A Liverpool interaction warning is a starting point for a decision, not the decision itself. The clinically useful information sits behind the colour: which medicines were checked, what interaction is expected, how strong the evidence is and what action the explanation actually supports. A screenshot without that context is a weak prescribing record.
The Liverpool HIV Drug Interactions resource, checked on 19 September 2026, provides a specialist route for evaluating combinations involving HIV medicines. Its prescribing resources also distinguish interaction explanations, treatment selectors and pharmacokinetic information. Those resources serve different questions; they should not be collapsed into a generic "safe or unsafe" label.
First check that the medicines are really the medicines
Start with an original fictional example. An adult with well-controlled HIV is offered a new inhaler. The primary-care record says "antiretroviral treatment from hospital", and the patient remembers the appearance of the tablet but not its name. A colleague finds a reassuring result for a familiar HIV medicine and proposes documenting that no important interaction was found.
The check is premature. The selected medicine has not been shown to be the patient's medicine. Confirm the active ingredients, including components of combination products, the route and the exact new preparation. Use the appropriate clinical records or pharmacy and specialist communication channels rather than a visual guess.
Also establish whether a listed medicine is current, recently stopped, proposed but not started, or taken intermittently. Record non-prescription products and supplements where relevant. A technically correct interaction result for the wrong treatment list is still an incorrect assessment of the situation.
Separate four questions that the warning may answer
The first question is identity: what pair, formulation and route does the entry cover? The second is mechanism: why might one medicine alter the effect or exposure of the other? The third is evidence: was the combination studied, reported clinically or assessed using pharmacological reasoning? The fourth is action: what does the entry recommend in the circumstances described?
Keeping these separate prevents a common shortcut. An explanation of a plausible mechanism is not automatically an instruction to discontinue treatment. Conversely, limited direct evidence does not automatically make a potentially serious interaction unimportant. The source's evidence assessment and action wording need to be read together.
In your notes, use four plain headings: checked combination, expected effect, basis of evidence, proposed response. Write the response in your own words after reviewing the full explanation. This is a practical reading method, not an alternative grading system for Liverpool's classifications.
Do not let the colour erase the conditions
Where the resource uses coloured categories, read the accompanying text and legend. Do not translate every warning into the same action, and do not describe an absence of an expected interaction as a guarantee of safety in every respect.
A drug-pair checker is not necessarily assessing whether the new medicine is indicated, whether the patient's symptoms require a different assessment or whether the patient can follow a proposed monitoring arrangement. Those questions remain even after the interaction has been understood.
For the fictional inhaler enquiry, the useful next step is to confirm the regimen, run the correct check and examine the relevant route-specific explanation. Liverpool's public resource collection explicitly includes material on corticosteroids and on non-oral steroids with strong metabolic inhibitors, as listed on 19 September 2026. That identifies a relevant place to investigate; it is not a conclusion about the unidentified combination in the example.
Ask whether the proposed response is feasible
Suppose the verified entry supports a managed combination in specified circumstances. The next question is not simply whether a monitoring plan can be written down. It is whether the responsible service can deliver it, whether the patient understands it and whether the relevant information will reach the clinicians involved.
Alternatively, the entry may support avoiding a combination or considering another option. That still requires an appropriate prescribing decision, particularly when one medicine is part of a specialist regimen. Do not independently alter an established HIV regimen to accommodate a new prescription without the necessary specialist input.
For a difficult enquiry, state exactly what remains unresolved: uncertainty about the preparation, a mismatch between the published population and the patient, conflicting product information, or an action that is not feasible locally. "The checker is amber" leaves the recipient to reconstruct the problem.
Document the decision, not just the search
A stronger fictional record would say: "Current treatment confirmed with the specialist medication record. Exact proposed preparation checked in the Liverpool resource on the recorded date. The interaction explanation and evidence assessment were reviewed. Specialist advice requested on the appropriate alternative or monitoring approach before finalising the new prescription; responsibility for contacting the patient remains with the prescribing team."
That wording deliberately leaves the actual clinical conclusion open. It is an example of documentation structure, not advice to withhold every medicine while awaiting routine correspondence. The urgency of the underlying condition must shape the communication and interim plan.
When a conclusion is reached, retain enough information to explain it: the combination, source date, key qualification, advice received and who will act. An imported warning becomes useful only when it is connected to an accountable plan.
What belongs in the learning record
This article is published by iatroX and includes its learning tools as a complement to Liverpool, not as a validated replacement for a specialist interaction database. The September 2026 iatroX brief describes question-specific Socratic tutoring that asks learners to explain their reasoning.
A focused learning question could be: "What information would make this interaction assessment more or less applicable?" That explores pharmacology and uncertainty without uploading a patient's medication record. The clinical combination should still be checked in the appropriate specialist resource.
For a clinician needing a drug-pair assessment, start with Liverpool and relevant product information. For a learner who cannot explain the mechanism or the qualification, targeted teaching may help. Neither task is improved by pretending that the two products perform the same job.
Frequently asked questions
Is the traffic-light category enough to document an interaction check?
No. Record the verified combination, the relevant explanation and the action taken, including any qualification or specialist advice.
Does a reassuring interaction result mean the prescription is appropriate?
Not necessarily. Indication, patient circumstances, other medicines and the feasibility of follow-up remain separate clinical questions.
Should I paste a patient's full medication record into a learning chatbot?
Use approved clinical systems for identifiable records and keep general educational prompts free of patient-identifiable information.