Text this: Estimating time-varying exposure-outcome associations using case-control data: logistic and case-cohort analyses