14.15: Bias in Epidemiological Studies

Biases can arise at various stages of research, from study design and data collection to analysis and interpretation. Recognizing and addressing these biases is essential to ensure the validity and reliability of epidemiological findings.Broadly speaking, biases in epidemiology fall into three main categories: selection bias, information bias, and confounding. A more detailed description of possible biases is:

1. Selection Bias: This occurs when the study population is not representative of the target population. For example, a survey only among urban dwellers when aiming to understand a national health issue.
2. Classification Bias: This happens when there is an error in assigning participants to different categories or groups, affecting the study's accuracy.
3. Confounding Bias: This arises when an extraneous variable correlates with both the dependent and independent variables, distorting the true relationship. For instance, if a study on smoking and lung cancer does not adjust for age, age could confound the results because older people might have both higher smoking rates and higher cancer rates.
4. Prevalence or Incidence Bias (Neyman Bias): Occurs if the initial study population excludes participants who already have the outcome due to early death or disease progression.
5. Admission Rate Bias (Berkson Bias): This type of bias is introduced when the study subjects are selected from hospitalized patients who may not to represent the general population.
6. Non-response Bias: Evident when individuals who do not participate in the study differ significantly from those who do. A common scenario could be a higher dropout rate among a particular demographic group.
7. Information or Wrong Classification Bias: This includes non-differential and differential biases stemming from how data on exposure or outcomes is collected.
8. Memory Bias: It is a specific type of information bias in which participants may not remember past events accurately, leading to misclassification.
9. Interviewer or Observer Bias: This occurs when the person collecting data has preconceived notions that affect the interpretation of responses.
10. Confounding Bias: A pivotal bias where external variables affect the primary variables under study, which can mislead the results unless properly controlled during analysis.

These biases highlight the importance of careful study design and execution in epidemiological research to minimize errors and provide reliable data. Each type of bias poses unique challenges, and their presence can weaken the credibility of study findings. Understanding the sources and mechanisms of bias equips researchers with the tools to design better studies and apply appropriate analytical adjustments. In this way, minimizing bias is not merely a technical task but a step toward ensuring that epidemiological research provides meaningful and actionable insights for public health.

A bias is a systematic tendency of a quantity's estimate or expected value to be skewed or far from the true value.

For example, a thermometer consistently measures body temperature 3 degrees lower, giving a biased temperature measurement.

Epidemiological studies encounter different types of biases at various stages.

Sampling or ascertainment bias occurs when a sample has a non-random subset of the population under study due to a higher or lower probability of selecting some members than others.

Selection bias involves selecting participants favoring one characteristic over another.

Studies based on questionnaires often face attrition bias when participants drop out of a study, especially during the follow-up, skewing the results.

Similarly, response and non-response biases occur when nonrespondents differ from respondents regarding the variable of interest or respond inaccurately.

Spectrum bias occurs when a diagnostic test is assessed using a non-representative patient population, resulting in an inflated perception of the test's accuracy and reliability.

Lastly, observer bias arises when a researcher subconsciously brings a skew in the data collection process or analysis.