Evidence-Based Practice & Research Literacy
Master the full EBP cycle: formulate clinical questions, locate and appraise evidence, understand study designs, interpret statistics, and apply findings to patient care.
Phase 2 certification expansion
This subject now anchors 10 research and EBP lessons, 150 clinical cases, 1500 application questions, and a 150-question Pre-Nursing Foundations Certification blueprint.
Nursing Research & EBP lesson sequence
Anatomy of a Research Article
Understanding each section's purpose — IMRAD framework
Research articles follow a standardized structure called IMRAD: Introduction, Methods, Results, and Discussion. Understanding what each section contains lets you efficiently extract information without reading every word.
Sections of a Research Article
Study Tip — IMRAD Speed-Reading Strategy
Read the abstract first. If relevant, jump directly to Methods (is the design sound?) and Results (what did they actually find?). Only then read Introduction and Discussion. This order saves time and prevents the introduction and discussion from biasing your interpretation of the data.
Research Designs
Choosing the right design for the right question
Study Designs — From Experimental to Observational
Internal Validity vs External Validity
Internal validity: did the study accurately measure what it claimed? (Are the results trustworthy for this sample?) External validity: can the results be generalized to other patients and settings? High internal validity is necessary but not sufficient — you also need to assess whether your patient population matches the study sample.
Research Designs — Self-Check
1/4Which study design provides the strongest evidence for causation?
Levels of Evidence & EBP Process
Evidence hierarchy and the PICO framework
Evidence Pyramid — Highest to Lowest
The PICO Framework
PICO is a framework for formulating clinical questions that can be answered through research. P = Patient/Population (Who is the patient or group?), I = Intervention (What treatment or action is being considered?), C = Comparison (What is the alternative — another treatment, placebo, or no treatment?), O = Outcome (What is the desired measurable result?). Example: In hospitalized elderly patients (P), does hourly rounding (I) compared to standard care (C) reduce fall rates (O)? A well-built PICO question guides your literature search and helps you find the most relevant evidence.
The Five Steps of EBP
- 1.Ask — Formulate a focused clinical question using PICO
- 2.Acquire — Search databases (PubMed, CINAHL, Cochrane) for the best evidence
- 3.Appraise — Critically evaluate the validity and relevance of the evidence
- 4.Apply — Integrate evidence with clinical expertise and patient values
- 5.Assess — Evaluate the outcome and share findings
Critical Appraisal
Evaluating whether a study is trustworthy and applicable
Critical appraisal asks three fundamental questions: (1) Is the study valid? (Are the methods sound?), (2) Are the results important? (Is the effect size clinically meaningful?), and (3) Are the results applicable? (Does this match your patient population?). No study is perfect — the question is whether the biases are large enough to invalidate the conclusions.
Step 1 — Is it Valid?
- Is the design appropriate for the question?
- Was randomization performed and concealed?
- Were groups similar at baseline?
- Were all participants accounted for at the end?
- Were assessors blinded to group assignment?
Step 2 — Are Results Important?
- How large is the effect? (ARR, NNT, effect size)
- How precise is the estimate? (Confidence interval)
- Is the statistical significance clinically meaningful?
- Are all relevant outcomes reported?
Step 3 — Is it Applicable?
- Is your patient similar to those studied?
- Are the outcomes relevant to your patient?
- Is the intervention feasible in your setting?
- What do your patient's values and preferences indicate?
Red Flags — Stop and Question
- Industry funding without independent replication
- Truncated Y-axis on graphs
- Only relative risk reported (no ARR/NNT)
- Surrogate outcomes only (not patient-important outcomes)
- Very small sample with imprecise CIs
Critical Appraisal — Self-Check
1/3When critically appraising a study, the FIRST question to ask is:
Identifying Bias in Research
Recognizing threats to validity
Selection Bias
Non-representative participants or non-random assignment. A fall prevention study that only includes alert, oriented patients excludes the highest-risk group — results cannot be applied to cognitively impaired patients.
Measurement Bias
Outcomes measured inconsistently or assessors know group assignment. Blinding (masking) prevents unconscious influence on assessment. Double-blinded RCTs (patient and assessor blinded) have the strongest protection.
Publication Bias
Studies with positive results more likely to be published, overestimating treatment effectiveness. Systematic reviews that search trial registries and grey literature help counteract this.
Attrition Bias
Uneven dropout between groups. Intention-to-treat analysis (analyze all participants in original groups regardless of completion) prevents this bias from distorting results.
Recall Bias
In case-control studies, participants with the outcome (cases) may remember exposures differently than controls. People with a disease may search harder for a cause.
Hawthorne Effect
Participants modify behavior because they know they are being observed. This is a form of measurement bias that can make interventions appear more effective during the study period than they truly are in practice.
Understanding Bias
Bias is any systematic error that distorts study results. Selection bias occurs when participants are not representative of the target population or are not randomly assigned. Measurement bias happens when outcomes are assessed inconsistently or when assessors know which group participants belong to. Publication bias arises because studies with positive results are more likely to be published, creating a skewed evidence base. Attrition bias occurs when participants drop out unevenly between groups. Understanding bias helps you evaluate whether a study's conclusions are trustworthy.
Applying Evidence to Practice
Translating research findings into clinical decisions
The three pillars of EBP are: (1) best available research evidence, (2) clinical expertise and judgment, and (3) patient values and preferences. Evidence without clinical judgment leads to cookbook medicine. Clinical judgment without evidence perpetuates outdated practices. Both without patient values violates autonomy.
Grading Recommendations
Applying Evidence — Self-Check
1/2Evidence-based practice integrates three components. Which is NOT one of them?
Descriptive Statistics
Summarizing and describing data distributions
Descriptive statistics summarize and describe the characteristics of a dataset. They tell you WHAT the data looks like without drawing conclusions about a population. Every research paper presents descriptive statistics in the Methods or Results section.
Key Descriptive Statistics
Inferential Statistics
Drawing conclusions about populations from samples
Inferential statistics use sample data to make inferences about larger populations. Every clinical research study uses inferential statistics to determine whether observed differences are real or due to chance.
Null Hypothesis (H₀)
States there is no difference or relationship. Research tries to reject H₀. Example: "This drug has no effect on blood pressure." We set up the null hypothesis to test against with statistics.
Alpha (α) Level & Significance
Alpha = the threshold for rejecting H₀. Conventional α = 0.05 (5% chance of false positive). If p < α, reject H₀ and conclude statistical significance. If p ≥ α, fail to reject H₀ — NOT the same as proving H₀ is true.
Type I & Type II Errors
Type I error (false positive, α): concluding there IS an effect when there is not. Type II error (false negative, β): concluding there is NO effect when there actually is. Power = 1 − β = probability of detecting a true effect.
Sample Size & Power
Larger samples detect smaller true effects (more statistical power). Underpowered studies miss real effects. Power of 80% means the study has an 80% chance of detecting a real effect if one exists. Power analysis before data collection determines required sample size.
P-Values & Confidence Intervals
The two most important statistics you will see in research
P-Value (Probability Value)
The probability of obtaining results at least as extreme as observed, assuming the null hypothesis is true. p < 0.05 = statistically significant. p = 0.049 and p = 0.051 are practically identical — the 0.05 cutoff is a convention, not a biological truth. P-values do NOT measure effect size, clinical importance, or the probability the treatment works.
Confidence Interval (CI)
A 95% CI provides the range within which we are 95% confident the true population value falls. Example: RR = 0.75 (95% CI: 0.60–0.94) — significant, CI does not cross 1.0. RR = 0.75 (95% CI: 0.55–1.10) — not significant, CI crosses 1.0 (for RR). For mean differences, a CI crossing zero = not significant.
Statistical vs Clinical Significance
Statistical significance (p < 0.05) means the result is unlikely due to chance alone, but it does NOT mean the result is clinically important. A study might find a statistically significant blood pressure reduction of 1 mmHg with a new drug — statistically real but clinically meaningless. Clinical significance asks: Is the effect large enough to matter to patients? Always look at effect size, confidence intervals, and clinical context — not just p-values.
Relative Risk & Odds Ratios
Measuring association between exposure and outcome
Relative Risk (RR)
Used in cohort studies and RCTs. RR = risk in exposed group / risk in unexposed group. RR = 1.0: no difference. RR > 1.0: increased risk. RR < 1.0: protective effect. Example: If 20% of smokers develop lung disease vs 4% of non-smokers, RR = 20/4 = 5.0 — smokers are 5× more likely.
Odds Ratio (OR)
Used in case-control studies. OR = odds of exposure in cases / odds of exposure in controls. OR ≈ RR when the outcome is rare (<10%). Like RR: OR = 1.0 means no association; OR > 1.0 means increased odds; OR < 1.0 means protective. OR can overestimate effect for common outcomes.
ARR, RRR & NNT — Clinical Interpretation
ARR (Absolute Risk Reduction) = control event rate − treatment event rate. RRR (Relative Risk Reduction) = ARR / control rate. NNT = 1/ARR. Example: 10% vs 5% event rate → ARR = 5%, RRR = 50%, NNT = 20. Relative numbers sound more impressive but NNT provides clinical context.
Clinical Interpretation of Statistics
Translating numbers into patient care decisions
Sensitivity vs Specificity
Sensitivity (SnNOUT): if the test is highly Sensitive and Negative, it rules OUT disease. Best for screening. Specificity (SpPIN): if the test is highly Specific and Positive, it rules IN disease. Best for confirmation. High sensitivity but low specificity = many false positives (confirm before treating). High specificity but low sensitivity = many false negatives (miss cases during screening).
Match the Statistical Concept
Terms
Definitions
Evidence-Based Practice & Research Literacy — Comprehensive Quiz
1/10A p-value of 0.03 means:
Pre-nursing comprehensive review
1/20Which organelle contains its own DNA and is inherited exclusively from the mother?