Effect of Electronic Prescribing Strategies on Medication Error and Harm in Hospital: a Systematic Review and Meta-analysis (2024)

As a library, NLM provides access to scientific literature. Inclusion in an NLM database does not imply endorsem*nt of, or agreement with, the contents by NLM or the National Institutes of Health.
Learn more: PMC Disclaimer | PMC Copyright Notice

J Gen Intern Med. 2019 Oct; 34(10): 2210–2223.

Published online 2019 Aug 8. doi:10.1007/s11606-019-05236-8

PMCID: PMC6816608

PMID: 31396810

Nadia Roumeliotis, MDCM, MSc,^1,² Jonathan Sniderman, MDCM,¹ Thomasin Adams-Webber,³ Newton Addo, MD,⁴ Vijay Anand, MD,⁵ Paula Rochon, MD, MPH,⁶ Anna Taddio, PhD,² and Christopher Parshuram, MB.ChB, D.Phil.^1,²

Author information Article notes Copyright and License information PMC Disclaimer

Associated Data

Supplementary Materials

Abstract

Background

Computerized physician order entry and clinical decision support systems are electronic prescribing strategies that are increasingly used to improve patient safety. Previous reviews show limited effect on patient outcomes. Our objective was to assess the impact of electronic prescribing strategies on medication errors and patient harm in hospitalized patients.

Methods

MEDLINE, EMBASE, CENTRAL, and CINAHL were searched from January 2007 to January 2018. We included prospective studies that compared hospital-based electronic prescribing strategies with control, and reported on medication error or patient harm. Data were abstracted by two reviewers and pooled using random effects model. Study quality was assessed using the Effective Practice and Organisation of Care and evidence quality was assessed using Grading of Recommendations Assessment, Development, and Evaluation.

Results

Thirty-eight studies were included; comprised of 11 randomized control trials and 27 non-randomized interventional studies. Electronic prescribing strategies reduced medication errors (RR 0.24 (95% CI 0.13, 0.46), I² 98%, n = 11) and dosing errors (RR 0.17 (95% CI 0.08, 0.38), I² 96%, n = 9), with both risk ratios significantly affected by advancing year of publication. There was a significant effect of electronic prescribing strategies on adverse drug events (ADEs) (RR 0.52 (95% CI 0.40, 0.68), I² 0%, n = 2), but not on preventable ADEs (RR 0.55 (95% CI 0.30, 1.01), I² 78%, n = 3), hypoglycemia (RR 1.03 (95% CI 0.62–1.70), I² 28%, n = 7), length of stay (MD − 0.18 (95% − 1.42, 1.05), I² 94%, n = 7), or mortality (RR 0.97 (95% CI 0.79, 1.19), I² 74%, n = 9). The quality of evidence was rated very low.

Discussion

Electronic prescribing strategies decrease medication errors and adverse drug events, but had no effect on other patient outcomes. Conservative interpretations of these findings are supported by significant heterogeneity and the preponderance of low-quality studies.

Electronic supplementary material

The online version of this article (10.1007/s11606-019-05236-8) contains supplementary material, which is available to authorized users.

KEY WORDS: electronic prescribing, CPOE, CDSS, medication error, preventable adverse drug events

INTRODUCTION

Information technology has a central role in twenty-first century healthcare.^¹ Electronic medical records support the implementation of computerized physician order entry and clinical decision support systems that are increasingly used with the intent of making prescribing safer.^²,³ Computerized physician order entry enables order entry, and clinical decision support systems matches patient-specific data with a computerized knowledge base to generate patient-specific recommendations.^⁴

Over the past decade, information technology and design of computerized order entry and clinical decision support systems have evolved considerably.^⁵ Although computerized clinical decision support systems may function independently to assist in drug-related recommendations, newer systems are integrated with computerized physician order entry to aid in weight- and age-based dosing calculation, renal dosing adjustment, screening for drug-drug interactions, administration scheduling, and therapeutic monitoring.^²,⁶,⁷

Previous systematic reviews on electronic prescribing found patient outcomes were infrequently reported^{²,⁷–¹¹} and the few studies suggesting benefit of computerized order entry and clinical decision support systems on prescribing error and adverse drug events were of very low-quality,^¹²–¹⁶ with very few randomized trials.^¹³,¹⁷

Given the increased potential of newer electronic prescribing systems, and their widespread adoption, re-evaluating their effect on patient-relevant outcomes is necessary. We therefore sought to evaluate the impact of newer electronic prescribing strategies, given their lack of evidence on patient safety. The objective of this study was to assess the effect of electronic prescribing strategies on medication errors and patient outcomes.

METHODS

Study Design

We conducted a systematic review evaluating the impact of electronic prescribing strategies on medication errors and patient outcomes in hospitals. The review was conducted according to the PRISMA guidelines^¹⁸ (Supplement Table ¹) and was prospectively registered (PROSPERO No. CRD42017055663).^¹⁹

Study Eligibility

We searched for eligible full-text studies published in English from January 1, 2007, to January 1, 2018, that were randomized studies, or prospective non-randomized interventional study designs. Included studies reported on patients in hospitals, in emergency departments, and in long-term care facilities. Eligible interventions were an electronic prescribing strategy, and these were compared with a control without electronic prescribing support. An electronic prescribing strategy was defined as a computerized clinical decision support system, or a computerized physician order entry with or without an embedded clinical decision support system. Reported outcomes had to include at least one of medication error or patient harm outcome.

We excluded studies that were retrospective or ambispective; compared two electronic prescribing strategies; involved multicomponent interventions (training on error reduction, teaching, prescribing reminders, reorganization); included outpatients/ambulatory clinics; and evaluated interventions where applications did not use patient-specific data and where outcomes were limited to administrative process.

Study Outcomes

The outcomes were medication error and patient harm. We defined medication error as any error in the process of ordering, transcribing, dispensing, administering, and monitoring of medications.^¹⁵ Dosing error was evaluated as a type of medication error.

The patient outcomes included harm and potential harm to the patient. These were the following: (1) adverse drug events (ADEs) and preventable ADEs, (2) a change in patient symptomatology, (3) receipt of inappropriate therapy and time to therapy, (4) clinical effect of therapy, (5) duration of therapy, (6) length of stay, and (7) death.^¹⁵ Harmful or unintended effects of the intervention were reported in each study.

Study Search and Selection

The search was conducted in MEDLINE (Ovid), EMBASE (Ovid), Cochrane CENTRAL (Ovid), and CINAHL (EBSCO) (TA-W) in February 2018. Search terms included database subject headings and text words for the following: Clinical Decision Support System, Computerized Physician Order Entry, hospital information system, electronic prescribing, computer assisted drug therapy, cohort, and clinical trial. Further adverse event–related keyword search included the following: safety, drug error, prescription errors, dosing error, medication error, and sentinel event (Supplement Methods ¹). Additional articles were obtained by screening bibliographic references of included articles, PubMed-related articles, and related systematic reviews. Conference abstracts were not included in the study selection. All citations were imported into EndNote.^²⁰ Two reviewers (NR, JS) independently evaluated the eligibility of the studies identified in the search. Disagreements were resolved by consensus between reviewers, or by a third reviewer (CP). Where data from a trial were distributed in more than one publication, the principal publication was selected unless it was prior to inclusion date, in which case the later article was chosen.

Data Abstraction and Quality

All study data was abstracted independently by two reviewers (NR, JS). For each included study, we abstracted study characteristics, country of origin, study design, setting, patient population, characteristics of the electronic prescribing strategy, study period, and outcomes. Interventions were categorized as stand-alone Clinical Decision Support Systems (CDSS) or Computerized Physician Order Entry (CPOE). CPOE functionality was further defined as without CDSS, embedded with limited CDSS (dosing limits and allergy), or advanced CDSS (decision support for weight-based dosing, renal dosing, or drug-drug interactions).

Randomized and non-randomized studies were evaluated for risk of bias using the Effective Practice and Organisation of Care tool from the Cochrane Collaboration; grading each category as “Low,” “Unclear,” or “High” Risk of Bias.^²¹ The quality of evidence for each outcome was assessed across studies by design using the Grades of Recommendation Assessment, Development and Evaluation (GRADE), with randomized controlled trials starting at high quality and non-randomized prospective studies starting at low quality.^²² GRADE Evidence Profile and Summary of Findings tables were created with GRADEpro.^²³

We sought contact with authors if study eligibility was unclear, or to complete and clarify missing data for included studies (Fig.1).

Effect of Electronic Prescribing Strategies on Medication Error and Harm in Hospital: a Systematic Review and Meta-analysis (3)

Open in a separate window

Figure 1

PRISMA study flow diagram. CPOE, computerized physician order entry; CDSS, clinical decision support system. *Ten of 14 authors contacted to confirm study eligibility. If author was not reached, and study eligibility remained unclear, study was excluded.^†Two of 5 authors contacted clarified data for quantitative analysis.

Data Management and Analysis

Studies were described by study design, by outcome, and by study population. Cohens’ kappa was used to quantify reviewer agreement for study inclusion. Patient outcomes were summarized using descriptive measures. Due to multiple outcomes, effects were classified according to statistical significance (p < 0.05 or 95% confidence interval not including 1).

Medication errors were abstracted as number of medication errors per number of drug prescriptions; however, if errors were reported in proportions, they were converted to absolute numbers. Where conversion to absolute numbers was not possible, the study was not included in meta-analysis. ADEs or preventable ADEs were reported as the number of events per total number of patients. Where rates were provided in events per 1000 patient-days, the event rate pre- and post-intervention was multiplied by the number of patient-days, and divided by 1000. This was done when the number of admissions and patient-days was known in both intervention periods.^²⁴ In non-randomized studies, when two or more time periods were evaluated after an intervention, the last intervention period was compared with their control and included in the quantitative analysis.

Forest plots were used to illustrate the findings of a quantitative analysis when 3 or more studies reported the same outcome, and calculated the relative risk (RR) or mean difference (MD) with 95% confidence interval. A random effects model was used for all meta-analyses, subtotalled by study design where appropriate. Subtotals were not combined if outcome definition differed, or if a study with 2 intervention groups was included in a subgroup. Meta-regression by year of publication was conducted when the number of studies was sufficient. We used I² to measure heterogeneity across studies and funnel plots with Egger tests to evaluate for publication bias. Forest and funnel plots were created with Review Manager^²⁵ which were then imported into GRADEpro. Egger tests and meta-regression with bubble plots were performed in RStudio.^²⁶

RESULTS

Study Selection

The review yielded 2832 citations from which 172 full-text articles were reviewed and 34 were included. Hand search of bibliographic review led to a further 25 full-text reviews and four additional included studies, resulting in 38 included studies (Fig. (Fig.1).1). The inter-rater agreement for inclusion of studies from full-text article review was good, with Cohen’s kappa 0.67 initially and 0.93 after discussion between reviewers.

Description of Studies

The 38 included studies were from 12 countries; 33 studies reported on 51,894 patients and five studies described only the number of prescriptions, admissions, or patient-days. The hospital settings included the intensive care unit, wards, emergency department, and operating room (Table (Table11).

Table 1

Included Studies (n = 38) by Design Type and Year of Publication

Author Year Country	Electronic strategy (CDSS or CPOE) vs control	Setting	Patients included	Intervention	Outcome		Significant reduction (p < 0.05)
Author Year Country	Electronic strategy (CDSS or CPOE) vs control	Setting	Patients included	Intervention	Medication error	Patient harm	Significant reduction (p < 0.05)
Randomized control trials (N = 11)
Adult
LeMeur^²⁷ 2007 France	CDSS vs usual care	N/S	Post-renal transplant patients N = 137	Electronic concentration controlled MMF dosing		1. Treatment failure 2. Mortality at 12 mos 3. Graft loss 4. Acute rejection	1. Yes 2. No 3. No 4. Yes
Pachler^²⁸ 2008 Austria	CDSS vs usual care	ICU	MV, ICU > 3days N = 50	Automated blood glucose (eMPC) for target glucose 80–110mg/dL with continuous insulin infusion		1. Hypoglycemia	1. No
Saager^²⁹ 2008 USA	CDSS vs paper protocol	Cardiac surgery OR	Cardiac surgery with diabetes N = 40	Automated blood glucose for target glucose 90–150mg/dL with continuous insulin infusion in OR		3. Hypoglycemic events	3. No
Blaha^³⁰ 2009 Czech Republic	CDSS vs Mathis/bath insulin protocol	ICU	Elective cardiac surgery N = 120	Automated blood glucose (eMPC) for target glucose 80–110mg/dL with continuous insulin infusion		2. Hypoglycemic events	2. No
Cordingley^³¹ 2009 UK	CDSS vs paper protocol	ICU	BG > 120mg/dL and MV for 72h, N = 34	Automated blood glucose (eMPC) for target glucose 80–110mg/dL with continuous insulin infusion		2. Hypoglycemic events	3. No
Newton^³² 2010 USA	CDSS vs paper protocol	ICU	BG > 140mg/dL or > 120mg/dL N = 157	Automated blood glucose for target glucose 80–120mg/dL with continuous insulin infusion		1. Hypoglycemia 2. Length of stay 3. Mortality	1. No 2. No 3. No
Wexler^³³ 2010 USA	CDSS vs usual care	Wards	Type 2 diabetics, BG > 180mg/dL N = 128	Blood glucose control for type 2 diabetics		1. Hypoglycemic events 2. Severe hypoglycemia	1. No 2. No
Dumont^³⁴ 2012 USA	CDSS vs paper protocol	Cardiac ICU	Cardiothoracic surgery patients N = 300	Automated blood glucose for target glucose 80–150mg/dL with continuous insulin infusion		1. Hypoglycemic events	1. No
Leibovici^³⁵ 2013 Israel	CDSS vs usual antibiotic guideline	Wards	Suspected sepsis patients on antibiotics N = 1683	Advanced CDSS for empiric antibiotic treatment		1. Mortality at 180days	1. No
Mann^³⁶ 2017 USA	CDSS vs paper protocol	ICU	Thermal burn, insulin > 6days N = 22	Automated blood glucose for target glucose 80–110mg/dL with continuous insulin infusion		1. Hypoglycemia	1. No
Pediatric
Geurts^³⁷ 2016 Netherlands	CDSS vs usual care	ER	Children < 5yrs, with vomiting and/or diarrhea N = 222	CDSS for rehydration; including fluid amount and route		1. Hospitalization 2. Readmission 3. Need for IV Therapy	1. No 2. No 3. No
Non-randomized interventional studies (N = 27)
Adult
Rohrig^³⁸ 2008 Germany	CDSS vs usual care	ICU	All patients with antibiotics N = 156	Advanced CDSS for empiric antibiotic choice		1. Adequate antimicrobial coverage	1. Yes
Okon^³⁹ 2009 USA	CDSS vs no alert	N/S	Inpatients with pain > 7/10 N = 5370	Alert CDSS for pain control re-assessment		1. Pain re-assessment 2. Pain resolution 3. Naloxone administration	1. Yes 2. Yes 3. Yes
VanDoormaal^⁴⁰ 2009 Netherlands	CDSS + CPOE vs hand-written	Wards	Internal, gastro, rheum, geriatric N = 1195	New computerized order entry with limited clinical decision support	1. Medication errors 2. Dose errors 3. Therapeutic errors 4. Transcription errors 5. Administrative errors	A. ADE B. Preventable ADE C. Length of stay	1. Yes 2. Yes 3. No 4. Yes 5. Yes	A. No B. Yes C. No
Bertsche^⁴¹ 2010 Germany	CDSS vs no CDSS	ICU	Patients with ≥ 8 drugs N = 265	Advanced CDSS for drug-drug interactions		1. Drug interaction 2. ADE 3. Mortality	1. Yes 2. Yes 3. No
Roberts^⁴² 2010 Australia	CDSS vs hand-written	Wards	Elderly N = 1001	Advanced CDSS for renal dosing adjustment of renally cleared drugs, based on GFR	Dose conformity:
Roberts^⁴² 2010 Australia	CDSS vs hand-written	Wards	Elderly N = 1001		1. Enoxaparin 2. Gentamycin 3. Vancomycin		1. Yes 2. Yes 3. No
Tafelski^⁴³ 2010 Germany	CDSS vs paper guideline	ICU	All patients admitted > 36h N = 186	Decision support for empiric sepsis treatment		1. Time to antibiotics 2. Antibiotic-free days	1. No 2. Yes
Nelson^⁴⁴ 2011 USA	CDSS vs usual care	ER	All patients N = 398 alerts	Surveillance alert for early detection of sepsis/infection		1. Administration of antibiotics 2. Time to antibiotics	1. No 2. No
Schwann^⁴⁵ 2011 USA	CDSS + POCEP vs usual care	OR	Surgical procedures N = 19744	Point of care prompts for intra-operative empiric antibiotic 60min prior to incision		1. Surgical site infection	1. Yes
Blankenship^⁴⁶ 2012 USA	CPOE vs hand-written	ER	Pain in ER N = 1238	Computerized order entry for pain control		1. Time to pain control	1. No
Cartmill^⁴⁷ 2012 USA	CPOE vs hand-written	ICU	ICU-ordered antibiotics N = 289	Computerized order entry for antibiotics		1. Time to antibiotic administration	1. No
Kooij^⁴⁸ 2012 Netherlands	CDSS vs usual care	Surgical	Elective, non-cardiac patients N = 2662	Decision support for post-operative nausea and vomiting based on risk factors		1. Post-operative nausea and vomiting	1. Yes
Zoni^⁴⁹ 2012 Spain	CDSS vs nurse history	Wards	Internal medicine N = 162	Electronic order reconciliation	1. Unintended discrepancies		1. Yes
Ali^⁵⁰ 2010 UK	CPOE vs hand-written	Cardiac ICU	All N = 613	New computerized order entry without CDSS	1. Prescription omissions 2. Dosing error		1.Yes 2. Yes
Davis^⁵¹ 2014 USA	CDSS + CPOE vs hand-written	Inpatient med/surg	All Rx = 24767	New computerized order entry system, with advanced CDSS	1. Medication error 2. Dosing error		1. Yes 2. Yes
Armada^⁵² 2014 Spain	CPOE vs hand-written	Cardiac ICU	All N = 187	New computerized order entry with limited CDSS	1. Prescription error 2. Wrong dosing		1. Yes 2. Yes
Micek^⁵³ 2014 USA	CDSS vs usual care	ICU	Patients with gram-negative antibiotic N = 3616	Alert for previous recent antibiotic category use or previous resistant gram-negative organism		1. Inappropriate antibiotics 2. ICU length of stay 3. Length of stay 4. ICU mortality 5. Hospital mortality	1. Yes 2. No 3. Yes 4. No 5. Yes
Nachtigall^⁵⁴ 2014 Germany	CDSS vs paper guideline	ICU	All patients admitted > 48h N = 1316	Decision support for antibiotic guideline adherence		1. Antibiotic-free days 2. Length of stay 3. Mortality	1. Yes 2. Yes 3. No
Aziz^⁵⁵ 2015 Pakistan	CDSS + CPOE vs hand-written	Oncology	All Rx = 9279	New computerized order entry for chemotherapy protocols, with advanced CDSS	1. Prescription error 2. Dosing error		1. Yes 2. Yes
Dean^⁵⁶ 2015 USA	CDSS vs paper guideline	ER	Pneumonia in ER N = 2450	Decision support for pneumonia management in ER (based on guidelines)		1. Mortality at 30days 2. Admission rate 3. Length of Stay	1. No 2. No 3. No
Haddad^⁵⁷ 2015 Saudi Arabia	CPOE + CDSS vs hand-written	ICU	MV + infusion of sedation or analgesia N = 279	CPOE with sedation protocol, including nursing pain and sedation scores and daily wakening		1. MV duration 2. ICU length of stay 3. ICU mortality 4. Hospital mortality	1. No 2. No 3. Yes 4. Yes
Kappen^⁵⁸ 2015 Netherlands	CDSS vs usual care	Surgical	Elective surgical inpatients N = 4228	Decision support for post-operative nausea and vomiting based on risk factors		1. Post-operative nausea and vomiting	1. Yes
Han^⁵⁹ 2016 USA	CPOE + CDSS vs hand-written	ICU	All patients N = 797	New electronic health record, with advanced CDSS	1. Medication error		1. Increased errors
Han^⁵⁹ 2016 USA	CPOE + CDSS vs hand-written	ICU	All patients N = 797	New electronic health record, with advanced CDSS		A. Length of ICU stay B. Length of stay C. Mortality		A. Yes B. Yes C. Yes
Pediatric
Holdsworth^⁶⁰ 2007 USA	CPOE + CDSS vs hand-written	Ward, ICU	All patients N = 2407	New computerized order entry, with advanced CDSS for pediatrics	1. Dosing error	A. ADE B. Preventable ADE	1. No	A. Yes B. Yes
Taylor^⁶¹ 2008 USA	CDSS + CPOE vs hand-written	NICU	All NICU patients Rx = 526	New computerized order entry, with advanced CDSS for pediatrics	1. Variance rate 2. Dose error 3. Dosing time error 4. Dose omissions		1. Yes 2. No 3. Yes 4. No
Walsh*^²⁴ 2008 USA	CDSS+ CPOE vs hand-written	Wards, NICU, PICU	Random selection 627 admissions	New computerized order entry, with advanced CDSS for pediatrics	1. Medication error 2. Serious medication error 3. Non-intercepted error	A. Preventable ADE	1. No 2. No 3. No	A. No
Warrick^⁶² 2011 UK	CPOE + CDSS vs hand-written	PICU	All PICU patients N = 624	New computerized order entry, with advanced CDSS for pediatrics	1. Prescribing errors		1. No
Garner^⁶³ 2015 USA	CPOE + CDSS vs hand-written	NICU	Antibiotics for late-onset sepsis N = 79	Computerized antibiotic prescription and adjustment, with CDSS for sepsis	1. Drug error 2. Potential drug error 3. Dosing error		1. Yes 2. Yes 3. Yes

Open in a separate window

Not all study outcomes listed above; only patient harm outcomes as defined in methods

ADE, adverse drug event; BG, blood glucose; CDSS, clinical decision support system; CPOE, computerized physician order entry; eMPC, enhanced Model Predictive Control Algorithm; ER, emergency room; GFR, glomerular filtration rate; ICU, intensive care unit; ID, identification; IR, interventional radiology; IV, intravenous; med/surg, medical/surgical patients; mos, months; MV, mechanical ventilation; N, number of patients in study; NICU, neonatal intensive care unit; N/S, not specified; OR, operating room; PICU, pediatric intensive care unit; POCEP, point of care electronic prompt; Rx, number of prescriptions; SIRS, systemic inflammatory response syndrome; yrs, years

*Interrupted time series analysis (ITS)

Study exclusions: (1) retrospective (2) compare two different CDSS or CPOE systems to each other, if they involved (3) a multicomponent intervention (training on error reduction, teaching, prescribing reminders, reorganization), (4) outpatients/ambulatory clinics; (5) were applications not linked to patient-specific data, (6) evaluated administrative process, or (7) compared a CPOE/CDSS to an existing CPOE

Design

Eleven (29%) randomized controlled trials were included, all of which reported on patient outcomes and none on medication error (Table (Table1).1). The units of randomization were wards (n = 1),^³⁵ providers (n = 1),^³³ and patients (n = 9).^{²⁷–³²,³⁴,³⁶,³⁷} The 27 (71%) non-randomized interventional studies included 23 controlled before-after studies, two interrupted time series,^²⁴,⁴⁰ and two interventional cohorts.^⁵¹,⁵³

Methodological Quality Assessment

The randomized controlled trials (RCTs) included in the study had low or unclear risk of bias (Supplement Table ²). Quality assessment of the included non-randomized studies varied from high to low risk of bias. Studies were heterogeneous with regard to study quality and risk of bias for each outcome (Supplement Table ³).

Interventions

The electronic prescribing strategies included 24 (63%) stand-alone clinical decision support systems and 14 (37%) computerized physician order entry systems, of which eight had advanced decision support built within them,^{²⁴,⁵¹,⁵⁵,⁵⁹–⁶³} three had limited decision support,^{⁴⁰,⁵²,⁵⁷} two had no decision support,^{⁴⁶,⁵⁰} and one did not specify.^⁴⁷

Of the stand-alone decision support systems, nine (38%) evaluated single drug dosing adjustment for insulin (n = 8)^{²⁸–³⁴,³⁶} and mycophenolate mofetil (n = 1)^²⁷; eight (33%) involved surveillance/treatment of infection,^⁴⁴ including pneumonia management,^⁵⁶ adherence to guidelines for antibiotic therapy,^⁵⁴ empiric antibiotic choice,^{³⁵,³⁸,⁴³} empiric antibiotics for surgery,^⁴⁵ or antibiotic adjustment^⁵³; and seven (29%) were for post-operative nausea and vomiting,^{⁴⁸,⁵⁸} rehydration for children,^³⁷ dose adjustment for renally cleared drugs,^⁴² drug-drug interactions,^⁴¹ pain control,^³⁹ and medication reconciliation.^⁴⁹

The advanced decision support systems within the computerized order entry included tools for detecting drug-drug interactions,^⁵¹,⁵⁹ pediatric weight-based dosing,^{²⁴,⁶⁰–⁶³} and specialized chemotherapy ordering.^⁵⁵ Given that the majority of computerized order entry systems had a decision support system of some form built within them, they were regarded as a single category for analyses.

Outcomes Evaluated

Thirteen (34%) studies reported on medication error (0 RCTs), 29 (76%) reported on a patient harm outcome (11 RCTs), and 3 (8%) reported both (0 RCTs).^{²⁴,⁴⁰,⁵⁹} Table Table22 summarizes the studies showing improvement in the outcomes (medication error and patient harm outcomes) according to the electronic intervention type.

Table 2

Table of Interventions and Outcomes, for Studies Included in the Review (N = 38 Included Studies)

Intervention, N = 38	Studies with improvement in Outcome
Intervention, N = 38	Mediation error (N = 13)	Patient harm^a (N = 29)
CPOE, no CDSS^b, n = 3	1/1^⁵⁰	0/2^{⁴⁶,⁴⁷}
CPOE + limited CDSS, n = 3	2/2^⁴⁰,⁵²	1/2^{⁴⁰,⁵⁷}
CPOE + advanced CDSS, n = 8	4/8^{²⁴,⁵¹,⁵⁵,⁵⁹–⁶³}	2/3^{²⁴,⁵⁹,⁶⁰}
CDSS alone, n = 24	2/2^⁴²,⁴⁹	9/22^{²⁷–³¹,³³–³⁸,⁴¹,⁴³–⁴⁵,⁴⁸,⁵³,⁵⁴,⁵⁶,⁵⁸}

Open in a separate window

CPOE, computerized physician order entry; CDSS, clinical decision support system

^aOutcomes listed in Table Table11

^bOne study did not specify if CDSS embedded within CPOE (Cartmill)

Medication Errors

The definitions of medication error in the studies included incomplete prescriptions, prescription correction, dose frequency error, error due to drug-drug interactions, transcription error, and errors in dispensing, administration, and monitoring (Fig.2a). Ten of 13 (77%) studies demonstrated a reduction in overall medication error rate (0 RCTs). Two (67%) of the 3 studies not showing a difference in medication error rates were in children^²⁴,⁶² and the other showed an increased overall error rate in adult patients.^⁵⁹ Meta-analysis for the effect of electronic prescribing on medication error showed a significant reduction in overall medication errors (RR 0.24 (95% CI 0.13, 0.46), I² 98%, n = 11), with high heterogeneity (0 RCT) (Fig. (Fig.2a).2a). Meta-regression analysis by year was significant (RR 0.68 (95% CI 0.56, 0.83), n = 11), with fewer medication errors in more recent publication years (bubble plot, Supplement Figure ²A). Two studies were not included in meta-analysis due to differences in unit of intervention or unit of analysis (Fig. (Fig.2a2a).^⁵⁹,⁶³ We rated GRADE quality of evidence as very low overall for the outcome medication error, with asymmetry in the funnel plot (Egger’s test, p = 0.003) suggesting publication bias (Supplement Figure ¹A).

Effect of Electronic Prescribing Strategies on Medication Error and Harm in Hospital: a Systematic Review and Meta-analysis (4)

Open in a separate window

Figure 2

Quantitative analysis using forest plot for the effect of electronic prescribing strategies on risk of a overall medication errors and b dosing errors. a Overall medication errors. b Medication dosing errors. RCT, randomized controlled trial; NRIS, non-randomized interventional study; M-H, Mantel-Haenszel random effects model. Counts are expressed as events (errors) per total number of prescriptions, except Terrell which is events (errors) per total number of renal dosing alerts. Studies are ordered by calendar year. a Medication error definitions: prescription incomplete (Ali); unintended discrepancies (Zoni); any error in drug ordering, transcribing, dispensing, administration or monitoring (Aziz, Walsh, Van Doormal); proportion of variance between ordered and administered meds (Taylor); any error including drug name, pharmacologic form, dosing, allergy, or interaction (Armada); any pharmacy intervention for wrong dose, drug, patient, drug interaction, allergy, missing medication, or wrong dosage form (Davis); incomplete, insufficient information, illegible, error of prescribing decision or other (Warrick); dosing error (Holdsworth); dosing within 30% above or below appropriate drug dose from gentamycin, vancomycin, and enoxaparin (Roberts). Garner et al. (NRIS) not included in meta-analysis as number of errors exceeded number of prescriptions (1.1 errors/prescription in control phase). Han et al. (NRIS) not included as medication errors expressed as number of errors per 1000 patient-days. Definition unspecified. b Dosing error definitions: incomplete or wrong dose (Ali); > 10% over- or underdosing for age and weight (Garner, Taylor); gentamycin/enoxaparin/vancomycin dosing conformity with 30% of dose (Roberts); error in dosage of dosing figures (Armada); error in strength, frequency, dosage (Aziz, Davis), or length (Van Doormal).

Dosing errors were reduced in 7 (78%) of the 9 studies reporting this outcome. All of the studies reporting on dosing error were non-randomized, and compared computerized physician order entry with advanced clinical decision support systems to hand-written prescriptions. Meta-analysis demonstrated a reduction in dosing errors (RR 0.17 (95% CI 0.08, 0.38), I² 96%, n = 9) with electronic versus no electronic strategy, with very high heterogeneity (Fig. (Fig.2b).2b). Meta-regression by year found fewer errors in more recently published studies (RR 0.73 (95% CI 0.61, 0.83), n = 9) (bubble plot, Supplement Figure ²B). The reduction in dosing error occurred in adults (RR 0.11 (95% CI 0.04, 0.32), I² 97%, n = 6) but was not significant in children (RR 0.55 (95% CI 0.22, 1.39), I² 64%, n = 3) (data not shown). We rated GRADE quality of evidence for the outcome of dosing error as very low, and the funnel plot shows asymmetry (Egger’s test, p = 0.01), suggesting possible publication bias (Supplement Figure ¹B).

Patient Harm Outcomes

Twenty-nine (76%) studies reported on patient harm outcomes; comprised of ADEs or preventable ADEs (n = 4),^{²⁴,⁴⁰,⁴¹,⁶⁰} mortality (n = 9),^{²⁷,³²,³⁵,⁴¹,⁵³,⁵⁴,⁵⁶,⁵⁷,⁵⁹} length of stay (n = 7),^{³²,⁴⁰,⁵³,⁵⁴,⁵⁶,⁵⁷,⁵⁹} hypoglycemia (n = 8),^{²⁸–³⁴,³⁶} treatment failure (n = 1),^²⁷ hospitalization and readmission (n = 1),^³⁷ time to therapy (n = 2),^{⁴⁴,⁴⁷} adequate therapy (n = 1),^³⁸ pain control (n = 2),^³⁹,⁴⁶ post-operative nausea and vomiting (n = 2),^{⁴⁸,⁵⁸} and new infection (n = 1).^⁴⁵

Four studies reported adverse drug events (ADEs) or preventable ADEs (0 RCT) (Fig.3a). Three of these studies (75%) screened patients for ADEs through pharmacist/physician review^{⁴⁰,⁴¹,⁶⁰} and 1 (25%) screened with incident reporting.^²⁴ Electronic prescribing strategies were associated with reduced ADE (RR 0.52 (95% CI 0.40, 0.68), I² 0%, n = 2), but not preventable ADE (RR 0.55 (95% CI 0.30, 1.01), I² 78%, n = 3), versus no electronic strategy. For ADE, the funnel plot did not show significant asymmetry but Egger’s test was significant (p = 0.046) (Supplement Figure ¹C). GRADE quality of evidence was rated as very low.

Effect of Electronic Prescribing Strategies on Medication Error and Harm in Hospital: a Systematic Review and Meta-analysis (5)

Open in a separate window

Figure 3

Meta-analysis using forest plot for effect of electronic prescribing strategies on a adverse drug events (ADE), b mortality, c length of hospital stay, and d hypoglycemic events. a Adverse drug events (ADE) and preventable ADE. b Mortality. c Length of stay (in days). d Hypoglycemic events (all RCTs). RCT, randomized controlled trial; NRIS, non-randomized interventional trial; M-H, Mantel-Haenszel random effects model. a Counts are expressed as events (ADE) per total number of patients. Subtotals not pooled due to duplication of studies in each subgroup. b Counts are presented as deaths per number of patients in each group. Mortality is presented as follows: 30-day mortality (Dean), 180-day mortality (Leibocivi), ICU mortality (Haddad), hospital mortality (Micek, Haddad, Newton, Han), overall mortality (Bertsche, Nachtigall at 2 time points), and 12-month mortality (LeMeur). c Counts are mean (SD) length of stay in days in each group, analyzed with mean difference in each group. Studies reported the following: hospital length of stay (Van Doormal, Haddad, Newton, Micek, Dean, Han) and ICU length of stay (Nachtigall). Dean results unadjusted and originally reported as median (95%CI) due to skewness. d Counts are patients with hypoglycemic events in each group. All studies of hypoglycemic events are RCTs. Subtotals are not pooled due to study duplication in reporting mild and severe hypoglycemic events. Dumont et al. not included in meta-analysis of hypoglycemic events as data were presented in hypoglycemic events per total glycemic measurements.

Nine studies reported mortality (3 RCTs) (Fig. (Fig.3b):3b): 7 (78%) evaluated CDSS alone (3 RCTs) and 2 (22%) evaluated CPOE with advanced CDSS (0 RCTs). Overall, there was no effect of computerized prescribing strategies on mortality (RR 0.97 (95% CI 0.79, 1.19), I² 74%, n = 9). We rated GRADE quality of evidence overall as low; in the RCTs, quality was rated as high, whereas in the non-randomized studies, we rated GRADE as low with high heterogeneity (Fig. (Fig.3b).3b). The funnel plot was symmetrical (Supplement Figure ¹D) and Egger’s test was not significant.

Length of stay was reported in 7 studies (1 RCT) (Fig. (Fig.3c).3c). The forest plot of mean difference (MD) in hospital length of stay (in days) showed reduced length of stay in the one RCT (MD − 6.40 (95% CI − 13.20, 0.40)), and no significant effect in non-randomized studies (MD 0.0 (95% CI − 1.25, 1.24), I² 95%, n = 6) or overall (MD − 0.18 (95% CI − 1.42, 1.05), I² 94%, n = 7). We rated GRADE quality of evidence as low overall. The funnel plot did not show asymmetry (Supplement Figure ¹E) and Egger’s test was not significant.

Eight RCTs evaluated the effect of CDSS for glycemic control (Fig. (Fig.3d).3d). Meta-analysis did not demonstrate an effect of automated CDSS on the incidence of mild hypoglycemic episodes (< 60mg/dL) (RR 1.03 (95% CI 0.62–1.70), I² 28%, n = 4), or severe hypoglycemic episodes (< 40mg/dL) (RR 0.79 (95% CI 0.30–2.09), I² 0%, n = 6). One study was not included in meta-analysis due to a difference in units.^³⁴ We rated GRADE quality of evidence for hypoglycemia as moderate. The funnel plot did not show asymmetry (Supplement Figure ¹F) and Egger’s test was not significant for hypoglycemia.

Amongst other outcomes assessed, two studies demonstrated improvement in post-operative nausea and vomiting with electronic prescribing strategies,^{⁴⁸,⁵⁸} one study in time to pain control,^⁴⁶ and one in frequency of pain assessment and naloxone administration,^³⁹ compared with no electronic strategy.

Harm Related to Intervention

Two studies reported an increase in medication errors after electronic intervention.^²⁴,⁵⁹ Walsh et al. conducted a time series analysis after CPOE implementation in children and found a decrease in serious non-intercepted medication errors immediately after implementation, followed by a non-significant increase in the following season.^²⁴ Han et al. found a significant increase in overall medication errors after implementation of an electronic health record with CPOE (p = 0.002) in an adult intensive care unit, which was attributed to errors in delayed drug administrations.^⁵⁹

DISCUSSION

This systematic review and meta-analysis of 38 prospective interventional studies, published since 2007, found that electronic prescribing strategies reduced medication errors, dosing errors, and adverse drug events, compared with no electronic strategy. However, evidence was very low-quality and studies had high risk of bias. Preventable adverse drug events were also reduced by electronic prescribing, although this did not achieve statistical significance. Other patient outcomes including length of stay, mortality, and hypoglycemia were not significantly altered by electronic prescribing. Studies were very heterogeneous; varying in size, settings, interventions, outcomes evaluated, and methodological quality.

This review complements findings of earlier systematic reviews of computerized prescribing strategies versus control that showed improved care processes,^⁸,⁹ adherence to guidelines,^⁶⁴ and time to target physiology,^⁷ without measurable differences in patient outcomes. More recent systematic reviews have suggested that there may be some effect on patient outcomes, although these are inconsistent.^¹⁴ Nuckols et al. evaluated the effect of CPOE systems and CDSS on errors and adverse drug events in studies published before 2013, and found they reduced preventable adverse drug events, regardless of CDSS sophistication.^¹⁶ A Cochrane Review was updated to 2011 and concluded that computerized advice led to better target physiology of specific medications, decreased thromboembolic events in outpatients, tended to reduce length of hospital stay, but did not change mortality.^⁷ The heterogeneity of interventions and outcomes, and predominance of low/very low quality of evidence in our review, are concordant with previous systematic reviews, as was the lack of effect on length of stay or mortality.

In addition to the above, new findings from our meta-analysis support an optimistic view of the potential of computerized systems. We reviewed studies from the last decade with the assumption that advancing prescribing technology may have translated into improvements in patient-related outcomes that were not found in earlier systematic reviews. This assumption was supported by the finding that more recent computerized prescribing strategies have a greater impact on medication and dosing error reduction. In addition, the newer prescribing strategies included in this review had a significant impact on adverse drug events, and possible impact on preventable adverse drug events, suggesting their translation to better clinical outcomes.

The mechanism by which contemporary electronic prescribing strategies reduce medication errors, and adverse drug events is not fully understood. The factors that might contribute to increased error reduction include the following: improvements in ordering and decision support technology, improved electronic health data to which the clinical decision support rules are applied, more sophisticated implementation and widespread adoption of these technologies, or a combination of all these. The reduction in medication and dosing error appears to be related to improved dosing for renal impairment, prescription completeness, and drug-drug interactions. Irrespective of the mechanism of error reduction—now shown in individual studies and meta-analyses spanning several decades, the increased magnitude of error reduction with newer technologies may now be transferred to harm reduction. Further understanding about the contributions of these potential mechanisms of effect may help inform the development of future systems.

There are several limitations to this study. First, the computerized interventions that aid in medical prescribing remain heterogeneous, from order entry without decision support to order entry with advanced decision support. These electronic systems varied greatly in their prescribing function, clinical use, technological development, and target population. This heterogeneity contributes to caution in the interpretation of results. Other systematic reviews on electronic prescribing have also highlighted this heterogeneity, some presenting only quantitative findings without meta-analysis,^{⁸,¹⁰,¹³,¹⁵} while others have combined these heterogeneous interventions to study their effect.^⁷,¹⁶ Second, reported outcomes were diverse, ranging from prescribing errors to patient symptoms, to adverse events, ventilation days, length of stay, and mortality. While electronic strategies may improve physiologic variables and symptomatology, effects on overall outcomes of hospitalization, length of stay, and death have not yet been clearly demonstrated. In addition to the limited quantity of studies per outcome, healthcare organizations and hospitals implement, modify, and study these prescribing strategies differently, contributing to further heterogeneity. Third, the modest number of studies, over a wide variety of hospital patients and settings, limited our ability to conduct further subgroup analysis and sensitivity analysis (with removal of very low-quality studies for example). Finally, the interpretation of findings should be tempered by the limited number of randomized trials in the modern era, of which only 1 (9%) showed clinical benefit from electronic prescribing strategies, and none evaluated medication errors.

An informal review identified a small number of ongoing RCTs evaluating the effects of electronic prescribing on prescribing errors and harm outcomes, such as medication-related falls (Clinicaltrials.gov: NCT03484793, NCT00297609, NCT00818285). Further large randomized trials are needed to increase the quality of the evidence supporting this multi-billion-dollar endeavor healthcare expense.

CONCLUSION

This systematic review of prospective studies found very low-quality evidence that current era electronic prescribing strategies reduced medication errors and adverse drug events in patients, compared with no strategy, in hospitals. The available evidence was heterogeneous, largely non-randomized studies, and provides early data to justify implementation and further evaluation of computerized strategies with higher quality evidence.

Electronic Supplementary Material

ESM 1^{(2.4M, docx)}

(DOCX 2517 kb)

Acknowledgments

The authors would like to thank Marla Campbell for creation of the database for data entries, Dr. Prakesh Shah for his valuable input in synthesizing the results, and Dr. Eleanor Pullenayegum for her help with analysis in R. In addition, we thank all authors who clarified design for inclusion, and authors Dr. J Han and Dr. N Dean who provided clarification of data.

Compliance with Ethical Standards

Conflict of Interest

N Roumeliotis has doctoral financial support from “Fonds de Recherche Quebec-Santé (FRQS)” as well as from the Canadian Critical Care Trials Group (CCCTG). The remaining authors have disclosed that they do not have any conflicts of interest.

Footnotes

Publisher’s Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

References

1. Crossing the Quality Chiasm . Crossing the Quality Chasm: A New Health System for the 21st Century. Washington DC: 2001 by the National Academy of Sciences; 2001. [Google Scholar]

2. Radley DC, Wasserman MR, Olsho LE, Shoemaker SJ, Spranca MD, Bradshaw B. Reduction in medication errors in hospitals due to adoption of computerized provider order entry systems. J Am Med Inform Assoc. 2013;20(3):470–6. doi:10.1136/amiajnl-2012-001241. [PMC free article] [PubMed] [CrossRef] [Google Scholar]

3. Kruse CS, Goetz K. Summary and frequency of barriers to adoption of CPOE in the U.S. J Med Syst. 2015;39(2). 10.1007/s10916-015-0198-2. [PMC free article] [PubMed]

4. Sim I, Gorman P, Greenes RA, Haynes RB, Kaplan B, Lehmann H, et al. Clinical decision support systems for the practice of evidence-based medicine. J Am Med Inform Assoc. 2001;8(6):527–34. doi:10.1136/jamia.2001.0080527. [PMC free article] [PubMed] [CrossRef] [Google Scholar]

5. Miller Kristen, Mosby Danielle, Capan Muge, Kowalski Rebecca, Ratwani Raj, Noaiseh Yaman, Kraft Rachel, Schwartz Sanford, Weintraub William S, Arnold Ryan. Interface, information, interaction: a narrative review of design and functional requirements for clinical decision support. Journal of the American Medical Informatics Association. 2017;25(5):585–592. doi:10.1093/jamia/ocx118. [PMC free article] [PubMed] [CrossRef] [Google Scholar]

6. Kuperman GJ, Bobb A, Payne TH, Avery AJ, Gandhi TK, Burns G, et al. Medication-related clinical decision support in computerized provider order entry systems: a review. J Am Med Inform Assoc. 2007;14(1):29–40. doi:10.1197/jamia.M2170. [PMC free article] [PubMed] [CrossRef] [Google Scholar]

7. Gillaizeau F, Chan E, Trinquart L, Colombet I, Walton RT, Rege-Walther M, et al. Computerized advice on drug dosage to improve prescribing practice. Cochrane Database Syst Rev. 2013;11:CD002894. doi:10.1002/14651858.CD002894.pub3. [PubMed] [CrossRef] [Google Scholar]

8. Bright TJ, Wong A, Dhurjati R, Bristow E, Bastian L, Coeytaux RR, et al. Effect of clinical decision-support systems: a systematic review. Ann Intern Med. 2012;157(1):29–43. doi:10.7326/0003-4819-157-1-201207030-00450. [PubMed] [CrossRef] [Google Scholar]

9. Garg AX, Adhikari NK, McDonald H, Rosas-Arellano MP, Devereaux PJ, Beyene J, et al. Effects of computerized clinical decision support systems on practitioner performance and patient outcomes: a systematic review. JAMA. 2005;293(10):1223–38. doi:10.1001/jama.293.10.1223. [PubMed] [CrossRef] [Google Scholar]

10. Pearson SA, Moxey A, Robertson J, Hains I, Williamson M, Reeve J, et al. Do computerised clinical decision support systems for prescribing change practice? A systematic review of the literature (1990-2007) BMC Health Serv Res. 2009;9:154. doi:10.1186/1472-6963-9-154. [PMC free article] [PubMed] [CrossRef] [Google Scholar]

11. Eslami S, Abu-Hanna A, de Keizer NF. Evaluation of outpatient computerized physician medication order entry systems: a systematic review. J Am Med Inform Assoc. 2007;14(4):400–6. doi:10.1197/jamia.M2238. [PMC free article] [PubMed] [CrossRef] [Google Scholar]

12. Thompson G, O’Horo JC, Pickering BW, Herasevich V. Impact of the electronic medical record on mortality, length of stay, and cost in the hospital and ICU: a systematic review and metaanalysis. Crit Care Med. 2015;43(6):1276–82. doi:10.1097/ccm.0000000000000948. [PubMed] [CrossRef] [Google Scholar]

13. Wolfstadt JI, Gurwitz JH, Field TS, Lee M, Kalkar S, Wu W, et al. The effect of computerized physician order entry with clinical decision support on the rates of adverse drug events: a systematic review. J Gen Intern Med. 2008;23(4):451–8. doi:10.1007/s11606-008-0504-5. [PMC free article] [PubMed] [CrossRef] [Google Scholar]

14. Ranji SR, Rennke S, Wachter RM. Computerised provider order entry combined with clinical decision support systems to improve medication safety: a narrative review. BMJ Qual Saf. 2014;23(9):773–80. doi:10.1136/bmjqs-2013-002165. [PubMed] [CrossRef] [Google Scholar]

15. Ammenwerth E, Schnell-Inderst P, Machan C, Siebert U. The effect of electronic prescribing on medication errors and adverse drug events: a systematic review. J Am Med Inform Assoc. 2008;15(5):585–600. doi:10.1197/jamia.M2667. [PMC free article] [PubMed] [CrossRef] [Google Scholar]

16. Nuckols TK, Smith-Spangler C, Morton SC, Asch SM, Patel VM, Anderson LJ, et al. The effectiveness of computerized order entry at reducing preventable adverse drug events and medication errors in hospital settings: a systematic review and meta-analysis. Syst Rev. 2014;3:56. doi:10.1186/2046-4053-3-56. [PMC free article] [PubMed] [CrossRef] [Google Scholar]

17. Reckmann MH, Westbrook JI, Koh Y, Lo C, Day RO. Does computerized provider order entry reduce prescribing errors for hospital inpatients? A systematic review. J Am Med Inform Assoc. 2009;16(5):613–23. doi:10.1197/jamia.M3050. [PMC free article] [PubMed] [CrossRef] [Google Scholar]

18. Moher D, Liberati A, Tetzlaff J, Altman DG. Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement. PLoS Med. 2009;6(7):e1000097. doi:10.1371/journal.pmed.1000097. [PMC free article] [PubMed] [CrossRef] [Google Scholar]

19. PROSPERO. In: International prospective register of systematic reviews. University of York, York. www.crd.york.ac.uk/PROSPERO/display_record.asp?ID=CRD42017055663.

20. Thompson Reuters. EndNote X7.1. USA2014.

21. (EPOC) CEPaOoC. Suggested risk of bias criteria for EPOC reviews. 2017.

22. Guyatt G, Oxman AD, Akl EA, Kunz R, Vist G, Brozek J, et al. GRADE guidelines: 1. Introduction-GRADE evidence profiles and summary of findings tables. J Clin Epidemiol. 2011;64(4):383–94. doi:10.1016/j.jclinepi.2010.04.026. [PubMed] [CrossRef] [Google Scholar]

23. GRADEpro. McMaster University: GDT:GRADEpro Guideline Development Tool.; 2015.

24. Walsh KE, Landrigan CP, Adams WG, Vinci RJ, Chessare JB, Cooper MR, et al. Effect of computer order entry on prevention of serious medication errors in hospitalized children. Pediatrics. 2008;121(3):e421–7. doi:10.1542/peds.2007-0220. [PubMed] [CrossRef] [Google Scholar]

25. Review Manager (RevMan) [computer program] Version 5.3 [database on the Internet]. Copenhagen:The Nordic Cochrane Centre, The Cochrane Collaboration 2014.

26. RStudio . R Studio Version 1.1.383. Boston: RStudio, Inc.; 2009. [Google Scholar]

27. Le Meur Y, Buchler M, Thierry A, Caillard S, Villemain F, Lavaud S, et al. Individualized mycophenolate mofetil dosing based on drug exposure significantly improves patient outcomes after renal transplantation. Am J Transplant. 2007;7(11):2496–503. doi:10.1111/j.1600-6143.2007.01983.x. [PubMed] [CrossRef] [Google Scholar]

28. Pachler C, Plank J, Weinhandl H, Chassin LJ, Wilinska ME, Kulnik R, et al. Tight glycaemic control by an automated algorithm with time-variant sampling in medical ICU patients. Intensive Care Med. 2008;34(7):1224–30. doi:10.1007/s00134-008-1033-8. [PubMed] [CrossRef] [Google Scholar]

29. Saager L, Collins GL, Burnside B, Tymkew H, Zhang L, Jacobsohn E, et al. A randomized study in diabetic patients undergoing cardiac surgery comparing computer-guided glucose management with a standard sliding scale protocol. J Cardiothorac Vasc Anesth. 2008;22(3):377–82. doi:10.1053/j.jvca.2007.09.013. [PubMed] [CrossRef] [Google Scholar]

30. Blaha J, Kopecky P, Matias M, Hovorka R, Kunstyr J, Kotulak T, et al. Comparison of three protocols for tight glycemic control in cardiac surgery patients. Diabetes Care. 2009;32(5):757–61. doi:10.2337/dc08-1851. [PMC free article] [PubMed] [CrossRef] [Google Scholar]

31. Cordingley JJ, Vlasselaers D, Dormand NC, Wouters PJ, Squire SD, Chassin LJ, et al. Intensive insulin therapy: enhanced Model Predictive Control algorithm versus standard care. Intensive Care Med. 2009;35(1):123–8. doi:10.1007/s00134-008-1236-z. [PubMed] [CrossRef] [Google Scholar]

32. Newton CA, Smiley D, Bode BW, Kitabchi AE, Davidson PC, Jacobs S, et al. A comparison study of continuous insulin infusion protocols in the medical intensive care unit: computer-guided vs. standard column-based algorithms. J Hosp Med. 2010;5(8):432–7. doi:10.1002/jhm.816. [PMC free article] [PubMed] [CrossRef] [Google Scholar]

33. Wexler DJ, Shrader P, Burns SM, Cagliero E. Effectiveness of a computerized insulin order template in general medical inpatients with type 2 diabetes: a cluster randomized trial. Diabetes Care. 2010;33(10):2181–3. doi:10.2337/dc10-0964. [PMC free article] [PubMed] [CrossRef] [Google Scholar]

34. Dumont C, Bourguignon C. Effect of a computerized insulin dose calculator on the process of glycemic control. Am J Crit Care. 2012;21(2):106–15. doi:10.4037/ajcc2012956. [PubMed] [CrossRef] [Google Scholar]

35. Leibovici L, Kariv G, Paul M. Long-term survival in patients included in a randomized controlled trial of treat, a decision support system for antibiotic treatment. J Antimicrob Chemother. 2013;68(11):2664–6. doi:10.1093/jac/dkt222. [PubMed] [CrossRef] [Google Scholar]

36. Mann EA, Jones JA, Wolf SE, Wade CE. Computer decision support software safely improves glycemic control in the burn intensive care unit: A randomized controlled clinical study. J Burn Care Res. 2011;32(2):246–55. doi:10.1097/BCR.0b013e31820aaebf. [PMC free article] [PubMed] [CrossRef] [Google Scholar]

37. Geurts D, de Vos-Kerkhof E, Polinder S, Steyerberg E, van der Lei J, Moll H, et al. Implementation of clinical decision support in young children with acute gastroenteritis: a randomized controlled trial at the emergency department. 2017;1. [PMC free article] [PubMed]

38. Rohrig R, Niczko EJ, Beutefuhr H, Bottger S, Klasen J, Fussle R, et al. Examination of computer assisted prescribing of an initial calculated antibiotic treatment. Stud Health Technol Inform. 2008;136:63–8. [PubMed] [Google Scholar]

39. Okon TR, Lutz PS, Liang H. Improved pain resolution in hospitalized patients through targeting of pain mismanagement as medical error. J Pain Symptom Manag. 2009;37(6):1039–49. doi:10.1016/j.jpainsymman.2008.06.009. [PubMed] [CrossRef] [Google Scholar]

40. van Doormaal JE, van den Bemt PM, Zaal RJ, Egberts AC, Lenderink BW, Kosterink JG, et al. The influence that electronic prescribing has on medication errors and preventable adverse drug events: an interrupted time-series study. J Am Med Inform Assoc. 2009;16(6):816–25. doi:10.1197/jamia.M3099. [PMC free article] [PubMed] [CrossRef] [Google Scholar]

41. Bertsche T, Pfaff J, Schiller P, Kaltschmidt J, Pruszydlo MG, Stremmel W, et al. Prevention of adverse drug reactions in intensive care patients by personal intervention based on an electronic clinical decision support system. Intensive Care Med. 2010;36(4):665–72. doi:10.1007/s00134-010-1778-8. [PubMed] [CrossRef] [Google Scholar]

42. Roberts GW, Farmer CJ, Cheney PC, Govis SM, Belcher TW, Walsh SA, et al. Clinical decision support implemented with academic detailing improves prescribing of key renally cleared drugs in the hospital setting. J Am Med Inform Assoc. 2010;17(3):308–12. doi:10.1136/jamia.2009.001537. [PMC free article] [PubMed] [CrossRef] [Google Scholar]

43. Tafelski S, Nachtigall I, Deja M, Tamarkin A, Trefzer T, Halle E, et al. Computer-assisted decision support for changing practice in severe sepsis and septic shock. J Int Med Res. 2010;38(5):1605–16. doi:10.1177/147323001003800505. [PubMed] [CrossRef] [Google Scholar]

44. Nelson JL, Smith BL, Jared JD, Younger JG. Prospective trial of real-time electronic surveillance to expedite early care of severe sepsis. Ann Emerg Med. 2011;57(5):500–4. doi:10.1016/j.annemergmed.2010.12.008. [PubMed] [CrossRef] [Google Scholar]

45. Schwann NM, Bretz KA, Eid S, Burger T, Fry D, Ackler F, et al. Point-of-care electronic prompts: an effective means of increasing compliance, demonstrating quality, and improving outcome. Anesth Analg. 2011;113(4):869–76. doi:10.1213/ANE.0b013e318227b511. [PubMed] [CrossRef] [Google Scholar]

46. Blankenship JF, Rogers L, White J, Carey A, Fosnocht D, Hopkins C, et al. Prospective evaluation of the treatment of pain in the ED using computerized physician order entry. Am J Emerg Med. 2012;30(8):1613–6. doi:10.1016/j.ajem.2011.11.003. [PubMed] [CrossRef] [Google Scholar]

47. Cartmill RS, Walker JM, Blosky MA, Brown RL, Djurkovic S, Dunham DB, et al. Impact of electronic order management on the timeliness of antibiotic administration in critical care patients. Int J Med Inform. 2012;81(11):782–91. doi:10.1016/j.ijmedinf.2012.07.011. [PubMed] [CrossRef] [Google Scholar]

48. Kooij FO, Vos N, Siebenga P, Klok T, Hollmann MW, Kal JE. Automated reminders decrease postoperative nausea and vomiting incidence in a general surgical population. Br J Anaesth. 2012;108(6):961–5. doi:10.1093/bja/aes024. [PubMed] [CrossRef] [Google Scholar]

49. Zoni AC, Duran Garcia ME, Jimenez Munoz AB, Salomon Perez R, Martin P, Herranz AA. The impact of medication reconciliation program at admission in an internal medicine department. Eur J Intern Med. 2012;23(8):696–700. doi:10.1016/j.ejim.2012.08.013. [PubMed] [CrossRef] [Google Scholar]

50. Ali J, Barrow L, Vuylsteke A. The impact of computerised physician order entry on prescribing practices in a cardiothoracic intensive care unit. Anaesthesia. 2010;65(2):119–23. doi:10.1111/j.1365-2044.2009.06134.x. [PubMed] [CrossRef] [Google Scholar]

51. Davis L, Brunetti L, Lee EK, Yoon N, Cho SH, Suh DC. Effects of computerized physician order entry on medication turnaround time and orders requiring pharmacist intervention. Res Soc Adm Pharm. 2014;10(5):756–67. doi:10.1016/j.sapharm.2013.11.004. [PubMed] [CrossRef] [Google Scholar]

52. Armada ER, Villamanan E, Lopez-de-Sa E, Rosillo S, Rey-Blas JR, Testillano ML, et al. Computerized physician order entry in the cardiac intensive care unit: effects on prescription errors and workflow conditions. J Crit Care. 2014;29(2):188–93. doi:10.1016/j.jcrc.2013.10.016. [PubMed] [CrossRef] [Google Scholar]

53. Micek ST, Heard KM, Gowan M, Kollef MH. Identifying critically ill patients at risk for inappropriate antibiotic therapy: a pilot study of a point-of-care decision support alert. Crit Care Med. 2014;42(8):1832–8. doi:10.1097/CCM.0000000000000337. [PubMed] [CrossRef] [Google Scholar]

54. Nachtigall I, Tafelski S, Deja M, Halle E, Grebe MC, Tamarkin A, et al. Long-term effect of computer-assisted decision support for antibiotic treatment in critically ill patients: A prospective ‘before/after’ cohort study. BMJ Open. 2014;4(12):(no pagination)(e005370). doi:10.1136/bmjopen-2014-005370. [PMC free article] [PubMed] [CrossRef] [Google Scholar]

55. Aziz MT, Ur-Rehman T, Qureshi S, Bukhari NI. Reduction in chemotherapy order errors with computerised physician order entry and clinical decision support systems. Health Inf Manage J. 2015;44(3):13–22. [PubMed] [Google Scholar]

56. Dean NC, Jones BE, Jones JP, Ferraro JP, Post HB, Aronsky D, et al. Impact of an electronic clinical decision support tool for emergency department patients with pneumonia. Ann Emerg Med. 2015;66(5):511–20. doi:10.1016/j.annemergmed.2015.02.003. [PubMed] [CrossRef] [Google Scholar]

57. Haddad SH, Gonzales CB, Deeb AM, Tamim HM, AlDawood AS, Al Babtain I, et al. Computerized physician order entry of a sedation protocol is not associated with improved sedation practice or outcomes in critically ill patients. BMC Anesthesiol. 2015;15:177. doi:10.1186/s12871-015-0161-2. [PMC free article] [PubMed] [CrossRef] [Google Scholar]

58. Kappen TH, Vergouwe Y, VanWolfswinkel L, Kalkman CJ, Moons KGM, Van Klei WA. Impact of adding therapeutic recommendations to risk assessments from a prediction model for postoperative nausea and vomiting. Br J Anaesth. 2015;114(2):252–60. doi:10.1093/bja/aeu321. [PubMed] [CrossRef] [Google Scholar]

59. Han JE, Rabinovich M, Abraham P, Satyanarayana P, Liao TV, Udoji TN, et al. Effect of electronic health record implementation in critical care on survival and medication errors. Am J Med Sci. 2016;351(6):576–81. doi:10.1016/j.amjms.2016.01.026. [PubMed] [CrossRef] [Google Scholar]

60. Holdsworth MT, Fichtl RE, Raisch DW, Hewryk A, Behta M, Mendez-Rico E, et al. Impact of computerized prescriber order entry on the incidence of adverse drug events in pediatric inpatients. Pediatrics. 2007;120(5):1058–66. doi:10.1542/peds.2006-3160. [PubMed] [CrossRef] [Google Scholar]

61. Taylor JA, Loan LA, Kamara J, Blackburn S, Whitney D. Medication administration variances before and after implementation of computerized physician order entry in a neonatal intensive care unit. Pediatrics. 2008;121(1):123–8. doi:10.1542/peds.2007-0919. [PubMed] [CrossRef] [Google Scholar]

62. Warrick C, Naik H, Avis S, Fletcher P, Franklin BD, Inwald D. A clinical information system reduces medication errors in paediatric intensive care. Intensive Care Med. 2011;37(4):691–4. doi:10.1007/s00134-010-2126-8. [PubMed] [CrossRef] [Google Scholar]

63. Garner SS, Cox TH, Hill EG, Irving MG, Bissinger RL, Annibale DJ. Prospective, controlled study of an intervention to reduce errors in neonatal antibiotic orders. J Perinatol. 2015;35(8):631–5. doi:10.1038/jp.2015.20. [PubMed] [CrossRef] [Google Scholar]

64. Eslami S, de Keizer NF, Abu-Hanna A. The impact of computerized physician medication order entry in hospitalized patients--a systematic review. Int J Med Inf. 2008;77(6):365–76. doi:10.1016/j.ijmedinf.2007.10.001. [PubMed] [CrossRef] [Google Scholar]

Articles from Journal of General Internal Medicine are provided here courtesy of Society of General Internal Medicine

Effect of Electronic Prescribing Strategies on Medication Error and Harm in Hospital: a Systematic Review and Meta-analysis (2024)

FAQs

What is the effect of e-prescribing implementation on reducing medication error in hospital? ›

Electronic prescribing significantly reduced the rate of medication errors. The electronic system was unable to decrease all the error types. More trusted systems with decision support might be needed for better results. Pharmacy workflow and patient outcome were enhanced by the electronic system.

Read On ›

Has electronic prescribing reduce medication errors? ›

Electronic prescribing strategies decrease medication errors and adverse drug events, but had no effect on other patient outcomes. Conservative interpretations of these findings are supported by significant heterogeneity and the preponderance of low-quality studies.

Discover More Details ›

What is the impact of electronic prescribing? ›

The e-prescribing system encourages prescribers to choose medications with lower copayments by using point-of-prescribing FDS. We found that clinicians who used the e-prescribing system prescribed a higher proportion of tier 1 medications.

What are the problems with ePMA? ›

The ePMA system also had limitations, the most important of which were its rigidity and the difficulty of introducing changes into it. For example, the system would not enable the printing of FP10 prescriptions for the dispensing of prescriptions outside of the hospital.

See Details ›

Does e-prescribing reduce errors and increase efficiency? ›

E-prescribing systems can also provide healthcare professionals with up-to-date information on medication interactions, allergies, and dosing guidelines, reducing the risk of medication errors and improving patient safety.

Find Out More ›

What are prescribing errors with electronic prescribing? ›

Compared with other studies on handwritten prescribing [29, 30], our study on prescribing errors in e-prescriptions demonstrated a lower error rate of 0.34% (1.9 per 1000 orders). These results were similar to the review research by Weingart SN et al.

Tell Me More ›

What is the effect of electronic prescribing on medication errors and adverse drug events a systematic review? ›

This systematic review and meta-analysis of 38 prospective interventional studies, published since 2007, found that electronic prescribing strategies reduced medication errors, dosing errors, and adverse drug events, compared with no electronic strategy.

Show Me More ›

What is a disadvantage of electronic prescribing? ›

As a result, e-prescribing may produce problems that can lead to patient harm, such as improper patient selection from electronic lists, alert fatigue, improper or difficult product selection, and interfacing challenges between prescribers and pharmacies. Read the full article, including suggestions for improvement.

Explore More ›

How does electronic prescribing affect patient safety? ›

Electronic prescribing (EP) in the hospital setting has the potential to improve safety through reduction of errors and adverse drug events. Evidence for the effects of EP on workflow and timesaving is mixed. Unintended consequences of the computerisation of prescribing are well documented.

What are the advantages and disadvantages of electronic prescribing? ›

Some of the most notable disadvantages are introduction of prescription errors, poor design features of e-prescribing software, and disruptions in pharmacy workflow. E-prescribing can potentially yield cost savings and improve efficiency and patient safety.

Show Me More ›

What are the 5 advantages associated with electronic prescribing? ›

Benefits of E-Prescribing

2) Prevent prescription drug errors. ...
3) Easily prescribe controlled substances. ...
4) Monitor controlled substance prescriptions. ...
5) Reconcile medication history quickly. ...
6) Meet meaningful use requirements. ...
7) Easily track prescription fulfillment. ...
8) Reduce lost prescriptions.

More items...

Read The Full Story ›

Why is e-prescribing important in healthcare? ›

E-prescribing enables a prescriber to electronically send an accurate, error-free and understandable prescription directly to a pharmacy from the point-of-care and is an important element in improving the quality of patient care.

See Details ›

What is EPMA analysis? ›

Electron probe microanalysis (EPMA) is a technique used to determine the chemical composition of materials at the micrometer scale. The instrument (“electron-probe,” or often simply known as the “probe”) is a variant of a scanning electron microscope (SEM).

Get More Info Here ›

What are the basic principles of EPMA analysis? ›

The electron probe microanalysis (EPMA) or “microprobe” technique is based on bombarding the sample with an electron beam. The sample then emits X-rays at wavelengths characteristic to the elements being analyzed.

How does EPMA improve patient safety? ›

For 18.4% of the low-harm incidents (n=59) and 20.3% (n=13) of the moderate-harm incidents, EPMA could reduce the likelihood of the incident occurring without configuration. Medication errors most likely to be reduced by EPMA were due to illegibility, multiple drug charts or missing drug charts.

How does e-prescribing help improve medication safety? ›

Illegible prescriptions can lead to confusion among pharmacists, resulting in dispensing errors. E-prescribing ensures that prescriptions are typed and clearly legible, minimizing the chances of misinterpretation and reducing medication errors caused by handwriting issues.

View Details ›

What are the effects of prescribing errors? ›

Prescription errors account for 70% of medication errors that could potentially result in adverse effects. A mean value of prescribing errors with the potential for adverse effects in patients of about 4 in 1000 prescriptions was recorded in a teaching hospital.