Chat GPT vs. Clinical Decision Support Systems in the Analysis of Drug–Drug Interactions

Clinical Pharmacology & Therapeutics, Journal Year: 2025, Volume and Issue: unknown

Published: Feb. 11, 2025

The current standard method for the analysis of potential drug–drug interactions (pDDIs) is time‐consuming and includes use multiple clinical decision support systems (CDSSs) interpretation by healthcare professionals. With emergence large language models developed with artificial intelligence, an interesting alternative arose. This retrospective study included 30 patients polypharmacy, who underwent a pDDI between October 2022 August 2023, compared performance Chat GPT established CDSSs (MediQ®, Lexicomp®, Micromedex®) in pDDIs. A multidisciplinary team interpreted obtained results decided upon relevance assigned severity grades using three categories: (i) contraindicated, (ii) severe, (iii) moderate. expert review identified total 280 clinically relevant pDDIs (3 contraindications, 13 264 moderate) CDSSs, 80 (2 5 73 GPT. almost entirely neglected risk to QTc prolongation (85 vs. 8), which could also not be sufficiently improved specific prompt. To assess consistency provided GPT, we repeated each query found inconsistent 90% cases. In contrast, acceptable comprehensible recommendations questions on side effects. identification cannot recommended currently, because were detected, there obvious errors inconsistent. However, if these limitations are addressed accordingly, it promising platform future.

Language: Английский

---

A comparative analysis of the drug interaction programmes amongst geriatric outpatients

Indian Journal of Physiology and Pharmacology, Journal Year: 2025, Volume and Issue: 0, P. 1 - 6

Published: Jan. 13, 2025

Objectives: Drug interaction programmes are considered imperative tools utilised by healthcare professionals to screen potential risks that may occur from drug combinations. However, the outcomes of analysing drug–drug interactions (DDIs) differ across each programme. It is crucial for clinician be aware varying results database and their limitations before utilising them. This study aimed compare in geriatric patients at an outpatient clinic a tertiary care hospital. Materials Methods: A retrospective was performed hospital Thailand. collected all prescriptions during November 2021 2022. The programs using Micromedex, Medscape Lexicomp were used detect assess severity DDIs. Results: participants recruited electronic medical records enrolment total 10,877 individuals. majority these male, with average age 74.3 (standard deviation 6.8) years. prevalence major DDIs Lexicomp, Micromedex 28.1%, 57.9% 18.2%, respectively. Only 1700 (15.6%) observed consistent three programmes. strength agreement amongst Kappa statistics 0.15, 0.35 0.61 ( P <0.01) major, moderate minor + no groups, Conclusion: degree among three-drug programmes, Medscape, minimal. To maintain uniformity information sources, it essential apply measures standardisation documentation.

Language: Английский

---

Drug-drug interactions with oral anticoagulants: information consistency assessment of three commonly used online drug interactions databases in Switzerland

Frontiers in Pharmacology, Journal Year: 2024, Volume and Issue: 15

Published: April 2, 2024

Background: Toxicity or treatment failure related to drug-drug interactions (DDIs) are known significantly affect morbidity and hospitalization rates. Despite the availability of numerous databases for DDIs identification management, their information often differs. Oral anticoagulants deemed at risk a leading cause adverse drug events, most which being preventable. Although many include involving anticoagulants, none specialized in them. Aim method: This study aims compare content four direct oral two vitamin K antagonists three major DDI used Switzerland: Lexi-Interact, Pharmavista, MediQ. It evaluates consistency terms differences severity rating systems, mechanism interaction, extraction documentation processes transparency. Results: revealed 2’496 six with discrepant classifications. Only 13.2% were common all databases. Overall concordance classification (high, moderate, low risk) was slight (Fleiss’ kappa = 0.131), while high-risk demonstrated fair agreement 0.398). The nature more consistent across Qualitative assessments highlighted process transparency, similarities references. Discussion: highlights discrepancies between commonly inconsistency how terminology is standardised incorporated when classifying these DDIs. also need creation specialised tools anticoagulant-related interactions.

Language: Английский

---

Update and Application of a Deep Learning Model for the Prediction of Interactions between Drugs Used by Patients with Multiple Sclerosis

Pharmaceutics, Journal Year: 2023, Volume and Issue: 16(1), P. 3 - 3

Published: Dec. 19, 2023

Patients with multiple sclerosis (MS) often take drugs at the same time to modify course of disease, alleviate neurological symptoms and manage co-existing conditions. A major consequence for a patient taking different medications is higher risk treatment failure side effects. This because drug may alter pharmacokinetic and/or pharmacodynamic properties another drug, which referred as drug-drug interaction (DDI). We aimed predict interactions that are used by patients MS based on deep neural network (DNN) using structural information input. further identify potential drug-food (DFIs), can affect efficacy safety well. DeepDDI, multi-label classification model specific DDI types, changes in pharmacological effects adverse events when two or more taken together. The original ~34 million trainable parameters was updated >1 DDIs recorded DrugBank database. Structure data food components were obtained from FooDB medication plans (n = 627) then searched pairwise between compounds. DeepDDI achieved accuracies 92.2% 92.1% validation testing sets, respectively. 312 small molecule prescription over-the-counter medications. In plans, we identified 3748 13,365 DeepDDI. At least one found most 509 81.2% DNN model). predictions revealed many would be increased bleeding bradycardic complications due if they start disease-modifying therapy cladribine 242 38.6%) fingolimod 279 44.5%), also numerous Bruton’s tyrosine kinase inhibitors clinical development MS, such evobrutinib 434 DDIs). Food sources related DFIs corn 5456 DFIs) cow’s milk 4243 DFIs). demonstrate learning techniques exploit chemical structure similarity accurately MS. Our study specifies pairs potentially interact, suggests mechanisms causing effects, informs about whether interacting replaced alternative avoid critical provides dietary recommendations who certain drugs.

Language: Английский

---

Polypharmacy in Multiple Sclerosis: Prevalence, Risks, and Mitigation Strategies

Current Neurology and Neuroscience Reports, Journal Year: 2023, Volume and Issue: 23(9), P. 521 - 529

Published: July 31, 2023

Language: Английский

---

COMPARATIVE EVALUATION OF ARTIFICIAL INTELLIGENCE AND DRUG INTERACTION TOOLS: A PERSPECTIVE WITH THE EXAMPLE OF CLOPIDOGREL

Ankara Universitesi Eczacilik Fakultesi Dergisi, Journal Year: 2024, Volume and Issue: 48(3), P. 22 - 22

Published: Aug. 5, 2024

Objective: The study aims to compare the ability of free artificial intelligence (AI) chatbots detect drug interactions with freely available interaction tools, using clopidogrel as an example. Material and Method: Lexicomp database was used a reference determine clopidogrel. ChatGPT-3.5 AI Bing were selected chatbots. Medscape Drug Interaction Checker, DrugBank Checker Epocrates Check tools. Accuracy score comprehensiveness calculated for each tool kappa coefficient assess inter-source agreement severity. Result Discussion: results most similar those obtained from chatbot. chatbot performed best, 69 correct accuracy 307. has highest overall 387 points comprehensiveness. In addition, found (0.201, fair agreement). However, some by need be improved they are incorrect/inadequate. Therefore, information tools should not clinical applications healthcare professionals patients change their treatment without consulting doctor.

Language: Английский

---

Prescribed Versus Taken Polypharmacy and Drug–Drug Interactions in Older Cardiovascular Patients during the COVID-19 Pandemic: Observational Cross-Sectional Analytical Study

Journal of Clinical Medicine, Journal Year: 2023, Volume and Issue: 12(15), P. 5061 - 5061

Published: Aug. 1, 2023

The study aimed to assess clinical pharmacology patterns of prescribed and taken medications in older cardiovascular patients using electronic health records (EHRs) (n = 704) (2019–2022). Medscape Drug Interaction Checker was used identify pairwise drug–drug interactions (DDIs). Prevalence rates DDIs were 73.5% 68.5% among drugs, respectively. However, the total number significantly higher (p < 0.05). Serious comprised 16% 7% all medications, respectively Median numbers between vs. Me 2, IQR 0–7 3, per record, polypharmacy compared with that drugs Women taking more had prevalence No sex-related differences observed list medications. ICD code U07.1 (COVID-19, virus identified) associated highest median DDI record. Further research is warranted improve EHR structure, implement patient engagement reporting adverse drug reactions, provide genetic profiling avoid potentially serious DDIs.

Language: Английский

---

Categorical Analysis of Database Consistency in Reporting Drug–Drug Interactions for Cardiovascular Diseases

Pharmaceutics, Journal Year: 2024, Volume and Issue: 16(3), P. 339 - 339

Published: Feb. 28, 2024

Drug–drug interactions (DDIs) can either enhance or diminish the positive negative effects of associated drugs. Multiple drug combinations create difficulties in identifying clinically relevant interactions; this is why electronic interaction checkers frequently report DDI results inconsistently. Our paper aims to analyze cardiovascular diseases by selecting drugs from pharmacotherapeutic subcategories interest according Level 2 Anatomical Therapeutic Chemical (ATC) classification system. We checked DDIs between 9316 pairs and 25,893 other then evaluated overall agreement on severity two checkers. Thus, we obtained a fair for category, as well (i.e., non-cardiovascular) categories, reflected Fleiss’ kappa coefficients κ=0.3363 κ=0.3572, respectively. The categorical analysis ATC-defined reveals that indicate levels varying poor (κ<0) perfect (κ=1). main drawback assessment it includes same subcategory, situation therapeutic duplication seldom encountered clinical practice. conclusion more insightful than approach, allows thorough investigation disparities databases better exposes factors influence different responses Using avoids potential inaccuracies caused particularizing an statistical heterogeneous dataset.

Language: Английский

---

Einfluss von Interaktionsdatenbanken auf die Arzneimitteltherapiesicherheit im Krankenhaus

Published: Jan. 1, 2024

Die Ergebnisse der Studie und die Diversität Datenbanken ist groß. Für 12 wurde ein Punktesystem mit elf Items entworfen, um Qualität einzelnen zu objektivieren. Keine Datenbank konnte alle Bewertungskriterien erfüllen. Der insgesamt schlechte Punktedurchschnitt Indikator für Mängel aktuell verfügbaren Datenbanken. Außerdem konnten wir einen Qualitätsunterschied zwischen kostenpflichtigen kostenfreien beweisen mussten im Zuge dieser Frage stellen, ob kostenfreie überhaupt nützlich sind. Zwischen den fallen Qualitätsunterschiede weniger gravierend aus, wenngleich Stärken Schwächen sich deutlich unterscheiden. häufigsten Wechselwirkungen wurden in allen großem Abstand rein psychiatrischen Interaktionspaaren erfasst. Dieses zeigt, wie wechselwirkungsreich Psychopharmaka sind dass psychiatrische Patienten besonders vulnerabel Nutzung digitaler Hilfsmittel scheint bei Betrachtung hohen Anzahl ausgegebener Warnmeldungen unabdingbar sein, dennoch existiert große Uneinheitlichkeit Bewertung Interaktionen. Vorstellung, zwei Kliniker zweier unterschiedlicher völlig unterschiedlichen Empfehlungen kommen, fällt nicht schwer. Gleichzeitig könnte Kooperation von Heilberuflern, unterschiedliche verwenden, Chance auf zusätzlichen Informationsgewinn Austausch erhöhen, was Umkehrschluss einer erhöhten Arzneimitteltherapiesicherheit resultiert. In Studien positive Effekt interdisziplinärer Zusammenarbeit bereits bewiesen werden. Zusammenfassend umfangreiche Differenzen untersuchten aufgezeigt Um Anforderungen des klinischen Alltags genügen, müssen digitale Unterstützungssysteme weiterentwickelt „ideale Datenbank“ gibt es bisher – das lässt durch unser beweisen. Alltag Patientensicherheit gewährleisten ausreichend. Gewährung sollte oberstes Ziel sein dieses erreichen, bedarf vieler Komponenten. Neben vor allem Weiterentwicklung auch zwischenmenschliche weiter gefördert Interdisziplinäre Sinne pharmazeutischer Dienstleistungen zur Medikationsanalyse könnten zusätzliches Instrument Vermeidung arzneimittelbezogener Probleme Zukünftig werden unsere am meisten optimaler weiterentwickelter Technologien, sowie wachsendem zwischenmenschlichem profitieren.

---

Potentially Inappropriate Medication: A Pilot Study in Institutionalized Older Adults

Healthcare, Journal Year: 2024, Volume and Issue: 12(13), P. 1275 - 1275

Published: June 26, 2024

Institutionalized older adults often face complex medication regimens, increasing their risk of adverse drug events due to polypharmacy, overprescribing, interactions, or the use Potentially Inappropriate Medications (PIM). However, data on and associated risks in this population remain scarce. This pilot study aimed characterize sociodemographic, clinical pharmacotherapeutic profiles, PIM among institutionalized elders residing Residential Structures for Elderly People (ERPI) Faro municipality, located Portuguese region Algarve. We conducted a cross-sectional non-randomized sample 96 participants (mean age: 86.6 ± 7.86 years) where trained researchers reviewed profiles identified potentially inappropriate medications using EU(7)-PIM list. Over 90% exhibited polypharmacy (≥5 medications), with an average 9.1 4.15 per person. About 92% had potential including major moderate interactions. More than 86% used at least one medication, most commonly central nervous system drugs. demonstrates that may be high medication-related problems. Implementing comprehensive review programs promoting adapted prescribing practices are crucial optimize improve well-being vulnerable population.

Language: Английский

---