Chart — Professional Practice
Levels of Evidence Chart
Evidence is ranked by how well its design resists bias — randomization and synthesis at the top, opinion at the bottom. The seven levels, what each design actually is, and an example you’d recognize in a journal.
Educational use only. Level is only the starting point of appraisal — a poorly conducted RCT can be weaker than an excellent cohort study; quality and applicability always get judged alongside design. This material supports nursing education and exam review. It is not medical advice and is not a substitute for clinical judgment, institutional policy, or medical direction. Always follow facility protocols and current provider orders.
The Pyramid
Level I · Systematic review / meta-analysis of RCTs
Level II · Randomized controlled trial (RCT)
Level III · Controlled trial without randomization (quasi-experimental)
Level IV · Case-control or cohort study
Level V · Systematic review of descriptive/qualitative studies
Level VI · Single descriptive or qualitative study
Level VII · Expert opinion / committee reports
Strongest (top) → weakest (bottom)
Level by Level
| Level | Design | What It Is | Example |
|---|---|---|---|
| I | Systematic review / meta-analysis of RCTs | All relevant randomized trials located, appraised, and statistically combined | A Cochrane review pooling 24 RCTs of chlorhexidine bathing to prevent CLABSI |
| II | Randomized controlled trial (RCT) | Participants randomly assigned to intervention or control — randomization controls unknown confounders | Patients randomized to early mobilization vs usual care after cardiac surgery |
| III | Controlled trial without randomization (quasi-experimental) | Intervention and comparison groups exist, but assignment isn't random | One unit adopts hourly rounding; a similar unit doesn't; fall rates compared |
| IV | Case-control or cohort study | Observational: groups followed forward (cohort) or compared backward from an outcome (case-control) | Following 5,000 nurses for a decade to link shift work and hypertension |
| V | Systematic review of descriptive/qualitative studies | Synthesis of non-experimental studies (meta-synthesis) | A synthesis of 18 qualitative studies on ICU family experiences |
| VI | Single descriptive or qualitative study | One non-experimental study — surveys, interviews, chart reviews | Interviews with 20 new grads about transition shock |
| VII | Expert opinion / committee reports | Authority without systematic data — the floor of the pyramid | An editorial or consensus statement by specialty leaders |
Exam Traps
- ✦"Strongest evidence" = systematic review/meta-analysis of RCTs — not a single RCT, however large.
- ✦Randomization is what separates Level II from III — spot the words 'randomly assigned.'
- ✦Cohort goes forward in time from exposure; case-control works backward from outcome.
- ✦Qualitative research is low on this hierarchy but answers questions RCTs can't — 'what is the patient experience?' has no randomized answer.
- ✦Expert opinion is still evidence (Level VII) — the answer 'not evidence at all' is the trap.
Related Resources
Standards & sources
Fact-checked Jun 21, 2026This page is written to align with ANA Code of Ethics & Scope/Standards of Practice · NCSBN · HIPAA (U.S. HHS). It is an educational summary, not a citation of any single document — always verify specific doses, values, and protocols against current guidelines and your facility policy. How we source content →