If the study population does not reflect the clinical population, which bias increases?

• A. Selection bias
• B. Information bias
• C. Confounding bias
• D. Observer bias

Answer: (A)

When the study population doesn’t resemble the clinical population, the main issue is that the sample isn’t representative of those who actually have or would receive the treatment. This creates selection bias, a systematic distortion of findings caused by how participants are chosen or retained. Because the enrolled group differs in important ways (such as severity of disease, age, or comorbidities), the observed association between exposure and outcome may not reflect what happens in the real-world clinical setting, reducing generalizability.

For example, if a new therapy is studied mainly in younger, healthier patients, the results might overstate safety and effectiveness for the broader population that includes older individuals and those with multiple health issues.

Information bias would involve errors in measuring or classifying exposure or outcome, not who is included in the study. Confounding bias arises when an external factor is linked to both exposure and outcome and isn’t controlled. Observer bias occurs when investigators’ expectations influence data assessment.