13.2 Evidence-Based Medicine & Study Designs

Evidence-Based Medicine & Study Designs

Evidence-Based Medicine (EBM) is the conscientious, explicit, and judicious use of current best evidence in making decisions about the care of individual patients. It integrates clinical expertise, patient values, and the best available scientific research. To practice EBM effectively, a clinician must evaluate the validity of clinical studies and understand their designs.

Primary Study Designs in Clinical Research

Clinical studies are broadly divided into experimental (where the investigator assigns the exposure or intervention) and observational (where the investigator observes natural exposures and outcomes).

1. Randomized Controlled Trials (RCTs) - Experimental

The RCT is the gold standard for establishing causal relationships between interventions and outcomes. Its strength lies in its design features:

• Randomization: Allocating participants to treatment or control arms by chance. This distributes both known and unknown confounding variables equally between the groups, eliminating selection bias.
• Blinding (Masking): Prevents bias in treatment delivery or outcome assessment.
  • Single-blind: Patient is unaware of their allocation.
  • Double-blind: Patient and treating clinician/assessor are unaware.
  • Triple-blind: Patient, clinician, and data analyst are unaware.
• Intention-to-Treat (ITT) Analysis: A fundamental principle where all participants are analyzed in the group to which they were originally randomized, regardless of whether they adhered to the protocol, received the treatment, or dropped out. ITT maintains the baseline comparability achieved by randomization and prevents attrition bias. It contrasts with per-protocol analysis, which only includes patients who fully completed the treatment, often leading to an overestimation of drug efficacy by ignoring those who stopped due to side effects or lack of efficacy.

2. Cohort Studies - Observational

Cohort studies follow a group of individuals defined by their exposure status (exposed vs. unexposed) forward in time to observe the development of the outcome. They can be prospective (following participants into the future) or retrospective (using historical medical records to reconstruct exposure and follow-up to the present).

• Strengths: Excellent for establishing temporal sequences (exposure precedes outcome); allows calculation of incidence and Relative Risk (RR); can study multiple outcomes of a single exposure.
• Weaknesses: Expensive and time-consuming; susceptible to attrition bias (loss to follow-up over time) and confounding.

3. Case-Control Studies - Observational

Case-control studies are retrospective, starting with the outcome. Researchers identify individuals with the disease (cases) and matching individuals without the disease (controls). They then look back in time (via interviews or records) to compare the frequency of prior exposures.

• Strengths: Highly efficient for studying rare diseases or diseases with long latency periods; relatively quick and inexpensive.
• Weaknesses: Extremely vulnerable to recall bias (diseased patients may remember past exposures more vividly than healthy controls) and selection bias (difficulty in selecting a representative control group). Measures the Odds Ratio (OR), not Relative Risk.

4. Cross-Sectional Studies - Observational

Often called a "prevalence study" or "snapshot," this design assesses exposure and outcome simultaneously in a population at a single point in time.

• Strengths: Quick, inexpensive, and useful for determining prevalence and generating hypotheses.
• Weaknesses: Cannot establish a temporal relationship (does the exposure precede the disease?), a limitation known as temporal ambiguity.

---

The Hierarchy of Evidence

In EBM, clinical evidence is ranked in a hierarchy based on the study design's susceptibility to bias. Decisions should be guided by the highest level of evidence available.

---

Core Calculations: NNT and NNH

To apply study results to clinical practice, physicians must translate relative risk reductions into absolute measures that reflect the real-world impact on patients.

Number Needed to Treat (NNT)

NNT is the number of patients that must be treated with a specific intervention to prevent one additional adverse outcome. It is the reciprocal of the Absolute Risk Reduction (ARR).

1. Calculate risk in the control group ($R_{\text{control}}$) and treatment group ($R_{\text{treatment}}$).
2. Calculate ARR: $\text{ARR} = R_{\text{control}} - R_{\text{treatment}}$
3. Calculate NNT: $\text{NNT} = \frac{1}{\text{ARR}}$

Clinical Rule: NNT must always be rounded UP to the nearest whole integer (e.g., $12.1 \rightarrow 13$), because you cannot treat a fraction of a patient, and rounding down would overestimate the benefit of the treatment.

Example: In a clinical trial for a new drug to prevent cardiovascular events, the incidence of myocardial infarction is 8% in the control group and 5% in the treatment group.

• $\text{ARR} = 0.08 - 0.05 = 0.03$ (or 3%)
• $\text{NNT} = 1 / 0.03 = 33.33 \rightarrow \mathbf{34}$. You need to treat 34 patients to prevent one myocardial infarction.

Number Needed to Harm (NNH)

NNH is the number of patients treated with an intervention before one patient experiences an adverse drug event or side effect. It is the reciprocal of the Absolute Risk Increase (ARI) or Attributable Risk.

1. Calculate ARI: $\text{ARI} = R_{\text{treatment}} - R_{\text{control}} \quad \text{(for the adverse event)}$
2. Calculate NNH: $\text{NNH} = \frac{1}{\text{ARI}}$

Clinical Rule: For patient safety, NNH is typically rounded DOWN to the nearest whole integer (e.g., $22.7 \rightarrow 22$) to avoid underestimating the potential for harm.

Example: In the same trial, 4% of patients in the treatment group experienced severe gastritis compared to 1% in the control group.

• $\text{ARI} = 0.04 - 0.01 = 0.03$
• $\text{NNH} = 1 / 0.03 = 33.33 \rightarrow \mathbf{33}$. For every 33 patients treated, one will develop severe gastritis.