification, which is an adaptation of the

National Coordinating Council for Medication Error Reporting

and Prevention taxonomy to the Spanish setting.12,13 The conse-

quences of ME‐MARs were rated using the adaptation of the

potential future risk matrix for ME‐MAR previously published by

our group.14

2. Technology‐induced error taxonomy:

• Classification of problems involving information technology15: ME‐

MARs were first divided into those that mainly involved human fac-

tors or technical problems, and then assigned to 1 or more sub-

classes. Human factors were defined as problems related to

human‐HIT interactions. We examined errors in the use of software

(use errors) as well as sociotechnical contextual variables (contrib-

uting factors) that contributed to incidents (eg, training, cognitive

load, and clinical workflow). Regarding technical problems, we

examined and characterized hardware and software issues.

• Classification of clinical errors16: We next sought to examine ME‐

MARs arising from the problems based on their underlying mecha-

nisms. A clinical error was an ME‐MAR with potential conse-

quences for a patient. They were classified into: errors that were

unique to eMAR application (class A), errors more likely with eMAR

(class B), errors more likely to cause harm with eMAR (class C),

errors that did no difference (class D).

The taxonomies were adapted to ME‐MAR by a research group,

which comprised 2 researchers and 3 pharmacists with expertise in

patient safety and management.

2.5 | Data analysis

Sample‐size analysis showed that 5294 observations (half this number

in each phase) would be needed to detect a difference in the ME‐MAR

rate from 15%10 to 12%. The calculation was based on an α of 0.05

and a β of 0.2, taking into account clustering by patient and a mean

of 7 administration doses per patient and shift.

The researchers independently examined the free‐text ME‐MAR

descriptions to classify them and assess their potential risk. They com-

pared their results and in the case of disagreement, the free‐text ME‐

MAR description was examined by the research team and a consensus

category was assigned. If an ME‐MAR was assigned to more than 1

category, the primary category (the one most directly related to poten-

tial consequences) was used in the analysis. Electronic Health Record Assignment

The ME‐MAR rates were calculated and compared by determining

the number of ME‐MARs identified per number of medication doses

prescribed for the preimplementation and postimplementation groups.

The chi‐square test or Fisher’s exact test was used to compare cate-

gorical data. Generalized estimating equation analysis was conducted

to compare error rates between phases, taking into account clustering

by patient. Ordered logit modelling and multinomial logistic regression

were conducted to analyse the differences in the potential future risk

of ME‐MAR between phases, the former for overall differences and

the latter by categories. A P value of <.05 was used as a cutoff for sta-

tistical significance. It was assumed that the implementation of the

eMAR application increased patient safety if the odds ratio (OR) or rel-

ative risk (RR) were less than 1. All statistical analyses were performed

using STATA v.12 software.

2.6 | Ethics

The study was approved by the Hospital’s Clinical Investigation Ethical

Committee.

3 | RESULTS

A total of 5456 MARs were observed (2835 preimplementation and

2621 postimplementation). Table 1 shows the medications involved

in MARs and the characteristics of the patients who received them.

Significant differences were found between the 2 phases in the medi-

cations involved in MARs. Medications were compared by ATC groups

or by classes of medications (P < .001).

3.1 | Medication errors in medication administration records (ME‐MAR)

Overall, ME‐MAR rates decreased from 48.0% (1362 ME‐MARs) in the

preimplementation phase to 36.9% (967 ME‐MARs) in the

postimplementation phase (P < .05). Electronic Health Record Assignment

3.1.1 | Classic medication error taxonomy

The same types of ME‐MAR were observed, except for “MAR with

incomplete information” and wrong medication, which was only

observed in the preimplementation phase (Table 2).

The most frequent type of ME‐MAR in both phases was “MAR at

the wrong time” (31.6% vs 30.2%). A subanalysis of this type of error

showed that nurses recorded medication administration before medi-

cation was provided significantly more frequently in the

preimplementation phase than in the postimplementation phase

(11.5% vs 6.9% [OR = 0.6, P = .001]). Nevertheless, the nurses

recorded medication administration after administration less fre-

quently in the preimplementation phase than in the

postimplementation phase (20.2% vs 23.2% [OR = 1.2, P = .24]).

The main cause of ME‐MARs in both phases was failure to follow

work procedures (92% [n = 1258] vs 94% [n = 906]).

The potential future risk of ME‐MAR significantly decreased in the

postimplementation phase (OR = 0.6, P = .007). Table 3 shows the ME‐

MARs classified by potential future risk categories.

In both phases, the drugs most frequently associated with ME‐

MAR were in ATC groups: “A: alimentary” (299 [22.0%] vs 226

[23.4%]), “C: cardiovascular” (223 [16.4%] vs 194 [20.1%]), and “N:

Nervous system” (206 [19.5%] vs 155 [16.6%]).

3.1.2 | Technology‐induced error taxonomy

All ME‐MARs were use errors because of human factors (Table 4). No

technical problems were observed. The contributing factors were as

890 VICENTE OLIVEROS ET AL.

 

 

follows: failure to carry out duty (92.8%, n = 897), lapse (3.4%, n = 33),

staffing/training (3.3%, n = 32), and integration with clinical workflow

(0.5%, n = 5). In total, 1.2% (n = 12) of the ME‐MARs were only

observed in the postimplementation phase (class A), 5 of which

(48%) were due to the integration of eMAR application in the CPOE

system. Electronic Health Record Assignment

3.2 | Medical unit

MARs were not recorded in the surgical unit in the

postimplementation phase. A subanalysis was conducted for the med-

ical unit (Appendix 2). A total of 1449 MARs were observed

preimplementation and 2621 postimplementation. Significant

TABLE 1 Characteristics of medication administration records and patients before and after the implementation of the electronic medication administration record application

Characteristics Preimplementation Postimplementation

Medication administration records

Shift

Morning_ n°/total n° (%) 1588/2835 (56.0) 1735/2621 (66.2)

Afternoon_ n°/total n° (%) 1247/2835 (44.0) 886/2621 (33.8)

Classification of ATC_n°/total n° (%)

A, Alimentary tract and metabolism 697 (24.6) 662 (25.3)

B, Blood and blood‐forming organs 315 (11.1) 294 (11.2)

C, Cardiovascular system 423 (14.9) 408 (15.6)

D, Dermatologicals 22 (0.8) 27 (1.0)

G, Genito‐urinary system and sex hormones 13 (0.5) 19 (0.7)

H, Systemic hormonal preparations, excluding sex hormones and insulins

49 (1.7) 120 (4.6)

J, Antiinfectives for systemic use 253 (8.9) 161 (6.1)

L, Antineoplastic and immunomodulating agents 4 (0.1) 0

M, Musculo‐skeletal system 89 (3.1) 14 (0.5)

N, Nervous system 670 (23.6) 599 (22.9)

R, Respiratory system 285 (10.1) 271 (10.3)

S, Sensory organs 8 (0.3) 43 (1.6)

V, Various 7 (0.3) 3 (0.1)

Class of medication2

Class 1 (low‐risk medication) 698 (24.6) 693 (26.4)

Class 2 (medium‐risk medication) 1335 (47.1) 1021 (39.0)

Class 3 (high‐risk medication) 802 (28.3) 907 (34.6)

Patients

Patients (no.) 409 340

Women no./total no. (%) 214/409 (52.3) 145/340 (42.7)

Age, years (means � SD) 72.5 � 15.9 80.0 � 10.2

Abbreviations: ATC, Anatomical and therapeutic classification. 2See definitions in Appendix S1.

TABLE 2 Types of medication errors in medication administration records

Preimplementation Postimplementation Type of ME‐MAR n° of ME‐MAR (% of doses) OR (p)

Incomplete information 34 (1.2) 0

MAR at the wrong time 897 (31.6) 791 (30.2) 0.9 (0.31)

Omission 387 (13.7) 158 (6.0) 0.4 (0.00)*

Wrong dose 13 (0.5) 12 (0.5) 0.9 (0.83)

Wrong formulation 13 (0.5) 2 (0.1) 0.2 (0.03)*

Wrong medication 1 (0.0) 0

Wrong route 4 (0.1) 1 (0.0) 0.3 (0.24)

Wrong time 13 (0.5) 3 (0.1) 0.2 (0.04)*

Abbreviations: ME‐MAR, medication errors in medication administration records; OR, odds ratio.

*Significant difference (P < .05).

VICENTE OLIVEROS ET AL. 891

 

 

differences were observed between phases in the medications

involved in the MARs phases. Medications were compared by ATC

groups or by classes of medications (P < .001).

The ME‐MAR rate in the medical unit decreased from 41.0% (594

ME‐MARs) to 36.9% (P < .05). The types of ME‐MAR and causes were

similar to that observed in the overall analysis. No significant differ-

ences in potential future risk were observed between the 2 phases

(OR = 0.8, P = .06).

4 | DISCUSSION

This study evaluated the impact of the implementation of an eMAR

application on patient safety. Although some studies have evaluated

HIT implementation, as far as we know, this study is the first to isolate

the effects of an eMAR application on patient safety. This approach is

justified by the fact eMAR is frequently implemented with other tech-

nologies, such as electronic prescribing systems, and their effects mea-

sured together.7

The implementation of the eMAR application was associated with

a significant decrease in ME‐MARs. However, the percentage of ME‐

MARs were unexpected. The difference between the ME‐MAR rates

and the ones predicted by the pilot study could be explained by the dif-

ferent methodology used.10 The data collection in the pilot study was

conducted the following day of MAR. Thus, the main type of error

MAR at the wrong time (MAR before or after medication administra-

tion) was not detected. Electronic Health Record Assignment

A small decrease in ME‐MARs has been observed after the eMAR

application implementation. Some researchers have already suggested

that HIT contributes very little to the overall rate of MEs.17 In line with

other studies, we also found that the benefit of implanting an eMAR

can be hindered by employee resistance, which may reduce or prevent

the effective use of the technology18 or related work processes that

are not effectively integrated with the eMAR.19 The

postimplementation phase began 3 months after implementation;

however, Munysia et al suggested that it may take more than 1 year

to integrate the use of a new electronic documentation system into

daily work.20 Moreover, the use of HIT improves outcomes over time

and achieves a safer system. Continuous evaluation and improvement

occurs over the dynamic and iterative life cycle of HIT.2. Electronic Health Record Assignment

4.1 | Classic medication error taxonomy

Similar types of errors were detected before and after the implementa-

tion of the eMAR application. The most frequent type of ME‐MAR in

both phases was MAR at the wrong time. In the preimplementation

phase, a large number of medication administrations were recorded

before medication was provided. This behaviour represented a breach

in the organization’s documentation protocol. Thus, some workflow

blocks were intentionally incorporated in the eMAR application to pre-

vent recording before providing medication. In the

postimplementation phase, it was found that although there was a sig-

nificant decrease in MAR before administration, there was an increase

in MAR after administration. We found that the use of the eMAR appli-

cation was of assistance in changing the nurses’ behaviour regarding

documentation; however, before the workflow blocks were intro-

duced, the risk of possible workarounds to intentional blocks had to

be assessed.21,22

It is considered that some aspects of medication administration

documentation, such as the accuracy and quality of information,

improve following eMAR implementation.7,23 In contrast to paper

MAR, eMAR has been associated with easier medication documenta-

tion, and improvements in the reliability of information on medication

dose and time, patient safety, teamwork, and administering medica-

tions in a timely manner.23 Some of these findings are in line with

those of the present study, since MAR with incomplete information

was only observed in the preimplementation phase and “wrong time”

errors significantly decreased. However, no differences were observed

between the 2 phases in MAR omission. Electronic Health Record Assignment

“Wrong medication” error disappeared, but this result was not sig-

nificative. This error was difficult to detect in both phase because our

study only identified the ME‐MAR when they did not match with the

medical prescription. It would be necessary for its detection to observe

the nurse during all the medication administration process.

The main cause of ME‐MAR in both phases was failure to follow

work procedures. The standard procedures are reviewed and evalu-

ated on an ongoing basis by a hospital commission. Nevertheless,

external factors such as distractions, interruptions, time pressure,

noise, and high workload, make their compliance difficult.24-26 It is

important to highlight that the eMAR application implementation

improved accuracy and quality of MAR, but it did not decreased the

external factors mentioned above.

TABLE 3 ME‐MAR classification by potential future risk categories

Preimplementation Postimplementation Potential future risk n° of ME‐MAR (% of doses) RR (P)

Very low 27 (1.0) 3 (0.1) 0.1 (0.00)*

Low 928 (32.7) 759 (29.0) 0.7 (0.00)*

Moderate 325 (11.5) 139 (5.3) 0.4 (0.00)*

High 82 (2.9) 66 (2.5) 0.5 (0.09)*

ME‐MAR: medication errors in medication administration records; RR: Rel- ative Risk;

*Significant difference (P < .05)

TABLE 4 Technology‐induced error taxonomy

Postimplementation

Types of ME‐MAR n° of ME‐MAR (% of doses)

Classification of problems involving information technology

Wrong entry 18 (1.9)

Partial entry 1 (0.1)

Did not enter 157 (16.2)

Workaround 791 (81.8)

Classification of clinical errors

A: Unique to eMAR application 12 (1.2)

B: More likely with eMAR application 649 (67.1)

D: No difference with eMAR application 306 (31.6)

Abbreviations: ME‐MAR, medication errors in medication administration records; eMAR, electronic medication administration record. Electronic Health Record Assignment

892 VICENTE OLIVEROS ET AL.

 

 

The classic ME taxonomy allows to classify the severity of MEs

that do not reach the patient as ME‐MAR.13 Before the incorporation

of “potential future risk,” the severity of the MEs was graded according

to the actual impact on the patient. ME‐MAR do not necessarily harm

the patient, but which could create the conditions to make them more

likely to occur.24,27

Overall, there was a significant decrease in potential future risk,

which suggests that an eMAR application can improve patient safety.

However, this assumption should be taken with caution because dif-

ferent factors could have influenced the results. For example, MARs

were only reviewed in the medical unit during the postimplementation

phase and significant differences were found between phases in the

medications involved. Electronic Health Record Assignment

4.2 | Technology‐induced error taxonomy

A search of the literature failed to find any specific classification for

eMAR‐induced errors. Thus, we chose 2 HIT‐induced error taxon-

omies15,16 to analyse ME‐MAR in the postimplementation phase.

All ME‐MARs were classified as human‐machine interaction errors

according to Magrabi classification.15 This result is in complete con-

trast to the findings of a study that used this classification28 and to

those of Magrabi et al, who suggested that 92% of the errors were

due to technical problems.29 Two aspects may explain this difference:

technical problems can be reduced by designing out error‐prone fea-

tures at the software users’ interface15; the eMAR application evalu-

ated was developed using a continuous usability evaluation, which

involved different healthcare professionals. Usability evaluation is 1

way of ensuring that HITs are adapted to the users and their tasks

and that they have a usable design.30,31

Magrabi classification15 allows the introduction of new categories

to account for problems in new scenarios; thus, two new categories

were added for errors involving human factors: workaround (use error)

and lapse (contributing factor). Most of the ME‐MARs were classified

as workaround. It was found that nurses overrode safety workflow

blocks intentionally introduced in the eMAR by working around the

block to prevent recording before providing the patient with medica-

tion. Vogelsmeir et al justified such workrounds on the ground that

nurses viewed blocks as cumbersome and time‐consuming.21 The next

most frequent error was the omission error “did not enter.” Medication

administration requires a high level of concentration and distances

between the patient’s bed and the centralized nursing station can

expose nurses to interruptions. Subsequently, they may forget to sign

the medication charts.19,32 Some recommendations for mitigating both

these errors include the implementation of a device at the patient’s

bedside, such as a desktop computer with or without a bar‐code

technology, or a wireless technology coupled to portable handheld

devices. These devices would make the MAR process easier and would

provide the nurses with real‐time MAR at the bedside; consequently,

the workaround and did not enter error rate may decrease. Electronic Health Record Assignment

The main contributing factor was failure to carry out duty. The

failure to follow standard operating procedures was included in this

category. The implementation of a new eMAR application that changes

the normal workflow highlights the need to develop strategies that

support and accelerate the integration of the new documentation

practice into the nurses’ routine activities and the need to train the staff

to promote user acceptance, good usability, and proficiency.3,20,33,34

According to the clinical error classification used,16 the majority of

the ME‐MARs found were the same as those found with the use of

paper records. However, more than a half were more likely to occur

with the eMAR application and a small percentage appeared after the

implementation of the eMAR application. As mentioned, MAR after

medication administration was more likely with the eMAR application

because of the workflow blocks introduced. Moreover, MAR omission

could also be induced by nurse records being entered both electroni-

cally and on paper,35 as was the case in the postimplementation phase.

Using a single system for health records enhances patient safety and

the coordination of care and has the potential to significantly improve

information sharing across the continuum of care.3

Although the percentage of errors unique to the eMAR

applicitaion was small, it is an important point to take into account.

Almost half of the ME‐MARs which occurred with the use of eMAR

application were due to the integration of the eMAR application in

the CPOE system. Doctors prescribed incorrectly without noticing that

the eMAR application was working in real time, and there was a stan-

dardization of administration times. This incorrect prescription

affected directly to eMAR, nurses could not record medication admin-

istration. Moreover, new MAR omission appeared because nurses for-

got to check medical prescription before medication ward rounds. We

believe that these errors would disappear with a training tailored to the

needs of doctors and nurses. The knowledge and skills of users are fun-

damental to safe use of HIT.3

4.3 | Strengths and limitations

We are aware that these findings cannot be completely extrapolated

to other settings, mainly because of the particular characteristics of

our application. Nevertheless, the strengths of the study reside in its

design: the impact of the eMAR application on patient safety was eval-

uated; the study included experts skilled in the detection of medication

errors; and 3 classifications were used to classify errors.

However, a long period passed between the 2 phases, and thus, it

cannot be ensured that the ME‐MARs were only due to the introduc-

tion of the application. The time of data collection was dictated by the

implementation of the eMAR, which experienced several delays. When

the study finished, eMAR application had not been implemented yet in

the surgical unit. Thus, during the postimplementation phase, the data

were only collected in the medical unit. A subanalysis of the medical

unit was conducted to diminish any possible effect. Electronic Health Record Assignment

5 | CONCLUSION

The use of an eMAR application significantly reduces the rate of med-

ication administration recording errors and their potential risk. The

main cause of ME‐MAR was failure to follow work procedures. Thus,

new strategies should be developed to integrate the use of an eMAR

application into nurses’ daily schedule and to improve working

procedures.

VICENTE OLIVEROS ET AL. 893

 

 

ACKNOWLEDGEMENT

The authors wish to thank Dr. Alfonso Muriel García, biostatistician

from Hospital Ramón y Cajal, for his contribution in the study design

and data analysis.

REFERENCES

1. Mitchell I, Schuster A, Smith K, Pronovost P, Wu A. Patient safety inci- dent reporting: A qualitative study of thoughts and perceptions of experts 15 years after ‘To Err is Human’. BMJ Qual Saf. 2016;25 (2):92‐99.

2. Intitute of Medicine. Health IT and Patient Safety: Building Safer Systems for Better Care. Washington, D.C.: The National Academies Press; 2012.

3. Magrabi F, Ong MS, Coiera E. Health IT for patient safety and improv- ing the safety of health IT. Stud Health Technol Inform. 2016;222:25‐36.

4. Institute of Medicine. In: Aspden P, Corrigan JM, Wolcott J, et al., eds. Patient Safety: Achieving a New Standard of Care. Washington, D.C.: National Academy Press; 2004.

5. Institute of Medicine. In: Aspden P, Wolcott JA, Bootman JL, et al., eds. Preventing Medication Errors: Quality Chasm Series. Washington, D.C.: National Academy Press; 2006.

6. Blumenthal D, Glaser JP. Information technology comes to medicine. N Engl J Med. 2007;356(24):2527‐2534.

7. Oliver K, Raban M, Baysari M, Westbrook J. Evidence briefings on interventions to improve medication safety: electronic medication administration record. Aust Comm Saf Qual Health Care. 2013;1(5). https://www.safetyandquality.gov.au/publications/evidence-briefings- on-interventions-to-improve-medication-safety-electronic-medication- administration-records/. Accessed March 21, 2017.

8. Magrabi F, Aarts J, Nohr C, et al. A comparative review of patient safety initiatives for national health information technology. Int J Med Inform. 2013;82(5):e139‐e148.

9. Weiner JP, Kfuri T, Chan K, Fowles JB. “e‐Iatrogenesis”: The most critical unintended consequence of CPOE and other HIT. J Am Med Inform Assoc. 2007;14(3):387‐388; discussion 389.

10. Vicente N, Delgado E, Pérez C, et al. Errors in medication administra- tion recording in a university hospital. Eur J Clin Pharm. 2013;15 (5):365‐368.

11. Landis JR, Koch GG. The measurement of observer agreement for categorical data. Biometrics. 1977;33(1):159‐174.

12. National Coordinating Council for Medication Error Reporting and Prevention. Taxonomy of medication errors. 1998. Available from http://www.nccmerp.org/taxonomy‐medication‐errors‐now‐available (access 06/01/2017).

13. Otero Lopez MJ, Castano Rodriguez B, Perez Encinas M, et al. Updated classification for medication errors by the Ruiz‐Jarabo 2000 Group. Farm Hosp. 2008;32(1):38‐52.

14. Vicente Oliveros N, Perez Menendez‐Conde C, Gramage Caro T, et al. Potential future risk of errors in medication administration recording. J Eval Clin Pract. 2016;22(5):745‐750.

15. Magrabi F, Baker M, Sinha I, et al. Clinical safety of England’s national programme for IT: A retrospective analysis of all reported safety events 2005 to 2011. Int J Med Inform. 2015a;84(3):198‐206.

16. Magrabi F, Liaw ST, Arachi D, et al. Identifying patient safety problems associated with information technology in general practice: An analysis of incident reports. BMJ Qual Saf. 2015b.

17. Walker JM, Carayon P, Leveson N, et al. EHR safety: The way forward to safe and effective systems. J Am Med Inform Assoc. 2008;15(3):272‐277.

18. Karsh BT. Beyond usability: Designing effective technology implemen- tation systems to promote patient safety. Qual Saf Health Care. 2004;13(5):388‐394.

19. Qian S, Yu P, Hailey DM. The impact of electronic medication adminis- tration records in a residential aged care home. Int J Med Inform. 2015;84(11):966‐973.

20. Munyisia EN, Yu P, Hailey D. Caregivers’ time utilization before and after the introduction of an electronic nursing documentation system in a residential aged care facility. Methods Inf Med. 2013;52(5):403–410.

21. Vogelsmeier AA, Halbesleben JR, Scott‐Cawiezell JR. Technology implementation and workarounds in the nursing home. J Am Med Inform Assoc. 2008;15(1):114‐119.

22. Siewert B, Hochman MG. Improving safety through human factors engineering. Radiographics. 2015;35(6):1694‐1705.

23. Moreland PJ, Gallagher S, Bena JF, Morrison S, Albert NM. Nursing sat- isfaction with implementation of electronic medication administration record. Comput Inform Nurs. 2012;30(2):97‐103.

24. Drach‐Zahavy A, Somech A, Admi H, et al. How do we learn from errors? A prospective study of the link between the ward’s learning practices and medication administration errors. Int J Nurs Stud. 2014;51(3):448‐457.

25. Hughes RG, Blegen MA. Medication Administration Safety. In: Hughes RG, ed. Patient. Safety and Quality: An Evidence‐Based Handbook for Nurses; 2008.

26. Pape TM, Guerra DM, Muzquiz M, et al. Innovative approaches to reducing nurses’ distractions during medication administration. J Contin Educ Nurs. 2005;36(3):108‐116.

27. Blair W, Smith B. Nursing documentation: Frameworks and barriers. Contemp Nurse. 2012;41(2):160‐168.

28. Warm D, Edwards P. Classifying health information technology patient safety related incidents—an approach used in Wales. Appl Clin Inform. 2012;3(2):248‐257.

29. Magrabi F, Ong MS, Runciman W, Coiera E. Using FDA reports to inform a classification for health information technology safety prob- lems. J Am Med Inform Assoc. 2012;19(1):45‐53.

30. Bastien JM. Usability testing: A review of some methodological and technical aspects of the method. Int J Med Inform. 2010;79(4):e18‐e23.

31. Schoeffel R. The concept of product usability. ISO Bulletin. 2003;34:6‐7.

32. Westbrook JI, Woods A, Rob MI, Dunsmuir WT, Day RO. Association of interruptions with an increased risk and severity of medication administration errors. Arch Intern Med. 2010;170(8):683‐690.

33. Shekelle PG, Pronovost PJ, Wachter RM, et al. The top patient safety strategies that can be encouraged for adoption now. Ann Intern Med. 2013;158(5 Pt 2):365‐368.

34. McComas J, Riingen M, Chae KS. Impact of an electronic medication administration record on medication administration efficiency and errors. Comput Inform Nurs. 2014;32(12):589‐595.

35. Magrabi F, Ong MS, Runciman W, Coiera E. An analysis of computer‐ related patient safety incidents to inform the development of a classifi- cation. J Am Med Inform Assoc. 2010;17(6):663‐670.

SUPPORTING INFORMATION

Additional Supporting Information may be found online in the

supporting information tab for this article.

How to cite this article: Vicente Oliveros N, Gramage Caro T,

Pérez Menendez‐Conde C, et al. Effect of an electronic medi-

cation administration record application on patient safety. J

Eval Clin Pract. 2017;23:888–894. https://doi.org/10.1111/

jep.12753

894 VICENTE OLIVEROS ET AL.

 

 

Copyright of Journal of Evaluation in Clinical Practice is the property of Wiley-Blackwell and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder’s express written permission. However, users may print, download, or email articles for individual use.