The recurring disappointment of pharmacogenomics in practice is not that the associations are weak. It is that the drugs with the clearest genotype–toxicity links are often given urgently, and a test that takes days cannot inform a decision taken in an hour.
The MT-RNR1 guideline quantifies this honestly: only about a fifth of aminoglycoside use is predictable enough to test in advance. That single figure explains why the same guideline argues for point-of-care testing in neonatal units rather than simply asking for more laboratory capacity.
It is a useful filter to apply to any proposed pharmacogenomic service. Ask where in the pathway the prescription is decided. Where that moment is elective — an oncology regimen planned at a multidisciplinary meeting, a mycobacterial course, a transplant protocol — laboratory testing is enough and the yield is real. Where it is emergent, only a bedside assay changes anything, and a service built on send-away testing will generate results that arrive after the harm.
- Elective and protocolised prescribing is where send-away pharmacogenomics earns its cost
- Emergency prescribing needs point-of-care testing or nothing — an unused result is not a neutral outcome, it is wasted spend
- A negative genotype does not license a drug the patient has another reason to avoid
- Where testing is unavailable, the fallback is an explicit alternative agent decision, documented, not an unexamined default
Why it matters
It explains why several well-evidenced pharmacogenomic tests have changed almost nothing in practice, and predicts which ones will.
Read the rest in the app
You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.
Scan to keep reading on your phone. No account needed to start.
Tomorrow morning, before your first patient
One edition a day for clinical pharmacology, written by the desk, every claim tied to its paper. Six minutes.
Get the app — free