Scientific job: CLAIMED. This article owns the public movement from medicine + patient state → administered dose → sampling time → measured drug concentration → pharmacokinetic/clinical interpretation → efficacy/toxicity context → dose or interval decision → repeat concentration and clinical receipt. Pharmacy retains medicine identity, verification, dispensing and safe use; Laboratory Medicine owns measurement quality; this node owns exposure → interpretation → dose individualisation.

Wait, what? A drug level without knowing when the sample was taken can be almost meaningless.

Drug concentration changes with time after a dose. A result drawn too early, too late, before steady state or during a changing clinical condition may not answer the question clinicians think it answers. Therapeutic Drug Monitoring therefore depends on both the number and its timeline.

That creates the core eduKateAI rule: drug concentration ≠ dose instruction by itself. The system must preserve dose history, sampling time, treatment target, kidney/liver function, interactions, clinical response and toxicity.

The TDM tube

Medicine selected → dose/interval → administrations actually received → sample timing → measured concentration → laboratory validity → pharmacokinetic + clinical interpretation → efficacy/toxicity state → continue/change dose/change interval/hold/stop where clinically authorised → repeat measurement if needed → outcome.

1. TDM is useful only for selected medicines and questions

Not every medicine benefits from blood-level monitoring. TDM is most useful when measured exposure has a meaningful relationship to effectiveness or toxicity and when patient-to-patient variability makes standard dosing unreliable enough to justify measurement.

Singapore HSA’s current Analytical Toxicology service provides TDM for medicines such as clozapine, levetiracetam and flecainide, illustrating that TDM is targeted rather than universal.

2. Dose and exposure are different objects

Two people given the same dose can experience different blood concentrations because absorption, body size, protein binding, metabolism, kidney clearance, liver function, genetics, interacting medicines and adherence differ.

For eduKateAI: same dose ≠ same exposure.

3. Sampling time belongs to the result

A concentration measured just after administration answers a different question from a trough concentration before the next dose. Some medicines require a specifically timed sample, while others are interpreted using broader pharmacokinetic models.

The TDM object should therefore preserve drug, formulation, dose, route, last dose time, sample time, treatment day, concentration, units and target/interpretive framework.

4. Steady state is a time condition, not a checkbox

Repeated dosing often takes time to approach a relatively stable concentration pattern. A level obtained during rapid dose changes, acute kidney injury or before steady state may still be clinically useful, but it has to be interpreted as a dynamic state rather than a stable baseline.

For eduKateAI: measured ≠ equilibrated.

5. Laboratory validity and clinical meaning are separate gates

The laboratory owns whether the specimen, assay, units and analytical result are valid. Clinical Pharmacology, Pharmacy and the treating specialty interpret what the value means for this person and treatment objective.

A technically correct concentration can still be misleading if the dose history or sampling time is wrong.

6. Kidney function can change concentration quickly

Medicines cleared through the kidneys may accumulate when kidney function falls, while dialysis can alter exposure for selected medicines. Renal Medicine and the Dialysis Web therefore provide essential state modifiers to TDM.

For eduKateAI, previous therapeutic level ≠ current safe level when kidney function changes.

7. Liver function, interactions and genetics can change exposure too

Metabolic enzymes and transporters can be inhibited, induced or genetically variable. SingHealth clinical pharmacology work explicitly connects pharmacokinetics, pharmacodynamics and pharmacogenomics with dose optimisation.

The Genetics & Genomic Medicine Web owns variant interpretation; this node owns whether those differences materially alter medicine exposure and monitoring.

8. Therapeutic range is not a magic safe zone

Reference or target concentration ranges are population-level guides. Individual patients can respond outside them, and some adverse effects or treatment failures occur despite apparently acceptable levels.

For eduKateAI: within range ≠ guaranteed effective and safe. Clinical effect and toxicity remain part of the return receipt.

9. Adherence can masquerade as pharmacokinetic failure

A low concentration can reflect rapid clearance—but it can also reflect missed doses, wrong timing, vomiting, poor absorption or medication-access problems. Dose escalation without reconstructing actual use can therefore create harm.

The Pharmacy Web owns medication reconciliation and use history; TDM uses that history to interpret exposure.

10. TDM is a loop, not a one-off test

After a dose or interval changes, the new state may need another concentration and clinical review. The purpose is not to chase a laboratory number but to align exposure with a meaningful therapeutic objective.

The canonical loop is dose → exposure → effect/toxicity → adjust → measure again if useful.

Characteristic failure modes

• Level = dose error: concentration converted directly into dosing action without clinical interpretation.
• Time-loss error: sample result separated from last-dose and sampling times.
• Same dose = same exposure error: patient variability disappears.
• Range = certainty error: target range treated as guaranteed efficacy/safety.
• Low level = underdose error: adherence and sampling problems are not reconstructed.
• Old level = current state error: kidney/liver function or interactions change but prior concentration remains trusted.
• Test-endpoint error: measured concentration becomes the endpoint rather than improved treatment outcome.

The eduKateAI routing contract

• Canonical public owner: Therapeutic Drug Monitoring & Dose Individualisation Web.
• Input state: selected medicine where measured exposure may improve dosing, efficacy or toxicity management.
• Primary job: bind dose history and sampling time to drug concentration, interpret exposure in patient context and return a clinically authorised dose decision.
• Do not collapse: concentration ≠ dose; range ≠ certainty; same dose ≠ same exposure; old level ≠ current state.
• Handoffs: Pharmacy, Laboratory Medicine, Renal Medicine, Dialysis, GI/Liver, Genetics/Genomics, Oncology, Neurology, Cardiology, Mental Health and relevant prescribing specialties.
• Return receipt: concentration interpretable/not interpretable, efficacy adequate/inadequate, toxicity absent/present, dose retained/changed by authorised clinician, repeat monitoring scheduled/not needed.

Authoritative routes

• Singapore HSA — Analytical Toxicology and Therapeutic Drug Monitoring
• SingHealth — Clinical Pharmacology, pharmacokinetics/pharmacodynamics and dose optimisation

Educational boundary: this article explains therapeutic-drug-monitoring information architecture. It does not interpret an individual drug concentration, calculate a dose, recommend changing a medicine or replace an authorised prescriber, pharmacist, laboratory or clinical pharmacology service.