Evidence-Based Prescribing: The Complete Clinician's Guide

Prescribing a medicine is the most common intervention in healthcare. In the United Kingdom alone, over 1.1 billion prescription items are dispensed annually in primary care, and every hospital admission generates multiple drug orders. Yet prescribing remains one of the most error-prone activities in medicine. Studies consistently show that between 7% and 10% of hospital prescriptions contain an error, and medication-related adverse events account for approximately 6.5% of all hospital admissions ^[3]. The majority of these are preventable.

This guide distils the principles, pharmacology, and practical wisdom that underpin safe, effective, evidence-based prescribing. Whether you are an F1 doctor writing your first inpatient drug chart or a consultant reviewing a complex polypharmacy regimen, these foundations apply to every prescription you write.

The Principles of Good Prescribing

The WHO 6-Step Model of Rational Prescribing

The World Health Organization's Guide to Good Prescribing ^[4] remains the most widely taught framework for rational prescribing worldwide. It provides a structured approach that guards against reflexive, habit-driven drug selection:

1. Define the patient's problem. Arrive at a specific diagnosis or, where that is not yet possible, a clearly defined symptom or syndrome requiring treatment. Prescribing without a working diagnosis is one of the most common root causes of inappropriate drug use.

2. Specify the therapeutic objective. What, precisely, are you trying to achieve? Cure (eradication of infection with antibiotics), symptom control (analgesia for pain), prevention (statins for cardiovascular risk reduction), or replacement (levothyroxine for hypothyroidism)? Each objective demands a different prescribing strategy.

3. Select an appropriate drug. Choose from your personal formulary of well-understood drugs, cross-referenced against current guidelines. Consider efficacy, safety profile, suitability for the individual patient, and cost.

4. Start the treatment. Write a clear, unambiguous prescription specifying the drug, dose, route, frequency, and duration. Provide the patient with information about what the drug does, how to take it, common side effects, and when to seek help.

5. Give information, instructions, and warnings. This step is frequently rushed or omitted entirely, yet it is the single most important determinant of adherence. Patients who understand why they are taking a medicine, and what to expect, are significantly more likely to take it correctly.

6. Monitor and review. Has the treatment achieved its objective? Has it caused harm? Does it need adjustment or discontinuation? Every prescription should have a defined review point.

The Prescribing Pyramid

The prescribing pyramid, widely used in UK medical education, extends the WHO model into a visual hierarchy. At its base sits the broadest competency: understanding the patient's condition and the therapeutic options. Each ascending layer narrows the focus — from choosing the drug, to choosing the dose and formulation, to writing the prescription correctly, to monitoring the outcome. The critical insight of the pyramid is that errors at any level undermine everything above it. Choosing the correct dose of the wrong drug is still a prescribing error.

Patient-Centred Prescribing

Evidence-based prescribing is not simply about matching a diagnosis to a guideline recommendation. It requires integrating three domains: the best available research evidence, the clinician's own expertise and judgement, and the patient's values, preferences, and circumstances ^[4]. A guideline may recommend an ACE inhibitor as first-line for hypertension, but the specific drug, the dose, and whether the patient will actually take it depend on factors no guideline can capture: renal function, concurrent medications, previous adverse reactions, ability to swallow tablets, daily routine, health beliefs, and financial constraints.

In practice, patient-centred prescribing means:

• Asking about previous medication experiences before prescribing something new
• Checking the patient's allergy and adverse reaction history — not just the allergy box, but the detail of what happened
• Considering the formulation (liquid, dispersible, modified-release, patch) in light of the patient's ability to take it
• Discussing the expected benefits and risks in language the patient understands
• Agreeing a plan that the patient is genuinely willing to follow

Shared Decision-Making in Drug Therapy

Shared decision-making has moved from aspiration to requirement. NICE expects it for all treatment decisions, and the GMC's Good Medical Practice mandates that doctors "work in partnership with patients" ^[2]. For prescribing, this means presenting the realistic options — including the option of not prescribing — and exploring what matters to the patient. A patient starting a statin for primary prevention has a different risk-benefit calculation from one starting it after a myocardial infarction. The evidence is the same; the conversation is different.

Decision aids, where available, improve the quality of shared decisions. For medications with narrow benefit-risk margins (anticoagulation in atrial fibrillation, for example), tools such as CHA2DS2-VASc and HAS-BLED scoring formalise the discussion and make the trade-offs explicit.

Pharmacokinetic Foundations for Prescribers

Every prescriber needs a working understanding of pharmacokinetics — not at the level of a pharmaceutical scientist, but enough to predict how a drug will behave in a given patient, anticipate interactions, and adjust doses rationally.

ADME in Clinical Context

Bioavailability and Why It Matters

Bioavailability (F) is the fraction of an administered dose that reaches the systemic circulation in unchanged form. Intravenous drugs have 100% bioavailability by definition. Oral bioavailability varies enormously: paracetamol approaches 90%, while oral morphine is approximately 30% due to extensive first-pass metabolism, and oral ciclosporin can be as low as 20-50% depending on the formulation. This is precisely why dose conversions between routes (IV to oral, or between oral formulations) must never be assumed to be 1:1.

First-Pass Metabolism

When a drug is absorbed from the gastrointestinal tract, it enters the portal venous system and passes through the liver before reaching the systemic circulation. During this first pass, hepatic enzymes can metabolise a significant proportion of the drug, reducing the amount that reaches its target. This is why some drugs (glyceryl trinitrate, lidocaine) are given sublingually, transdermally, or intravenously to bypass first-pass metabolism entirely. Liver disease can paradoxically increase the bioavailability of high first-pass drugs by reducing hepatic extraction — a patient with cirrhosis may experience toxicity at standard oral doses of propranolol or morphine.

Half-Life and Steady State

The elimination half-life (t1/2) is the time taken for the plasma concentration of a drug to fall by 50%. It determines dosing frequency and the time to reach steady state. At a constant dosing rate, steady state is reached after approximately 4-5 half-lives. This has direct clinical implications:

• Warfarin (t1/2 ~40 hours) takes approximately one week to reach steady state after a dose change. Checking an INR the day after changing the dose tells you almost nothing about the eventual effect.
• Digoxin (t1/2 ~36 hours) reaches steady state in about a week. A digoxin level taken before steady state is misleading.
• Amiodarone (t1/2 ~50 days) takes months to reach steady state and months to wash out. Interactions with amiodarone develop slowly and persist long after it is stopped.

Protein Binding and Free Drug Levels

Most drugs in the bloodstream exist in equilibrium between a protein-bound fraction (usually bound to albumin or alpha-1-acid glycoprotein) and a free (unbound) fraction. Only free drug is pharmacologically active — it crosses membranes, reaches receptors, and is available for metabolism and excretion. For the majority of drugs, this is clinically irrelevant because the total and free fractions rise and fall proportionally. However, for drugs that are highly protein-bound (>90%) and have a narrow therapeutic index, changes in protein binding become significant.

The classic example is phenytoin. In a patient with normal albumin, a total phenytoin level of 60 micromol/L corresponds to a therapeutic free level. In a hypoalbuminaemic patient, the same total level may represent a toxic free concentration. Corrected phenytoin calculations (or direct measurement of free phenytoin levels) are essential in critically ill patients, those with liver disease, and patients with nephrotic syndrome.

The CYP450 System: Clinical Relevance

The cytochrome P450 enzyme superfamily is responsible for Phase I metabolism of approximately 75% of all drugs in clinical use ^[1]. Three isoenzymes are responsible for the majority of clinically important interactions:

Phase I and Phase II Metabolism

Phase I reactions (oxidation, reduction, hydrolysis) introduce or expose a functional group on the drug molecule. These reactions can produce active metabolites (diazepam to desmethyldiazepam), inactive metabolites, or occasionally more toxic metabolites (paracetamol to NAPQI, the hepatotoxic intermediate).

Phase II reactions (conjugation with glucuronic acid, sulphate, glutathione, or acetyl groups) generally produce inactive, water-soluble metabolites ready for renal excretion. Importantly, Phase II reactions are generally preserved in liver disease even when Phase I reactions are impaired. This is why drugs relying solely on Phase II metabolism (lorazepam, oxazepam, temazepam — the "LOT" benzodiazepines) are preferred in patients with hepatic impairment over drugs requiring Phase I metabolism (diazepam, chlordiazepoxide, midazolam).

Drug Selection: Choosing the Right Medication

Evidence Hierarchy for Drug Selection

Not all evidence is equal. When choosing between therapeutic options, the hierarchy of evidence provides a framework for weighing the available data:

1. Systematic reviews and meta-analyses of randomised controlled trials (the highest level)

2. Individual randomised controlled trials (particularly large, well-conducted, double-blind trials)

3. Cohort studies and case-control studies

4. Case series and case reports

5. Expert opinion and clinical experience (the lowest level, but not negligible)

In practice, guidelines from NICE, SIGN, and specialist societies synthesise this evidence into actionable recommendations. The prescriber's role is to apply these recommendations to the individual patient — which sometimes means departing from the guideline with documented justification.

NICE Guidelines and Formulary Recommendations

In the UK, NICE guidelines and the British National Formulary (BNF) ^[2] are the twin pillars of prescribing guidance. NICE provides condition-specific treatment pathways with graded recommendations; the BNF provides drug-specific information including doses, interactions, side effects, and monitoring requirements. Local formularies further refine these recommendations based on cost-effectiveness, local resistance patterns, and procurement agreements. Prescribing outside the local formulary is not prohibited but should be justified and documented.

For Indian prescribers, the National List of Essential Medicines (NLEM) and state formulary guidelines serve an analogous function, while WHO Essential Medicines Lists provide a global baseline.

Cost-Effectiveness Considerations

Two drugs with equivalent efficacy and safety profiles may differ ten-fold in cost. Prescribing the cheaper option is not penny-pinching — it is an ethical obligation. The NHS drug tariff and local formulary preferentially list the most cost-effective options. Generic prescribing (by approved name rather than brand name) is standard practice in the UK and should be the default unless there is a specific reason to prescribe by brand (modified-release formulations with different bioavailability, such as lithium, theophylline, and some antiepileptics).

Patient Factors in Drug Selection

Guidelines recommend treatments for populations. Prescribers treat individuals. The following patient factors must be considered for every prescription:

• Comorbidities: An ACE inhibitor is first-line for hypertension — but not if the patient has bilateral renal artery stenosis or a history of angioedema. A beta-blocker is excellent for rate control in atrial fibrillation — but contraindicated in severe asthma.
• Allergies and adverse reactions: True allergy (anaphylaxis, urticaria) is different from intolerance (nausea, headache). Mislabelling a patient as "allergic" to penicillin when they experienced mild gastrointestinal upset denies them access to one of the safest and most effective antibiotic classes.
• Adherence factors: A once-daily medication is more likely to be taken than a four-times-daily one. A patient who struggles with polypharmacy may benefit from combination tablets. A patient who cannot swallow tablets needs a liquid or dispersible alternative.
• Patient preferences: Some patients have strong views about certain drug classes (steroids, opioids, psychiatric medications). Exploring these preferences and addressing concerns improves adherence and outcomes.

Therapeutic Alternatives and Step-Up/Step-Down Approaches

Many conditions are treated using a stepwise approach. Asthma management (BTS/SIGN guidelines), hypertension (NICE CG136), and type 2 diabetes (NICE NG28) all follow step-up models where first-line agents are tried before escalating to more potent or complex regimens. Equally important — and frequently neglected — is stepping down. If a patient's asthma has been well controlled for three months on a high-dose inhaled corticosteroid, the guideline explicitly recommends reducing to the lowest effective dose. Stepping down reduces side effects, cost, and pill burden without sacrificing control.

Dose Optimisation

Starting Dose, Titration, and Maintenance

The starting dose is not always the effective dose. Many drugs require gradual titration to the target or maximum tolerated dose. This is particularly true for:

• ACE inhibitors and ARBs in heart failure (start low, uptitrate every 2-4 weeks to target dose)
• Beta-blockers in heart failure (start at the lowest dose, double at 2-weekly intervals)
• Antidepressants (SSRIs may be started at half the usual dose in anxious patients)
• Opioids in chronic pain (start low, titrate slowly, review frequently)

The principle is "start low, go slow" — but do not stop at subtherapeutic doses. Many patients remain on starting doses of ACE inhibitors and beta-blockers in heart failure when the evidence for mortality benefit comes from trials using target doses.

Weight-Based Dosing

Some drugs are dosed by body weight, particularly in paediatrics and for drugs with narrow therapeutic indices. Common examples include low-molecular-weight heparins (enoxaparin 1.5 mg/kg once daily or 1 mg/kg twice daily for treatment), aminoglycosides (gentamicin 5-7 mg/kg once daily, using ideal body weight in obesity), and many chemotherapy agents. Using actual body weight for gentamicin in an obese patient will cause toxicity; using ideal body weight for enoxaparin will cause underdosing. Know which weight measure applies to which drug.

Renal Dose Adjustment (eGFR-Based)

Renal impairment affects drug clearance, and many drugs require dose reduction or avoidance when the estimated glomerular filtration rate (eGFR) falls. The BNF provides renal dosing guidance for each drug, typically stratified by eGFR bands: >50, 30-50, 15-30, and <15 mL/min/1.73m2.

Common examples:

• Metformin: Reduce dose at eGFR 30-45; stop at eGFR <30 (risk of lactic acidosis)
• DOACs: Apixaban and rivaroxaban require dose reduction at CrCl 15-29 mL/min; edoxaban requires dose reduction at CrCl 15-50 mL/min
• Digoxin: Reduce dose and monitor levels more frequently at eGFR <60
• Opioids: Morphine metabolites accumulate in renal failure — use oxycodone or fentanyl instead

Hepatic Dose Adjustment (Child-Pugh)

Hepatic dose adjustment is less standardised than renal adjustment because the liver's metabolic reserve is difficult to quantify. The Child-Pugh score (based on bilirubin, albumin, INR, ascites, and encephalopathy grade) classifies liver disease as A (mild), B (moderate), or C (severe), and drug data sheets often reference these categories. In general:

• Drugs with high hepatic extraction (propranolol, morphine, verapamil) have increased bioavailability in liver disease because first-pass metabolism is reduced
• Drugs requiring Phase I metabolism may accumulate; Phase II metabolism is relatively preserved
• Drugs that are hepatotoxic (methotrexate, statins at high doses, isoniazid) should be used with caution or avoided
• Protein binding is reduced (lower albumin), increasing the free fraction of highly bound drugs

Age-Related Pharmacokinetic Changes

Ageing reduces renal mass and function (even with a "normal" serum creatinine in an elderly patient, eGFR may be significantly reduced), hepatic blood flow and enzyme activity, total body water (increasing the concentration of water-soluble drugs), and lean body mass (while body fat increases, prolonging the half-life of lipophilic drugs). These changes collectively mean that older patients are more sensitive to many drugs at standard adult doses. The default should be to start at the lower end of the dosing range and titrate cautiously.

Drug Interactions: A Practical Framework

Pharmacokinetic Interactions

Pharmacokinetic interactions alter the concentration of a drug reaching its target. The most clinically important involve CYP enzyme inhibition or induction:

Other pharmacokinetic interactions include displacement from protein binding sites (generally transient and self-correcting), altered renal excretion (lithium levels rise when NSAIDs or ACE inhibitors reduce renal lithium clearance), and altered absorption (chelation of tetracyclines or quinolones by calcium, iron, or antacids).

Pharmacodynamic Interactions

Pharmacodynamic interactions occur when two drugs exert effects on the same physiological system, regardless of any change in drug levels:

• Additive: Two drugs with the same effect produce a combined effect equal to the sum of their individual effects. Example: prescribing an NSAID alongside an anticoagulant — both increase bleeding risk through different mechanisms.
• Synergistic: The combined effect exceeds the sum of individual effects. Example: trimethoprim plus sulfamethoxazole (co-trimoxazole) — sequential folate pathway blockade produces greater antibacterial effect than either drug alone.
• Antagonistic: One drug reduces the effect of another. Example: a beta-2 agonist (salbutamol) is antagonised by a non-selective beta-blocker (propranolol) — a potentially dangerous interaction in an asthmatic patient.

High-Risk Drug Combinations

Certain drug combinations carry sufficiently high risk that they should be avoided or, if absolutely necessary, managed with intensive monitoring:

• Methotrexate + trimethoprim: Both inhibit folate metabolism; the combination causes potentially fatal pancytopenia. Avoid if at all possible.
• Simvastatin + clarithromycin (or other potent CYP3A4 inhibitors): Risk of rhabdomyolysis. Contraindicated.
• ACE inhibitor + potassium-sparing diuretic + potassium supplement: Triple threat for hyperkalaemia.
• SSRI + tramadol (or other serotonergic drugs): Risk of serotonin syndrome.
• Warfarin + NSAID: Increased bleeding risk through both pharmacokinetic (CYP2C9 inhibition by some NSAIDs) and pharmacodynamic (impaired platelet function, gastrointestinal mucosal damage) mechanisms.
• QT-prolonging drugs in combination: Amiodarone + macrolide + ondansetron — each prolongs the QT interval; the combination can cause torsades de pointes.

Using Interaction Checkers Effectively

Electronic interaction checkers (BNF interactions tool, Stockley's Interactions, Lexicomp, Medscape) are essential clinical tools, but they require interpretation. Most checkers grade interactions by severity and evidence quality. The challenge is alert fatigue — a patient on 10 medications will generate dozens of interaction alerts, the majority of which are theoretical or manageable. The skill lies in identifying the alerts that require action (dose adjustment, enhanced monitoring, or drug substitution) versus those that can be acknowledged and monitored.

When to Proceed Despite an Interaction

Not every drug interaction mandates avoidance. Sometimes the interacting combination is the best treatment option, and the interaction can be managed:

• Prescribing a proton pump inhibitor with clopidogrel (the interaction is modest, and the gastric protection may be clinically necessary)
• Using amiodarone in a patient on warfarin (reduce warfarin dose by approximately one-third and monitor INR closely)
• Combining an ACE inhibitor with a potassium-sparing diuretic in heart failure (the combination is evidence-based, but potassium must be monitored)

Document the rationale, implement monitoring, and communicate the plan to the patient and the team.

Therapeutic Drug Monitoring (TDM)

Drugs Requiring TDM

Therapeutic drug monitoring is reserved for drugs where there is a clear relationship between plasma concentration and therapeutic or toxic effect, where the therapeutic window is narrow, and where clinical assessment alone is insufficient to guide dosing ^[1]. The classic TDM drugs are:

• Warfarin — monitored via INR rather than drug level; target INR 2.0-3.0 for most indications, 2.5-3.5 for mechanical mitral valves
• Lithium — target trough level 0.4-0.8 mmol/L for maintenance (up to 1.0 mmol/L in acute mania); toxicity occurs above 1.5 mmol/L and is life-threatening above 2.0 mmol/L
• Aminoglycosides (gentamicin, amikacin) — once-daily dosing regimens use trough levels (pre-dose) to guide dose interval adjustment; target depends on local protocol
• Vancomycin — trough levels (target 10-20 mg/L for most infections; AUC/MIC-guided dosing increasingly recommended)
• Digoxin — target trough level 0.5-1.0 nanogram/mL (lower end associated with mortality benefit in heart failure); toxicity above 2.0 nanogram/mL
• Phenytoin — target total level 10-20 mg/L (40-80 micromol/L); free level measurement required in hypoalbuminaemia
• Theophylline — target 10-20 mg/L; toxicity includes arrhythmias and seizures
• Ciclosporin and tacrolimus — trough levels monitored to prevent rejection (too low) or nephrotoxicity (too high)

When to Measure Levels

Timing of sampling is critical. Levels measured at the wrong time are uninterpretable:

• Trough levels (immediately before the next dose) are used for most TDM drugs. They represent the lowest concentration in the dosing interval.
• Digoxin levels should be taken at least 6 hours after the last dose (ideally 12 hours) to allow distribution to complete.
• Lithium levels should be taken 12 hours post-dose.
• Gentamicin trough levels in once-daily regimens are taken 18-24 hours post-dose, depending on local protocol.
• Levels should generally be taken at steady state (4-5 half-lives after starting or changing dose), unless toxicity is suspected, in which case an immediate level is appropriate.

Interpreting Trough vs Peak Levels

A trough level tells you whether the drug concentration drops too low between doses (risking therapeutic failure) or remains too high (risking toxicity). For aminoglycosides in traditional multiple-daily dosing, both peak (1 hour post-infusion) and trough levels are measured: the peak reflects efficacy (concentration-dependent killing), while the trough reflects safety (nephrotoxicity correlates with sustained high trough levels). Modern once-daily aminoglycoside regimens simplify this to a single trough-based nomogram.

Adjusting Doses Based on Levels

When a drug level is outside the target range, dose adjustment must account for:

• Whether the patient was at steady state when the level was taken
• Whether the timing of the sample was correct
• Whether there are reversible factors (dehydration increasing lithium levels; a new CYP inhibitor increasing phenytoin levels)
• The drug's pharmacokinetics — phenytoin, in particular, has saturable (zero-order) kinetics at therapeutic doses, meaning that small dose increases can produce disproportionately large rises in plasma concentration

Prescribing in Special Populations

Renal Impairment

Renal impairment is one of the most common reasons for dose adjustment, and failure to adjust doses is one of the most common causes of drug toxicity in hospital practice.

• NSAIDs — reduce renal blood flow; avoid in eGFR <30 (ideally avoid in all significant renal impairment)
• Metformin — risk of lactic acidosis; reduce at eGFR 30-45, stop below 30
• Lithium — renally excreted; dose reduction and more frequent monitoring required
• ACE inhibitors/ARBs — may be used in renal disease (they are renoprotective in diabetic nephropathy) but require monitoring of potassium and creatinine, particularly at initiation and dose changes
• Gentamicin and vancomycin — extend dosing interval based on levels

The BNF and many drug manufacturers base renal dosing recommendations on creatinine clearance (CrCl) calculated using the Cockcroft-Gault equation, which uses actual body weight. The laboratory-reported eGFR (CKD-EPI equation) is normalised to body surface area and may overestimate renal function in small or cachectic patients and underestimate it in large patients. For most clinical purposes, eGFR is sufficient. For drugs with critical renal dosing thresholds (DOACs, aminoglycosides), calculate CrCl specifically.

Hepatic Impairment

Hepatic dose adjustment is less precise than renal adjustment, but the principles are well established.

• NSAIDs — risk of renal failure, gastrointestinal bleeding, and fluid retention
• Opioids (especially morphine) — increased bioavailability and reduced clearance; use with extreme caution at reduced doses; avoid codeine entirely (unpredictable activation)
• Methotrexate — hepatotoxic
• Sedatives and anxiolytics — increased sensitivity; if a benzodiazepine is needed, use one of the "LOT" drugs (lorazepam, oxazepam, temazepam) that undergo Phase II metabolism only
• Statins — increased risk of hepatotoxicity; use at reduced doses with monitoring

Pregnancy and Breastfeeding

Prescribing in pregnancy requires balancing the risk of drug exposure to the developing foetus against the risk of untreated disease in the mother. The old FDA letter-based risk categories (A, B, C, D, X) have been replaced by narrative descriptions in many jurisdictions because the letter grades were misleading and oversimplified.

• Sodium valproate — neural tube defects, cognitive impairment; the Pregnancy Prevention Programme applies to all women of childbearing potential
• Isotretinoin — severe birth defects; mandatory pregnancy prevention measures
• Methotrexate — abortifacient and teratogenic; must be stopped at least 3 months before conception
• Warfarin — embryopathy in the first trimester (nasal hypoplasia, stippled epiphyses); CNS abnormalities with exposure later in pregnancy
• ACE inhibitors and ARBs — renal dysgenesis and oligohydramnios in the second and third trimesters; avoid throughout pregnancy
• Lithium — Ebstein's anomaly (lower risk than historically believed, approximately 1 in 1,000, but still significant)
• Carbamazepine, phenytoin — neural tube defects (lower risk than valproate)

Elderly Patients

Older adults bear a disproportionate burden of drug-related harm. Polypharmacy (commonly defined as five or more regular medications) is almost universal in the over-75s, and inappropriate polypharmacy — where one or more drugs are unnecessary, duplicated, or causing more harm than benefit — is the norm rather than the exception.

• STOPP: Long-term benzodiazepines (falls risk, cognitive impairment); PPIs beyond 8 weeks without clear indication; loop diuretics for ankle oedema without heart failure; anticholinergic drugs in patients with cognitive impairment
• START: ACE inhibitor in heart failure with reduced ejection fraction; statin post-myocardial infarction; bisphosphonate in patients on long-term corticosteroids; influenza and pneumococcal vaccination

1. Identify the drug to be considered for deprescribing

2. Assess the current indication, benefits, and harms

3. Assess the feasibility of stopping (some drugs require tapering — beta-blockers, corticosteroids, antidepressants, benzodiazepines, opioids)

4. Plan the withdrawal schedule with the patient

5. Monitor for relapse or withdrawal effects

Paediatric Prescribing

Children are not small adults. Their pharmacokinetics differ qualitatively, not just quantitatively, from adults, and these differences change continuously as the child grows.

• Neonates have immature hepatic and renal function, reduced protein binding, and a higher proportion of body water, all of which alter drug handling
• Infants and young children have faster metabolic rates (per kilogram) than adults for some drugs, requiring proportionally higher doses
• Adolescents approach adult pharmacokinetics but may still require weight-based dosing for some drugs

Prescribing Errors and How to Prevent Them

Prescribing errors are common, costly, and largely preventable. The EQUIP study found that 8.9% of medication orders in UK hospitals contained a prescribing error, with 2.0% considered potentially serious ^[3].

Common Error Types

• Wrong drug: Often caused by look-alike/sound-alike drug names or selection errors in electronic prescribing systems (picking the wrong drug from a dropdown)
• Wrong dose: The most common error type, ranging from simple decimal point errors (10-fold overdoses) to failure to adjust for renal impairment
• Wrong route: Oral drugs given intravenously (vincristine administered intrathecally is invariably fatal; methotrexate intended for oral use given parenterally at the oral dose)
• Wrong frequency: Prescribing a twice-daily drug once daily, or vice versa
• Wrong patient: Particularly in hospital wards where drug charts may be at the end of the bed
• Omission errors: Failure to prescribe essential regular medications (VTE prophylaxis, insulin, anti-epileptic drugs) on admission to hospital
• Drug interactions and allergies missed: Prescribing a drug that interacts dangerously with the patient's existing medications, or prescribing a drug to which the patient has a documented allergy

Systems Approach to Error Prevention

Individual vigilance is necessary but insufficient. A systems approach to prescribing safety includes:

• Standardised prescribing processes: Structured drug charts, mandatory allergy documentation, standard dosing protocols
• Clinical decision support: Electronic prescribing systems with interaction checks, allergy alerts, dose range checks, and renal dosing calculators
• Pharmacist review: Clinical pharmacists reviewing all prescriptions on ward rounds or within 24 hours of admission
• Prescribing education: Structured training for all prescribers, including simulation of high-risk prescribing scenarios
• Incident reporting and learning: A culture of reporting near-misses and errors without blame, with systematic analysis and feedback

Electronic Prescribing Safety Features

Electronic prescribing and medicines administration (ePMA) systems reduce certain error types (illegible handwriting, incomplete prescriptions) but introduce others (alert fatigue, wrong-patient selection, copy-forward errors). Effective use of ePMA requires:

• Engaging with clinical decision support alerts rather than overriding them reflexively
• Verifying the patient identity before prescribing
• Reviewing the full medication list, not just the current order
• Being aware of the system's limitations (it may not detect all interactions, particularly with non-prescription drugs)

Double-Checking High-Risk Medications

Certain drugs carry such high risk that independent double-checking (by a second prescriber, pharmacist, or nurse) is mandated in most healthcare organisations. These typically include insulin, opioids, anticoagulants, potassium chloride concentrate, cytotoxic agents, and high-risk intravenous infusions.

TALL Man Lettering and Look-Alike/Sound-Alike Drugs

TALL man lettering uses uppercase letters to highlight the differences between look-alike drug names: hydrOXYzine vs hydrALAZINE, chlorproMAZINE vs chlorproPAMIDE, predniSONE vs prednisoLONE. While the evidence for its effectiveness is mixed, it is widely adopted as one component of a multi-layered safety strategy.

Antimicrobial Stewardship

Antimicrobial resistance is one of the greatest threats to global health. Every antibiotic prescription exerts selective pressure that promotes resistance. Prescribers have a direct responsibility to use antibiotics judiciously.

Principles of Antibiotic Prescribing

• Do not prescribe antibiotics for self-limiting viral infections. Upper respiratory tract infections, most cases of acute bronchitis, and many cases of acute sinusitis do not benefit from antibiotics.
• Take cultures before starting antibiotics wherever possible, particularly for bloodstream, urinary tract, and respiratory infections. Empirical therapy can be rationalised (narrowed or stopped) once culture results are available.
• Use the narrowest effective spectrum. Broad-spectrum antibiotics (co-amoxiclav, piperacillin-tazobactam, carbapenems) should be reserved for situations where narrow-spectrum agents are inappropriate.
• Prescribe for the shortest effective duration. Evidence increasingly supports shorter courses for many infections: 5 days for community-acquired pneumonia (if clinically improving), 3-5 days for uncomplicated UTI, 5 days for cellulitis.

Empirical vs Targeted Therapy

Empirical therapy is based on the likely causative organisms and local resistance patterns, prescribed before culture results are available. It should cover the most probable pathogens for the clinical syndrome. Targeted therapy is guided by culture and sensitivity results, allowing de-escalation to the narrowest effective agent. The "start smart, then focus" approach — mandated in NHS hospitals — formalises this two-step process.

IV to Oral Switch Criteria

Intravenous antibiotics should be switched to oral equivalents as soon as clinically appropriate. Standard switch criteria (the OPAT/BSAC criteria) include:

• Patient clinically improving (temperature normalising, inflammatory markers falling)
• Able to tolerate oral medication (not vomiting, functioning gastrointestinal tract)
• An appropriate oral agent available with adequate bioavailability for the infection being treated
• No indication for continued IV therapy (endocarditis, meningitis, severe sepsis requiring high tissue concentrations)

Early IV-to-oral switch reduces line infections, enables earlier discharge, and reduces cost without compromising outcomes.

Resistance Patterns and Local Guidelines

Antibiotic resistance patterns vary by region, hospital, and even ward. Local antibiotic guidelines, developed by microbiology and pharmacy teams based on local resistance surveillance, should always take precedence over generic textbook recommendations. These are typically available on hospital intranets and should be the first resource consulted when prescribing empirical antibiotics.

Controlled Drugs and Legal Prescribing

Controlled drugs are subject to additional legal requirements designed to prevent misuse, diversion, and dependence. In the UK, controlled drugs are classified under the Misuse of Drugs Act 1971 and regulated by the Misuse of Drugs Regulations 2001.

UK Controlled Drug Schedules

• Schedule 2: Opioids (morphine, oxycodone, fentanyl, methadone, diamorphine), amphetamines, cocaine. Full controlled drug prescription requirements; must be stored in a CD cupboard; a register must be maintained.
• Schedule 3: Barbiturates, buprenorphine, midazolam, temazepam, tramadol. CD prescription requirements apply (except for temazepam); no register requirement (except for phenobarbital in some settings).
• Schedule 4: Benzodiazepines (except temazepam and midazolam), zopiclone, zolpidem, anabolic steroids. Standard prescription requirements; no CD cupboard or register required.
• Schedule 5: Low-strength preparations containing small amounts of controlled drugs (codeine linctus, co-codamol, pholcodine). Minimal additional requirements.

Prescription Requirements for Controlled Drugs

Schedule 2 and 3 controlled drug prescriptions must include:

• The patient's full name, address, and date of birth (or age if under 12)
• The name and form of the preparation
• The strength (where appropriate)
• The dose
• The total quantity in words and figures
• The prescriber's signature and date

Electronic prescribing systems handle many of these requirements automatically, but prescribers must verify that all mandatory fields are completed.

Safe Prescribing of Opioids

Opioid prescribing requires particular care:

• Assess pain systematically before prescribing (not all pain requires opioids)
• Start at the lowest effective dose; titrate gradually
• Prescribe laxatives prophylactically (opioid-induced constipation is almost universal)
• Consider the patient's opioid tolerance status — an opioid-naive patient is at much higher risk of respiratory depression
• Plan an exit strategy: set a review date, plan for tapering, and discuss expectations with the patient from the outset
• In chronic non-cancer pain, the evidence for long-term opioid benefit is weak, and the harms (dependence, hyperalgesia, hormonal disruption, cognitive impairment) are substantial

Benzodiazepine Prescribing Guidelines

Benzodiazepines should be prescribed for the shortest possible duration (ideally 2-4 weeks) because tolerance and dependence develop rapidly. Long-term benzodiazepine use is associated with cognitive impairment, falls, and road traffic accidents. Withdrawing benzodiazepines after long-term use requires gradual dose reduction over weeks to months to avoid withdrawal seizures. The British National Formulary recommends switching to an equivalent dose of diazepam (which has a longer half-life, allowing smoother tapering) and reducing by approximately 10% of the dose every 1-2 weeks.

MedNext Formulary: Your Prescribing Companion

Evidence-based prescribing depends on ready access to reliable, up-to-date drug information at the point of care. The MedNext Formulary has been designed with this need in mind.

Frequently Asked Questions

What is the most common prescribing error?

Dose errors are consistently the most common prescribing error type across all healthcare settings. Within hospitals, the most frequent subtypes are incorrect dose for the patient's renal function, omission of a regular medication on admission, and ten-fold dose errors (often caused by decimal point misplacement). In primary care, prescribing a drug to which the patient has a documented allergy and failing to review long-term medications are also prominent.

How do I check for drug interactions quickly?

Use a structured approach: when adding a new drug to a patient's regimen, check it against all existing medications using the BNF interactions tool, Stockley's Interactions (available online via MedicinesComplete), or the MedNext Formulary interaction checker. Prioritise interactions involving narrow therapeutic index drugs. In acute settings, if time is limited, focus on the most dangerous interactions: anticoagulants, immunosuppressants, antiepileptics, antiarrhythmics, and lithium.

When should I refer to a clinical pharmacist?

Consider involving a clinical pharmacist when managing complex polypharmacy (particularly in elderly patients), when a patient has significant renal or hepatic impairment requiring multiple dose adjustments, when prescribing drugs requiring therapeutic drug monitoring, when managing drug interactions that cannot be avoided, and when planning a deprescribing strategy. In hospital practice, clinical pharmacists review all medication charts — but highlighting specific concerns proactively leads to better outcomes than waiting for routine review.

How do I approach deprescribing?

Start by identifying the drug with the worst benefit-to-harm ratio in the patient's regimen. Consider drugs that were started for a condition that has resolved, drugs that are duplicated (two antihypertensives from the same class, for example), drugs causing identifiable side effects, and prophylactic drugs where the original indication no longer applies. Use the STOPP criteria as a systematic screening tool. Discuss the plan with the patient — many patients are relieved to be taking fewer tablets. Taper drugs that require it (beta-blockers, corticosteroids, opioids, benzodiazepines, antidepressants) and monitor for recurrence of the original condition.

What are the most dangerous drug interactions?

The drug interactions most likely to cause serious harm or death include:

• Methotrexate + trimethoprim — fatal pancytopenia
• Warfarin + miconazole oral gel — massive INR rise and life-threatening bleeding
• Potassium supplements + potassium-sparing diuretics in renal impairment — fatal hyperkalaemia
• Simvastatin + clarithromycin — rhabdomyolysis
• Lithium + NSAIDs — lithium toxicity
• Intrathecal vincristine (strictly a route error, not an interaction, but one of the most devastating medication safety failures in medical history)

In each case, the interaction is well-documented, predictable, and preventable through systematic prescribing checks.

---