Sage Journals: Discover world-class research

Abstract

Background:

Closed-loop electronic medication management systems are effective measures for preventing medication errors (MEs). However, there is limited evidence supporting this, and few studies have evaluated the long-term effects of these systems on safe medication.

Objective:

To evaluate the long-term effects of implementing a closed-loop medication order executive system on the safe clinical use of medications.

Design:

A quasi-experimental design.

Method:

Data from 2017 to 2023 were extracted and retrospectively analyzed. The primary outcome indicator was the ME rate. Secondary outcome indicators were the accuracy of order verification and patient identification and the implementation rate of fresh medicine dispensing. The autoregressive integrated moving average (ARIMA) model in time-series analysis was used to evaluate the level and trend changes in ME rates using SPSS 25.0 before and after system implementation. Root cause analysis and descriptive statistics were used to assess changes in types, stages, and causes of ME rates. The independent samples t-test was used to analyze secondary outcomes.

Results:

Overall, 295 MEs were reported with a mean of 0.26 ± 0.26 ME rates per month during 2017–2023. The ARIMA model showed a decrease in the average level of ME rates after system implementation, with no statistically significant decrease in the long term, and a significant drop in the ME rate in the short and medium term (p < 0.01). Nurses’ administration accounted for more than 60% of errors post-implementation, and lack of communication was a prominent factor. The accuracy of order verification and patient identification and the implementation rate of fresh medicine dispensing all increased after implementation.

Conclusion:

Adopting a closed-loop executive system is beneficial for ensuring patient medication safety, but a single integrated system does not completely eliminate MEs. Optimizing system functionalities and applying structured handoff tools are recommended to meet clinical needs and enhance system usability.

Plain language summary

The study used some quality management indicators to evaluate the effect of a medication order executive system on the safety of medication administration

Why was the study done? Medication errors (MEs) may cause harm to patients. A medication order execution system can effectively prevent MEs, but its long-term effects have not been fully identified. What did the researchers do? The research team collected some quality management indicators at the Jiangsu Province Hospital, such as ME rates and accuracy of patient identification, from pre-implementation to post-implementation in the short, medium, and long term. What did the researchers find? The use of the system significantly reduced ME rates, especially in the short term. The accuracy and timeliness of execution of medication orders were improved. What do the findings mean? The system can ensure the safety of medication administration to a certain extent, but the prevention and control of human risks are limited. The system can still be continuously improved by optimizing the functions of the system and using tools to further reduce the occurrence of adverse events.

Keywords

closed-loop electronic medication management system medication error patient safety

Introduction

Drug therapy is the most commonly used and basic clinical treatment. Execution of medication orders directly affects the progress of patients’ conditions. Nurses, who are the final executors and gatekeepers of medication orders, are susceptible to medication errors (MEs) owing to the complicated medication administration process and heavy workload, as well as disruptions in human and system-related factors.¹ In the United Kingdom, 237 million MEs occur each year, with more than half occurring at the administration stage.² More than a quarter of MEs are considered serious or fatal,³ directly affecting patient outcomes and leading to prolonged hospital stay, increased financial burdens, and even patient disability or death. Accordingly, medication safety has become a challenge to global patient safety.³

At present, computerized physician order entry (CPOE), electronic medication administration records (eMAR), and other information technology tools are recognized as effective measures for enhancing medication safety.^4,5 However, some studies report that long-term use of technologies such as CPOE has not reduced MEs, but rather caused new problems,^6,7 further threatening patient medication safety. A closed-loop medication order execution system, hereafter referred to as a closed-loop execution system, integrates CPOE, barcode medication management, and eMAR technologies, forming a closed-loop link at each stage of medical order entry, prescription verification, and administration, and further monitors the whole process of medical orders from ordering to execution in real time.⁸ However, there is limited evidence on the effect of a closed-loop execution system in preventing MEs.^8–10

The previous study has described the design of closed-loop electronic medication management systems (EMMSs) and its impact on medication safety in the U.S. and Finnish hospitals in terms of system functionalities and medication processes.¹⁰ However, few studies have quantitatively assessed the long-term effect of EMMS application on medication safety.¹¹ Furthermore, data from these studies may not apply to all countries and information systems. Lindén-Lahti⁹ has found that the system still needs to be improved with regard to medication ordering and prescribing. Therefore, this study aims to investigate the long-term impact of local closed-loop execution system implementation on medication safety.

Methods

The reporting of this study followed the framework for enhanced reporting of interrupted time-series studies statement¹² (Supplemental Material 1).

Study setting

Jiangsu Province Hospital (JPH) is a tertiary hospital in Nanjing, China, with more than 4500 open beds, 6,226,800 annual outpatient and emergency visits, 210,800 discharges, and an average hospital length of stay of 6–7 days. The number of inpatients per month is about 18,000 and continues to rise as JPH expands. The hospital applied the closed-loop execution system including a hospital information system (HIS), automatic dispensing machine, personal digital assistant (PDA), and mobile nursing trolley, in January 2019. With the integration of an electronic medical record (EMR) system, a laboratory information system, and a bedside intelligent interactive system, among other systems, the closed-loop execution system conducts the clinical diagnosis and treatment care work of the whole hospital. This integration achieves core business integration of healthcare, clinical processes, and charging. The study was conducted in the hospital’s inpatient department, which is equipped with an intravenous configuration center. The department piloted the centralized configuration of long-term intravenous medication, and oral medication was packaged using a pill-packing machine.

All medications prescribed during a patient’s hospitalization were included, excluding medicines from the outpatient department, emergency department, and home. Given the lack of clear start or end time on “doctor’s orders” that allow use of medications as needed—Latin pro re nata (PRN) and failure in the entry of verbal orders, these orders were excluded. Table 1 describes the inclusion and exclusion criteria for data collection.

Table 1.

Inclusion and exclusion criteria.

Inclusion criteria	Exclusion criteria
Medications prescribed during hospitalization	1. Narcotic drugs
	2. PRN medications
	3. Verbal and written orders
	4. Orders prioritised first as needed urgently—Latin STATim (STAT)
	5. Medication orders placed within the perioperative setting
	6. Medicines from outpatient, emergency, and home

System design and workflow

A quasi-experimental design was used to assess the effect of the closed-loop execution system on safe medications for patients. Training for the use of the closed-loop execution system was provided to all nurses and other healthcare professionals (as required) before implementation. The system was piloted in some wards from January to February 2019. Table 2 shows the workflow before and after system implementation. More workflow details are found in Supplemental Material 2.

Table 2.

Workflow before and after the implementation of a closed-loop execution system.

Workflow before implementation	Workflow after implementation
1. Medical advice entry
Doctors log into HIS to enter medical orders directly	Doctors input their work numbers and passwords to log into HIS and enter medical orders, and the system automatically blocks allergy medications and overdoses
2. Check medical orders
The nurse checks and confirms medical orders	The nurse checks medical orders; if they are correct, the nurse confirms them in batch, and if they are wrong, the nurse returns them
3. Pharmacist audit
Pharmacists manually review the rational use of medications and call doctors to modify the order if an error is found	Pharmacists review the irrational prescriptions screened by the system and the parts that cannot be reviewed by the system and send a message to doctors on the system
4. Distribution
Distribution personnel delivers medications to the ward and the nurse checks the medication based on the medical order execution sheet	After delivery to the ward, the nurse and distribution personnel scan the label or barcode of the sheet for handover
5. Dispensing
Intravenous medications are dispensed by two nurses after checking; oral medications are checked by one nurse based on medical order execution sheets	Medication dispensing is performed in the order of setting, checking, and dispensing; two nurses scan high-alert medications with at least one nurse having the qualification for configuration
6. Administration
The nurse scans the medication bar-code and patient’s wristband for identity verification, and clicks “Execution” after confirming that there is no error; oral medications are crosschecked by medical orders, and recorded on the sheet after execution	All medications are verified by scanning labels and wristband; if the system sets an alarm, the nurse must confirm related information and re-scan; otherwise, the order cannot be executed; execution information is recorded in real time; oral medication is given by scanning twice to confirm that the patient has taken the medicine
7. Medication inspection
Time is set to remind nurses to inspect; nurses manually record information in a sheet after inspection	For special medications (e.g., high-alert medications), the system automatically reminds nurses to carry out inspections and records information by scanning barcodes
8. Medication education
Nurses memorize drug effects and adverse effects and give verbal instructions	Nurses can view all tests and medication information on the system

HIS, hospital information system.

Data collection and analysis

The primary and secondary outcome indicators were extracted from the adverse drug event reporting system and quality management platform for retrospective audits between January 2017 and December 2023. The pre-and post-implementation of the closed-loop execution system was categorized into four phases: pre-implementation (January 2017 to December 2018); post-implementation short term (January 2019 to December 2019), mid-term (January 2020 to December 2021), and long term (January 2022 to December 2023).

In general, the likelihood of MEs increased with the number of patients on medication. Therefore, ME rates were used to replace MEs to reduce the impact of patient numbers. The monthly ME rate was calculated as the total number of MEs/total number of patients on medication × 1000‰. To assess the impact of the system implementation, a time-series analysis (TS) was conducted using the autoregressive integrated moving average (ARIMA) model. The month of the year was used as the independent variable and the monthly ME rate as the dependent variable. Given that ME rates showed a declining trend over time, which was a non-stationary series, the data were preprocessed at a difference of 1 and the series was then stationary after preliminary adjustment. Autocorrelation and seasonality of the preprocessed data were assessed using Ljung Box Q fit statistic and visual inspection of autocorrelation (ACF) and partial autocorrelation (PACF) plots. Ljung Box Q fit statistic and Visual inspection of ACF and PACF plots did not indicate the presence of autocorrelation. Inspection of ACF plots for periodicity or cyclical fluctuations at 12 lags indicated no obvious seasonality. All analyses were performed in the SPSS 25.0 forecasting module and p < 0.05 indicated statistical significance.

A root cause analysis group was set up, comprising one nursing manager, two quality management members, three clinical nurses, and two researchers, to identify all possible causes of ME using brainstorming and fishbone diagrams, which were then classified to determine the root causes of ME. Finally, the root causes were divided into four categories: (1) system failures or defects, (2) personal factors (e.g., personal negligence and/or lack of knowledge and skills), (3) lack of communication (e.g., healthcare communication, medication handover, and/or nurse-patient communication), and (4) workflow problems. MEs reported before and after system implementation were entered into Excel 16.0, and descriptive statistical analysis was used to summarize the stages, error type,¹³ and root causes. The ME rate for each error type was calculated as the total number of MEs in that category/total number of patients on medication at the same time × 10,000‱, and categorical data were expressed as reporting rate (composition ratio).

Secondary outcome indicators included accuracy of order verification (total number of correct order verification/total number of order verification × 100%), accuracy of patient identification (total number of correct patient identifications/total number of all patient identifications × 100%), and implementation rate of fresh medicine dispensing.¹⁴ The monthly data for each stage of medication order execution from 2017 to 2023 were extracted from the quality indicator monitoring module and imported into IBM SPSS Statistics for Windows, Version 25.0. (Armonk, NY: IBM Corp) for statistical analysis. Continuous variables were expressed as mean ± standard deviation ( $\bar{X} \pm s$ ), and independent-sample t-tests were used for comparison between groups.

Results

ME rates

During 2017–2023, 295 MEs were reported with a mean of 0.26 ± 0.26 ME rates per month. Figure 1 shows a general decrease in ME rates after system implementation. The mean ME rate reported before the implementation of the system was 0.57 and the mean ME rates after the implementation were 0.19, 0.24, and 0.07 in the short, medium, and long term, respectively. The ARIMA model analysis shows that the average level of ME rate reported in the short, medium, and long term declined after implementation (Table 3) (p < 0.01). There was no significant change in the trend for long-term decline in ME rates, but in the short-term application, ME rates decreased significantly, with a rate of 5% per month (p < 0.01). In the medium term, the trend of ME rates did not decline but rather increased (p < 0.01).

Figure 1.

Level changes in ME rate over time.

Table 3.

Level and trend changes in ME rates: results from ARIMA analysis.

Indicators	Short term		Medium term		Long term
	Estimate (SE)	p	Estimate (SE)	p	Estimate (SE)	p
Level change	−0.099	0.000	−0.153	0.000	−0.147	0.000
Trend change	−0.051	0.002	0.014	0.005	−0.011	0.064

ARIMA, autoregressive integrated moving average; ME, medication error.

To ensure result reliability, the following two methods were used for verifying the robustness of the results. (1) Replacement test method: a general linear model was used for the test, and the results showed that ME rates decreased with the implementation of the system. Robust standard errors were used for the regression as the results of White’s heteroskedasticity test indicated the presence of heteroskedasticity problem (p < 0.01). The results were validated against the parameter estimates of the robust standard error (p < 0.01). Visual inspection of scatter plots of heteroskedasticity predicted values versus standardized residuals showed random distribution of residuals, indicating no significant heteroskedasticity problem. (2) Replacement of explanatory variables: using the monthly ME to replace ME rate in a time-series analysis, the specific predictive values obtained were slightly different, but an overall decreasing trend was consistent, indicating that the predictive results of the previous ARIMA model are still valid, that is, system implementation can reduce MEs.

Types and causes of ME rates

Table 4 shows the distribution of ME rates reported in the short, medium, and long term before and after implementation by stage and error type. Once the system was applied for a long time, the rate of MEs in prescription, delivery, dispensing, and administration sessions decreased, with a significant decrease in the reported ME rate in the prescription session, whereas the reported ME rate in the patient administration session remained essentially unchanged. Nurse administration sessions reported lower ME rates after system implementation, but their weightage (>60%) increased compared with other error types, gradually contributing a higher proportion to overall errors. In terms of error types, most errors decreased or disappeared, with wrong patient and administration omissions being the main error types after long-term implementation.

Table 4.

Distribution of ME rates by stage and type before and after implementation.

Stage of medication process	ME types	Pre-intervention	Post-intervention
Stage of medication process	ME types	Pre-intervention	Short term	Medium term	Long term
Prescribing	Wrong patient	0.47 (8.59)	0.11 (5.56)	0.03 (1.54)	0 (0)
	Wrong drug	0.27 (4.91)	0 (0)	0 (0)	0 (0)
	Wrong dose	0.37 (6.75)	0.11 (5.56)	0.03 (1.54)	0 (0)
	Wrong route	0.17 (3.07)	0.05 (2.78)	0 (0)	0 (0)
	Wrong frequency	0.2 (3.68)	0.11 (5.56)	0 (0)	0 (0)
	Wrong time	0.03 (0.61)	0 (0)	0 (0)	0 (0)
	Wrong diluent	0.2 (3.68)	0 (0)	0 (0)	0 (0)
	Unreasonable order (e.g., allergy medications)	0.07 (1.23)	0 (0)	0 (0)	0 (0)
	Incomplete free-text order	0 (0)	0.05 (2.78)	0 (0)	0 (0)
	Order duplicated	0.07 (1.23)	0 (0)	0 (0)	0 (0)
	Order omission	0 (0)	0 (0)	0.03 (1.54)	0 (0)
	Total	1.83 (33.74)	0.43 (22.22)	0.09 (4.62)	0 (0)
Dispensing	Wrong drug	0.77 (14.11)	0 (0)	0.17 (9.23)	0.07 (9.68)
	Wrong size	0.47 (8.59)	0.22 (11.11)	0.23 (12.31)	0.07 (9.68)
	Wrong quantity	0.1 (1.84)	0 (0)	0.03 (1.54)	0.02 (3.23)
	Wrong diluent	0.07 (1.23)	0.05 (2.78)	0.06 (3.08)	0.02 (3.23)
	Omission	0.1 (1.84)	0 (0)	0 (0)	0.02 (3.23)
	Expired drug	0.07 (1.23)	0 (0)	0.06 (3.08)	0 (0)
	Wrong label	0.03 (0.61)	0 (0)	0 (0)	0 (0)
	ADM problem	0.1 (1.84)	0 (0)	0 (0)	0.02 (3.23)
	Total	1.7 (31.29)	0.27 (13.89)	0.55 (29.23)	0.22 (32.26)
Distribution	Wrong location	0.03 (0.61)	0 (0)	0.03 (1.54)	0 (0)
	Damaged drug	0.03 (0.61)	0 (0)	0 (0)	0 (0)
	Total	0.07 (1.23)	0 (0)	0.03(1.54)	0 (0)
Administration	Wrong patient	0.77 (14.11)	0.54 (27.78)	0.44 (23.08)	0.11 (16.13)
	Wrong drug	0.2 (3.68)	0.05 (2.78)	0.2 (10.77)	0.04 (6.45)
	Wrong dose	0.2 (3.68)	0.38 (19.44)	0.12 (6.15)	0.07 (9.68)
	Wrong route	0.13 (2.45)	0 (0)	0.12 (6.15)	0.02 (3.23)
	Wrong time	0.13 (2.45)	0.05 (2.78)	0.15 (7.69)	0.02 (3.23)
	Wrong rate	0 (0)	0.11 (5.56)	0.03 (1.54)	0.07 (9.68)
	Wrong sequence	0.03 (0.61)	0 (0)	0 (0)	0 (0)
	Omission	0.27 (4.91)	0.05 (2.78)	0.15 (7.69)	0.09 (12.9)
	Medication pump problem	0.1 (1.84)	0 (0)	0 (0)	0 (0)
	Total	1.83 (33.74)	1.19 (61.11)	1.19 (63.08)	0.42 (61.29)
Patient administration	Wrong dose	0 (0)	0 (0)	0.03 (1.54)	0.02 (3.23)
	Wrong frequency	0 (0)	0 (0)	0 (0)	0.02 (3.23)
	Missed oral medications	0 (0)	0.05 (2.78)	0 (0)	0 (0)
	Total	0 (0)	0.05 (2.78)	0.03 (1.54)	0.04 (6.45)

ADM, automatic dispensing machine; ME, medication error.

Table 5 shows the distribution of ME rates by cause reported before and after system implementation. System failures, personal problems, lack of communication, and workflow problems decreased after implementation, with personal problems showing the most obvious decrease in the long term but remaining to be the leading cause of short- and long-term MEs. Compared with other causes of MEs, problems related to lack of communication showed a less pronounced downward trend and resulted in a higher proportion of MEs after implementation.

Table 5.

Distribution of ME rates by cause before and after implementation.

Causes	Pre-intervention	Post-intervention
Causes	Pre-intervention	Short term	Medium term	Long term
Technical failures and flaws in system	0.77 (14.11)	0.27 (13.89)	0.17 (9.23)	0.09 (12.9)
Personal factors	2.4 (44.17)	0.7 (36.11)	0.7 (36.92)	0.24 (35.48)
Lack of communication	0.63 (11.66)	0.38 (19.44)	0.29 (15.38)	0.16 (22.58)
Workflow problems	1.63 (30.06)	0.59 (30.56)	0.73 (38.46)	0.2 (29.03)

ME, medication error.

Secondary outcome indicators

Table 6 shows secondary outcome indicators reported before and after system implementation. Compared with pre-intervention, the accuracy of patient identification increased significantly in the short, medium, and long term. The accuracy of order verification remained relatively stable in the short term, but exhibited a significant change over the long term, and the implementation rate of fresh medicine dispensing increased significantly. The implementation rate of fresh medicine dispensing was opposite.

Table 6.

Secondary outcome indicators before and after implementation.

Secondary outcomes	Pre-intervention	Post-intervention
Secondary outcomes	Pre-intervention	Short term	Medium term	Long term
Accuracy of order verification	98.92 ± 0.99	99.32 ± 0.27	99.37 ± 0.47	99.7 ± 0.16**
Accuracy of patient identification	98.79 ± 0.85	99.43 ± 0.61*	99.84 ± 0.3**	100**
Implementation rate of fresh medicine dispensing	96.26 ± 1.77	99.38 ± 0.42**	99.74 ± 0.2**	99.75 ± 0.07

p < 0.05. **p < 0.01.

Discussion

Effect of a closed-loop execution system on safe medication administration

The findings of the study showed that a closed-loop execution system had a positive effect on ensuring the safety in the medication administration, and this effect was more significant in the short term and medium term. However, Burkoski et al.¹¹ explored the impact of the closed-loop medication administration system on MEs and found that about 3 years of system implementation caused a reduction in the average level of MEs, but the decline was not significant (p = 0.186). This may be closely associated with the design of each hospital’s system, which is highlighted in studies and reports on informatization construction in healthcare organizations. A report from China showed that nearly 30% of hospitals’ EMR systems and medical prescription systems are not integrated to date, and approximately 40% of hospitals do not have allergy history management and allergy history alert functions.¹⁵ The Study in U.S. and Finnish hospitals has also reported that systems, such as Apotti, have shortcomings in medication ordering and prescribing, in that systems are not capable of splitting long-term medical orders and executing them over time.¹⁰ This suggests that there is still a difference between current systems and “closed-loop management.” In this study, the system ensures safe medication management by integrating patient allergy information, automated review of prescription completeness and safety, and intelligent splitting of prescriptions, which provides a reference for optimizing other systems. Considering the trend change and workflow problems in the medium term and long term, there is also a need to support EMMS design from a safety viewpoint by comprehensively analyzing system functionality and medication processes using the System Human Error Reduction and Prediction Approach or System Theoretical Accident Models and Processes,¹⁶ which are regarded as the main measures to prevent MEs.¹⁷

Root cause analysis and management measures

Although the integrated application of PDA, eMAR, and other technologies showed reduced ME rates to some extent with regard to administration, the difference was not substantial, but rather the proportion of ME rates resulting from nurse administration sessions increased.¹⁸ Among the reported errors, wrong patient and administration omission were the main MEs. Although the system provides alarms for identity verification and reminders for unexecuted medical orders, the data in Table 4 show that related errors have not been completely eliminated. This may be related to workarounds and procedures that result in the ignoring of alarms and reminders. For example, nurses choose to scan wristbands in batches after medication administration to minimize interruptions.^18,19 This, in turn, results in failures and the system is unable to safeguard five powers of medication administration. Furthermore, sometimes reminders are obscured by other alarms,²⁰ and nurses cannot always be at the side of the system to check information, so omission is inevitable. To avoid missing reminders or interruptions, the system has made improvements, such as including changing the visual display and modifying the scope and content of alerts,²¹ to remind nurses to process medical orders on time. While these measures optimize workflow and reduce the occurrence of ME rates, they are inconsistent and difficult to roll out, thus Institute for Safe Medication Practices recommends measures to be layered for each stage of the medication administration process to reduce the risk of MEs.

Through systematic pre-training and widespread clinical application, ME rates caused by various factors decreased to some extent, and the proportion of ME rates resulting from lack of communication increased. Root cause analysis showed that communication problems were observed mainly in three major domains: healthcare communication, medication handover, and nurse–patient communication, with medication handovers being the most serious problem. The analysis found that errors and omissions during nursing transitions caused more than 70% of MEs,²² with medication information gaps being the most obvious.²³ Multiple and frequent changes in patient medications, lack of training in effective communication skills, and lack of handover tools are risk factors for ME during patient handovers.²⁴ After the implementation of the system, nurses typically used informational nursing record sheets for patient medication information-related communication. However, the previous study has confirmed that this approach is undesirable²³ because the records cannot contain or accurately document important medication information and it may be easier to convey medication information using a structured approach during handovers. The Situation-Background-Assessment-Recommendation (SBAR) protocol is currently the most widely used handoff tool. Hada et al.^25,26 applied it to assess the effect of standardized bedside nursing handovers on MEs. The results showed that structured handoff tools improved the quality of information transfer and reduced the occurrence of MEs. By contrast, the study has found no significant effect of structured tools on ME.²⁷ Therefore, more randomized controlled studies are required to explore causality.

Strengthening systematic control of personal problems at all stages

A number of studies in China and worldwide have shown that ME can occur at all stages of drug therapy, with prescription and administration being the most common.^28,29 As shown in Table 6, after the implementation of HIS and execution of scanning the code and handing over, the accuracy of order verification and patient identification as well as the timeliness of medication administration improved. Nurses could identify incorrect prescriptions at an earlier stage to avoid incorrect administration and reduce the risk of MEs, and at the same time, integrated information management was beneficial for nurses to understand patients’ status. However, on comparing the data in Table 4, we found that the error rates fluctuated less before and after the implementation of the system in some processes that relied on manual checking and signature records, such as nurses’ administration of medication and patients’ medication, compared with the process of prescription. This suggests that the effect of this system implementation is limited with regard to preventing and controlling human error, which is consistent with the findings of previous studies.^28,30,31 The previous study has found that more than half of serious MEs occur and cause harm to patients if personal problems are not managed in a timely manner.⁹ Westbrook et al.³⁰ found that the addition of targeted decision support tools to the system can effectively improve nurse adherence to medication safety procedures. Thus, there is a need to introduce sophisticated information-based early warning and identification tools to reinforce the management of human error on an ongoing basis.

Limitations

Given that ME is based on voluntary, spontaneous, and anonymous reporting, there may have been a low detection rate and poor accuracy of ME reported, which could lead to lower ME rates. To comprehensively understand the execution process, we retrieved objective data from various execution stages and compared them with ME reports that were voluntarily uploaded. We found that there was a deviation between the actual operation and the content of the reports. This indicated that real-time on-site investigations are needed to discover and collect medication-related data at a later stage to ensure the reliability of the data. Another limitation of the current study is that the study did not verify the hypothesis of a causal relationship between system implementation and reduction in ME rates by comparing with a control group. Accordingly, it was difficult to identify whether the reduction in errors was because of the new system or because of changes in the medication process after implementation.

Conclusion

The implementation of a closed-loop execution system can reduce ME rates and improve the timeliness of medication dispensing and use. However, it cannot completely eliminate MEs. The results of the current study highlight the need to continuously and systematically analyze the functions of the existing systems and clinical processes. We can enhance the usability of the system to meet clinical needs and overcome adverse events caused by personal problems by measures such as improving the system’s functionality and applying structured handoff tools.

Supplemental Material

sj-docx-1-taw-10.1177_20420986241288421 – Supplemental material for Effect of a closed-loop medication order executive system on safe medication administration at a tertiary hospital: a quasi-experimental study

Supplemental material, sj-docx-1-taw-10.1177_20420986241288421 for Effect of a closed-loop medication order executive system on safe medication administration at a tertiary hospital: a quasi-experimental study by Xuwen Yin, Haiyan Song, Jieyu Lu, Jing Yang, Rong Wang, Zheng Lin, Shudi Jiang, Hui Yuan, Xumei Wang, Dongmei Xu, Chunhong Gao, Yuan Zhou, Jiayi Xu, Chen Chen, Chenyu Gu, Qingqing Diao, Fang Li and Zejuan Gu in Therapeutic Advances in Drug Safety

Supplemental Material

sj-docx-2-taw-10.1177_20420986241288421 – Supplemental material for Effect of a closed-loop medication order executive system on safe medication administration at a tertiary hospital: a quasi-experimental study

Supplemental material, sj-docx-2-taw-10.1177_20420986241288421 for Effect of a closed-loop medication order executive system on safe medication administration at a tertiary hospital: a quasi-experimental study by Xuwen Yin, Haiyan Song, Jieyu Lu, Jing Yang, Rong Wang, Zheng Lin, Shudi Jiang, Hui Yuan, Xumei Wang, Dongmei Xu, Chunhong Gao, Yuan Zhou, Jiayi Xu, Chen Chen, Chenyu Gu, Qingqing Diao, Fang Li and Zejuan Gu in Therapeutic Advances in Drug Safety

Supplemental Material

sj-docx-3-taw-10.1177_20420986241288421 – Supplemental material for Effect of a closed-loop medication order executive system on safe medication administration at a tertiary hospital: a quasi-experimental study

Supplemental material, sj-docx-3-taw-10.1177_20420986241288421 for Effect of a closed-loop medication order executive system on safe medication administration at a tertiary hospital: a quasi-experimental study by Xuwen Yin, Haiyan Song, Jieyu Lu, Jing Yang, Rong Wang, Zheng Lin, Shudi Jiang, Hui Yuan, Xumei Wang, Dongmei Xu, Chunhong Gao, Yuan Zhou, Jiayi Xu, Chen Chen, Chenyu Gu, Qingqing Diao, Fang Li and Zejuan Gu in Therapeutic Advances in Drug Safety

Footnotes

Thanks to MogoEdit for providing professional English editing (

) and the JPH quality management platform for providing related data.

Declarations

ORCID iDs

Xuwen Yin

Zejuan Gu

Supplemental material

Supplemental material for this article is available online.

References

Tolley

Watson

Heed

, et al. The impact of a novel medication scanner on administration errors in the hospital setting: a before and after feasibility study. BMC Med Inform Decis Mak 2022; 22: 86.

Elliott

Camacho

Campbell

, et al. Prevalence and economic burden of medication errors in the NHS in England: rapid evidence synthesis and economic analysis of the prevalence and burden of medication error in the UK. Report, The University of Sheffield. 2024.

World Health Organization. Patient safety, https://www.who.int/zh/news-room/fact-sheets/detail/patient-safety (2023, accessed 3 April 2024).

Roumeliotis

Sniderman

Adams-Webber

, et al. Effect of electronic prescribing strategies on medication error and harm in hospital: a systematic review and meta-analysis. J Gen Intern Med 2019; 34: 2210–2223.

Manias

Kusljic

Interventions to reduce medication errors in adult medical and surgical settings: a systematic review. Ther Adv Drug Saf 2020; 11: 2042098620968309.

Kinlay

Zheng

Burke

, et al. Medication errors related to computerized provider order entry systems in hospitals and how they change over time: a narrative review. Res Social Adm Pharm 2021; 17: 1546–1552.

Palojoki

Saranto

Reponen

, et al. Classification of electronic health record-related patient safety incidents: development and validation study. JMIR Med Inform 2021; 9: e30470.

Franklin

O’Grady

Donyai

, et al. The impact of a closed-loop electronic prescribing and administration system on prescribing errors, administration errors and staff time: a before-and-after study. Qual Saf Health Care 2007; 16: 279–284.

Lindén-Lahti

Kivivuori

S-M

Lehtonen

, et al. Implementing a new electronic health record system in a university hospital: the effect on reported medication errors. Healthcare (Basel) 2022; 10: 1020.

10.

Shermock

Schepel

LL.

Closed-loop medication management with an electronic health record system in U.S. and Finnish hospitals. Int J Environ Res Public Health 2023; 20: 6680.

11.

Burkoski

Yoon

Solomon

, et al. Closed-loop medication system: leveraging technology to elevate safety. Nurs Leadersh (Tor Ont) 2019; 32: 16–28.

12.

López Bernal

. The use of interrupted time series for the evaluation of public health interventions. Doctoral, London School of Hygiene & Tropical Medicine, London, 2018.

13.

Tan

Zeng

Cao

, et al. Quality appraisal and content analysis of clinical practice guidelines on prevention and management of medication errors. Chinese J Nurs 2019; 54: 867–873.

14.

Chunhong

Xiaolin

The application of mobile information technology on raising the implementation rate of fresh medicine-dispensing. Int J Nurs 2016; 35: 1172–1174.

15.

JiangSu Hospital Information Management Association. 2019 Jiangsu Provincial Hospital informatics research report, https://www.chima.org.cn/Html/News/Articles/4824.html (2019, accessed 30 March 2024).

16.

Awad

Amon

Baillie

, et al. Human factors and safety analysis methods used in the design and redesign of electronic medication management systems: a systematic review. Int J Med Inform 2023; 172: 105017.

17.

Kim

Suh

HS.

Priorities in the prevention strategies for medication error using the analytical hierarchy process method. Healthcare (Basel) 2022; 10: 512.

18.

Drach-Zahavy

Somech

Admi

, 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: 448–457.

19.

Fraczkowski

Matson

Lopez

KD.

Nurse workarounds in the electronic health record: an integrative review. J Am Med Inform Assoc 2020; 27(7): 1149–1165.

20.

Taft

Rudd

Thraen

, et al. “Are we there yet?” Ten persistent hazards and inefficiencies with the use of medication administration technology from the perspective of practicing nurses. J Am Med Inform Assoc 2023; 30: 809–818.

21.

Kinlay

LMR

Zheng

, et al. Electronic medication management systems: analysis of enhancements to reduce errors and improve workflow. Appl Clin Inform 2021; 12: 1049.

22.

Abdellatif

Bagian

Barajas

, et al. Communication during patient hand-overs. JCJQPS 2007; 33: 439–442.

23.

Braaf

Rixon

Williams

, et al. Medication communication during handover interactions in specialty practice settings. J Clin Nurs 2015; 24: 2859–2870.

24.

Alcalá Minagorre

Domingo Garau

Salmerón Fernández

, et al. Safe handoff practices and improvement of communication in different paediatric settings. An Pediatr (Engl Ed) 2023; 99: 185–194.

25.

Hada

Coyer

Jack

Nursing bedside clinical handover: a pilot study testing a ward-based education intervention to improve patient outcomes. J Australas Rehabil Nurses Assoc 2020; 21: 9–18.

26.

Hada

Jones

Jack

, et al. Translating evidence-based nursing clinical handover practice in an acute care setting: a quasi-experimental study. Nurs Health Sci 2021; 23: 466–476.

27.

Hada

Coyer

Shift-to-shift nursing handover interventions associated with improved inpatient outcomes—falls, pressure injuries and medication administration errors: an integrative review. Nurs Health Sci 2021; 23: 337–351.

28.

Xin

Yongxin

Lining

, et al. Risk identification of medication errors in clinical drug therapy. Chinese Pharm Affairs 2017; 31: 423–429.

29.

World Health Organization. Medication without harm: policy brief, https://www.who.int/publications/i/item/9789240062764 (2023, accessed 7 March 2024).

30.

Westbrook

Sunderland

Woods

, et al. Changes in medication administration error rates associated with the introduction of electronic medication systems in hospitals: a multisite controlled before and after study. BMJ Health Care Inform 2020; 27(3): e100170.

31.

Van de Vreede

McGrath

de Clifford

Review of medication errors that are new or likely to occur more frequently with electronic medication management systems. Aust Health Rev 2019; 43: 276–283.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.30 MB

0.02 MB

0.00 MB