Effect of Computerized Physician Order Entry and Clinical Decision Support System on Adverse Drug Events Prevention in the Emergency Department: A Systematic Review

Development of machine-learning models using pharmacy inquiry database for predicting dose-related inquiries in a tertiary teaching hospital

BACKGROUND

Drug-related problems (DRPs) are a significant concern in healthcare. Pharmacists play a vital role in detecting and resolving DRPs to improve patient safety. A pharmacy inquiry program was established in a tertiary teaching hospital to document inquiries about physicians' orders, aimed at preventing potential DRPs or providing medication information during order reviews.

OBJECTIVE

We aimed to develop machine-learning models using a pharmacy inquiry database to predict dose-related inquiries based on prescriptions and patient information.

METHODS

This retrospective study analyzed 20,393 pharmacy inquiries collected between January 2018 and February 2023. Data included prescription information (drug ingredient, dose, unit, and frequency), patient characteristics (age, sex, weight, and department), and renal function. The inquiries were categorized into two classes: dose-related inquiries (e.g., wrong dose and inappropriate regimen) and non-dose-related inquiries (e.g., inappropriate drug form and administration route). Six machine-learning models were developed: logistic regression, support vector classifier, decision tree, random forest, extreme gradient boosting, and categorical boosting. To evaluate the performance of the models, the area under the receiver operating characteristic curve and the accuracy were compared.

RESULTS

The CatBoost model achieved the highest performance (sensitivity: 0.92; accuracy: 0.79). The SHapley Additive exPlanations values highlighted the importance of features in the model predictions, drug ingredients, units, and renal function, in that order. Notably, lower renal function positively contributed to the prediction of dose-related inquiries. Additionally, the subsequent feature importance among drug ingredients showed that drugs such as acetylsalicylic acid, famotidine, metformin, and spironolactone strongly influenced the prediction of dose-related inquiries.

CONCLUSION

Machine-learning models that use pharmacy inquiry data can effectively predict dose-related inquiries. Further external validation and refinement of the models are required for broader applications in healthcare settings. These findings provide valuable guidance for healthcare professionals and highlight the potential of machine learning in pharmacists' decision-making.

The effect of electronic error-reporting forms on nurse's stress and the rate of error-reporting

OBJECTIVES

The patient safety culture includes a systematic approach that promotes safe care for patients and the leadership that supports it. Medical errors threaten patient safety. A significant portion of medical errors is committed by nurses. Although error-reporting provides valuable information to prevent errors, most nurses do not report their errors due to their high level of stress. This study was to investigate the effect of electronic error-reporting forms on nurses' stress and the rate of error-reporting.

METHODS

The nurses' level of stress was compared when using paper error-reporting and 6 months after using electronic forms. A revised version of the Coudron questionnaire was completed by 186 nurses. Data were analyzed by SPSS 23 using Wilcoxon test. The number of reported errors in paper and electronic media was compared over the same period.

RESULTS

Implementation of the electronic error-reporting form reduced the job stress of nurses by 22.22 points (p=.00) and increased the error-reporting rate by 12.86% (p<.05).

CONCLUSIONS

Although nurse's stress significantly decreases after implementing electronic error-reporting forms, their level of stress is still high and they are still at risk for physical and mental problems. Using methods like modifying the error-reporting form will increase the error-reporting rate.

Computerized clinical decision support systems for prescribing in primary care: main characteristics and implementation impact-protocol of an evidence and gap map

BACKGROUND

Computerized clinical decision support systems are used by clinicians at the point of care to improve quality of healthcare processes (prescribing error prevention, adherence to clinical guidelines, etc.) and clinical outcomes (preventive, therapeutic, and diagnostics). Attempts to summarize results of computerized clinical decision support systems to support prescription in primary care have been challenging, and most systematic reviews and meta-analyses failed due to an extremely high degree of heterogeneity present among the included primary studies. The aim of our study will be to synthesize the evidence, considering all methodological factors that could explain these differences, and build an evidence and gap map to identify important remaining research questions.

METHODS

A literature search will be conducted from January 2010 onwards in MEDLINE, Embase, the Cochrane Library, and Web of Science databases. Two reviewers will independently screen all citations, full text, and abstract data. The study methodological quality and risk of bias will be appraised using appropriate tools if applicable. A flow diagram with the screened studies will be presented, and all included studies will be displayed using interactive evidence and gap maps. Results will be reported in accordance with recommendations from the Campbell Collaboration on the development of evidence and gap maps.

DISCUSSION

Evidence behind computerized clinical decision support systems to support prescription use in primary care has so far been difficult to be synthesized. Evidence and gap maps represent an innovative approach that has emerged and is increasingly being used to address a broader research question, where multiple types of intervention and outcomes reported may be evaluated. Broad inclusion criteria have been chosen with regard to study designs, in order to collect all available information. Regarding the limitations, we will only include English and Spanish language studies from the last 10 years, we will not perform a grey literature search, and we will not carry out a meta-analysis due to the predictable heterogeneity of available studies.

SYSTEMATIC REVIEW REGISTRATION

This study is registered in Open Science Framework https://bit.ly/2RqKrWp.

Medication Safety in the Emergency Department: A Study of Serious Medication Errors Reported by 101 Hospitals From 2011 to 2020

Background: Although serious medication errors are uncommon, accounting for only 0.4% (166 of 46,568) of medication errors reported to the Pennsylvania Patient Safety Reporting System (PA-PSRS) in 2020, their effects can be devastating for patients and their loved ones. Methods: We queried PA-PSRS for reports of serious events classified as medication errors that occurred in the emergency department (ED) from January 1, 2011, to December 31, 2020. We performed a descriptive analysis to identify trends among patient sex, patient age, event harm score, event day of the week, and event time of day. We also manually coded and analyzed events based on the prescribed medication(s) and medication class(es), stage of the medication-use process at which the error occurred, and medication error type. Results: We identified 250 reports of serious medication error events in PA-PSRS that occurred in the ED from 2011 to 2020. Reports more often specified that the patient was female (61.2%; 153 of 250), and patients ranged in age from 8 months to 96 years, with a median patient age of 55 years. Events were significantly more likely to occur Friday through Sunday versus Monday through Thursday (p = .0214) and in the p.m. hours versus a.m. hours, (p = .0007). The most common prescribed medications mentioned in reports were epinephrine, insulin, hydromorphone, sodium chloride, heparin, propofol, diltiazem, ketamine, and morphine. Events occurred most often at the prescribing stage of the medication-use process (42.0%; 105 of 250), and the most common medication error type was a wrong dose (42.0%; 105 of 250). Conclusion: Potential safety strategies to consider to reduce the risk of medication errors in the ED include stocking epinephrine autoinjectors, using clinical decision support at the ordering/prescribing stage of the process, and adding an emergency medicine pharmacist as part of an interdisciplinary emergency medicine team.