Clinical Nurse Specialist Evidence-Based Practice & Research — Questions and Answers
Question 1: A CNS wants to implement a new evidence-based practice change on a unit. Using the Iowa Model of Evidence-Based Practice, what is the first step when the CNS identifies a problem-focused trigger?
- Immediately gather evidence from literature and implement the strongest recommendation
- Determine if the topic is a priority for the organization before investing resources in an EBP project (Correct answer)
- Form an interprofessional team to appraise the evidence
- Pilot the change on one unit before spreading organization-wide
Correct answer: Determine if the topic is a priority for the organization before investing resources in an EBP project
The Iowa Model begins with determining organizational priority — not all problems warrant full EBP projects; alignment with organizational priorities ensures resource support and sustainability.
The Iowa Model of Evidence-Based Practice (Titler et al., University of Iowa Hospitals and Clinics): Step 1: Identify triggers — knowledge-focused (new research, new guidelines) or problem-focused (risk management data, adverse events); Step 2: Is the topic a priority? (decision point) — consider: patient volume affected, financial impact, nursing practice scope, strategic alignment; if No, continue to monitor for sufficient evidence; if Yes, proceed to Step 3: Form an EBP team; Step 4: Assemble, appraise, and synthesize the evidence; Step 5: Is the evidence sufficient? (decision point) — if No, conduct research; if Yes, Step 6: Design practice change and pilot; Step 7: Evaluate and disseminate. The Iowa Model's explicit priority-checking is its distinguishing feature — it prevents investing organizational resources in low-priority projects. Compare to PARIHS (Promoting Action on Research Implementation in Health Services) which focuses on evidence + context + facilitation as implementation determinants.
Question 2: A CNS is critically appraising a randomized controlled trial studying a new wound care dressing. The trial reports a relative risk reduction (RRR) of 50% but the absolute risk reduction (ARR) is 2%. How should the CNS interpret these findings for clinical decision-making?
- RRR of 50% confirms the dressing is highly clinically significant and should be adopted immediately
- ARR of 2% means 50 patients need treatment to prevent one additional wound healing failure — NNT should guide clinical and economic decisions (Correct answer)
- Both RRR and ARR support implementation since both are statistically significant
- A 2% ARR is always too small to be clinically meaningful and the dressing should not be adopted
Correct answer: ARR of 2% means 50 patients need treatment to prevent one additional wound healing failure — NNT should guide clinical and economic decisions
ARR and NNT provide clinically meaningful context that RRR alone obscures — an NNT of 50 means treating 50 patients to prevent one outcome, which must be weighed against cost and burden.
Critical appraisal of treatment effects requires understanding multiple metrics: (1) Relative Risk Reduction (RRR) = (CER - EER)/CER: here, RRR = 50% (impressive sounding); however, if baseline event rate (CER) = 4% and treatment event rate (EER) = 2%, then ARR = 4-2% = 2%; (2) Absolute Risk Reduction (ARR) = CER - EER = 2%: the absolute benefit per patient is 2 in 100; (3) Number Needed to Treat (NNT) = 1/ARR = 1/0.02 = 50: must treat 50 patients to prevent 1 additional wound failure; (4) Clinical significance vs. statistical significance: p-value only tests whether the difference could be due to chance; clinical significance requires evaluation of NNT, treatment burden, cost, and patient values; NNT of 50 may be acceptable for a life-threatening outcome (stroke, death) but questionable for wound healing if the intervention is costly or burdensome; (5) The CNS uses ARR and NNT to communicate realistic expected benefits to stakeholders and inform resource allocation decisions. Marketers often report RRR to make modest effects appear dramatic.
Question 3: A CNS is leading a journal club and asks staff to evaluate the internal validity of a cohort study examining the association between nurse staffing ratios and patient falls. Which threat to internal validity is most relevant to this study design?
- Selection bias — patients with higher fall risk may be preferentially cared for on lower-staffed units (Correct answer)
- Publication bias — negative studies are less likely to be published
- Hawthorne effect — staff know they are being observed and change behavior
- Reporting bias — falls may be under-documented on some units
Correct answer: Selection bias — patients with higher fall risk may be preferentially cared for on lower-staffed units
Confounding by indication (selection bias) is the primary threat — sicker, higher-fall-risk patients may cluster on units with different staffing due to acuity-based assignments, making staffing appear causative.
Internal validity in cohort studies — major threats: (1) Confounding: extraneous variable associated with both exposure (staffing ratio) and outcome (falls) that explains the association without causality. Confounding by indication: higher-acuity patients (higher fall risk) may be placed on units with lower staffing ratios (or lower staffing is a response to higher acuity units), creating a spurious association between low staffing and falls; (2) Selection bias: systematic differences in how participants are selected into exposure groups; (3) Information bias: differential misclassification — fall reporting rates may differ by unit culture; (4) Temporality: exposure must precede outcome — cohort studies generally preserve this; (5) Dose-response relationship, biological plausibility support causality. Mitigation: multivariate regression adjusting for patient acuity (APACHE/SOFA), case mix, unit type; propensity score matching. Publication bias affects meta-analyses, not individual studies. The CNS teaches staff to distinguish association from causation using Hill's criteria.
Question 4: A CNS wants to understand nurses' lived experience of implementing a rapid response system. Which research methodology is most appropriate?
- Randomized controlled trial to measure RRS impact on patient outcomes
- Phenomenological qualitative study to explore nurses' lived experiences of RRS implementation (Correct answer)
- Retrospective cohort study comparing outcomes before and after RRS implementation
- Systematic review of RRS implementation literature across multiple institutions
Correct answer: Phenomenological qualitative study to explore nurses' lived experiences of RRS implementation
Phenomenology examines lived experience and subjective meaning — the appropriate qualitative methodology for understanding nurses' experiences of a new practice change.
Research methodology selection requires matching the method to the question: (1) Quantitative questions: 'Does X cause/affect Y?' or 'How many/much?' use RCT (causation), cohort/case-control (association), cross-sectional (prevalence); (2) Qualitative questions: 'What is the meaning/experience/process of X?' use Phenomenology (lived experience, Husserl/Heidegger), Grounded Theory (theory development from data, Glaser and Strauss), Ethnography (culture/social processes), Case Study (bounded system); (3) This question ('nurses' lived experience') calls for Phenomenological approach: purposive sampling, unstructured/semi-structured interviews, thematic analysis, bracketing researcher assumptions, seeking essence of experience; (4) Qualitative rigor: trustworthiness criteria (Lincoln and Guba): credibility (member checking, prolonged engagement), transferability (thick description), dependability (audit trail), confirmability (peer debriefing). Understanding nurses' experience identifies implementation barriers and facilitators that quantitative outcome data cannot — the CNS uses this for future implementation planning.
Question 5: When appraising a systematic review and meta-analysis, which statistical measure quantifies the heterogeneity across included studies?
- Odds ratio (OR) — measures strength of association across studies
- I-squared statistic — quantifies the proportion of variation across studies due to heterogeneity rather than chance (Correct answer)
- Number needed to treat (NNT) — synthesizes treatment effects across studies
- Confidence interval width — indicates precision of the pooled estimate
Correct answer: I-squared statistic — quantifies the proportion of variation across studies due to heterogeneity rather than chance
The I-squared statistic quantifies heterogeneity in meta-analysis: values of 25%, 50%, and 75% represent low, moderate, and high heterogeneity respectively.
Meta-analysis statistical concepts: (1) Heterogeneity: variation in study results beyond what is expected by chance; sources: clinical (different populations, interventions, outcomes — PICO variation), methodological (study design, risk of bias), statistical (random variation); (2) I-squared statistic (Higgins et al., 2003): measures the percentage of variability due to heterogeneity vs. sampling error; interpretation: 0-25% = low heterogeneity (pooling is appropriate), 26-50% = moderate, 51-75% = high, >75% = very high (pooling is questionable); (3) Q test (Cochran's Q): tests whether observed heterogeneity is more than expected by chance; p <0.10 = significant heterogeneity (liberal threshold); (4) Forest plot: visual display of individual study effects and pooled estimate; diamond width = confidence interval of pooled effect; (5) If high heterogeneity: consider random-effects model (vs. fixed-effects), subgroup analyses, or meta-regression to explain heterogeneity; question whether pooling is appropriate. The CNS uses I-squared to assess whether meta-analytic findings are robust or driven by heterogeneous studies.
Question 6: A CNS is implementing EBP using the PARIHS framework (Promoting Action on Research Implementation in Health Services). The evidence is strong, but implementation is failing. According to PARIHS, which other factors should the CNS assess?
- Patient acuity and unit census as barriers to implementation
- Context (organizational culture, leadership, evaluation processes) and facilitation (support for implementation process) as determinants of success (Correct answer)
- The strength of the research design and statistical power of the evidence
- Staff nurse educational level and years of experience
Correct answer: Context (organizational culture, leadership, evaluation processes) and facilitation (support for implementation process) as determinants of success
PARIHS: successful EBP implementation = f(evidence, context, facilitation). Strong evidence with poor context (culture, leadership) or inadequate facilitation explains implementation failure.
PARIHS (Kitson et al., 1998, updated i-PARIHS 2015): Successful Implementation (SI) = f(Evidence, Context, Facilitation): (1) Evidence: not just research — includes clinical experience and patient preferences; strength on continuum from weak to strong; (2) Context: the environment into which evidence is placed; sub-elements: organizational culture (learning vs. blame culture), leadership (supportive vs. transformational vs. absent), evaluation processes (data feedback, monitoring); weak context = bureaucratic culture, autocratic leadership, lack of feedback; strong context = learning culture, transformational leadership, continuous evaluation; (3) Facilitation: the support mechanism enabling implementation; styles: task-focused (enabling, holistic, 'this is how') vs. enabling (empowering teams, building capacity); facilitation roles: external facilitator (expert), internal facilitator (unit-based), combination most effective; (4) i-PARIHS added innovation characteristics (adaptability of the intervention) and the recipient (individual/team/organizational innovation readiness). If evidence is strong but implementation fails, the CNS diagnoses whether the context is resistant (culture, leadership, no feedback loop) or whether facilitation is insufficient (no champions, no support structure).
Clinical Nurse Specialist Evidence-Based Practice & Research
A CNS wants to implement a new evidence-based practice change on a unit.
Using the Iowa Model of Evidence-Based Practice, what is the first step when the CNS identifies a problem-focused trigger?