Prescribing safely: the habits that prevent most medication harm
Most prescribing errors are not exotic. They are a routine drug given to the wrong patient at the wrong moment.
A striking amount of harm in hospitals comes from medicines, and a large share of it is preventable. What is worth noticing is how mundane the causes are. Very little of it involves rare drugs or obscure interactions. It is an anti-inflammatory given to someone with poor kidneys, a sedative given to a patient with liver disease, a dose that was correct last year and is now too large because the patient is older and drier, or a new symptom treated as a new disease when it was a side effect all along.
That is encouraging, because it means the protective habits are few and learnable. This chapter is about those habits rather than about memorising lists.
🩺 Where this lives: The single most useful question in clinical medicine may be "could a drug be causing this?" — and it is chronically underasked. A new tremor, a fall, confusion in an older person, a cough, a rash, constipation, a rising potassium: each has a long differential, and in a patient taking several medicines the drug chart belongs near the top of it. Asking the question costs nothing. Not asking it produces the prescribing cascade, where every side effect earns its own new prescription.
💡 A note on lists. This chapter deliberately does not give exhaustive lists of enzyme inducers, inhibitors or interacting drug pairs, and gives no doses. Those lists are long, incomplete wherever they are printed, and change over time — memorising six inducers still leaves you missing the seventh. What protects patients is knowing the direction of an effect and having the habit of checking a current formulary before prescribing. That is what is taught here.
What the body does to the drug
Absorption, distribution, metabolism and excretion are worth learning not as a classification but as a map of where things go wrong. Most clinically important interactions occur at metabolism, and most dose adjustments in sick patients are needed because of excretion. If you remember only two things: the liver is where drugs interfere with each other, and the kidney is where they accumulate.
Inducers and inhibitors
THE DIRECTION, AND WHY IT FEELS BACKWARDS
ENZYME INDUCER
→ more metabolising enzyme
→ the affected drug is broken down FASTER
→ its level FALLS
→ TREATMENT FAILURE
Classic consequences: loss of anticoagulant control,
failure of hormonal contraception, loss of seizure or
transplant drug control.
ENZYME INHIBITOR
→ metabolising enzyme is blocked
→ the affected drug is broken down SLOWER
→ its level RISES
→ TOXICITY at an unchanged dose
Classic consequence: bleeding in a patient stable on
warfarin for years, after a short course of another drug.
MEMORY HOOK: INDUCE = REDUCE. INHIBIT = INCREASE.
(The verb describes what happens to the ENZYME; the effect
on the DRUG LEVEL is the opposite.)
TIMING MATTERS TOO:
Induction is SLOW — new enzyme protein must be
synthesised, so the effect builds over days to weeks
and persists for a while after the inducer is stopped.
Inhibition is often FAST — sometimes within a dose or
two.
So stopping an inducer can cause DELAYED toxicity as the
enzyme level falls back, which is easily missed.
THE DRUGS THAT SUFFER MOST are those with a NARROW
THERAPEUTIC INDEX — where the toxic level is close to the
effective one. Warfarin, digoxin, lithium, phenytoin,
aminoglycosides and immunosuppressants are the recurring
examples. For these, ANY new drug deserves a check.
Renal and hepatic impairment
💡 Exam angle: the recurring stem is a patient with impaired kidneys given something renally cleared or directly nephrotoxic. Two specific pairings appear repeatedly: NSAIDs in renal impairment, which reduce renal blood flow exactly when it is already compromised, and sedatives in liver disease, which may precipitate hepatic encephalopathy. Note also that there is no simple blood test of hepatic drug-metabolising capacity — liver enzymes indicate damage, not clearance — so caution rather than calculation is the rule.
Patients at particular risk
PREGNANCY — RESIST THE TWO OPPOSITE ERRORS
ERROR 1: prescribing without checking. Teratogenic risk is
greatest during ORGANOGENESIS in the first trimester,
which is often before pregnancy is recognised — so the
question "could this patient be pregnant?" belongs
before the prescription, not after.
Drugs contraindicated in pregnancy that appear elsewhere
in this syllabus include ACE inhibitors and ARBs,
warfarin, statins and radioiodine.
ERROR 2: stopping everything out of caution. UNTREATED
MATERNAL ILLNESS ALSO HARMS THE FETUS. Uncontrolled
epilepsy, asthma, diabetes, hypothyroidism, severe
infection and severe mental illness all carry real risk
to both mother and baby.
The correct approach is neither reflex prescribing nor
reflex stopping: check each drug against a current source,
and weigh the risk of the drug against the risk of the
untreated condition — with specialist input where the
condition is serious.
ALLERGY: always ask, and always ask WHAT HAPPENED. A rash
in childhood, nausea, and anaphylaxis are three different
things recorded under the same word, and treating simple
intolerance as true allergy can deny a patient the best
drug for a serious infection.
The prescribing cascade
The cascade is worth recognising because it is so easy to join halfway through. A patient arrives on eight medicines, several of which exist to treat the effects of the others, and each individual step looked reasonable at the time. Breaking it requires one deliberate habit: when a new symptom appears in someone taking medication, review the drug chart before adding to it.
Clinical reasoning: four presentations
🔍 Case 1 — bleeding after a short course
PresentationA man stable on warfarin for four years is treated for an infection. Ten days later he presents with bruising, epistaxis and a markedly raised INR. His warfarin dose has not changed.
Key clueA new drug added to a narrow-therapeutic-index drug.
ReasoningEnzyme inhibition slows warfarin metabolism, so its level rises and the INR climbs at an unchanged dose. The stability of the previous four years is exactly what makes the new drug the suspect.
AnswerManage the raised INR and bleeding per protocol, identify the interacting drug, and check interactions before adding anything to warfarin in future. This is the standard interaction question.
🔍 Case 2 — a tremor treated as a new disease
PresentationAn older woman on several medicines develops a tremor and slowed movement. Parkinson's disease is diagnosed and treatment started. She then becomes nauseated, and an antiemetic is added. She becomes more unsteady and falls.
PatternA prescribing cascade in progress.
ReasoningDrug-induced parkinsonism is a recognised adverse effect of several drug classes and is reversible on stopping the culprit. Each subsequent prescription treated the consequences of the last.
AnswerReview the whole drug chart and ask whether a drug could be causing the original tremor, before adding anything further. Stop or substitute the culprit and reassess.
🔍 Case 3 — the same dose, a different patient
PresentationA 78-year-old has taken the same dose of a renally cleared drug for years without difficulty. After several days of vomiting and poor intake she becomes confused, and the drug level is high.
Key pointThe prescription did not change; the clearance did.
ReasoningDehydration and age-related decline reduce renal clearance, so a renally excreted drug accumulates on an unchanged dose. This is one of the commonest mechanisms of medication harm in older people.
AnswerWithhold the drug, rehydrate, check renal function and the drug level, and review the dose against current renal function. Sick-day guidance should be given for the future.
🔍 Case 4 — an allergy that was not
PresentationA patient needing treatment for a serious infection has "penicillin allergy" recorded. Asked what happened, she describes nausea and loose stools as a child. A broader-spectrum alternative is chosen automatically.
Key questionWhat actually happened?
ReasoningNausea is an intolerance, not an allergy. Mislabelling leads to broader-spectrum, sometimes less effective and more toxic alternatives, and drives resistance.
AnswerTake a proper allergy history — what happened, how soon, how severe — and correct the record. Genuine anaphylaxis is an absolute contraindication; childhood nausea is not.
Commonly confused
Confusion
The distinction
Why it matters
Inducer vs inhibitor
Induce = reduce the level; inhibit = increase it
The direction is the whole question.
Speed of induction vs inhibition
Days to weeks versus often rapid
Explains delayed effects after stopping an inducer.
Narrow vs wide therapeutic index
How close toxic is to effective
Determines which interactions actually matter.
Renal vs hepatic impairment
Accumulation versus reduced metabolism
Different drugs, different adjustments.
Liver enzymes vs liver function
Enzymes show damage, not clearance
No simple test of metabolising capacity.
Allergy vs intolerance
Anaphylaxis versus nausea or a childhood rash
Mislabelling denies patients the best drug.
New disease vs adverse effect
Ask whether a drug could cause it
Prevents the prescribing cascade.
Stopping all drugs in pregnancy vs reviewing
Untreated illness also harms the fetus
Both reflexes are errors.
Rapid revision
MUST-KNOW FACTS
1. Most medication harm involves ordinary drugs, not exotic ones.
2. The liver is where drugs interact; the kidney is where they accumulate.
3. ENZYME INDUCER → faster metabolism → level FALLS → treatment failure.
4. ENZYME INHIBITOR → slower metabolism → level RISES → toxicity.
5. Memory hook: INDUCE = REDUCE · INHIBIT = INCREASE.
6. Induction is SLOW (days to weeks); inhibition is often rapid.
7. Stopping an inducer can cause DELAYED toxicity.
8. Interactions matter most for NARROW THERAPEUTIC INDEX drugs.
9. Narrow index examples: warfarin, digoxin, lithium, phenytoin.
10. Check interactions before adding ANY drug to warfarin.
11. Renal impairment: renally cleared drugs ACCUMULATE.
12. Reduce the dose or lengthen the interval in renal impairment.
13. AVOID NSAIDs in renal impairment — they reduce renal blood flow.
14. Hepatic impairment: reduced metabolism, clotting factors and albumin.
15. SEDATIVES in liver disease may precipitate ENCEPHALOPATHY.
16. Liver enzymes indicate damage, NOT drug-metabolising capacity.
17. Older patients: declining renal function, polypharmacy, sensitivity.
18. START LOW, GO SLOW in the elderly.
19. Children need WEIGHT-based dosing — they are not small adults.
20. Teratogenic risk is greatest in the FIRST TRIMESTER.
21. Ask "could this patient be pregnant?" BEFORE prescribing.
22. Pregnancy-contraindicated in this syllabus: ACE inhibitors, ARBs, warfarin, statins, radioiodine.
23. But UNTREATED MATERNAL ILLNESS ALSO HARMS THE FETUS.
24. Do not reflexively stop everything in pregnancy — review each drug.
25. Always ask about allergy — and ask WHAT HAPPENED.
26. Nausea is intolerance; anaphylaxis is allergy.
27. Mislabelled allergy drives broader-spectrum use and resistance.
28. THE PRESCRIBING CASCADE: side effect → mistaken for disease → new drug.
29. Ask "COULD A DRUG BE CAUSING THIS?" for every new symptom.
30. Review the drug chart BEFORE adding to it.
31. Check a CURRENT formulary rather than relying on a memorised list.
💡 Exam angle: questions here are usually scenarios rather than facts. A stable patient becomes toxic after a new drug (inhibition), a treatment stops working after a new drug (induction), a long-standing dose becomes toxic after an illness (reduced clearance), or a symptom gets a new prescription instead of a drug review (cascade). Recognising which of those four is happening answers most of them.
Syllabus points
Where in the pharmacokinetic path things go wrong
Enzyme induction and inhibition: the direction
Why induction is slow and inhibition fast
Narrow therapeutic index drugs
Prescribing in renal impairment
Why NSAIDs are avoided in renal impairment
Prescribing in hepatic impairment
Sedatives and hepatic encephalopathy
The older patient and polypharmacy
Pregnancy: the two opposite errors
Allergy versus intolerance
The prescribing cascade and how to break it
Create a free account to tick topics off, take notes as you read, watch the video lessons and get a day-by-day study plan built around your exam date.