Berkson

Berkson refers to a phenomenon observed in statistical studies, particularly in case-control studies, where the observed association between two conditions can be distorted by the fact that the study population is selected based on one of the conditions. Specifically, if a condition is a prerequisite for inclusion in the study, then the observed association between that condition and another independent factor will be weakened or even reversed. This occurs because the selection bias inherent in case-control studies can lead to an unrepresentative sample of the underlying population.

The Berkson paradox, also known as Berkson's bias or Berkson's fallacy, was first described by the American physician and epidemiologist Joseph Berkson in 1946. He noted that when studying the relationship between two diseases, if one disease is a requirement for hospitalization, and therefore for inclusion in the study, then the other disease will appear to be negatively associated with the first, even if there is no true relationship in the general population. This is because individuals with both diseases are more likely to be hospitalized and thus overrepresented in the study compared to individuals with only one disease.

Understanding and accounting for Berkson's paradox is crucial for accurate interpretation of epidemiological data. Researchers often employ various statistical techniques to mitigate the effects of this bias, such as using different study designs or adjusting for the selection criteria when analyzing their results. It highlights the importance of carefully considering the sampling methods and potential biases when drawing conclusions from observational studies.