Scientific research aims to uncover truths about the world, but various factors can distort findings. One such factor is susceptibility bias, a systematic error that can skew study results, especially in health and medicine. Understanding this bias is important for interpreting scientific conclusions and making informed decisions.
What Susceptibility Bias Is
Susceptibility bias occurs when a pre-existing characteristic influences both the likelihood of receiving an exposure (e.g., treatment) and the study outcome. This means study groups may not be equivalent, leading to misleading associations. It can make an exposure appear to cause an outcome, when a pre-existing factor is the actual driver.
It is sometimes considered a form of selection bias or confounding by indication, where the reason for receiving a treatment is related to the outcome itself. It differs from other selection biases, like sampling bias, which relates to participant selection for population representation. Susceptibility bias involves systematic differences in initial risk or predisposition to the outcome between study groups, even if the sample is representative. For example, sicker patients might be given a more aggressive treatment, which could then appear less effective than it is, simply because those receiving it were already in a worse state.
How Susceptibility Bias Appears
It frequently appears in medical and epidemiological research when patients are not randomly assigned to treatments. Patients with more advanced cancer might be chosen for a new, aggressive treatment, while those with milder disease receive less aggressive options. If the treated group has a higher baseline illness severity, the new treatment might appear less effective or harmful compared to standard care, simply because these patients were already sicker, not due to the treatment’s efficacy.
Another example involves medications prescribed for chronic conditions. Patients with higher cholesterol might be prescribed a specific medication to lower their levels. If these patients then experience heart disease, the medication could be mistakenly blamed for causing the heart disease, when the underlying high cholesterol was the true predisposing factor.
Similarly, in a study on hormone replacement therapy (HRT) and cardiovascular disease, women receiving HRT were initially found to have a lower risk of heart disease; however, this apparent protective effect disappeared when researchers accounted for factors like healthier lifestyles and higher socioeconomic status among HRT users.
Strategies to Mitigate Susceptibility Bias
Researchers employ strategies to reduce susceptibility bias and enhance findings reliability. Randomized Controlled Trials (RCTs) are a primary method. In an RCT, participants are randomly assigned to an exposure group (e.g., new treatment) or a control group (e.g., placebo or standard care). This random assignment distributes known and unknown pre-existing characteristics, including susceptibilities, evenly across groups, minimizing systematic differences.
When RCTs are not feasible, such as in observational studies, statistical adjustment methods account for known susceptibilities. Techniques like regression analysis statistically control for confounding variables—factors related to both exposure and outcome—to isolate the true effect.
Propensity score matching is another statistical method. Researchers estimate the probability of a participant receiving an exposure based on pre-existing characteristics. Participants with similar propensity scores are then matched across exposure and control groups, creating comparable groups and reducing susceptibility bias.