Liu, Tianming

Abstract Purpose Large language models (LLMs) have proven performance for certain diagnostic tasks; however, limited studies have evaluated their consistency in recommending appropriate medication regimens for a given diagnosis. Medication management is a complex task that requires synthesis of drug formulation and complete order instructions for safe use. Here, the performance of GPT-4o, an LLM available with OpenAI’s ChatGPT, was tested on 3 medication management tasks. Methods GPT-4o performance was tested on 3 medication tasks: identifying available formulations for a given generic drug name, identifying drug-drug interactions (DDIs) for a given medication regimen, and preparing a medication order for a given generic drug name. For each experiment, the model’s raw text response was captured exactly as returned and evaluated using clinician evaluation in addition to standard LLM metrics, including Term Frequency-Inverse Document Frequency (TF-IDF) vectors, normalized Levenshtein similarity, and Recall-Oriented Understudy for Gisting Evaluation (ROUGE-1/ROUGE-L) F1 score between each response and its reference string. Results For the first task of drug-formulation matching, GPT-4o had 49% accuracy for generic medications being matched to all available formulations, with an average of 1.23 omissions per medication and 1.14 hallucinations per medication. For the second task of drug-drug interaction identification, the accuracy was 54.7% for identifying the DDI pair. For the third task, GPT-4o generated order sentences containing no medication or abbreviation errors in 65.8% of the cases. Conclusion Model performance for basic medication tasks was consistently poor. This evaluation highlights the need for domain-specific training through clinician-annotated datasets and a comprehensive evaluation framework for benchmarking performance.

Abstract Background The accuracy and safety of generating medication orders by large language models (LLMs) must be demonstrated. Without standardization, performance evaluation is limited to time and resource-intensive clinician grading. This evaluation aimed to develop a standardized medication format that supports automated performance evaluation (MedMatch). Methods First, a survey of 40 medication prompts was given to clinicians to assess agreement in medication order communication. Second, a clinician panel developed a standardized medication format (MedMatch) for oral and intravenous medications. Third, a clinician-annotated dataset of medication prompts and standardized answers in the MedMatch format was developed for LLM testing. Finally, LLMs were retested with the same dataset, adjusted to exclude route information, to evaluate the appropriate categorization of medication route. Results The formal medication orders consistently showed low omission rates and high overlap for all entities, compared to the verbal and brief written communication types. Lexical overlap results demonstrated pattern norms amongst clinicians with entities appearing most commonly in positions 1-5 in the order of drug name, dose, unit, route, and frequency. In the second survey, the formal written group performed the highest with 78.3% of prompts considered appropriate as a computer-generated response. LLM accuracy on MedMatch order standardization was highest in oral solid (64.2-72.5%), intravenous intermittent (72.5-84.3%), and intravenous push (62.7-74.5%) categories. LLMs performed the worst at categorizing medication orders accurately into intravenous push (18-61%) and intravenous intermittent (51-100%) routes. Conclusions Standardized format for computer-based outputs may support automated performance analysis and enhance the clarity of medication communication.

Abstract Background Large language models (LLMs) are increasingly used in medication-related tasks despite limited evidence supporting their accuracy and safety. LLMs are vulnerable to adversarial attacks and may confabulate in response to fabricated information in the prompt. Given the substantial risk of harm from medication errors, the purpose of this study was to establish a benchmark task to evaluate LLM performance in distinguishing fabricated from real medications. Methods Two datasets (brand and generic), each consisting of 250 medication lists were developed, each including 4-6 medications and one fabricated medication (specifically a Pokémon character), with complete dosing information provided. Six LLMs were evaluated (GPT-5-Chat, GPT-4o-mini, DrugGPT, Gemma-3-27B-IT, Llama-3.3-70B-Instruct, and Qwen3-32B). For each medication list, models were queried for dosing information and disease indication across three experimental conditions: standard decoding, standard decoding with mitigation prompting, and deterministic decoding (temperature = 0). Each experiment was performed in triplicate. Outputs were deemed confabulations if the LLM failed to recognize and reject the fabricated medication. The primary outcome was the rate of confabulations. Exact paired-permutation tests were used to compare confabulation rates across LLMs and prompting approaches. Results Confabulation rates under standard conditions ranged from 2.7% to 99.6%. DrugGPT demonstrated the lowest baseline confabulation rates (2.7-6.4%), whereas other models demonstrated substantially higher rates (47.2-99.6%). Mitigation prompting significantly reduced confabulation rates compared to standard and deterministic decoding across all tasks and datasets (p < 0.001): GPT-5-Chat achieved confabulation rates of 0% with mitigation prompting from baseline rates of 66.4-88.5%. Deterministic decoding did not substantially impact confabulation rates. Conclusions LLMs performing medication-related tasks are susceptible to confabulation in response to invalid or fabricated inputs, raising important safety concerns for clinical use. While mitigation prompting reduced confabulation rates, this study underscores the importance of robust safeguards when deploying LLMs in medication-related decision-making.

Abstract Background The purpose of this study was to develop a series of benchmarks specifically for large language model (LLM) performance in the identification of drug-drug interactions (DDIs). Methods We evaluated six LLMs using a clinician-designed dataset of 750 DDI scenarios. Tasks included: (1) a pointwise two-drug classification task (2) a pairwise three-drug discrimination task and (3) a listwise 4–6 drug selection task. Zero-shot prompting was applied and performance was assessed using precision, recall, F1 score, accuracy, and self-consistency. Results Model performance varied by task structure and interaction severity. GPT-5-Chat demonstrated the highest recall and F1 score in the pointwise task (80.7% and 80.2%), whereas DrugGPT achieved superior accuracy in the pairwise (93.7%) and listwise task (92.6%). All models had improved performance with more severe compared to less severe DDIs (accuracy 72.0-99.2% vs. 58.3-86.7%). All models tended to have higher self-consistency in the pairwise and listwise experiments compared to the pointwise experiment (68.8-93.6% vs. 31.0-74.2%). No model exhibited uniformly high reliability across all tasks. Conclusions Current LLMs show inconsistent capabilities in identifying DDIs. Performance degrades as reasoning space expands with limited stability. These findings emphasize the need for multi-format evaluation frameworks and reliability-aware assessment when considering LLMs for medication-safety applications.