19.3 Evidence-Based Practice, Research Terminology, Quality Improvement & Patient Safety

The Evidence-Based Practice Process

Evidence-based practice (EBP) combines the best research evidence, clinical expertise, and patient values and preferences.

1. Ask a focused question in PICOT format: Population, Intervention, Comparison, Outcome, Time. Example: In postpartum women with preeclampsia (P), does home blood pressure monitoring (I), compared with office visits (C), increase completed blood pressure checks (O) within 10 days of discharge (T)?
2. Acquire the evidence (databases, guidelines).
3. Appraise it for validity, results, and applicability.
4. Apply it, integrating patient preferences.
5. Assess outcomes and disseminate the findings.

Hierarchy of Evidence (Strongest to Weakest for Intervention Questions)

1. Systematic reviews and meta-analyses of randomized controlled trials (RCTs)
2. Well-designed RCTs
3. Controlled trials without randomization
4. Cohort and case-control studies
5. Systematic reviews of descriptive or qualitative studies
6. Single descriptive or qualitative studies
7. Expert opinion and committee reports

USPSTF grades: A (high certainty of substantial benefit), B (moderate to substantial benefit), C (small benefit, so individualize), D (no benefit or harm outweighs benefit, so recommend against), and I (insufficient evidence).

---

Study Designs

---

Research Terminology on the Blueprint

Reliability (Consistency)

• Test-retest: Similar results when repeated over time.
• Interrater: Agreement between observers (Cohen's kappa).
• Internal consistency: Items on a scale measure the same concept (Cronbach's alpha of 0.70 or higher is acceptable).

Validity (Accuracy)

• Measurement validity: Content validity (covers the domain), construct validity (measures the theoretical concept), and criterion validity (agrees with a gold standard, either concurrent or predictive).
• Internal validity: Confidence that the intervention, not bias or confounding, caused the outcome. Threats include selection bias, confounding, attrition, history, maturation, and testing effects.
• External validity: Generalizability to other populations and settings.

A measure can be reliable but not valid (consistently wrong), but it cannot be valid without being reliable.

Significance

• Statistical significance: A p-value below 0.05 means that, if there were truly no effect, results at least this extreme would occur less than 5% of the time. A 95% confidence interval that excludes 1 (for a ratio) or 0 (for a difference) is significant.
• Clinical significance: Whether the effect is large enough to matter to patients. A large study can make a trivial difference statistically significant.
• Type I error (alpha): Concluding there is an effect when there is none (false positive).
• Type II error (beta): Missing a real effect (false negative). Power = 1 − beta, typically set at 0.80, and rises with sample size and effect size.

Variables and Measurement Levels

• Independent variable (the intervention or exposure), dependent variable (the outcome), and confounders (related to both).
• Levels: nominal (categories, such as blood type), ordinal (ranked, such as pain mild/moderate/severe), interval (equal intervals, no true zero, such as temperature in °F), and ratio (true zero, such as weight).

---

Screening and Treatment Effect Math

Worked example: If preeclampsia occurs in 10% of controls and 6% of treated patients, ARR = 4%, relative risk reduction = 40%, and NNT = 1 ÷ 0.04 = 25. This is also why cfDNA has a lower positive predictive value in a young patient: the test's sensitivity and specificity are unchanged, but the condition is less prevalent.

---

Research Ethics and Utilization

• Belmont Report principles: Respect for persons (informed consent), beneficence, and justice.
• Institutional review boards protect participants; federal rules give additional protections to pregnant people, fetuses, prisoners, and children.
• Research utilization and implementation: Models such as the Iowa Model and Stetler Model guide translating evidence into practice, with pilot testing, staff education, and outcome evaluation.

---

Quality Improvement

• Donabedian framework: Structure (resources, staffing), process (what is done, such as the percentage of postpartum hypertensive patients with a blood pressure check within 10 days), and outcome (results, such as postpartum stroke rates), plus balancing measures that capture unintended effects.
• Model for Improvement: What are we trying to accomplish? How will we know a change is an improvement? What change can we make? Test changes with Plan-Do-Study-Act (PDSA) cycles, starting small and tracking data on run charts.
• QI vs. research: QI applies known evidence locally to improve care; research generates generalizable knowledge and requires IRB oversight. Projects that plan to publish generalizable findings may need IRB review.

---

Patient Safety

• Swiss cheese model: Harm occurs when holes in multiple defenses line up, so fix systems rather than blaming individuals.
• Just Culture: Console human error, coach at-risk behavior, and hold reckless behavior accountable.
• Root cause analysis: Required by The Joint Commission after sentinel events (for example, maternal death or severe maternal morbidity, wrong-site surgery, or a retained foreign object); it seeks system causes and an action plan.
• National Patient Safety Goals: Use two patient identifiers, reconcile medications, report critical results promptly, practice hand hygiene, and handle high-alert medications (insulin, magnesium sulfate, oxytocin, anticoagulants, opioids) with double checks.
• Disclosure: Communicate openly with patients after an adverse event, with an apology and an explanation of corrective steps (communication-and-resolution programs such as CANDOR).

Communication and Interprofessional Practice

• SBAR: Situation, Background, Assessment, Recommendation for concise handoffs and escalation.
• TeamSTEPPS tools: CUS ("I am Concerned, I am Uncomfortable, this is a Safety issue"), the two-challenge rule, call-outs, check-back (closed-loop communication), and structured handoffs (I-PASS).
• Chain of command: Escalate unresolved safety concerns up the clinical hierarchy.
• Simulation drills for obstetric hemorrhage, shoulder dystocia, and eclampsia improve team performance.
• Interprofessional Education Collaborative competencies: Values and ethics, roles and responsibilities, interprofessional communication, and teams and teamwork.