Professional Role · Topic 17 of 21

Evidence-Based Practice & Levels of Evidence

Professional role · evidence-based practice

Evidence-Based Practice & Levels of Evidence

Best evidence plus clinical expertise plus patient values: ask a PICOT question, rank the evidence, appraise it, then pilot and evaluate the change.

Key numbers

Melnyk evidence hierarchy: Levels I–VIILevel I = SR/MA of RCTs

01 · The rulePICOT frames the clinical question

Write the question in PICOT format — Population, Intervention (or issue of interest), Comparison, Outcome, Time — e.g., in ventilated ICU adults, does daily sedation interruption versus continuous sedation shorten ventilation during the ICU stay?

02 · The ruleMelnyk evidence hierarchy: Levels I–VII

Melnyk levels for intervention questions: I systematic review/meta-analysis of RCTs; II single RCT; III non-randomized controlled (quasi-experimental); IV cohort or case-control; V systematic review of descriptive/qualitative; VI single descriptive/qualitative; VII expert opinion or committee reports.

03 · DoMelnyk's seven EBP steps (0–6)

Melnyk's steps: 0 cultivate a spirit of inquiry; 1 ask a PICOT question; 2 search for evidence; 3 critically appraise it; 4 integrate it with expertise and patient preferences; 5 evaluate outcomes; 6 disseminate.

The rules

Know these

GRADE: high, moderate, low, very low certaintyGRADE rates certainty (quality) of evidence as high, moderate, low or very low; randomized trials without important limitations start as high and observational studies start as low.

GRADE downgrades for bias, inconsistency, indirectness, imprecisionGRADE rates certainty down for risk of bias, inconsistency, indirectness, imprecision and publication bias, and up for a large effect, a dose–response gradient, or plausible confounding that would shrink the observed effect.

Johns Hopkins hierarchy: Levels I–VJohns Hopkins levels: I experimental/RCT; II quasi-experimental; III nonexperimental and qualitative; IV clinical practice guidelines and consensus panels; V experiential and non-research evidence (QI projects, literature reviews, case reports, expert opinion).

Appraise: valid, important, applicable to my patients?Rapid critical appraisal asks three questions of each study: Are the results valid? What are the results, and are they important? Will the results help me care for my patients?

Match the study design to the questionEvidence hierarchies rank intervention questions; higher levels carry less risk of bias, but some designs fit some questions better (qualitative for experience, cohort for prognosis or harm), so hierarchies aid judgment, not replace it.

What it is

Core idea

EBP = evidence + expertise + patient values must knowEvidence-based practice integrates the best available research evidence with clinical expertise and the patient's values and preferences; research evidence alone does not justify a practice change.

Management

Orders

Johns Hopkins model: Practice question, Evidence, TranslationThe Johns Hopkins EBP model runs a three-phase PET process: Practice question (team and PICO), Evidence (search, appraise level and quality, synthesize), and Translation (action plan, implement, evaluate, disseminate).

Iowa model: three go/no-go decision pointsThe Iowa model starts from a problem- or knowledge-focused trigger and passes three decision points: Is this topic a priority for the organization? Is there sufficient evidence? Is the change appropriate for adoption in practice?

Pilot the change before system-wide adoptionBefore adopting a change across the system, pilot it on one unit: choose outcomes, collect baseline and post-pilot data, evaluate process and outcomes, and modify the practice; adopt only if the pilot supports it.

Insufficient evidence: conduct research, not a practice changeWith insufficient evidence, the Iowa model routes the team to conduct research — or, if a decision is needed now, to other evidence (case reports, expert opinion, scientific principles) — rather than adopt a change.

Melnyk or Johns Hopkins hierarchy?

Commonly confused

Number of levels	7 (I–VII)
Level I	Systematic review or meta-analysis of RCTs
Cohort and case-control studies	Level IV
Expert committee reports	Level VII (opinion of authorities, expert committees)
Process model	Seven steps (Step 0–6)

Commonly confused

Tell apart

Recommendation strength is not evidence certaintyGRADE recommendations are strong ("we recommend") or weak/conditional ("we suggest"); strength also weighs values, resources and feasibility, so low-certainty evidence can still support a strong recommendation (e.g., prompt antibiotics in sepsis).

More detail

Reference

Show 3 reference facts

AGREE II is the international tool for appraising the quality and reporting of clinical practice guidelines before adopting their recommendations locally.

The 2017 Iowa Model-Revised, built from a survey of 431 users and validated by 299 EBP conference participants, expanded piloting, implementation, patient engagement and sustaining change.

Disseminate EBP results through institutional EBP rounds, local and national conference presentations, peer-reviewed journals and professional newsletters; evaluation data then feed new triggers.

Pearl

Start with Step 0 — a spirit of inquiry — and never stop at Step 3: appraised evidence that is never implemented, evaluated and shared changes nothing.

Sources · University of Colorado Strauss Health Sciences Library — Levels of Evidence (Melnyk & Fineout-Overholt 2015 Box 1.3; Johns Hopkins EBP Model 2022 Appendix D) · Melnyk BM, Fineout-Overholt E, Stillwell SB, Williamson KM. Evidence-based practice: step by step — the seven steps of evidence-based practice. Am J Nurs 2010;110(1):51–53 · Columbus State Community College Library — Evidence-Based Practice (summarizes Melnyk et al. 2010 steps) · Titler MG et al. The Iowa Model of Evidence-Based Practice to Promote Quality Care. Crit Care Nurs Clin North Am 2001;13:497–509 (decision points, triggers, pilot; quotes Sackett’s EBM definition) · GRADE Handbook (Schünemann H, Brożek J, Guyatt G, Oxman A, eds.) — GRADE Working Group · Johns Hopkins Welch Medical Library — Johns Hopkins Nursing Evidence-Based Practice Model (PET process) · Iowa Model Collaborative. Iowa Model of Evidence-Based Practice: Revisions and Validation. Worldviews Evid Based Nurs 2017;14:175–182 · AGREE Enterprise — AGREE II instrument (Brouwers M et al. CMAJ 2010) · Starter sheet built from the verified AGACNP fact map (facts reviewed 2026-10-15).