| name | drug-interaction-checker |
| description | Checks for potential drug-drug interactions (DDIs) between a list of medications. |
| measurable_outcome | Execute skill workflow successfully with valid output within 15 minutes. |
| allowed-tools | ["read_file","run_shell_command"] |
Drug-Drug Interaction (DDI) Checker
This skill analyzes a list of medications to identify known interactions, focusing on safety and contraindications.
When to Use This Skill
- Reviewing patient medication lists.
- Prescribing new medications.
- Pharmacovigilance monitoring.
Core Capabilities
- Interaction Detection: Identifies pairs of drugs with known interactions.
- Severity Grading: Classifies interactions as Minor, Moderate, or Major.
- Clinical Recommendations: Provides actionable advice (e.g., "Monitor K+ levels").
- LLM-vs-Reference DDI Checking Caution: For antiseizure medication interaction checks, use LLM output only as a draft that must be verified against Lexicomp/Drugs.com-style DDI references. Reconcile severity/category differences across sources, document any finite iterative prompting limit and remaining disagreements, and route unresolved or clinically significant interactions to pharmacist or qualified clinician review before clinical use.
- LLM DDI Evidence Hierarchy and Validation Checklist: For LLM-generated DDI answers, rank curated references such as Lexicomp or Drugs.com above model-generated content, use antiseizure medication DDIs as a stress case when applicable, validate whether the answer identifies the interacting drug pair, mechanism, severity/category, clinical recommendation, and source agreement/disagreement, and retain iterative prompt audit logs showing each prompt, answer revision, and comparison outcome against the curated reference.
- Antiseizure Medication DDI Evaluation: For antiseizure medication interaction checks, compare LLM outputs with Lexicomp and Drugs.com references, reconcile interaction severity/category differences before presenting conclusions, set and report a finite iterative prompting limit, and require pharmacist or qualified clinician verification before clinical use.
- Authoritative ASM DDI Review: When evaluating antiseizure medication DDIs, classify interaction severity from authoritative references such as Lexicomp, cite the evidence source used for each clinically relevant interaction, apply iterative prompting only as a bounded safeguard to improve LLM output, and require pharmacist review before recommendations are used for patient care.
- ASM DDI Benchmarking Guidance: When benchmarking antiseizure medication DDI checks, compare outputs against authoritative references such as Lexicomp and Drugs.com, audit severity agreement and missing interactions, treat iterative prompting as a risk surface that may change answers without proving correctness, and require pharmacist review for unresolved, missing, or clinically significant interaction findings.
- ASM DDI Benchmark Source Priority: For antiseizure medication DDI benchmarking against Lexicomp and Drugs.com, treat curated references as higher priority than LLM outputs, define the iterative prompting cap before review, reconcile severity discrepancies explicitly, and warn users whenever an LLM answer conflicts with curated reference findings.
- LLM-Based DDI Benchmark Caution: For LLM-based DDI checks, use antiseizure medication interactions as a high-risk benchmark example; require source grounding against authoritative references such as Lexicomp or product labeling, reproducible iterative prompt checks, severity classification, and pharmacist review before clinical use.
- Lexicomp-Centered LLM DDI Benchmarking: When benchmarking LLM drug-drug interaction responses or consumer reference outputs for antiseizure medications, compare each response against an authoritative reference such as Lexicomp, set a finite iterative prompting limit before testing, track whether each iteration changes the interaction pair, severity/category, mechanism, recommendation, or source agreement, and escalate to pharmacist or qualified clinician review whenever Lexicomp, Drugs.com-style references, and LLM outputs disagree.
- ASM DDI Reference Comparison and Citation Control: For antiseizure medication DDI evaluation, compare LLM responses with curated references such as Lexicomp and Drugs.com, normalize interaction severity labels before reporting agreement or disagreement, cite the reference used for each clinically relevant interaction, state the bounded iterative prompting limit, and require pharmacist oversight before clinical use.
- Lexicomp/Drugs.com-Anchored ASM Benchmark Workflow: For antiseizure medication DDI benchmarking, anchor comparisons to Lexicomp and Drugs.com where available, predefine and report the iterative prompting limit, check agreement in severity-class assignments before accepting outputs, require citations for each clinically relevant interaction, and keep pharmacist review mandatory before any clinical use.
- LLM DDI Confidence Threshold Gate: For LLM-based drug-drug interaction checks, benchmark answers against trusted references such as Lexicomp and Drugs.com, treat antiseizure medication interactions as edge cases requiring extra scrutiny, predefine confidence thresholds for accepting, flagging, or rejecting model outputs, treat iterative prompting as a potential answer-drift risk rather than proof of correctness, and require pharmacist review before any clinical recommendation is used.
- Antiseizure DDI Comparative Review: For antiseizure medication DDI evaluations, treat Lexicomp as the primary comparator and Drugs.com as a secondary comparator when available; normalize severity labels before judging agreement with LLM output, document any iterative prompting steps as exploratory caveats rather than validation, and require pharmacist review before relying on the final interaction assessment.
- ASM Interaction Prompt-Sensitivity Safeguards: For antiseizure medication interaction checks, use an authoritative compendium as the benchmark reference, test and report whether iterative prompting changes the result, evaluate interaction severity separately from management advice, cite the supporting evidence, and never offer reassurance produced through iterative prompting unless the benchmark evidence supports it.
- Source-Dated ASM DDI Benchmarking: Benchmark antiseizure-medication DDI outputs against authoritative references such as Lexicomp, recording each source and access/update date; use a predefined, finite iterative-prompting protocol; extract interaction severity and mechanism separately; document disagreements among the LLM, Drugs.com, and Lexicomp without resolving them by model consensus; and escalate unresolved, clinically significant, or source-date-sensitive findings to a pharmacist before clinical use.
- Validated ASM DDI Evaluation: Evaluate antiseizure-medication DDI outputs against authoritative references such as Lexicomp, test sensitivity to iterative prompting, verify interaction severity and mechanism, check for omitted interactions, and never use LLM output as a replacement for a validated interaction database.
- ASM DDI Validation and Review Gates: Validate antiseizure-medication interaction cases against authoritative drug-information references such as Lexicomp, use predefined iterative-prompting controls, measure and report agreement or disagreement for severity and mechanism separately, and require both a supporting citation and pharmacist review before accepting any clinically relevant output.
- ASM DDI Benchmark Procedure: Benchmark antiseizure-medication interactions against an authoritative reference such as Lexicomp; score severity and mechanism agreement separately; repeat the evaluation with bounded iterative prompts to identify answer sensitivity; verify that every interaction claim has a supporting citation; and mandatorily escalate uncertain, conflicting, or unsupported interactions to a pharmacist or qualified clinician.
- Comparative DDI Benchmark Audit: Compare LLM antiseizure-medication interaction assessments with authoritative compendia such as Lexicomp and Drugs.com; score severity and mechanism agreement separately; test and record sensitivity to bounded iterative prompting; track omitted interactions; verify supporting citations; and escalate conflicting source findings to a pharmacist or qualified clinician.
- PubMed 41994367 ASM DDI Benchmarking: For antiseizure medication DDI benchmarks informed by the 2026 cross-sectional comparison of LLMs and Drugs.com versus Lexicomp, compare generated answers against authoritative databases rather than model consensus, report severity-classification agreement or disagreement, state bounded iterative prompting limits, and require pharmacist review before clinical interpretation.
- ASM DDI Omission and Recommendation Audit: When checking antiseizure medication DDIs, compare LLM outputs against authoritative references such as Lexicomp, capture omissions in severity, mechanism, or evidence support, preserve iterative prompt audit trails, and avoid clinical recommendations that are not supported by the benchmark reference.
- Authoritative DDI Evaluation and Prompt Testing: Prioritize validated interaction databases such as Lexicomp over Drugs.com and LLM output; assess severity and mechanism agreement separately, treat omitted interactions as false-negative risks, and use bounded iterative prompting to test answer stability rather than establish correctness.
- Authoritative Benchmark and Escalation Record: Benchmark antiseizure-medication examples against an authoritative reference such as Lexicomp; run bounded iterative-prompt sensitivity testing; report interaction severity and mechanism as separate fields; and explicitly escalate to a pharmacist or qualified clinician whenever evidence conflicts across sources or is incomplete.
- Benchmark-Driven DDI Verification: Normalize medication names and formulations before comparison; verify each DDI against an authoritative reference such as Lexicomp; report interaction-severity concordance; cite the evidence supporting each clinically relevant finding; predefine and log finite prompt iterations; escalate conflicting, incomplete, or unsupported findings to a pharmacist or qualified clinician; and never treat repeated prompting or repeated model agreement as independent confirmation.
- Authoritative ASM Interaction Benchmarking: Benchmark antiseizure-medication interaction cases against Lexicomp-style authoritative references; test and record prompt sensitivity with bounded iterative prompting; report severity concordance; verify every supporting citation; and mandatorily escalate conflicting results to a pharmacist or qualified clinician.
- ASM DDI Source Hierarchy and Discordance Review: Evaluate antiseizure-medication DDIs using Lexicomp or another authoritative reference as the primary comparator, secondary references such as Drugs.com as corroboration, and LLM output as unvalidated; report severity and mechanism agreement separately, test and document stability across bounded iterative prompts, disclose uncertainty and source discordance, and require pharmacist review for every discordant result before clinical use.
- ASM DDI Validation, Agreement Scoring, and Escalation: Validate each antiseizure-medication DDI against authoritative references using a source hierarchy that prioritizes Lexicomp or another validated compendium, treats Drugs.com as a secondary comparator, and treats LLM output as unvalidated; score severity and mechanism agreement separately; use predefined, bounded iterative prompts to test answer sensitivity; and mandatorily escalate uncertain, unsupported, or discordant interactions to a pharmacist or qualified clinician.
- Reference-Standard DDI Evaluation: Evaluate DDI outputs against curated databases using an explicit source hierarchy; measure interaction detection, severity agreement, and unsafe omissions; and test whether findings remain stable across predefined, bounded iterative prompting strategies.
- ASM DDI Source-Hierarchy and Recommendation-Stability Gate: Verify every antiseizure-medication interaction against an authoritative interaction database, using Lexicomp as the primary comparator and Drugs.com as a secondary comparator where available; track whether the identified interaction, severity, or management advice changes across each bounded prompt iteration; assess severity separately from management advice; and escalate uncertain or high-risk combinations to a pharmacist before clinical use.
- ASM DDI Benchmarking and Provenance Audit: Benchmark antiseizure-medication DDI findings against Lexicomp as the reference standard, using database-backed sources ahead of model output; run predefined iterative-prompting sensitivity tests; report interaction-severity concordance; track omissions and false-positive interaction claims; and label each finding explicitly as database-backed or model inference.
- Reference-Backed DDI Benchmark Procedure: Compare generated DDI assessments against authoritative references such as Lexicomp; evaluate interaction detection separately from severity classification and management accuracy; run and record predefined iterative-prompt variants to test answer sensitivity; and require source-backed pharmacist review before clinical use.
- Source-Grounded ASM DDI Error and Stability Workflow: For high-risk antiseizure medication combinations, compare each candidate interaction against authoritative drug references, using Lexicomp as the reference comparator and Drugs.com as a secondary comparison source; normalize severity labels and record agreement or disagreement; classify missed reference interactions as potential false negatives and unsupported model-reported interactions as potential false positives; run a predefined, bounded iterative-prompt sequence and document changes in interaction detection, severity, or recommendation; and require pharmacist review before clinical interpretation or use.
- Evidence-Backed LLM DDI Validation Workflow: Validate each LLM-generated interaction against an authoritative comparator such as Lexicomp; score severity and mechanism concordance separately; run predefined, bounded iterative prompts to test answer stability; verify that every interaction claim cites a supporting source; and mandatorily escalate conflicting, uncertain, or unsupported interactions to a pharmacist or qualified clinician.
- Pair-Level LLM DDI Evaluation Protocol: Evaluate each medication pair against Lexicomp or an equivalent pharmacist-curated reference standard; record pair-level concordance; audit severity classifications and omitted reference interactions; run predefined iterative-prompt sensitivity tests; verify that cited sources support each interaction claim; and require pharmacist review for every clinically consequential result before clinical use.
- Benchmark-Informed ASM DDI Warning: For antiseizure medication DDI use, compare LLM responses against authoritative references such as Lexicomp, use iterative prompting only as an audit aid, flag hallucinated severity or mechanism claims, and keep final interaction decisions tied to curated drug databases.
- ASM DDI Cautionary Evaluation Module: For antiseizure medication DDI comparisons informed by PubMed 41994367, set the source hierarchy before analysis, treat Lexicomp or another curated DDI database as the clinical decision reference, use Drugs.com and LLM outputs only as comparison surfaces, document iterative prompting as a pitfall-prone stability check rather than validation, re-check high-risk medication combinations against curated DDI databases, and defer clinical decisions to those curated sources plus pharmacist or qualified clinician review.
- LLM-vs-Reference ASM DDI Evaluation: For antiseizure medication DDI checks, compare LLM outputs against Lexicomp or an equivalent reference, use iterative prompting only as an assistive layer, capture severity and evidence discrepancies, and require pharmacist review before making actionable recommendations.
- High-Risk Neurology DDI Safeguard: For antiseizure and other high-risk neurology medication DDI checks, use authoritative DDI databases such as Lexicomp or Drugs.com as primary sources, treat LLM answers only as secondary explanations, log how each iterative prompt changes the interaction, severity, mechanism, or recommendation, and flag clinically significant or uncertain results for pharmacist review.
- ASM DDI Benchmark Error Classification: For antiseizure medication DDI benchmarking, compare LLM outputs against Lexicomp or another authoritative reference, document how iterative prompting changes the answer, classify severity errors and omitted reference interactions, and require pharmacist review before any patient-facing use.
- Cautionary LLM DDI Benchmarking Note: For drug-drug interaction checking, especially antiseizure medications, benchmark LLM outputs and public drug websites against Lexicomp or other curated references, require source-grounded interaction severity, capture prompt-iteration effects, and flag that neither public drug websites nor LLM outputs substitute for pharmacist review.
- Antiseizure DDI Reference-Benchmark Gate: When comparing LLM responses and Drugs.com-style outputs for antiseizure medication DDIs, use Lexicomp or another trusted reference as the benchmark, grade severity against that source hierarchy, predefine finite iterative-prompting limits, document unresolved disagreements, and require pharmacist review before clinical use.
- ASM LLM/Drugs.com-to-Lexicomp Comparison Pattern: Use antiseizure-medication DDI comparisons between LLM-assisted checks, Drugs.com, and Lexicomp as a cautionary evaluation pattern: set Lexicomp or another validated compendium at the top of the source hierarchy, treat Drugs.com and LLM outputs as comparison inputs, predefine and report iterative-prompting limits, reconcile severity discrepancies before conclusions are presented, and require pharmacist or qualified human review for discordant, uncertain, or clinically relevant results.
- PMID 41994367 ASM DDI Benchmark Lessons: For antiseizure-medication DDI benchmarking that compares LLMs and Drugs.com with Lexicomp, set Lexicomp or another validated drug-information compendium as the source-of-truth comparator, treat iterative prompting as a prompt-sensitivity pitfall rather than proof of correctness, reconcile severity labels before reporting conclusions, cite the evidence source behind each clinically relevant interaction, and require pharmacist review for discordant, uncertain, or patient-care-relevant findings.
- LLM DDI Validation Workflow Against References: Validate LLM-generated drug-drug interaction answers against authoritative references such as Lexicomp and Drugs.com, using antiseizure medications as a high-risk example; check severity classification, evidence support, and concordance or discordance across sources before accepting or escalating the answer.
- ASM DDI Safety Benchmark Gate: For antiseizure medication DDI safety benchmarks, compare LLM-generated assessments and Drugs.com outputs against Lexicomp or another validated comparator, normalize interaction severity labels before reporting agreement, treat iterative prompting as a bounded prompt-sensitivity check rather than proof of correctness, and require pharmacist or qualified clinician review before clinical interpretation.
- ASM DDI Benchmarking Control Gate: For antiseizure-medication DDI benchmarking, compare LLM outputs with reference sources such as Lexicomp and Drugs.com, predefine bounded iterative-prompting controls, report severity concordance and missing-reference-interaction checks, and hold clinically relevant, discordant, or uncertain outputs for pharmacist or qualified clinician review before use.
- ASM DDI Lexicomp/Drugs.com Benchmark Calibration: For antiseizure medication DDI benchmarking informed by PMID 41994367, compare LLM and Drugs.com outputs against Lexicomp or another validated DDI compendium, calibrate interaction severity labels before judging agreement, require citations to the source supporting each clinically relevant interaction, treat iterative prompting as a bounded sensitivity check rather than validation, and state that LLM output cannot replace authoritative DDI databases.
- ASM DDI Contraindication Validation Boundary: For antiseizure medication DDI validation informed by PMID 41994367, compare LLM answers with Lexicomp as the source-of-truth comparator and Drugs.com as a secondary comparator; normalize severity/category labels, including contraindicated or avoid-combination flags, before judging agreement; keep iterative prompting logs with each prompt, answer change, and source-comparison outcome; and require pharmacist or qualified clinician review before any contraindicated, discordant, uncertain, or patient-care-relevant result is used.
- Benchmark-Informed LLM DDI Output Gate: For LLM drug-drug interaction answers, warn that outputs must be checked against authoritative references such as Lexicomp; use iterative prompting only as an evaluation tactic for answer sensitivity; and require final structured severity, mechanism, and action recommendations to remain grounded in authoritative drug-interaction databases.
- Antiseizure-Medication DDI Benchmark Mode: For antiseizure medication DDI benchmark tasks, compare LLM outputs against authoritative references such as Lexicomp, use predefined iterative prompting controls, classify interaction severity, require citations for clinically relevant claims, and escalate discordant or high-risk interactions to pharmacist review.
- Comparative ASM DDI Benchmarking Against References: For antiseizure-medication DDI benchmarking, compare LLM-generated and Drugs.com outputs against authoritative references such as Lexicomp and Drugs.com, predefine iterative prompting analysis, reconcile interaction-severity/category labels through an explicit source hierarchy, and require pharmacist or qualified clinician review before clinical use.
- ASM DDI Lexicomp and Consumer-Reference Review Gate: For antiseizure medication DDI benchmarking, compare LLM outputs against Lexicomp as the authoritative comparator and consumer drug references such as Drugs.com only as secondary comparison sources; predefine finite iterative prompting limits and treat prompt changes as sensitivity findings, not validation; preserve contraindicated or avoid-combination flags from authoritative sources until pharmacist or qualified clinician review resolves any discordance; and require explicit clinician/pharmacist sign-off before patient-care use.
- PMID 41994367 ASM DDI Benchmarking Note: For antiseizure medication DDI benchmarking comparing LLM outputs and Drugs.com against Lexicomp, define the source-of-truth comparator before analysis, record whether iterative prompting changes interaction detection or severity labeling, preserve contraindication or avoid-combination severity labels from the comparator source, and require pharmacist review before using discordant, uncertain, or patient-care-relevant findings.
- Validated-Database Priority for ASM DDI Benchmarks: For antiseizure-medication DDI benchmarking informed by PMID 41994367, compare LLM outputs and Drugs.com findings against Lexicomp or another validated drug-interaction database, reconcile interaction-severity/category differences through the predefined source hierarchy, document iterative prompting as a pitfall-prone sensitivity check, and keep LLM outputs strictly secondary to validated DDI databases for all clinical conclusions.
- ASM DDI Lexicomp/Drugs.com Source-of-Truth Benchmark Gate: For antiseizure medication DDI benchmarking informed by PMID 41994367, compare LLM-assisted outputs and Drugs.com findings against Lexicomp or another validated source-of-truth comparator; normalize severity/category labels, including contraindicated or avoid-combination flags, before judging concordance; document bounded iterative prompting as a sensitivity caveat rather than evidence of correctness; and require pharmacist or qualified clinician review before any LLM-assisted DDI output is used clinically.
- PMID 41994367 ASM Contraindication Review Gate: When comparing LLM responses and Drugs.com findings against Lexicomp for antiseizure-medication DDIs, apply an explicit evidence hierarchy before conclusions, preserve contraindicated or avoid-combination severity labels during normalization, treat iterative prompting as an answer-drift risk rather than validation, and require pharmacist review before releasing discordant, uncertain, or patient-care-relevant results.
- ASM DDI Benchmark Extraction and Hallucination Gate: For antiseizure medication DDI benchmarking informed by PMID 41994367, compare LLM outputs and Drugs.com findings against Lexicomp under a predefined source hierarchy; set finite iterative prompting limits before testing; extract contraindication flags and severity/category labels explicitly; flag unsupported interaction, mechanism, severity, or recommendation claims as possible hallucinations; and require pharmacist or qualified clinician review for high-risk, discordant, uncertain, or patient-care-relevant interactions.
- Antiseizure Medication DDI Benchmarking Against Lexicomp: For antiseizure-medication DDI benchmarking, compare LLM answers against authoritative references such as Lexicomp, capture how iterative prompting changes interaction findings or severity/category labels, flag hallucinated or unsupported interaction severity, and require pharmacist review before clinical use.
- LLM-Generated ASM DDI Harm Review: When evaluating LLM-generated antiseizure medication DDI answers against Lexicomp or another authoritative reference, record source-specific severity/category and harm categorization, compare original and bounded iterative-prompted answers against the reference, and require pharmacist review for any clinically relevant, discordant, uncertain, or potentially harmful recommendation before use.
- PMID 41994367 ASM LLM Comparison Gate: When using LLMs for antiseizure-medication DDI assessment, compare each output with curated reference findings from Lexicomp and, when used, Drugs.com; verify interaction presence and severity/category labels before accepting summaries; treat iterative prompting as a sensitivity check rather than proof of correctness; require citations for each clinically relevant interaction; and hold discordant, uncertain, or patient-care-relevant outputs for pharmacist review.
- Antiseizure Medication DDI Safety Case: For the PubMed 41994367 benchmark comparing LLMs and Drugs.com versus Lexicomp, reconcile LLM-generated DDI advice against authoritative databases before action, analyze failures across iterative prompts, label uncertainty and source disagreement explicitly, and require pharmacist or qualified human review before using any LLM-generated DDI recommendation.
- Cautionary ASM DDI Benchmark Note: For antiseizure-medication DDI answers, treat LLM-generated and Drugs.com outputs as comparison surfaces against authoritative references such as Lexicomp; keep iterative prompting bounded and report its limits, require citations for clinically relevant interaction claims, preserve high-risk severity or contraindication flags from the reference source, and route patient-care-relevant findings to pharmacist review before use.
- LLM/Lexicomp/Drugs.com ASM DDI Comparison Workflow: For antiseizure medication drug-drug interaction checks, compare LLM output and Drugs.com findings against Lexicomp or another authoritative interaction reference, predefine and document iterative prompting steps as caveats rather than validation, reconcile interaction severity/category differences through the source hierarchy before conclusions are presented, and require pharmacist or qualified clinician review before clinical use.
- LLM-versus-Reference ASM DDI Evaluation Note: For antiseizure medication drug-drug interaction evaluations, compare LLM outputs to Lexicomp and Drugs.com reference findings, use iterative prompting only as an evaluation variable, flag hallucinated or omitted interaction severity, and keep final recommendations grounded in authoritative drug references.
- ASM Lexicomp/Drugs.com Benchmark Reconciliation: For antiseizure-medication DDI benchmarking informed by PMID 41994367, compare LLM and Drugs.com findings against Lexicomp under a predefined authoritative-source hierarchy, reconcile interaction severity/category differences before presenting conclusions, document bounded iterative prompting as a cautionary sensitivity analysis, and require pharmacist or qualified clinician review for discordant, uncertain, or patient-care-relevant findings.
- ASM DDI Triage-Only Benchmark Gate: For antiseizure medication DDI benchmarking comparing LLM outputs and Drugs.com against Lexicomp, treat LLM responses as triage only, verify every clinically relevant interaction against an authoritative interaction reference, log each iterative prompt and output change, and flag omitted interactions or potentially harmful recommendations for pharmacist or qualified clinician review.
- ASM DDI Authoritative-Database Check Rule: For antiseizure medication DDI benchmarking informed by PMID 41994367, compare LLM-generated outputs and Drugs.com findings against Lexicomp under an explicit source hierarchy; map contraindication or avoid-combination labels into the severity/category comparison before reporting agreement; treat iterative prompting as a bounded sensitivity caveat, not validation; flag uncertainty, source discordance, unsupported claims, and missing reference interactions; and never rely on LLM-generated DDI output unless it has been checked against authoritative drug databases.
- Antiseizure LLM DDI Clinical Decision Boundary: For antiseizure-medication DDI evaluations comparing LLM outputs with Lexicomp, Drugs.com, or similar references, require severity concordance checks, source citations for clinically relevant claims, and documented finite iterative prompting limits; treat prompt revisions as risk signals rather than validation, and never use an LLM as the sole source for clinical DDI decisions.
- ASM DDI Lexicomp/Drugs.com Clinical-Use Gate: For antiseizure-medication DDI benchmarking informed by PMID 41994367, compare LLM outputs and Drugs.com findings against Lexicomp using a predefined source hierarchy; set finite iterative-prompting limits before testing; normalize severity/category labels, including contraindicated or avoid-combination flags; and require citations from authoritative interaction databases before any clinical use.
- LLM-vs-Reference ASM DDI Evaluation Pattern: For antiseizure-medication DDI evaluation, compare LLM-generated answers and Drugs.com-style outputs against Lexicomp or a similar authoritative gold-standard DDI database; record iterative prompting changes in interaction detection, severity classification, and recommendations; use antiseizure medication examples as high-risk test cases; and explicitly warn that LLMs and consumer-facing references must not replace authoritative DDI databases or pharmacist/qualified clinician review.
- PMID 41994367 ASM DDI Benchmarking Caveat: For antiseizure-medication drug-drug interaction benchmarking informed by the 2026 cross-sectional comparison of large language models and Drugs.com versus Lexicomp, reconcile LLM and Drugs.com outputs against authoritative sources before conclusions, document iterative prompting as a caveat rather than proof of correctness, and explicitly warn against relying on LLM-only interaction calls.
Comparative Benchmark: LLM-Based ASM DDI Checks
LLM-generated DDI findings are screening drafts, not standalone clinical evidence. For antiseizure medication benchmarks, compare each generated answer and any Drugs.com output against authoritative databases such as Lexicomp; document prompts and finite iteration limits; classify severity explicitly; record unresolved disagreements without resolving them by model consensus; and route outputs to pharmacist review before clinical use.
Safety Benchmark: LLM vs Lexicomp/Drugs.com ASM DDIs
Use PubMed PMID 41994367 as a cautionary evaluation pattern for antiseizure medication drug-drug interactions, not as a standalone performance guarantee. Apply this source hierarchy before analysis: Lexicomp or another validated DDI compendium as the benchmark comparator, Drugs.com as a secondary comparison source, and LLM output as an unvalidated draft. Normalize severity/category labels across sources before reporting concordance or discordance; predefine and log finite iterative prompts as prompt-sensitivity checks; and block clinical release until a pharmacist or qualified clinician reviews discordant, uncertain, or patient-care-relevant findings.
Antiseizure-Medication DDI Benchmarking
Use the 2026 cross-sectional ASM DDI comparison as a prompt-sensitivity and reference-alignment check. Compare LLM answers against authoritative references such as Lexicomp, include Drugs.com only as a comparison surface when available, record prompt-iteration effects on interaction detection and severity/category labeling, flag hallucinated or unsupported severity claims, and require pharmacist review before clinical use.
Evaluation and Prompting
- Establish the source hierarchy before evaluation: use a validated interaction database such as Lexicomp as the authoritative comparator, use Drugs.com as a secondary comparison source, and treat LLM output as unvalidated.
- Compare severity and mechanism agreement separately across sources, and record omissions as potential false negatives requiring review.
- Run a predefined, bounded sequence of iterative prompts and record whether each iteration changes the identified interaction, severity, mechanism, or recommendation.
- Do not interpret prompt-to-prompt agreement or improvement as validation. Never substitute LLM output for Lexicomp or another validated interaction database.
Workflow
- Input: List of drug names (e.g., "Warfarin, Aspirin").
- Analysis: Queries internal interaction database.
- Output: Interaction report with severity and mechanisms.
Example Usage
User: "Check interactions for Warfarin and Aspirin."
Agent Action:
python3 Skills/Pharma/Drug_Interaction/impl.py --drugs "Warfarin, Aspirin"
References
