GC029 Am I Prescribing The Right Drug
A clinical decision-making framework that guides physicians in evaluating whether a selected medication is the most appropriate choice for a patient based on indication, efficacy, safety profile, patient-specific factors, and evidence-based guidelines.
Am I Prescribing the Right Drug?
This lecture by Prof. Bernard Cheung (Division of Clinical Pharmacology & Therapeutics, HKU) is a foundational clinical pharmacology session that every medical student must internalize before touching a drug chart. The core message is deceptively simple: prescribing is not just about choosing a drug — it's about choosing the right drug, for the right patient, at the right dose, via the right route, at the right time, while weighing efficacy, safety, tolerability, and cost-effectiveness. Errors kill people. This lecture systematically walks through:
- What a prescription actually is (legal document, not a casual note)
- Who can prescribe
- Elements of a valid prescription
- Ten principles of good prescribing (Aronson 2017)
- Prescription errors — who is at risk, who makes them, what types exist
- The 5 Rights of medication administration
- The Swiss Cheese Model of error
- Personalized medicine — pharmacogenetics, benefit-risk analysis
- Cost-effectiveness analysis (including ICER calculation)
- Quality prescribing framework
- The Prescribing Skills Assessment (PSA) tool
This lecture is highly examinable — it covers principles that underpin every drug question in MCQs, SAQs, minicases, and OSCEs. It integrates with GC 070 (Is this the best drug for me?), GC 079 (Prescribing in older people), GC 043/Block A (Drugs and the Kidney), and the Block A Clinical Pharmacology lectures.
A prescription (from Latin praescribere = "to write before") is a written drug order completed BEFORE the drug is prepared for the patient. Rx comes from Latin Recipere (= "to take"). [1]
Why This Matters
A prescription is a legal document. It carries medicolegal weight. An incomplete, ambiguous, or incorrect prescription can lead to patient harm and legal liability. Every element exists for a safety reason.
Who Can Write Prescriptions?
Registered doctors are allowed by law to write prescriptions. Other professionals are allowed, with limited rights, to prescribe. [1]
In Hong Kong, this includes dentists (for dental procedures) and in some jurisdictions, nurse prescribers or pharmacist prescribers with specific training and limited formularies. But for exam purposes: doctors have the broadest prescribing rights.
2. Elements of a Prescription
The elements of a prescription are: Date, Patient's identity, Rx/"please take", Drug name, Formulation, Dosage amount and frequency, Route, Additional directions, and Signature. [1]
| Element | Why It's Required | Common Error |
|---|---|---|
| Date | Establishes when the order was made; legal requirement; important for duration-limited drugs | Missing date → pharmacist cannot dispense |
| Patient's identity | Prevents wrong-patient errors | Using bed number instead of name + DOB/ID |
| Rx / "Please take" | Formal instruction to dispense | Omission |
| Drug name | Must be unambiguous; prefer generic names | Illegible handwriting; brand vs. generic confusion |
| Formulation | Tablet ≠ capsule ≠ syrup ≠ injection | Wrong formulation → wrong bioavailability |
| Dosage + frequency | Determines therapeutic vs. toxic levels | Missing frequency; "daily" vs. "BD" |
| Route | Oral, IV, IM, SC, topical, inhaled, rectal, etc. | Giving IV drug orally or vice versa |
| Additional directions | e.g., "with food," "before meals" | Missing → reduced efficacy or increased side effects |
| Signature | Legal requirement; identifies prescriber | Unsigned prescriptions are invalid |
The form shown in the lecture is more than a template: its layout is a sequence of safety checks [1].
- Complete the allergy box explicitly (for example, a recorded allergy or "NKDA" according to local policy); do not leave it blank.
- Attach/verify the patient's identification label and indicate whether the patient is an inpatient or outpatient.
- The prescriber completes the Rx body, then signs and records their name/code and date.
- The pharmacy trail records who filled, checked, and issued the medicine—three distinct accountability points.
- The pictured paper form is a downtime form, used when the computer system is unavailable, and states that the prescription is valid for one month from issue. Follow the current local form and policy rather than assuming every prescription has the same validity period.
Formulations include: Tablets, Capsules, Syrup, Injection, Drops, Inhaler, Suppository. [1]
Why this matters: route and formulation change absorption, onset, duration, and sometimes the appropriate dose. Prescribing only "morphine 10 mg", without route or formulation, is therefore ambiguous and potentially dangerous.
For PRN (as needed) prescriptions, in addition to the standard elements, you must specify: Indication, Maximum dose to give, Minimum interval between doses, Maximum dose in 24 hours. [1]
Exam Trap: PRN Prescribing
A common exam scenario asks what additional information a PRN order requires. State the indication, maximum single dose, minimum dose interval, and maximum total dose in 24 hours. Any numerical paracetamol example is illustrative only: the permitted maximum must be checked for the individual patient and local product guidance, especially with low body weight, frailty, liver disease, malnutrition, or concurrent paracetamol-containing products.
State the total duration of treatment or total amount for: Antibiotics, Steroids, Opioids, Hypnotics. [1]
Why these four specifically?
- Antibiotics: Use the guideline-indicated duration or a documented stop/review date; unnecessarily prolonged exposure increases adverse effects, C. difficile infection, and selection pressure for resistance.
- Steroids: Define the duration and review plan. Whether tapering is required depends on dose, duration, regimen, and risk of hypothalamic-pituitary-adrenal suppression—not every short course needs a taper.
- Opioids: Addiction risk; need clear end-point
- Hypnotics: Dependence and tolerance; should be short-term only
Ten Principles of Good Prescribing: [1]
- Be clear about the reasons for prescribing
- Take into account the patient's medication history before prescribing
- Take into account other factors that might alter the benefits and harms of treatment
- Take into account the patient's ideas, concerns, and expectations
- Select effective, safe, and cost-effective medicines individualized for the patient
- Adhere to national guidelines and local formularies where appropriate
- Write unambiguous legal prescriptions using the correct documentation
- Monitor the outcomes of treatment, both beneficial and adverse
- Communicate and document prescribing decisions and the reasons for them
- Prescribe within the limitations of your knowledge, skills, and experience
Let me unpack these because they are directly examinable:
| Principle | First-Principles Explanation | Clinical Example |
|---|---|---|
| 1. Clear reasons | Every drug must have an indication. Don't prescribe "just in case." | Don't give antibiotics for viral URTI |
| 2. Medication history | Know what the patient is already on → avoid duplications, interactions | Adding ACEI when patient already on ARB → dual RAAS blockade → hyperkalemia, AKI |
| 3. Other factors | Comorbidities, pregnancy, renal/hepatic function, age | NSAIDs in CKD → further renal damage [2] |
| 4. Patient ICE | Shared decision-making. Patient may refuse or have concerns | Patient worried about weight gain on insulin → discuss alternatives |
| 5. Effective + safe + cost-effective | The "quality prescribing" triad plus cost | Generic metformin vs. branded → same efficacy, lower cost |
| 6. Guidelines/formularies | Evidence-based practice; don't reinvent the wheel | Following STOPP/START criteria in elderly [3] |
| 7. Unambiguous prescriptions | Prevents misinterpretation | Writing "units" not "U" (U can be misread as 0) |
| 8. Monitor outcomes | Both therapeutic response AND adverse effects | Checking INR on warfarin; checking K+ on ACEI |
| 9. Communicate + document | Other team members need to know your reasoning | Drug chart notes, discharge summaries |
| 10. Prescribe within limits | If unsure, ask a senior. Don't guess. | Junior doctor unsure about chemotherapy → refer to oncologist |
4. Prescription Errors
The slide lists children, people with intellectual disability, older adults, nursing-home residents, and ICU patients as groups vulnerable to prescribing errors. [1]
The original slide uses historical language for intellectual disability; the person-first wording above is preferred.
Why these groups?
- Children: Weight-based dosing required; small errors → proportionally large overdoses; limited ability to report symptoms
- People with intellectual disability or cognitive/communication impairment: may need supported communication, medication reconciliation, and help reporting adverse effects
- Elderly: Polypharmacy, altered pharmacokinetics, cognitive impairment
- Nursing homes: Multiple carers, communication gaps, polypharmacy
- ICU: Multiple IV infusions, critically ill with altered physiology, high-risk drugs
Older patients may take multiple drugs, have chronic illnesses (e.g., hypertension, diabetes, COPD), be more susceptible to adverse effects, have cognitive problems, and require assistance with medicines. [1]
The slide's term COAD is retained here as its modern equivalent, COPD.
This connects directly to GC 079 (Prescribing in Older People), which gives the prescribing process as:
The right drug → The right dose → The right route → Check allergy → Check RFT → Aware of side effects → Start low & Go slow [3]
High Yield: Prescribing in the Elderly
"Start low, go slow" is useful when a lower starting dose and gradual titration are pharmacologically appropriate, but it must not mean undertreating an urgent condition. Review renal function before prescribing renally cleared medicines and use the drug-specific dosing metric required by the product information or local formulary (some drugs use estimated creatinine clearance rather than laboratory-reported eGFR).
In the historical local MIR series shown, July accounted for 16.5%, 10.2%, and 11.9% of reports in 2004, 2005, and 2006, respectively. [1]
The teaching interpretation is that July coincides with new house officers starting work. Treat this as a historical local safety signal supporting induction, supervision, and system checks—not as a current incidence estimate or proof of causation.
House officers were involved in 42.6% of the historical MIRs displayed. [1]
The chart shows the highest absolute report count for house officers, followed by medical officers. Because the slide gives neither the number of prescriptions written by each grade nor reporting propensity, it cannot establish an error rate or justify blaming one group. Its useful message is to strengthen transition-period supervision and downstream checks.
Two types of errors: (1) Lack of knowledge of what is right or wrong; (2) Has the knowledge but makes an unintentional mistake. [1]
This is a crucial distinction:
- Knowledge-based errors (Type 1): You didn't know metformin is contraindicated in eGFR < 30. Solution: education.
- Execution errors/slips (Type 2): You knew the right dose but wrote the wrong number because you were distracted. Solution: systems (checklists, electronic prescribing, double-checks).
Both types are addressed by the Swiss Cheese Model.
Patients are entitled to 5 rights: Right patient, Right medication, Right dose, Right time, Right route. [1]
| Right | How to Ensure It | Common Failure |
|---|---|---|
| Right patient | Check name + DOB/ID number (at least 2 identifiers); never use bed/room number | Drug given to wrong patient after bed swap |
| Right medication | Use generic name; check allergy status; verify indication | Sound-alike/look-alike drugs (e.g., chlorproMAZINE vs. chlorproPAMIDE) |
| Right dose | Weight-based in children; adjust for renal/hepatic function; check BNF | 10x overdose due to decimal point error |
| Right time | Correct frequency; time-critical drugs (e.g., antibiotics) | Missing doses; wrong timing relative to meals |
| Right route | Verify formulation matches route | IV potassium given too fast → cardiac arrest |
Double-Check Identity
Ask for patient's name and one other detail such as DOB or ID number. Check at least 2 of these (name, DOB, ID number). Do not use bed or room number. [1]
Why Not Bed Number?
Bed numbers change constantly — patients are moved, swapped, transferred. A bed number is an attribute of the location, NOT the patient. Using it as an identifier has caused fatal errors (e.g., giving chemotherapy to the wrong patient). Always use patient-intrinsic identifiers: name, DOB, hospital ID number.
A medication error is a failure in the treatment process that leads to, or has the potential to lead to, harm to the patient. (Ferner et al. Drug Saf. 2006, Reason. BMJ. 2000) [1]
The Swiss Cheese Model (James Reason) conceptualizes safety as multiple layers of defence (like slices of Swiss cheese). In the slide, a hazard travels from left to right toward the patient; each barrier can stop it, but harm occurs when temporary weaknesses (the holes) line up across several layers.
The layers in medication safety include:
Drug Safety — everyone is a stakeholder: Pharmaceuticals, Hospitals & Clinics, Prescriber, Pharmacist, Nurse or carer, Patient, Regulatory bodies & government, Drug store. [1]
| Layer | Role in Error Prevention |
|---|---|
| Regulatory bodies/government | Drug approval, safety monitoring, post-marketing surveillance |
| Pharmaceutical manufacturers | Drug design, labelling, packaging, and safety information |
| Hospitals | Protocols, electronic prescribing, formulary restrictions |
| Prescriber | Knowledge, careful prescribing, checking allergies/interactions |
| Pharmacist | Verifying prescriptions, flagging errors, dose-checking |
| Nurse/carer | Final check before administration; verifying 5 Rights |
| Patient | Questioning unfamiliar drugs; reporting side effects |
The stakeholder diagram is a chain, not only a list: upstream decisions by regulators and manufacturers shape the medicine reaching hospitals, prescribers, pharmacists and carers, while community drug stores provide another route to the patient. No single layer is perfect; safety comes from independent barriers catching what earlier layers missed.
Severity levels: S0 = error discovered before administration; S1 = did not result in patient injury; S2 = increasing monitoring but no change in vital signs and no injury; S3 = increasing monitoring, change in vital signs but no ultimate injury; S4 = required antidote or reversal or transfer to higher level of care; S5 = permanent injury; S6 = death. [1]
| Level | Description | Example |
|---|---|---|
| S0 | Caught before reaching patient | Pharmacist catches wrong dose |
| S1 | Reached patient, no harm | Extra dose of paracetamol, patient fine |
| S2 | Increased monitoring, no harm | Wrong antibiotic → monitored, no clinical effect |
| S3 | Vital sign change, no permanent harm | Hypotension from antihypertensive overdose |
| S4 | Antidote/reversal or transfer to a higher level of care needed | Naloxone and escalation of care after clinically significant opioid toxicity |
| S5 | Permanent injury | Renal failure from nephrotoxic drug |
| S6 | Death | Fatal anaphylaxis to known allergen |
In the displayed dataset, S0 dominates (roughly 220 reports), with progressively fewer S1-S3 events and no S4-S6 reports shown. This is descriptive of that dataset only; it does not imply that severe medication harm cannot occur. The goal is to intercept hazards at S0.
The lecture's pie chart classifies incidents by the stage of the medication-use process, rather than by clinical subtype [1]:
| Stage | Share of incidents shown |
|---|---|
| Ordering | 49% |
| Administration | 26% |
| Dispensing | 14% |
| Transcribing | 11% |
The four labelled slices sum to 100%. Almost half of the reports arose at the prescribing decision/order stage, while the Swiss-cheese layers downstream could still intercept them before administration.
9. Personalized Medicine
The benefits and the risks of a drug may be different in different patients. "What's one man's meat is another man's poison." [1]
This is the fundamental concept underlying personalized/precision medicine. A drug that works well in one patient may be ineffective or toxic in another.
Complex interplay between genetic and environmental factors affects the fate and effect of drugs. Factors include: Genetics, Demographics (sex, age, body weight), Concomitant conditions (e.g., renal failure), Interactions with food and other drugs — all affecting pharmacokinetics and pharmacodynamics. [1]
The slide is a two-axis matrix: pharmacokinetics (what the body does to the drug) and pharmacodynamics (what the drug does to the body) are the rows; genetics, demographics, concomitant conditions, and food/drug interactions are the columns. Any column may alter either row, so individualisation is not synonymous with genetic testing.
| Factor | Pharmacokinetic Effect | Pharmacodynamic Effect |
|---|---|---|
| Age | ↓ renal/hepatic clearance in elderly | ↑ sensitivity to benzodiazepines in elderly |
| Body weight | ↑ Vd for lipophilic drugs in obese | — |
| Renal failure | ↓ clearance of renally-excreted drugs | ↑ sensitivity to CNS effects (e.g., gabapentin accumulation) |
| Genetics | CYP2C19 poor metabolizer → ↓ clopidogrel activation | HLA-B*5801 → allopurinol hypersensitivity |
| Drug interactions | CYP3A4 inhibition by ketoconazole → ↑ statin levels | Additive sedation: opioid + benzodiazepine |
| Food | Grapefruit juice inhibits CYP3A4 → ↑ drug levels | High-vitamin-K diet → ↓ warfarin effect |
Using pre-treatment characteristics including sex, age, etc.; Using initial treatment response as a guide; Genetics/mutations; Receptor status. [1]
Examples:
- Pre-treatment characteristics: age, body size, renal/hepatic function, comorbidity, pregnancy potential, and interacting medicines can change drug choice or dose.
- Initial response: assess adherence, dose, duration, benefit and adverse effects after an indication-appropriate trial before continuing, titrating, switching, or stopping.
- Genetics/mutations: actionable variants may predict metabolism, toxicity, or response.
- Receptor/biomarker status: for example, HER2 or hormone-receptor status can determine whether a targeted breast-cancer therapy is appropriate.
Key pharmacogenetic tests: [1]
| Drug | Gene Test | Why |
|---|---|---|
| Carbamazepine | HLA-B*15:02 | Markedly increased risk of SJS/TEN in a treatment-naive carrier; avoid carbamazepine when positive if an alternative is available [6,7] |
| Allopurinol | HLA-B*58:01 | Increased risk of severe cutaneous adverse reactions; who should be tested depends on ancestry, risk, and local guidance |
| Abacavir | HLA-B*57:01 | Predicts abacavir hypersensitivity; a positive result precludes use |
| Azathioprine / 6-Mercaptopurine | TPMT and NUDT15 | Reduced function increases thiopurine-related myelosuppression risk and changes the recommended starting dose [4,8] |
| Clopidogrel | CYP2C19 | Poor metabolizers cannot activate the prodrug → reduced antiplatelet effect → ↑ cardiovascular events [5] |
| Warfarin | VKORC1, CYP2C9 | Determines dose sensitivity; VKORC1 variants → lower dose needed; CYP2C9 poor metabolizers → slower clearance |
| Statins | SLCO1B1 | Variants → ↑ risk of myopathy, especially with simvastatin |
The slide's gene network is deliberately many-to-many: several medicines can share a metabolising enzyme (notably CYP2D6 or CYP2C19), and one medicine can be influenced by several genes. Avoid the false "one gene = one drug" model. The table is a set of actionable examples, not a claim that every test is universally routine.
Carbamazepine and HLA-B*15:02
The lecture labels HLA-B15:02 testing "mandatory." Current Hong Kong Drug Office patient guidance specifically advises testing patients of Han Chinese or Thai origin before carbamazepine; local protocols may cover other at-risk ancestries [6]. CPIC recommends avoiding carbamazepine in a treatment-naive HLA-B15:02-positive patient when an alternative is available [7]. Do not generalise this to every person from "Southeast Asia"; check ancestry, prior exposure, indication, and the current local protocol.
TPMT and Azathioprine — deeper explanation from supporting notes [4]:
- Azathioprine is metabolized by three pathways. TPMT converts active metabolite 6-MP into inactive 6-MMP.
- If TPMT is deficient (1 in 300 people), most azathioprine is shunted toward the toxic 6-TGTP pathway → bone marrow suppression, neutropenic fever.
- NUDT15 loss-of-function variants are particularly relevant in East Asian and Hispanic populations; current CPIC guidance uses both TPMT and NUDT15 phenotype to choose a thiopurine starting dose [8].
- Also check if patient is on xanthine oxidase inhibitors (allopurinol, febuxostat) → these block another clearance pathway (XO), further increasing toxic metabolite accumulation.
- Genotyping/phenotyping guides the starting dose but does not replace blood-count and liver-function monitoring.
11. Benefit-Risk Analysis
Benefit-risk matrix: [1]
- Good efficacy + Low risk of harm → PRESCRIBE
- Good efficacy + High risk of harm → WEIGH BENEFIT AGAINST RISK
- Poor efficacy + Low risk of harm → DON'T PRESCRIBE (X)
- Poor efficacy + High risk of harm → DON'T PRESCRIBE (X)
For severe disease (e.g., cancer), even drugs with low benefit-to-risk ratio may be worth prescribing. For mild disease (e.g., dermatitis), only high benefit-to-risk drugs should be used. [1]
| Severe Disease (e.g., Cancer) | Mild Disease (e.g., Dermatitis) | |
|---|---|---|
| High benefit-to-risk | Prescribe | Prescribe |
| Low benefit-to-risk | Consider only if the expected overall benefit is favourable, no better option exists, and the patient makes an informed choice with an appropriate monitoring plan | Usually do not prescribe |
Key principle: disease severity changes what magnitude of risk may be acceptable, but it never makes a low benefit-to-risk treatment automatically appropriate. Alternatives, reversibility of harm, uncertainty, monitoring, and the patient's goals all matter.
The approach to informing patients depends on severity AND frequency of adverse effects: [1]
| Common | Rare | |
|---|---|---|
| Severe | Should not be prescribed | Prescribe with caution and warn the patient |
| Mild | Prescribe but inform the patient | Prescribe — no need to inform routinely |
This grid is a useful lecture heuristic, not a modern consent rule. A common severe adverse effect does not automatically prohibit a life-saving treatment, and a rare mild effect is not automatically irrelevant. Current GMC guidance requires a dialogue about material risks and reasonable alternatives: materiality depends both on what a reasonable person in the patient's position would consider significant and on what matters to that particular patient [9]. Prioritise likely effects, serious effects, effects that could alter the patient's decision or daily life, warning symptoms, and what to do if they occur.
12. Cost-Effectiveness Analysis
Costs are not just drug price. They include: cost of administering the drug, cost of monitoring, and cost of adverse effects. Costs depend on perspective (who is paying). [1]
Benefits depend on perspective. Future benefits are less valuable than immediate ones (discounting). A useful measure of benefit is QALY (Quality-Adjusted Life Year). [1]
QALY explained from first principles:
- 1 QALY = 1 year of perfect health
- If a treatment gives you an extra year of life but at 50% quality (e.g., bedridden), that's 0.5 QALY
- Treatment may improve survival (extend life) AND/OR increase quality of life (better function even if no extra years)
In the slide's utility-versus-time graph, QALYs are the area under the health-related quality-of-life curve. A new treatment can add area by raising quality, extending survival, or both. This quality adjustment is conceptually separate from discounting, the economic convention that gives future costs and outcomes less present value.
The Incremental Cost-Effectiveness Ratio (ICER) = ΔCost / ΔEffect
Example: Antibiotic A costs 80 with 80% efficacy. Per 100 patients: A costs 8,000 → 80 cured. Extra 2,000/20 = $100 per extra cure. [1]
The exam question answer is C: $100 per additional chest infection cured.
ICER Calculation — Exam Must-Know
ICER = (Cost_B - Cost_A) / (Effect_B - Effect_A)
In this example: (60) / (0.80 - 0.60) = 100 per additional cure
Or equivalently per 100 patients: (6000) / (80 - 60) = 100
A treatment is considered cost-effective if ICER falls below the "willingness-to-pay threshold" (varies by country/healthcare system).
The cost-effectiveness plane plots incremental effectiveness on the x-axis and incremental cost on the y-axis. The slide's diagonal line represents an illustrative willingness-to-pay threshold of $50,000 per QALY: a north-east point below that line would meet that example threshold, while one above it would not. This is not a universal threshold and should not be treated as a current Hong Kong decision rule.
The four quadrants are:
| Quadrant | ΔCost | ΔEffect | Decision |
|---|---|---|---|
| NE (upper right) | More costly | More effective | Depends on ICER vs. threshold |
| SE (lower right) | Less costly | More effective | Dominant — always adopt |
| NW (upper left) | More costly | Less effective | Dominated — never adopt |
| SW (lower left) | Less costly | Less effective | Depends on ICER vs. threshold |
The following historical "league table" slide ranks old interventions by incremental cost per QALY using values published around 2000. It teaches how cost-effectiveness comparisons are displayed; the figures are not current prices or current recommendations.
Quality prescribing considers: Efficacy, Safety, Tolerability, and Cost-effectiveness. [1]
| Pillar | Definition | Example |
|---|---|---|
| Efficacy | Does the drug work for the intended indication? | Ramipril reduces mortality in HFrEF |
| Safety | What are the serious adverse effects? | Clozapine → agranulocytosis (requires monitoring) |
| Tolerability | What are the common, bothersome side effects? | Metformin → GI upset (many patients stop it) |
| Cost-effectiveness | Is the benefit worth the cost? | Generic simvastatin vs. branded rosuvastatin |
Efficacy ≠ Tolerability: A drug can be very effective but poorly tolerated (e.g., full-dose aspirin for analgesia — effective but GI bleeding). Conversely, a drug can be well-tolerated but have poor efficacy for a given indication.
The lecture emphasizes that drug safety is a shared responsibility:
Drug Safety stakeholders: Pharmaceuticals, Hospitals & Clinics, Prescriber, Pharmacist, Nurse or carer, Patient, Regulatory bodies & government, Drug store. [1]
This maps directly onto the Swiss Cheese Model — each stakeholder is a "slice" that can catch errors.
The PSA is an online training and assessment package to promote better prescribing skills, bridging the gap between basic pharmacology knowledge and clinical application, using realistic clinical cases with assessment-scoring-feedback cycles. [1]
The lecture's circular diagram presents 8 sections, 60 items, 120 minutes, and 200 marks as a historical format snapshot [1]. The current official learner portal still describes eight question types and a standard 120-minute assessment, but candidates should verify the current item and mark structure for their sitting [10].
| Section | Content | What It Tests |
|---|---|---|
| 1 | Prescribing | Writing prescriptions correctly |
| 2 | Prescription Review | Identifying errors in existing prescriptions |
| 3 | Planning Management | Selecting appropriate drug therapy |
| 4 | Providing Information | Counseling patients about drugs |
| 5 | Calculation Skills | Drug dose calculations |
| 6 | Adverse Drug Reactions | Recognizing and managing ADRs |
| 7 | Drug Monitoring | Interpreting drug levels, lab monitoring |
| 8 | Data Interpretation | Using clinical data to guide prescribing |
Specialties covered: MED, SURG, ELD (elderly), PED, PSYCH, O&G, GP [1].
What the Illustrated PSA Case Is Testing
The final case slide shows a 71-year-old man with acute pulmonary oedema/heart failure, chronic medicines including oral furosemide, an eGFR of 36, and a once-only IV furosemide prescription entry [1]. The point is not to memorise the pictured 50 mg dose as a universal answer. It is to synthesise the presentation, observations, investigation images, current loop-diuretic exposure, renal function, allergy status and indication, then complete every field: medicine, dose, route, frequency, duration/once-only status, date, time, and prescriber identity. Select and monitor the dose according to the actual patient and current acute-heart-failure guidance.
16. Integration with Related Lectures
The elderly prescribing checklist from GC 079 directly extends the principles in this lecture:
- STOPP criteria — drugs that should be STOPPED in elderly (potentially inappropriate)
- START criteria — drugs that should be STARTED in elderly (potentially beneficial but often omitted)
- Beers Criteria — AGS list of potentially inappropriate medications in older adults
Key renal thresholds from STOPP/START V3 [3]:
| Drug | eGFR Threshold | Risk |
|---|---|---|
| Digoxin ≥ 125µg/day long-term | < 30 | Digoxin toxicity |
| Dabigatran | < 30 | Bleeding |
| Rivaroxaban/Apixaban/Edoxaban | < 15 | Bleeding |
| NSAIDs | < 50 | Renal deterioration |
| Metformin | < 30 | Lactic acidosis |
| Spironolactone/Eplerenone | < 30 | Hyperkalemia |
| Bisphosphonates | < 30 | Acute renal failure |
| Nitrofurantoin | < 45 | Drug toxicity |
| Methotrexate | < 30 | Methotrexate toxicity |
Three principles of drug prescribing in renal impairment:
- Avoid further nephrotoxic insult (e.g., NSAIDs, aminoglycosides)
- Attention to correct dose (use the renal estimate specified for that medicine; this may be eGFR or estimated creatinine clearance)
- Beware of side effects amplified in renal failure (e.g., ethambutol → optic neuritis; acyclovir, quinolones, imipenem → CNS side effects)
- CYP450 enzyme interactions (CYP3A4 metabolizes ~50% of drugs)
- Type A (augmented, dose-dependent) vs. Type B (bizarre, idiosyncratic) adverse drug reactions
- CYP2C19 and clopidogrel — poor metabolizers need alternative antiplatelet therapy
17. Exam Intelligence
| Question Type | Example Stem |
|---|---|
| MCQ | "Which of the following is NOT one of the 5 Rights?" |
| MCQ | "A 70-year-old with eGFR 25 is prescribed metformin. What is the concern?" |
| SAQ | "List the elements of a valid prescription" |
| SAQ | "Calculate the incremental cost-effectiveness ratio given..." |
| SAQ | "Name 3 pharmacogenetic tests and the drugs they relate to" |
| Minicase | "A new HO prescribes amoxicillin to the wrong patient. Discuss using the Swiss Cheese Model" |
| OSCE | "Write a prescription for this patient" / "Review this prescription and identify errors" |
| Trap | Why Students Fall For It | Correct Answer |
|---|---|---|
| Using bed number to identify patient | Seems convenient | Never — use name + DOB/ID |
| Forgetting PRN parameters | Write drug + dose only | Must include indication, max dose, min interval, max 24h dose |
| Confusing cost-effectiveness with cheapest drug | Cheapest isn't always best | ICER considers both cost AND effectiveness |
| Overgeneralising carbamazepine screening | Treating all "Asian" patients as one population | Check HLA-B*15:02 according to ancestry and current local protocol; in Hong Kong guidance specifically names Han Chinese and Thai patients [6] |
| Ignoring renal function when prescribing | "The drug is commonly used" | Check renal function and use the drug-specific dosing metric; some medicines require creatinine clearance rather than eGFR |
From past Fourth Summative papers, prescribing questions commonly test:
- Prescription writing errors (SAQ/minicase format)
- Dose adjustment in renal impairment
- Drug interactions and ADRs
- Pharmacogenetics (HLA-B1502, HLA-B5801, TPMT)
- STOPP/START criteria in elderly
- Cost-effectiveness calculations
High Yield Summary
- Prescribing is a legal clinical act: record date, patient identity, drug, formulation, dose, frequency, route, additional directions, and signature.
- Protect the 5 Rights—patient, medication, dose, time, and route—and use at least two patient identifiers, never a bed number.
- A PRN order also needs its indication, maximum single dose, minimum interval, and maximum 24-hour dose.
- Medication safety is a system property: independent Swiss-cheese barriers should intercept hazards before they reach the patient.
- Individualise by pharmacokinetics, pharmacodynamics, organ function, interactions, response, preferences, and actionable biomarkers—not genetics alone.
- High-yield pharmacogenetic pairs include HLA-B15:02/carbamazepine, HLA-B58:01/allopurinol, HLA-B*57:01/abacavir, TPMT + NUDT15/thiopurines, and CYP2C19/clopidogrel.
- Quality prescribing balances efficacy, safety, tolerability, and cost-effectiveness; ICER = Δcost/Δeffect.
- Check renal function and use the dosing estimate required for the medicine. In older adults, start cautiously when appropriate, titrate to effect, and do not undertreat urgent disease.
Active Recall - Am I Prescribing the Right Drug?
[1] Lecture slides: GC 029. Am I prescribing the right drug.pdf [2] Senior notes: Block A - Drugs and the Kidney.pdf [3] Lecture slides: GC 079. Prescribing in older people.pdf; GC 079 (supp-2)STOPP-START-V3.pdf [4] Senior notes: Block A - Chronic diarrhoea_ irritable bowel syndrome and inflammatory bowel disease.pdf (Azathioprine/TPMT section); Block A - Patients with non-viral chronic liver diseases.pdf [5] Senior notes: Learning_Points_All_Lectures.txt (Clinical Pharmacology sections) [6] Hong Kong Department of Health Drug Office. Oral antiepileptic drugs (HLA-B*15:02 advice for carbamazepine). [7] CPIC. Guideline for HLA genotype and use of carbamazepine and oxcarbazepine. [8] CPIC. Thiopurine dosing based on TPMT and NUDT15 genotypes: 2025 update. [9] General Medical Council. Decision making and consent and deciding if it is safe to prescribe. [10] British Pharmacological Society Assessment. PSA learner portal.
GC028 Accelerating Chest Pain: Acute Coronary
Acute coronary syndrome encompasses a spectrum of conditions—unstable angina, NSTEMI, and STEMI—caused by sudden reduction in coronary blood flow, typically due to atherosclerotic plaque rupture and thrombosis, presenting with accelerating or new-onset chest pain requiring urgent evaluation and management.
GC030 An Old Man With Bone Pain And Anaemia
An elderly patient presenting with bone pain and anemia, most likely due to multiple myeloma — a malignant plasma cell neoplasm causing lytic bone lesions, marrow infiltration, and monoclonal immunoglobulin production.