Drug Interaction Screening¶
Screening assessment — instantiates Antagonism Screening and Separation
Checks whether medications, treatments, or supplements create harmful combined effects before they are co-prescribed or co-used.
Drug Interaction Screening is the clinical assessment that, before two or more therapies are used together, checks the proposed combination against known interaction knowledge, identifies the pharmacologic mechanism at play, grades how dangerous the interaction is, and returns a graded recommendation — avoid, substitute, adjust the dose, or co-use under defined monitoring. Its defining move is that it is a human-facing assessment with a graded, conditional verdict: unlike a machine gate that simply permits or blocks, its most valuable output is the middle path — "this pair is acceptable with increased monitoring" — which keeps a needed therapy in play while surveilling the residual risk. It is a check performed on a specific proposed combination, in a specific patient context, at the moment of prescribing.
Example¶
A patient stabilized on warfarin, a narrow-therapeutic-index anticoagulant, is newly prescribed an antibiotic for an infection. At the point of order entry the pharmacy's screening flags the pair: the antibiotic potentiates warfarin's effect — by inhibiting the liver enzymes that clear it and disrupting the gut flora that supply vitamin K — so the combination raises bleeding risk. That mechanism is the antagonism condition; the screen classifies the severity as major, backed by well-established pharmacology.[n1] The recommendation is graded rather than binary: either substitute an antibiotic with no such interaction, or continue the pair with more frequent INR checks and a downward warfarin dose adjustment until the course finishes. The pharmacist chooses, documents the decision, and — if the pair is kept — sets the monitoring that watches the accepted risk. The dangerous combination is caught before the first co-administered dose, and the patient still gets treated.
How it works¶
The assessment matches the proposed set of therapies against an interaction knowledge base. For each hit it retrieves the mechanism — pharmacokinetic (one drug changing another's absorption, metabolism, or clearance) or pharmacodynamic (additive or opposing effects on the same system) — and classifies severity, from contraindicated down to minor, tagged with the quality of the underlying evidence. It then returns a proportionate recommendation and, for any pair accepted under caution, institutes the monitoring (laboratory values, symptom checks) that watches the residual risk over the course. Overrides are logged, so a clinician's decision to proceed leaves an auditable trail rather than a silent dismissal.
Tuning parameters¶
- Alert threshold — which severity tiers actually interrupt the prescriber. A low threshold surfaces everything but drives alert fatigue; a high one is quiet but can miss a moderate interaction that mattered for this patient.
- Knowledge-base breadth vs. specificity — a broad base catches rare pairs but fires on many theoretical, clinically-irrelevant ones; a curated base is quieter but can have gaps.
- Context sensitivity — whether the screen accounts for dose, route, renal/hepatic function, and age, or flags the pair abstractly regardless of whether it bites at this dose.
- Override friction — how much justification a clinician must supply to proceed past an alert; more friction deters unsafe overrides but slows every legitimate one.
- Monitoring intensity — how closely an accepted-with-caution pair is watched (which labs, how often), traded against burden on patient and system.
When it helps, and when it misleads¶
Its strength is catching dangerous combinations at the point of decision, before harm, with a graded response that neither waves everything through nor bans a useful therapy outright — the "keep it, but monitor" path preserves value that a blunt block would destroy.
Its signature failure is alert fatigue: fire too many low-value alerts and clinicians learn to click through all of them, including the one that mattered.[n2] The knowledge base can also be incomplete, and a context-blind screen flags interactions that are real in a textbook but harmless at this patient's dose, eroding trust. The classic misuse is a hard, undismissable alert on every theoretical pair, which trains the whole staff to override reflexively. The discipline that keeps it honest is to tier alerts by severity and evidence, tune them to patient context, and monitor override patterns so the screen's own noise is treated as a defect to fix, not a badge of thoroughness.
How it implements the components¶
compatibility_check— the assessment is the check: for a specific proposed set it asks whether the therapies can be co-used safely and records the answer with its evidence and rationale.antagonism_condition— it names the pharmacologic mechanism (enzyme inhibition, additive effect on a shared system) that makes the pair harmful, so the response fits the cause rather than a bare label.interaction_severity_rating— it grades each hit from contraindicated to minor, so alerting and the recommended response are proportional to the danger.residual_risk_monitoring— for pairs kept under caution, it sets up the ongoing labs and symptom checks that watch the accepted residual risk over the treatment course.
It assesses, grades, and monitors but does not itself impose a machine hard-block (exclusion_rule — Dependency Conflict Detection and Material Incompatibility Rule), stagger the therapies apart on a schedule (sequencing_rule — Schedule Conflict Prevention), or route a role conflict to accountable approval (exception_review_path — Separation-of-Duties Review). Its nearest twin is Dependency Conflict Detection — both catch a harmful pairing before combination — but Dependency Detection is an automated resolver that computes co-activation (co_activation_detector) and fails closed, whereas Drug Interaction Screening is a clinician-facing assessment that classifies severity and can permit the pair under monitoring.
Related¶
- Instantiates: Antagonism Screening and Separation — it is the archetype's clinical screen, turning interaction knowledge into a point-of-care decision.
- Consumes: Conflict Matrix — the curated interaction knowledge base it checks against is a compiled register of known-antagonistic pairings.
- Sibling mechanisms: Conflict Matrix · Dependency Conflict Detection · Material Incompatibility Rule · Schedule Conflict Prevention · Separation-of-Duties Review · Compatibility Matrix · Quarantine or Isolation Protocol
Editorial Notes¶
Form Classification¶
Form family: Assessment, Review & Assurance
Rationale: Drug Interaction Screening operates as a bounded evaluation of existing evidence or work that produces a finding or disposition because it checks whether medications, treatments, or supplements create harmful combined effects before they are co-prescribed or co-used.
Independent corroboration: The frozen evidence defines Drug Interaction Screening as 'Checks whether medications, treatments, or supplements create harmful combined effects before they are co-prescribed or co-used', so its operative form is Assessment, Review & Assurance.
Review outcome: Independent reviewer agreement; high confidence.
Origin Attribution¶
Primary origin: Pharmacology & Toxicology
Origin pattern: Single lineage
Present-day reach: Specialized
Rationale: Clinical pharmacology cohered mechanism- and severity-based screening for harmful pharmacokinetic and pharmacodynamic interactions before co-use.
Related originating lineages:
- Medicine & Healthcare — Clinical prescribing practice institutionalized patient-specific screening, alert tiering, pharmacist review, substitution, and monitoring at order time.
Review resolution: Pharmacology is primary because interaction mechanisms and exposure effects are the screen's substance; clinical medicine supplies prescribing context and action thresholds.
Review outcome: Reconciled after independent review; high confidence.
Notes¶
[n1] The cytochrome P450 (CYP450) enzyme family metabolizes a large share of common drugs; when one drug inhibits or induces the CYP pathway that clears another, blood levels of the second drug swing, which is a principal source of pharmacokinetic drug–drug interactions. It is the standard worked example of why a mechanism, not just a flag, belongs in the screen. ↩
[n2] Alert fatigue — the desensitization that sets in when a system issues too many low-value alerts, so that operators begin to ignore or reflexively override them, including the rare critical one. It is a well-documented hazard of clinical decision-support systems and the reason severity-tiering and context-tuning are part of a good screen rather than optional polish. ↩