Levels of evidence in nursing research rank studies by how well their design protects against bias. Higher levels give stronger grounds for changing practice; lower levels still matter, but carry less weight.
This guide explains the most widely taught hierarchy, how the Johns Hopkins model compares, how to place a study correctly, and how to present your appraisal in an evidence table.
The Seven Levels of Evidence at a Glance
The seven-level hierarchy taught in many nursing programs, associated with Melnyk and Fineout-Overholt, runs from systematic reviews of randomized controlled trials at Level I to expert opinion at Level VII.
The higher the level, the more confident you can be that an intervention caused the outcome. But level is only half the story: you must also appraise each study's quality.
Why Levels of Evidence Exist
Evidence-based practice asks nurses to base care on the best available research. Not every study answers a question equally well, so a hierarchy helps you weigh conflicting findings.
The ranking reflects one thing above all: how well the design controls for bias and chance. Randomization, control groups and pooling many studies all reduce the risk that a result is a fluke or the product of hidden differences between groups.
A hierarchy does not say lower-level studies are worthless. It says that when strong and weak designs disagree about an intervention's effect, the stronger design should usually carry more weight.
The Seven-Level Hierarchy, Highest to Lowest
| Level | Type of evidence | Why it sits here |
|---|---|---|
| I | Systematic review or meta-analysis of randomized controlled trials; guidelines based on such reviews | Pools several rigorous trials, reducing the influence of any one study |
| II | At least one well-designed randomized controlled trial (RCT) | Randomization balances known and unknown differences between groups |
| III | Controlled trial without randomization (quasi-experimental) | Has a comparison group, but groups may differ in important ways |
| IV | Case-control or cohort studies | Observes real-world associations but cannot rule out confounding |
| V | Systematic review of descriptive and qualitative studies | Synthesizes several studies, but of non-experimental designs |
| VI | A single descriptive or qualitative study | Rich insight into experience or patterns, not cause and effect |
| VII | Expert opinion and reports of expert committees | Based on experience rather than systematic data |
Check which version your program uses. Some textbooks and schools use five levels or different labels, and your marker will expect their scheme.
The Johns Hopkins Evidence Levels
Many nursing programs and hospitals use the Johns Hopkins Evidence-Based Practice model, which has five levels and pairs each one with a quality rating of A (high), B (good) or C (low or major flaws).
| Johns Hopkins level | Evidence included |
|---|---|
| Level I | Experimental studies, including RCTs, and systematic reviews of RCTs |
| Level II | Quasi-experimental studies, and systematic reviews of these with or without RCTs |
| Level III | Non-experimental studies, including qualitative studies, and their systematic reviews |
| Level IV | Opinions of respected authorities or nationally recognized expert committees, including guidelines |
| Level V | Literature reviews, quality improvement and case reports, and expert opinion based on experience |
The pairing matters. A Level I study rated C may deserve less trust than a well-conducted Level III study rated A. Always report both the level and the quality.
Matching Evidence Level to Your Question
The hierarchy above is built for intervention questions: does X work better than Y? Other kinds of question are best answered by other designs.
| Question type | Example | Strongest usual design |
|---|---|---|
| Intervention or therapy | Does chlorhexidine bathing reduce bloodstream infections? | Systematic review of RCTs, then RCTs |
| Prognosis or risk | Do smokers have slower wound healing after surgery? | Cohort studies |
| Etiology or harm | Is night-shift work linked to medication errors? | Cohort or case-control studies |
| Meaning or experience | How do new mothers experience breastfeeding support? | Qualitative studies and their syntheses |
| Diagnosis | How accurate is a screening tool for delirium? | Diagnostic accuracy studies |
So a qualitative study can be the best evidence for a meaning question, even though it sits low on the intervention hierarchy. Explain this in your paper when you rely on it.
How to Place a Study on the Hierarchy
Students often mislabel studies by trusting the title. Read the methods section and ask these questions in order.
- Is it a review of other studies? If it has a systematic search and appraisal, it is a systematic review; check what designs it includes.
- Did researchers assign the intervention? If yes, it is experimental or quasi-experimental.
- Was assignment random? Random allocation means an RCT; otherwise quasi-experimental.
- Did researchers only observe? Then it is non-experimental: cohort, case-control, cross-sectional or descriptive.
- Are the data words rather than numbers? Interviews and focus groups indicate a qualitative design.
Phrases That Hint at Each Level
- "Participants were randomly assigned" suggests an RCT.
- "Pre-test and post-test with a comparison unit" suggests quasi-experimental.
- "Followed over time" or "retrospective chart review" suggests a cohort study.
- "Semi-structured interviews" and "thematic analysis" suggest qualitative research.
- "We searched CINAHL, PubMed and Cochrane" suggests a systematic review.
Building an Evidence Table
An evidence table records your appraisal in one place and becomes the backbone of your literature review. Use the same columns for every source.
Example evidence table row.
Citation: Author et al. (2023). Design: Cluster RCT. Sample and setting: 12 medical wards, 1,100 older inpatients. Intervention: Scripted hourly rounding versus usual care. Outcome: Falls per 1,000 patient days. Findings: Falls lower in the intervention group. Level and quality: Level II, quality A. Limitations: Nurses could not be blinded; single health system.
The row above is illustrative. In your own table, record what each study actually reports, and keep limitations specific rather than generic.
Using Levels in Your Paper
Levels of evidence should shape how confidently you write. Strong, consistent Level I and II evidence supports a firm recommendation. A handful of Level VI studies supports only a cautious one.
- Lead with the highest-level, best-quality evidence on each theme.
- Use lower-level studies to add context, patient perspective or local relevance.
- Say plainly when the evidence base is thin or conflicting.
- Match your verbs to the strength: "demonstrates" for strong evidence, "suggests" for weaker.
Common Levelling Mistakes and How to Fix Them
| Mistake | Fix |
|---|---|
| Calling any study with a comparison group an RCT | Check whether allocation was genuinely random |
| Reporting level without quality | Add a quality rating from your appraisal tool |
| Treating a narrative review as a systematic review | Look for a documented search strategy and inclusion criteria |
| Basing a practice change on one small study | Look for consistent findings across several studies |
| Mixing hierarchies within one paper | Choose one scheme, name it and apply it throughout |
| Labeling a pilot study as Level II | Check sample size and design; many pilots are quasi-experimental |
Before you finalize your evidence table, reread the methods section of every source and confirm its label. Markers check this closely, and one mislabelled study can weaken an otherwise strong review.
How EssayHorse Helps with Levels of Evidence
Send your PICOT question, your program's hierarchy or appraisal tool, the rubric and any required sources. A writer with a nursing background writes a custom paper with each study leveled, appraised and synthesized by theme.
Use it to check your own labeling and see how levels shape a recommendation. Papers are written from scratch and plagiarism-checked, and revisions within your brief are free.
Need an Evidence-Based Nursing Paper?
Send your PICOT question, rubric and appraisal tool. A writer with a nursing background levels and synthesizes the evidence for you to study.
Order Your Nursing PaperFree revisions · Full refund if late · Written from scratch for your order
Frequently Asked Questions
A systematic review or meta-analysis of randomized controlled trials, along with guidelines built on such reviews, is usually ranked highest for intervention questions.
A single qualitative study is Level VI in the seven-level hierarchy and Level III in Johns Hopkins. For questions about experience and meaning, they can still be the best available evidence.
It depends on what it is based on. Guidelines drawn from systematic reviews of RCTs rank highly; consensus-based guidelines rank lower. Johns Hopkins places guidelines at Level IV.
Level reflects the study design. Quality reflects how well that design was carried out. A poorly run RCT can be weaker than a well-run cohort study.
No. A systematic review uses a documented search, inclusion criteria and appraisal. A narrative literature review does not, so it sits much lower.
Your rubric decides. Many programs want most sources at Levels I to IV, with lower levels used for context.
Use the one your program teaches or names in the rubric. If none is specified, state which hierarchy you used and cite it.
They are non-experimental and descriptive, so they usually sit at Level VI in the seven-level hierarchy, or Level III in Johns Hopkins. They describe patterns at one point in time but cannot show cause.
Yes, but at a low level. Johns Hopkins places them at Level V because they are designed to improve local practice rather than produce generalizable findings.
Yes. Where little research exists, expert committees and professional standards guide practice. Note the limitation when you rely on them.