Endpoint Selection in Clinical Trials

## Why Endpoint Selection Matters

The primary endpoint is the single most important measurement in a clinical trial. It defines what the drug must achieve to be considered effective, determines the sample size, and is the basis for the regulatory approval decision. A poorly chosen endpoint can doom an effective drug to a failed trial or allow an ineffective drug to appear successful.

## Types of Endpoints

### Clinical Endpoints

Clinical endpoints directly measure how a patient feels, functions, or survives. Overall survival (time to death) is the most definitive clinical endpoint because it is objective, clinically meaningful, and unambiguous. Other examples include myocardial infarction, stroke, hospitalization, and sustained disease remission.

### Surrogate Endpoints

Surrogate endpoints are laboratory measurements or physical signs that substitute for clinical endpoints. They are used when clinical endpoints would require impractically large trials or excessively long follow-up. Examples include HbA1c for diabetes (surrogate for microvascular complications), blood pressure for hypertension (surrogate for stroke/MI), and viral load for HIV (surrogate for AIDS progression).

The FDA accepts surrogate endpoints through three tiers: validated surrogates (strong evidence of clinical correlation), reasonably likely surrogates (supporting evidence but not definitive, eligible for accelerated approval), and candidate surrogates (insufficient evidence, not accepted for approval).

### Patient-Reported Outcomes (PROs)

PROs capture the patient's own assessment of their condition without clinician interpretation. Pain scores (NRS, VAS), quality of life instruments (SF-36, EQ-5D), and symptom diaries are common PROs. The FDA requires that PRO instruments be validated for the specific disease population.

## Composite Endpoints

Composite endpoints combine multiple events into a single outcome (e.g., MACE = cardiovascular death + non-fatal MI + non-fatal stroke). They increase event rates and reduce required sample sizes but can be misleading if the drug affects only the least clinically important component.

## Multiplicity and Hierarchical Testing

When trials have multiple endpoints, the probability of a false-positive finding increases. Hierarchical (fixed-sequence) testing controls the family-wise error rate by testing secondary endpoints only if the primary endpoint achieves statistical significance.

## Key Takeaways

- The primary endpoint defines trial success and determines sample size
- Clinical endpoints measure direct patient benefit; surrogates are indirect measures
- Surrogate endpoints enable smaller, faster trials but carry translational risk
- Composite endpoints increase statistical power but may obscure treatment effects
- Multiplicity adjustments prevent false-positive conclusions from multiple testing