Reference no: EM133308520

A case-control study was conducted among women enrolled in Medicaid to estimate the association between cocaine or methamphetamine use and the risk of stroke. The investigators attempted to identify all incident strokes (both fatal and non-fatal) in women aged 15-44 years old from 2017-2019. Stroke cases were identified from the following sources: hospital admissions, hospital discharge records, emergency department logs, and requests for out-of-plan hospitalizations. Controls were selected at random from women enrolled in Medicaid aged 15-44 years old. If a patient with stroke had died or was unable to communicate verbally, the investigators attempted to interview a proxy. Study staff conducted face-to-face interviews of both stroke patients and control subjects to obtain information on the use of cocaine and methamphetamine in the week before the clinical presentation of stroke for cases, or a corresponding date for controls. Use of cocaine/amphetamines was reported by 20 of the 347 cases and 7 of the 1021 controls.

Question

In the Discussion section, the authors write "our findings... might be explained by a greater likelihood that women with stroke reported drug use than women who were controls." What epidemiologic concept are the authors referring to? Explain how this might have biased the observed association between cocaine/amphetamine use and stroke (i.e., the direction of the bias).