Evaluation of Clinical Relevance of Drug-Drug Interaction Alerts Prior to Implementation

Artificial intelligence to assist decision-making on pharmacotherapy: A feasibility study

Background

Artificial intelligence (AI) has the capability to analyze vast amounts of data and has been applied in various healthcare sectors. However, its effectiveness in aiding pharmacotherapy decision-making remains uncertain due to the intricate, patient-specific, and dynamic nature of this field.

Objective

This study sought to investigate the potential of AI in guiding pharmacotherapy decisions using clinical data such as diagnoses, laboratory results, and vital signs obtained from routine patient care.

Methods

Data of a previous study on medication therapy optimization was updated and adapted for the purpose of this study. Analysis was conducted using R software along with the tidymodels extension packages. The dataset was split into 74% for training and 26% for testing. Decision trees were selected as the primary model due to their simplicity, transparency, and interpretability. To prevent overfitting, bootstrapping techniques were employed, and hyperparameters were fine-tuned. Performance metrics such as areas under the curve and accuracies were computed.

Results

The study cohort comprised 101 elderly patients with multiple diagnoses and complex medication regimens. The AI model demonstrated prediction accuracies ranging from 38% to 100% for various cardiovascular drug classes. Laboratory data and vital signs could not be interpreted, as the effect and dependence were unclear for the model. The study revealed that the issue of AI lag time in responding to sudden changes could be addressed by manually adjusting decision trees, a task not feasible with neural networks.

Conclusion

In conclusion, the AI model exhibited promise in recommending appropriate medications for individual patients. While the study identified several obstacles during model development, most were successfully resolved. Future AI studies need to include the drug effect, not only the drug, if laboratory data is part of the decision. This could assist with interpreting their potential relationship. Human oversight and intervention remain essential for an AI-driven pharmacotherapy decision support system to ensure safe and effective patient care.

Drug-Drug Interactions and Actual Harm to Hospitalized Patients: A Multicentre Study Examining the Prevalence Pre- and Post-Electronic Medication System Implementation

INTRODUCTION

Drug-drug interactions (DDIs) have potential to cause patient harm, including lowering therapeutic efficacy. This study aimed to (i) determine the prevalence of potential DDIs (pDDIs); clinically relevant DDIs (cDDIs), that is, DDIs that could lead to patient harm, taking into account a patient's individual clinical profile, drug effects and severity of potential harmful outcome; and subsequent actual harm among hospitalized patients and (ii) examine the impact of transitioning from paper-based medication charts to electronic medication management (eMM) on DDIs and patient harms.

METHODS

This was a secondary analysis of the control arm of a controlled pre-post study. Patients were randomly selected from three Australian hospitals. Retrospective chart review was conducted before and after the implementation of an eMM system, without accompanying clinical decision support alerts for DDIs. Harm was assessed by an expert panel.

RESULTS

Of 1186 patient admissions, 70.1% (n = 831) experienced a pDDI, 42.6% (n = 505) a cDDI and 0.9% (n = 11) an actual harm in hospital. Of 15,860 pDDIs identified, 27.0% (n = 4285) were classified as cDDIs. The median number of pDDIs and cDDIs per 10 drugs were 6 [interquartile range (IQR) 2-13] and 0 (IQR 0-2), respectively. In cases where a cDDI was identified, both drugs were 44% less likely to be co-administered following eMM (adjusted odds ratio 0.56, 95% confidence interval 0.46-0.73).

CONCLUSION

Although most patients experienced a pDDI during their hospital stay, less than one-third of pDDIs were clinically relevant. The low prevalence of harm identified raises questions about the value of incorporating DDI decision support into systems given the potential negative impacts of DDI alerts.

Interprofessional Evaluation of a Medication Clinical Decision Support System Prior to Implementation

BACKGROUND

 Computerized physician order entry (CPOE) and clinical decision support systems (CDSS) are widespread due to increasing digitalization of hospitals. They can be associated with reduced medication errors and improved patient safety, but also with well-known risks (e.g., overalerting, nonadoption).

OBJECTIVES

 Therefore, we aimed to evaluate a commonly used CDSS containing Medication-Safety-Validators (e.g., drug-drug interactions), which can be locally activated or deactivated, to identify limitations and thereby potentially optimize the use of the CDSS in clinical routine.

METHODS

 Within the implementation process of Meona (commercial CPOE/CDSS) at a German University hospital, we conducted an interprofessional evaluation of the CDSS and its included Medication-Safety-Validators following a defined algorithm: (1) general evaluation, (2) systematic technical and content-related validation, (3) decision of activation or deactivation, and possibly (4) choosing the activation mode (interruptive or passive). We completed the in-depth evaluation for exemplarily chosen Medication-Safety-Validators. Moreover, we performed a survey among 12 German University hospitals using Meona to compare their configurations.

RESULTS

 Based on the evaluation, we deactivated 3 of 10 Medication-Safety-Validators due to technical or content-related limitations. For the seven activated Medication-Safety-Validators, we chose the interruptive option ["PUSH-(&PULL)-modus"] four times (4/7), and a new, on-demand option ["only-PULL-modus"] three times (3/7). The site-specific configuration (activation or deactivation) differed across all participating hospitals in the survey and led to varying medication safety alerts for identical patient cases.

CONCLUSION

 An interprofessional evaluation of CPOE and CDSS prior to implementation in clinical routine is crucial to detect limitations. This can contribute to a sustainable utilization and thereby possibly increase medication safety.

An Evaluation of Drug-Drug Interaction Alerts Produced by Clinical Decision Support Systems in a Tertiary Hospital

Introduction Drug-drug interactions (DDIs) have the potential to harm patients. Hence, DDI alerts are meant to prevent harm; as a result, their usefulness is reduced when most alerts displayed to providers are ignored. This study aims to explore the rates and reasons for overriding alerts of DDI. Methods This is a retrospective study of DDI alert overrides that occurred between January 2020 and December 2020 within the inpatient medical records at a tertiary hospital, Medina City, Kingdom of Saudi Arabia. Results A total of 7,098 DDI alerts were generated from inpatient settings, of which 6,551(92.2%) were overridden by the physicians at the point of prescribing. "Will Monitor as Recommended" (33%) was the most common reason for the override, followed by 'Will Adjust the Dose as Recommended (27.1%)," "The Patient Has Already Tolerated the Combination" (25.7%), and "No Overridden Reason Selected" (13.0%). Discussion The DDI alert overriding is still high and is comparable to other studies. However, this study reveals that physicians are ready to deal with the consequences of around 58% of DDI alerts. Additionally, 13% of physicians were not willing to report the reason for overriding. This indicates an urgent need to review and restructure the DDI alert system. Conclusion The DDI alert override rates are high, and this is undesirable. It is recommended to revise the DDI alert system. Future studies should dig deep for real reasons for overriding and seek inputs from all stakeholders, including developing actionable metrics to track and monitor DDI alerting system.

Drug-drug Interactions with Warfarin: A Systematic Review and Meta-analysis

AIM

The objective of this paper is to systematically review the literature on drug-drug interactions with warfarin, with a focus on patient-important clinical outcomes.

METHODS

MEDLINE, EMBASE, and the International Pharmaceutical Abstract (IPA) databases were searched from January 2004 to August 2019. We included studies describing drug-drug interactions between warfarin and other drugs. Screening and data extraction were conducted independently and in duplicate. We synthesized pooled odds ratios (OR) with 95% confidence intervals (CIs), comparing warfarin plus another medication to warfarin alone. We assessed the risk of bias at the study level and evaluated the overall certainty of evidence using GRADE.

RESULTS

Of 42,013 citations identified, a total of 72 studies reporting on 3,735,775 patients were considered eligible, including 11 randomized clinical trials and 61 observational studies. Increased risk of clinically relevant bleeding when added to warfarin therapy was observed for antiplatelet (AP) regimens (OR=1.74; 95% CI 1.56, 1.94), many antimicrobials (OR=1.63; 95% CI 1.45, 1.83), NSAIDs including COX-2 NSAIDs (OR=1.83; 95% CI 1.29, 2.59), SSRIs (OR=1.62; 95% CI 1.42, 1.85), mirtazapine (OR=1.75; 95% CI 1.30, 2.36), loop diuretics (OR=1.92; 95% CI 1.29, 2.86), and others. We found a protective effect of proton pump inhibitors (PPIs) against warfarin-related gastrointestinal (GI) bleedings (OR=0.69; 95% CI 0.64, 0.73). No significant effect on thromboembolic events or mortality of any drug group used with warfarin was found, including single or dual AP regimens.

CONCLUSIONS

This review found low to moderate certainty evidence supporting the interaction between warfarin and a small group of medications, which result in increased bleeding risk. PPIs are associated with reduced hospitalization for upper GI bleeding for patients taking warfarin. Further studies are required to better understand drug-drug interactions leading to thromboembolic outcomes or death.

Clinically relevant potential drug-drug interactions in intensive care patients: A large retrospective observational multicenter study

PURPOSE

Potential drug-drug interactions (pDDIs) may harm patients admitted to the Intensive Care Unit (ICU). Due to the patient's critical condition and continuous monitoring on the ICU, not all pDDIs are clinically relevant. Clinical decision support systems (CDSSs) warning for irrelevant pDDIs could result in alert fatigue and overlooking important signals. Therefore, our aim was to describe the frequency of clinically relevant pDDIs (crpDDIs) to enable tailoring of CDSSs to the ICU setting.

MATERIALS & METHODS

In this multicenter retrospective observational study, we used medication administration data to identify pDDIs in ICU admissions from 13 ICUs. Clinical relevance was based on a Delphi study in which intensivists and hospital pharmacists assessed the clinical relevance of pDDIs for the ICU setting.

RESULTS

The mean number of pDDIs per 1000 medication administrations was 70.1, dropping to 31.0 when considering only crpDDIs. Of 103,871 ICU patients, 38% was exposed to a crpDDI. The most frequently occurring crpDDIs involve QT-prolonging agents, digoxin, or NSAIDs.

CONCLUSIONS

Considering clinical relevance of pDDIs in the ICU setting is important, as only half of the detected pDDIs were crpDDIs. Therefore, tailoring CDSSs to the ICU may reduce alert fatigue and improve medication safety in ICU patients.

Clinical Decision Support and Implications for the Clinician Burnout Crisis

OBJECTIVES

This survey aimed to review aspects of clinical decision support (CDS) that contribute to burnout and identify key themes for improving the acceptability of CDS to clinicians, with the goal of decreasing said burnout.

METHODS

We performed a survey of relevant articles from 2018-2019 addressing CDS and aspects of clinician burnout from PubMed and Web of Science™. Themes were manually extracted from publications that met inclusion criteria.

RESULTS

Eighty-nine articles met inclusion criteria, including 12 review articles. Review articles were either prescriptive, describing how CDS should work, or analytic, describing how current CDS tools are deployed. The non-review articles largely demonstrated poor relevance and acceptability of current tools, and few studies showed benefits in terms of efficiency or patient outcomes from implemented CDS. Encouragingly, multiple studies highlighted steps that succeeded in improving both acceptability and relevance of CDS.

CONCLUSIONS

CDS can contribute to clinician frustration and burnout. Using the techniques of improving relevance, soliciting feedback, customization, measurement of outcomes and metrics, and iteration, the effects of CDS on burnout can be ameliorated.

Drug-drug interactions in patients undergoing chemoradiotherapy and the impact of an expert team intervention

Background Several studies have examined the drug-drug interaction patterns in different patient populations and treatment settings; however, there is a need, particularly in the field of oncology and radiotherapy, for evaluating methods targeted towards preventing potential drug-drug interactions. One of the measures proposed is identifying potential interactions using computer programs and their evaluation by pharmacologists or clinical pharmacists, thereby providing clinically relevant information to the treating physician regarding the required prescription changes. Objective To determine the prevalence of potential drug-drug interactions in patients receiving chemoradiotherapy and assess the usefulness of expert team recommendations in minimizing interactions. Setting Patients admitted to the radiotherapy and oncology ward of a tertiary care teaching hospital in Karnataka, India. Method We conducted a prospective, cross-sectional study of prescriptions written for patients receiving chemoradiotherapy. Prescriptions containing two or more drugs, at least one of the drugs being an anticancer drug, were analyzed. They were screened for potential drug-drug interactions using the Lexicomp® drug interaction software. The interactions were classified as X, drug combination to be avoided; D, modification of therapy to be considered; and C, therapy to be monitored, as per the Lexicomp criteria. Main outcome measure The number of drug-drug interactions detected that were accepted by the treating radio-oncologist as requiring prescription change before and after the prescription review by an expert team. Results Two hundred twenty-three prescriptions were screened for the presence of drug-drug interactions; 106 prescriptions (47.53%) containing 620 drugs and 211 drug-drug interactions were identified. Of the 211 interactions identified, 6.64% (14/211), 18.48% (39/211), and 74.88% (158/211) drug-drug interactions belonged to category X, D, and C, respectively. Twenty-seven (50.94%) of the 53 category X and D interactions identified were accepted the oncologist as requiring a change in the prescription; an additional 13 (24.53%) interactions were identified as significant by the expert team, and 11 (84.62%) of these were accepted by the oncologist. Conclusion A system of alerting the treating physician to a potential drug-drug interaction leads to avoidance of prescription of the interacting drug combination, and the assistance by an expert team adds significantly to avoidance of clinically relevant drug interactions.

Evaluating the Impact of Interruptive Alerts within a Health System: Use, Response Time, and Cumulative Time Burden

BACKGROUND

Health systems often employ interruptive alerts through the electronic health record to improve patient care. However, concerns of "alert fatigue" have been raised, highlighting the importance of understanding the time burden and impact of these alerts on providers.

OBJECTIVES

Our main objective was to determine the total time providers spent on interruptive alerts in both inpatient and outpatient settings. Our secondary objectives were to analyze dwell time for individual alerts and examine both provider and alert-related factors associated with dwell time variance.

METHODS

We retrospectively evaluated use and response to the 75 most common interruptive ("popup") alerts between June 1st, 2015 and November 1st, 2016 in a large academic health care system. Alert "dwell time" was calculated as the time between the alert appearing on a provider's screen until it was closed. The total number of alerts and dwell times per provider per month was calculated for inpatient and outpatient alerts and compared across alert type.

RESULTS

The median number of alerts seen by a provider was 12 per month (IQR 4-34). Overall, 67% of inpatient and 39% of outpatient alerts were closed in under 3 seconds. Alerts related to patient safety and those requiring more than a single click to proceed had significantly longer median dwell times of 5.2 and 6.7 seconds, respectively. The median total monthly time spent by providers viewing alerts was 49 seconds on inpatient alerts and 28 seconds on outpatient alerts.

CONCLUSION

Most alerts were closed in under 3 seconds and a provider's total time spent on alerts was less than 1 min/mo. Alert fatigue may lie in their interruptive and noncritical nature rather than time burden. Monitoring alert interaction time can function as a valuable metric to assess the impact of alerts on workflow and potentially identify routinely ignored alerts.

Interactions of cobicistat and ritonavir in patients with HIV and its clinical consequences

INTRODUCTION

The prescription of antiretroviral treatment (ART) that contains pharmacokinetic enhancers such as ritonavir and cobicistat is frequent. The objective of this stdy was to analyze the potential interactions of ART that include these molecules in their formulation with the patient's home medication, as well as the clinical management of those potentially serious.

METHODS

Prospective study conducted in the pharmacy care clinic of a third level hospital between January and December of 2018. Those HIV+patients with an ART containing cobicistat or ritonavir were included in the study. Potential interactions between ART and concomitant medication were analysed in three databases (Micromedex®, Drugs.com and Liverpool), the interventions carried out were detailed, and adverse drug reactions analysed.

RESULTS

968 patients were included with a total of 2,148 prescriptions (274 different medications). A total of 86 interventions were performed regarding potential interactions in patients. The most frequent were substitutions of corticoid treatments, treatment suspensions and closer monitoring of treatments. A total of possible adverse drug reactions were analysed. The degree of agreement in the severity classification of the interactions for cobicistat and ritonavir was good among the three databases. It was remarkable Micromedex® as the most complete because it has more registered medications.

CONCLUSION

The interactions between ART with pharmacokinetic enhancers in its composition and concomitant medication is frequent and requires a significant variety of interventions. The check of interactions in different databases is recommended since they can cause adverse drug reactions.

Preventing potential drug-drug interactions through alerting decision support systems: A clinical context based methodology

BACKGROUND

The effectiveness of the clinical decision support systems (CDSSs) is hampered by frequent workflow interruptions and alert fatigue because of alerts with little or no clinical relevance. In this paper, we reported a methodology through which we applied knowledge from the clinical context and the international recommendations to develop a potential drug-drug interaction (pDDI) CDSS in the field of kidney transplantation.

METHODS

Prescriptions of five nephrologists were prospectively recorded through non-participatory observations for two months. The Medscape multi-drug interaction checker tool was used to detect pDDIs. Alongside the Stockley's drug interactions reference, our clinicians were consulted with respect to the clinical relevance of detected pDDIs. We performed semi-structured interviews with five nephrologists and one informant nurse. Our clinically relevant pDDIs were checked with the Dutch "G-Standard". A multidisciplinary team decided the design characteristics of pDDI-alerts in a CDSS considering the international recommendations and the inputs from our clinical context. Finally, the performance of the CDSS in detecting DDIs was evaluated iteratively by a multidisciplinary research team.

RESULTS

Medication data of 595 patients with 788 visits were collected and analyzed. Fifty-two types of interactions were most common, comprising 90% of all pDDIs. Among them 33 interactions (comprising 77% of all pDDIs) were rated as clinically relevant and were included in the CDSS's knowledge-base. Of these pDDIs, 73% were recognized as either pseudoduplication of drugs or not a pDDI when checked with the Dutch G-standard. Thirty-three alerts were developed and physicians were allowed to customize the appearance of pDDI-alerts based on a proposed algorithm.

CONCLUSION

Clinical practice contexts should be studied to understand the complexities of clinical work and to learn the type, severity and frequency of pDDIs. In order to make the alerts more effective, clinicians' points of view concerning the clinical relevance of pDDIs are critical. Moreover, flexibility should be built into a pDDI-CDSS to allow clinicians to customize the appearance of pDDI-alerts based on their clinical context.