11.3 Evidence-Based Medicine & Clinical Practice Guidelines

Evidence-Based Medicine & Clinical Practice Guidelines

Why Evidence-Based Medicine Matters for the PANCE

The NCCPA PANCE blueprint requires candidates to understand the principles of evidence-based medicine (EBM). EBM is the integration of the best research evidence with clinical expertise and patient values. PANCE questions under this category evaluate your ability to identify the hierarchy of scientific evidence, interpret diagnostic test characteristics (sensitivity, specificity, PPV, NPV), calculate epidemiological risk measures (relative risk, odds ratio, absolute risk reduction, number needed to treat), recognize study bias, and interpret clinical significance using p-values and confidence intervals.

---

The Hierarchy of Evidence

Not all clinical studies carry the same weight. Evidence is graded in a hierarchical structure, from the strongest evidence to the weakest:

1. Systematic Reviews and Meta-Analyses: The highest level of evidence. These studies pool and analyze data from multiple independent studies (typically randomized controlled trials) to establish a consensus on a clinical question.
2. Randomized Controlled Trials (RCTs): The gold standard for evaluating therapeutic interventions. Participants are randomly assigned to a treatment or control group to minimize bias and establish cause-and-effect relationships.
3. Cohort Studies: Observational and longitudinal. Researchers follow a group of exposed and unexposed individuals over time to see who develops the outcome. Cohort studies can be prospective or retrospective and are used to calculate Relative Risk (RR).
4. Case-Control Studies: Observational and retrospective. Researchers identify individuals with an outcome (cases) and individuals without the outcome (controls) and look backward in time to compare the frequency of exposure. Useful for rare diseases and used to calculate the Odds Ratio (OR).
5. Cross-Sectional Studies: Observational. Assess exposure and outcome simultaneously in a population at a single point in time. These studies measure disease prevalence.
6. Case Series and Case Reports: Descriptive studies tracking a small group of patients (series) or a single patient (report) with an unusual presentation or treatment outcome. No control group.
7. Animal Research and In Vitro Studies: The lowest tier of evidence, not directly generalizable to human patients.

---

Statistical Measures in Diagnostic Testing

To evaluate the clinical utility of diagnostic tests, clinicians must understand the relationships between true/false positives and true/false negatives.

Sensitivity (SnNout)

Sensitivity is the probability that a test is positive in a patient who has the disease: $\text{Sensitivity} = \frac{TP}{TP + FN}$ Clinical Pearl: A test with high sensitivity has very few false negatives. Therefore, if a highly sensitive test is negative, we can confidently rule OUT the disease (SnNout). Highly sensitive tests are ideal for screening (e.g., ELISA for HIV).

Specificity (SpPin)

Specificity is the probability that a test is negative in a patient who does not have the disease: $\text{Specificity} = \frac{TN}{TN + FP}$ Clinical Pearl: A test with high specificity has very few false positives. Therefore, if a highly specific test is positive, we can confidently rule IN the disease (SpPin). Highly specific tests are ideal for confirming a diagnosis after a positive screening test (e.g., Western blot or PCR for HIV).

Predictive Values and the Impact of Prevalence

Unlike sensitivity and specificity (which are inherent characteristics of the test itself), predictive values depend heavily on the prevalence of the disease in the population being tested:

• Positive Predictive Value (PPV): The probability that a patient with a positive test actually has the disease: $\text{PPV} = \frac{TP}{TP + FP}$ Prevalence Impact: As disease prevalence increases, the PPV of a test increases, and the NPV decreases.
• Negative Predictive Value (NPV): The probability that a patient with a negative test actually does not have the disease: $\text{NPV} = \frac{TN}{TN + FN}$ Prevalence Impact: As disease prevalence decreases, the NPV increases, and the PPV decreases.

---

Statistical Measures of Risk and Treatment Effect

To evaluate the effectiveness of a clinical intervention, you must be able to perform and interpret the following calculations:

Relative Risk (RR) vs. Odds Ratio (OR)

• Relative Risk (RR): The ratio of the risk of developing a disease in an exposed group compared to the risk in an unexposed group. Used in cohort studies: $\text{RR} = \frac{\text{Risk in exposed}}{\text{Risk in unexposed}}$
• Odds Ratio (OR): The ratio of the odds of exposure in the diseased group (cases) compared to the odds of exposure in the non-diseased group (controls). Used in case-control studies: $\text{OR} = \frac{\text{Odds of exposure in cases}}{\text{Odds of exposure in controls}}$
• Interpretation: An RR or OR of 1.0 indicates no difference in risk or odds between groups. An RR/OR >1.0 indicates increased risk; <1.0 indicates decreased risk (protective effect).

Absolute Risk Reduction (ARR) and Number Needed to Treat (NNT)

• Absolute Risk Reduction (ARR): The absolute difference in event rates between the control group (untreated/placebo) and the experimental group (treated): $\text{ARR} = \text{Control Event Rate (CER)} - \text{Experimental Event Rate (EER)}$
• Number Needed to Treat (NNT): The number of patients who must receive the treatment to prevent one additional adverse outcome. It is the reciprocal of the ARR: $\text{NNT} = \frac{1}{\text{ARR}}$ Rule: NNT must always be rounded UP to the next whole number (e.g., if NNT = 16.2, round to 17).
• Number Needed to Harm (NNH): The number of patients exposed to a risk factor or treatment to cause harm to one patient. It is the reciprocal of the Absolute Risk Increase (ARI): $\text{NNH} = \frac{1}{\text{ARI}}$ Rule: NNH must always be rounded DOWN to the next whole number.

---

Interpreting Guidelines, Significance, and Bias

Statistical vs. Clinical Significance

• p-value: The probability that the observed results occurred by chance alone. A p-value of <0.05 is the standard threshold for statistical significance, indicating a <5% probability that the result is due to chance.
• Confidence Intervals (CIs): A range of values that is likely to contain the true population parameter.
  • Ratio Measures (RR, OR): If the 95% CI for a relative risk or odds ratio includes 1.0 (e.g., 95% CI: 0.85 to 1.45), the result is not statistically significant because 1.0 represents no difference.
  • Difference Measures (ARR, mean difference): If the 95% CI for a difference measure includes 0 (e.g., 95% CI: -1.2% to 3.5%), the result is not statistically significant because 0 represents no difference.

Research Biases

• Selection Bias: Error in choosing participants, making groups non-comparable (mitigated by randomization).
• Recall Bias: Common in retrospective case-control studies where diseased individuals are more likely to remember past exposures than healthy controls.
• Confounding: A third variable that is associated with both the exposure and the outcome, distorting the true relationship. Controlled by randomization, matching, or statistical stratification.