A Nearly Unavoidable Mechanism for Collider Bias with Index-Event Studies

Abstract

Factors suspected of causing certain chronic diseases and death are often associated with lower mortality among those with disease. For end-stage renal disease, examples include high cholesterol and homocysteine. Here, we consider obesity, thought to cause both end-stage renal disease and premature mortality, but which is associated with lower mortality among end-stage renal disease patients. Such seeming paradoxes could reflect collider (index event) bias due to selection of a diseased population for study. However, previous descriptions are incomplete, as they posit an uncontrolled factor causing both end-stage renal disease (the index event) and death. Here, we explicitly note that death can precede end-stage renal disease onset. The target population is obese persons with end-stage renal disease, effects of interest are seemingly controlled direct effects, the usual estimator is a conditional risk ratio, and remaining at risk until the onset of end-stage renal disease is a collider. Collider bias is then expected if any mortality risk factor is uncontrolled, even if no factor also affects end-stage renal disease. The bias is similar to, but differs from, that associated with competing risks. Because control of every mortality risk factor is implausible, bias of the standard estimator is practically unavoidable. Better awareness of these issues by clinicians and researchers is needed if observational research is to usefully guide care of this vulnerable patient population.

Smoking and subsequent human papillomavirus infection: a mediation analysis

“…Past smoking status, however, may be causally relevant to HPV antibodies, and current smoking status is likely a good proxy (41). If we assume that past smoking behavior causes HPV infection, conditioning on infection status at baseline to obtain our at risk population opens the possibility of selection bias through an unmeasured risk factor for HPV infection (42, 43). This bias, however, is expected to be weak unless the unmeasured risk factor is particularly strong, and our model included all relevant risk factors available to us.…”

Acute Myocardial Infarction in Women

“…163 Lipoproteins levels are associated with long-term cardiovascular risk; however, in the AMI setting, there is a lipid paradox: Patients with significantly lower triglycerides and low-density lipoprotein cholesterol levels have higher in-hospital 164 and 30-day mortality rates. 165 This seeming paradox may be due to competing risks of collider (index event) bias resulting from the selection of a diseased population 166 such as older age and higher rates of DM in those with lower lipoprotein levels in the acute setting. Sex specific data examining lipids at admission and AMI outcomes are lacking.…”

Field Synopsis of the Role of Sex in Stroke Prediction Models

“…For example, prognosis among acute stroke patients is largely determined by age and stroke severity, captured in scales such as GCS and National Institutes of Health Stroke Scale, and sex is likely to play a much less influential role. Similarly, the relative infrequency of sex's inclusion in models of outcome events after stroke (including stroke recurrence) may also be understood in light of the potential for index event bias, which can generate paradoxical findings when the index and recurrent events have common risk factors, and studies select patients who have experienced the index event (ie, incident stroke) 59, 60, 61. The selection of patients with a first stroke influences the association between (both measured and unmeasured) stroke risk factors and sex in patients who are included in the study in ways that could obscure the predictive effects of sex on the incidence of subsequent strokes or other outcomes.…”