16.3 Evidence-Based Practice, Clinical Inquiry & Quality Improvement

16.3 Evidence-Based Practice, Clinical Inquiry & Quality Improvement

Clinical Scenario: In an 18-bed trauma-surgical intensive care unit (TSICU), the quarterly dashboard reveals that the central line-associated bloodstream infection (CLABSI) rate has risen to 3.6 per 1,000 line-days (local improvement target < 0.8). The lead AGACNP spearheads a clinical inquiry initiative evaluating whether transitioning from standard dressings to a bundle combining 2% chlorhexidine gluconate-impregnated gel dressings with alcohol disinfection caps reduces CLABSI rates. Navigating this initiative requires formulating an answerable clinical question, appraising literature, distinguishing quality improvement from research, executing rapid-cycle Plan-Do-Study-Act (PDSA) tests, and analyzing run charts to distinguish common cause from special cause variation.

Evidence-Based Practice (EBP) & The 5-Step Process

Defined by David Sackett and colleagues, Evidence-Based Practice (EBP) is the conscientious integration of the best external research evidence with clinical expertise and patient values and preferences within a specific clinical context.

The clinical inquiry workflow follows the 5-Step EBP Process:

1. Ask: Formulate an answerable, structured clinical practice question.
2. Acquire: Systematically search biomedical databases (PubMed, CINAHL, Cochrane Library) for top-tier evidence.
3. Appraise: Critically evaluate evidence for internal validity (methodological rigor), external validity (generalizability), and clinical significance (magnitude of effect).
4. Apply: Integrate evidence with clinical judgment, institutional resources, and patient values to implement practice changes.
5. Assess: Evaluate post-implementation outcomes, monitoring for clinical efficacy, balancing measures, and sustainability.

Formulating Clinical Questions: The PICO / PICOT Framework

Clinicians utilize the PICOT framework to structure searchable inquiry questions:

• P (Population/Problem): Adult ICU patients with central venous catheters.
• I (Intervention): 2% chlorhexidine gel dressings combined with alcohol disinfection caps.
• C (Comparison): Standard transparent semipermeable polyurethane dressings without chlorhexidine.
• O (Outcome): Incidence of laboratory-confirmed CLABSI per 1,000 line-days.
• T (Timeframe): Duration of central catheterization in the ICU.

The Hierarchy of Evidence Pyramid

The strength of clinical evidence is ranked using the established Hierarchy of Evidence (Melnyk & Fineout-Overholt framework):

Critical Appraisal: Statistical vs. Clinical Significance & Effect Metrics

Appraising clinical literature requires evaluating validity and quantitative effect metrics:

• Internal Validity: Degree to which study findings are free from methodological bias, confounding, and measurement error.
• External Validity: Generalizability of study findings to outside patient populations.
• Statistical vs. Clinical Significance: A p-value measures how incompatible the data are with a specified null model; p < 0.05 is a conventional significance threshold. Clinical significance asks whether the effect is large enough to matter to patients.
• 95% Confidence Intervals (CI): Reflects effect precision. For ratio metrics (Relative Risk [RR], Odds Ratio [OR], Hazard Ratio [HR]), if the 95% CI encompasses 1.0, the result is not statistically significant (p ≥ 0.05). For difference metrics (Absolute Risk Reduction, mean difference), if the 95% CI encompasses 0, the result is not statistically significant.

Clinical Effect Calculations

Loading diagram...

• Relative Risk (RR): Ratio of event risk in the experimental group to the control group: RR = EER / CER.
• Odds Ratio (OR): Ratio of odds of an event in the exposed group to the unexposed group.
• Absolute Risk Reduction (ARR): Absolute difference in event rates: ARR = CER - EER.
• Number Needed to Treat (NNT): Number of patients needed to receive experimental therapy to prevent one adverse outcome: NNT = 1 / ARR. Always rounded UP to the nearest whole integer. Lower NNT indicates greater clinical potency.
• Number Needed to Harm (NNH): Number of patients exposed resulting in one adverse event: NNH = 1 / ARI (where ARI = EER - CER). Always rounded DOWN to the nearest whole integer. Higher NNH indicates safer therapy.

Quality Improvement (QI) vs. Clinical Research

Distinguishing QI from human subjects research is an essential professional responsibility:

• Research: A systematic investigation designed to develop or contribute to generalizable knowledge. Methods may be quantitative or qualitative and need not involve blinding or control groups. Applicable human-subjects rules determine IRB review, exemptions and consent; an IRB may approve consent or documentation waivers when regulatory criteria are met.
• Quality Improvement: Work directed at improving local care, processes and outcomes, often using iterative tests. Activities limited to local improvement may be nonresearch rather than exempt research. A project can have both QI and research purposes; obtain the institution’s determination and required oversight before proceeding. Intent to publish alone does not make QI research.

The Model for Improvement & PDSA Rapid Cycles

The IHI Model for Improvement addresses three core questions followed by testing cycles:

1. What are we trying to accomplish? Establish a SMART Aim (e.g., 'Reduce TSICU CLABSI from 3.6 to < 1.0 per 1,000 line-days in 6 months').
2. How will we know that a change is an improvement? Measure Outcome Measures (CLABSI rate), Process Measures (dressing bundle compliance rate), and Balancing Measures (contact dermatitis incidence).
3. What changes can we make that will result in improvement? Select evidence-based interventions.

The Plan-Do-Study-Act (PDSA) Cycle

• Plan: Formulate the change test, establish predictions, and plan data collection.
• Do: Carry out the test on a small scale (e.g., pilot bundle on Beds 1–4 for one week); document barriers.
• Study: Analyze post-implementation data, compare to predictions, and summarize lessons.
• Act: Decide to Adopt hospital-wide, Adapt based on feedback, or Abandon and test an alternative.

Lean Six Sigma & Statistical Process Control (SPC)

Lean Six Sigma eliminates waste and reduces clinical process variation:

• Process Mapping: Visualizing workflow steps to eliminate delays and handoff waste.
• Ishikawa (Fishbone) Diagram: Categorizes root causes of quality defects across People, Methods, Machines/Equipment, Materials, Measurement, and Environment.
• Pareto Chart: Bar chart and cumulative percentage line based on the Pareto 80/20 Principle: ~80% of errors stem from ~20% of root causes.
• Run Charts: Plot results over time against a baseline median, without control limits. Apply a specified rule set consistently: commonly six consecutive points on one side of the median indicate a shift, and five increasing or decreasing points indicate a trend. Some conventions use longer sequences.
• Control Charts: Add a process-specific centerline and statistically calculated control limits, commonly 3-sigma limits. Use rules appropriate to the chart and data; a point beyond a control limit can signal special cause.
• Interpretation: Common-cause variation belongs to the system and generally requires system redesign. A nonrandom signal warrants investigation of possible changes. Review data quality, exposure denominators, concurrent changes and balancing measures before attributing clinical improvement to an intervention.

Nurse-Sensitive Quality Indicators & Benchmarking

The National Database of Nursing Quality Indicators (NDNQI) benchmarks nursing-sensitive outcomes such as CAUTI, CLABSI, ventilator-associated events, hospital-acquired pressure injuries, and patient falls, alongside workforce measures. Compare similar units using the same indicator definition, reporting period, and denominator; a local improvement target is not automatically a national benchmark.