A CredibleMeds category describes the evidence linking a medicine with QT prolongation or torsades de pointes. It does not calculate the probability that a particular patient will develop an arrhythmia. The practical question is therefore not simply whether a medicine appears on a list, but what that classification means in the circumstances of its proposed use.
Three categories, three different statements
In its category explanation, published on 17 March 2016 and checked on 19 September 2026, CredibleMeds distinguishes Known, Possible and Conditional Risk. Known Risk means that QT prolongation and an association with torsades are established during recommended use. Possible Risk identifies QT prolongation without sufficient evidence of torsades during recommended use. Conditional Risk identifies torsades in particular circumstances, or a medicine's ability to create circumstances that facilitate it.
These are not three numerical bands of absolute risk. CredibleMeds also cautions against ranking medicines within a category. A separate list addresses medicines to avoid in congenital long QT syndrome; it should not be treated as interchangeable with the general classification.
An easy but consequential misreading is to translate "possible" into "probably harmless" and "conditional" into "less important". Neither follows. The condition that matters may already be present in the person being assessed.
Start with the patient, then inspect the entry
Imagine a fictional patient attending a medicines review after an intercurrent illness. The repeat list looks familiar, but an additional medicine was prescribed elsewhere. The patient cannot remember its name, and the most recent blood results predate the illness.
A superficial review would look up each familiar prescription, find no new concern and move on. A better first task is to reconcile the actual medicines and establish which clinical information remains current. An interaction involving the additional medicine cannot be assessed reliably while its identity is unknown.
The point of the example is information quality, not a recommendation about an unnamed medicine. Record what is confirmed, what is reported by the patient and what must be obtained from another source. A risk category applied to an incomplete medicine list can create misplaced reassurance.
CredibleMeds' clinical-factor resource, checked on 19 September 2026, separately considers factors associated with prolonged QT or torsades, including bradycardia and electrolyte disturbances. Its evidence assessments should prompt a contextual review rather than be converted into an unofficial additive score.
Distinguish four questions before deciding what to do
First, is the concern about QT prolongation, torsades, a pharmacokinetic interaction or a condition that makes an adverse event more likely? Those concepts overlap without being identical.
Second, does the information concern the exact medicine and formulation being considered? Similar names, combination products and incomplete histories can derail the lookup before any clinical interpretation begins.
Third, what is the applicable prescribing information and local process? A specialist classification can inform a decision, but it does not replace the relevant product information, patient assessment or advice from an appropriate pharmacist or specialist.
Fourth, what decision must be made now? Initiating treatment, reviewing an established prescription, investigating symptoms and reconciling a discharge list are different tasks. They may require different information and different urgency even when the same medicine is involved.
A useful record separates those questions instead of compressing them into "QT risk checked". That phrase confirms almost nothing about what was actually considered.
A documentation example that preserves uncertainty
For the fictional review, an educational record might read:
"Medicine reconciliation remains incomplete because the recently issued product has not been identified. The current CredibleMeds classification and product information will be checked against the confirmed list. Relevant recent clinical findings and investigations will be reviewed before an individual plan is agreed. Responsibility for obtaining the missing information and communicating the decision has been assigned."
This is a writing example, not a completed patient record. Replace each general statement with facts only after they are established. Do not copy it into a clinical record as evidence that checks have occurred.
Notice what the wording avoids. It does not call the combination safe, claim a normal tracing that has not been reviewed, or attribute a symptom to a medicine without assessment. It states what blocks the decision and who will resolve the uncertainty.
What to do when another source seems to disagree
Before concluding that two resources conflict, compare their questions. One may classify evidence of torsades, while another describes a precaution, an interaction mechanism or a manufacturer warning. Different outputs can be compatible because they serve different purposes.
Then check the publication or review date, the population and any qualifying circumstances. Preserve the exact point of disagreement in your notes. "The sources differ" is less useful than "one source describes an interaction affecting exposure, while the other does not address that combination".
Where the disagreement remains clinically material, seek appropriate advice rather than averaging the statements. An AI-generated explanation may help identify concepts to investigate, but it cannot settle an unresolved prescribing question merely by sounding more coherent than the source documents.
Turning the review into learning
This article is published by iatroX and considers its educational role alongside a specialist resource. CredibleMeds is the more directly relevant starting point for understanding its own classifications. iatroX should not be presented as an independently validated replacement for that database.
The useful learning task comes afterwards: explain why a medicine classification is not an individual risk estimate, or why a change in clinical context can alter the interpretation of a familiar prescription. According to the iatroX methodology available on 19 September 2026, Ask-iatroX provides source-linked clinical reference with retrieval and checking processes. Those are design features, not a guarantee that every answer is correct.
For examination preparation, a related pharmacology question can test whether the distinction transfers to a different vignette. For practice, record the actual source, relevant context and agreed action. Neither task is improved by memorising a traffic-light colour without understanding what it represents.
Frequently asked questions
Does Conditional Risk mean a medicine is safer than one labelled Possible Risk?
Not necessarily: the categories describe different evidence and circumstances rather than a simple ascending risk score. The relevant condition may already apply to the patient.
Does a CredibleMeds entry tell me whether to stop a medicine?
No: an individual decision requires clinical context, appropriate prescribing information and a plan for the condition being treated. A classification alone is not a complete management instruction.
Can an AI answer replace checking the current classification?
It should not replace the specialist source when the classification is central to the decision. Use an explanation to clarify the reasoning, then inspect the underlying information.