L2 - Hierarchy of Evidence & Pharmacist Decision-Making

Learning Objectives

• By the end of the lecture students should be able to:
  • Describe core principles of evidence-based practice (EBP) and its application in pharmacy.
  • Identify and differentiate the levels of scientific evidence that form the evidence hierarchy.
  • Evaluate strengths and limitations of each evidence type: expert opinion, case reports, observational studies, randomized controlled trials (RCTs), systematic reviews, and meta-analyses.
  • Interpret evidence in order to optimise quality use of medicines (QUM) – ensuring effectiveness, safety, judiciousness, and appropriateness.
  • Recognise how evidence shapes clinical guidelines and fuels decision-support tools in real-world practice.

Evidence-Based Practice (EBP): Definition & Pillars

• Process of combining three streams:
  • Best available research evidence (hierarchy-rated).
  • Clinician’s own expertise & experience.
  • Patient values, preferences, and specific clinical context.
• Goal: Produce collaborative, patient/population-centred decisions for a given healthcare scenario.
• Practical realities:
  • Some therapeutic areas have abundant high-quality evidence, others rely on lower-tier data or consensus.
  • Off-label or rare-condition prescribing in hospitals often needs expert panels, drug-committee approvals, and local policies in lieu of large trials.
  • Complementary medicines illustrate the grey zone where claims outpace rigorous data; pharmacists must appraise and balance evidence with patient preference.

Why Evidence Quality Matters in Pharmacy

• Effective use: Achieve intended clinical outcomes through evidence-guided drug and dose selection.
• Safe use: Pre-empt interactions, contraindications, adverse effects—prevent medication-related harm.
• Judicious use: Avoid unnecessary polypharmacy, over-treatment, wasted resources.
• Appropriate use: Tailor therapy to patient characteristics (age, comorbidities, pregnancy, renal/hepatic function, social factors) and logistical context.
• Professional responsibility: Critically appraise literature, understand bias/rigor, communicate uncertainty honestly.

The Evidence Hierarchy (Visual Pyramid Concept)

1. Systematic Reviews & Meta-analyses – synthesised, highest level.
2. Randomized Controlled Trials (RCTs) – controlled, prospective intervention.
3. Observational Studies – cohort, case-control, cross-sectional.
4. Case Reports / Case Series.
5. Expert Opinion & Narrative Review.

Detailed Levels of Evidence

Expert Opinion

• Authored by recognised specialists; fills gaps when robust studies absent (e.g., early COVID-19 airway management paper on high-flow nasal oxygen).
• Strengths:
  • Rapid, context-rich guidance in emerging or rare scenarios.
  • Utilises clinicians’ tacit knowledge.
• Limitations:
  • No systematic methodology; high risk of bias & conflict of interest.
  • May reflect individual perspective rather than generalisable data.

Case Reports / Case Series

• In-depth description of a single (or few) patient(s) with novel presentation, adverse event, or therapeutic approach.
• Example: BMJ report of remdesivir-associated bradycardia recommending baseline ECG & continuous monitoring.
• Strengths: Spotlights rare or early-signal phenomena; hypothesis-generating.
• Limitations: N=1, lacks control/comparator; publication & observation bias; cannot establish causality.

Consensus Statements (e.g., Delphi Studies)

• Structured process (Delphi rounds) to achieve agreement among experts.
• COVID-19/TB global consensus example: Experts scored relevance (1–5) for statements on viral–TB interactions; iterative anonymous questionnaires until consensus achieved.
• Useful when: trials infeasible, urgent guidance needed, complex scope-of-practice questions.
• Still subject to expert bias but more methodical than informal opinion.

Observational Studies

• No investigator-imposed intervention—researchers simply "observe".
• Types:
  • Cohort (prospective or retrospective): follow exposed vs. non-exposed groups over time.
  • Case-control: start with outcome (case) vs. no outcome (control) and look retrospectively for exposures.
  • Cross-sectional: snapshot to assess prevalence/associations at a single time point.
• Strengths:
  • Feasible when RCT unethical or impractical (rare events, long latency).
  • Can leverage existing dispensing or electronic health record (EHR) data.
• Limitations:
  • Confounding variables (comorbidities, concomitant meds) and selection bias uncontrolled.
  • Cannot definitively prove causality.

Randomized Controlled Trials (RCTs)

• Participants randomly allocated to intervention vs. control (often placebo or standard-of-care), preferably double-blinded.
• Key features:
  • Randomisation → minimises selection bias.
  • Blinding → minimises performance & detection bias.
  • Pre-defined primary & secondary outcomes (efficacy, safety, survival, quality-of-life…).
• Sample-size (power) calculation: determines n needed to detect a pre-specified effect size (e.g., \Delta =10\% difference) with adequate statistical power.
• Strengths:
  • Gold standard for establishing efficacy & safety; cornerstone for regulatory approvals (TGA, PBS, FDA).
• Limitations:
  • Resource-intensive (time, funding, staff, monitoring, audits, pharmacy blinding procedures).
  • Strict inclusion/exclusion criteria may limit generalisability to paediatrics, pregnancy, obesity, multimorbidity, etc.

Systematic Reviews

• Rigorous, reproducible method to locate, screen, appraise, and synthesise all studies addressing a focused research question (often framed in PICO).
• Steps:
  1. Register protocol on PROSPERO for transparency.
  2. Define PICO:
  • P – Population/Problem (e.g., middle-aged male amputees with phantom limb pain).
  • I – Intervention/Exposure (e.g., gabapentin).
  • C – Comparator (e.g., placebo).
  • O – Outcome (e.g., pain reduction).
  1. Develop database search strategy (e.g., Medline, Embase) with keywords/MeSH; publish strategy.
  2. PRISMA flow diagram: records number of articles identified, screened, excluded (with reasons), and finally analysed.
  3. Quality assessment of included studies (risk-of-bias tools).
• Outputs: qualitative narrative synthesis and/or quantitative meta-analysis.
• Limitations: overall conclusions limited by quality & heterogeneity of included studies; sparse data may preclude meta-analysis.

Meta-analysis

• Statistical pooling of results from multiple comparable studies within a systematic review.
• Generates a single summary effect size (e.g., risk ratio RR, odds ratio OR, mean difference \Delta\mu) with 95\% confidence interval.
• Uses weighting: larger, more precise studies contribute more influence.
• Forest plot interpretation:
  • Individual study squares (effect, CI) + overall diamond.
  • Centre line denotes RR = 1 (no effect) or \Delta\mu = 0 (difference).
  • Heterogeneity quantified by I^2 statistic (e.g., I^2 = 25\% low, >75\% high).
• Strengths: Improves power, clarifies direction/size of effect across settings.
• Limitations: Garbage-in/garbage-out (sensitive to included study bias); misleading if heterogeneity extreme or publication bias unaddressed.

Translation of Evidence into Practice

• Clinical Practice Guidelines (CPGs):
  • Synthesise evidence + expert consensus → graded recommendations (e.g., WHO TB guidelines, national therapeutic guidelines, hospital policies).
  • Range: international umbrella guidance to local procedure-specific protocols.
• Decision-support Tools:
  • Diagnostic algorithms, treatment pathways, risk calculators (e.g., Type 2 Diabetes Risk; CVD absolute risk calculators).
  • Built and validated using datasets from multiple evidence sources; applicability limited to validation population.
• Pharmacist Role:
  • Navigate guidelines, understand levels of evidence underpinning each statement.
  • Apply or adapt recommendations to individual patient considering context and values.

Summary & Key Takeaways

• Evidence exists on a continuum—from anecdote to meta-analysis; hierarchy guides confidence but does not render lower levels useless.
• Critical appraisal requires understanding study design, bias, confounders, statistical significance, and clinical relevance.
• Pharmacists integrate high-quality evidence with clinical judgement and patient factors to realise QUM.
• Systematic reviews/meta-analyses provide top-tier synthesis but must be scrutinised for heterogeneity and bias (I^2, risk-of-bias tools).
• RCTs remain gold standard for efficacy/safety but are costly and exclude many patient groups→ complementary evidence needed.
• Observational and lower-tier evidence crucial for rare events, early signals, and real-world practice—interpret with caution.
• Rigorous methodology (PROSPERO registration, PRISMA, Delphi) and transparent reporting protect against bias and enhance trust in conclusions.
• Ultimately, evidence synthesis underpins clinical guidelines and decision-support tools that shape daily prescribing, dispensing, and counselling activities.