The changing "natural history" of symptomatic coronary artery disease: basis versus bias

Area of Science:

• Cardiology
• Epidemiology
• Medical Statistics

Background:

• National statistics and randomized trials show 20-30% reduction in ischemic heart disease mortality.
• Recent "natural history" studies suggest greater survival improvements in symptomatic coronary patients.
• Uncritical comparison of these studies may lead to overestimation of true survival gains.

Purpose of the Study:

• To identify and explain biases in natural history studies of coronary heart disease.
• To caution against overestimating survival improvements by comparing patients across different studies and eras.
• To advocate for careful scrutiny of natural history studies for selection biases.

Main Methods:

• Analysis of potential biases in the interpretation of natural history studies.
• Identification of lead-time bias, referral bias, and incidence-prevalence bias.
• Conceptual framework for comparing patient cohorts from different time periods.

Main Results:

• Lead-time bias: Earlier diagnosis increases survival duration irrespective of treatment efficacy.
• Referral bias: Differences in patient symptom severity and anatomy between study eras.
• Incidence-prevalence bias: Prevalence studies overrepresent slower-progressing disease survivors.

Conclusions:

• Natural history studies, while accurate intrinsically, are prone to biases when compared across different patient groups or time periods.
• Lead-time, referral, and incidence-prevalence biases can significantly inflate perceived survival improvements.
• Researchers must critically evaluate natural history studies for these biases before comparing findings to avoid overestimating therapeutic or diagnostic advancements.