New Methods for assessing Cumulative effects of Time-Varying exposures, with applications in pharmaco-epidemiology and cardiovascular epidemiology

An accurate assessment of the safety or effectiveness of drugs in longitudinal cohort studies requires defining an etiologically correct time-varying exposure model, which specifies how previous drug use affects the hazard of the event of interest [1]. To account for the dosage, duration and timing of past exposures, we proposed a flexible weighed cumulative exposure (WCE) model [2,3], where the cumulative effect of past exposure history is modeled as a weighed sum of past values of exposure (e.g. drug doses), observed until the time when the risk is being assessed [2,3]. The weights reflect the relative importance weights to past doses, depending on the time elapsed since the dose was taken [2]. The estimated WCE is then included as a time-dependent covariate in the Cox’s PH model. Likelihood ratio tests are used to compare w(τ–t) against the standard un-weighted cumulative dose model, and to test the Ho of no association [3]. We have extended the WCE model further to (i) flexible Marginal Structural modeling (MSM) with Inverse Probability of Treatment (IPT) weights [4], to account for time-varying covariates that act as both confounders mediators of the treatment effect; and to (ii) pharmacovigilance [5]. The accuracy of the estimates and tests is assessed in simulations [3,4].
I will illustrate the applications of the WCE model to investigate the potential associations of (a) glucocorticoids and infections [6], (b) benzodiazepines and fall-related injuries [7], and (c) didanosine and cardiovascular risks in HIV (MSM analysis) [8].

Recently, we focus on further extension to flexible modeling of cumulative effects of past values of continuous risk/prognostic factors, with possibly non-linear (NL) effects (on log hazard). The potential insights that may offered by the new flexible NL WCE model will be illustrated by re-assessing the impact of past values of SBP on cardiovascular morbidity/mortality in the Framingham Heart Study.

1. Abrahamowicz M, Beauchamp M-E, Sylvestre M-P. Comparison of alternative models for linking drug exposure with adverse effects. Statistics in Medicine 2012; 31(11-12):1014–1030.
2. Abrahamowicz M, Bartlett G, Tamblyn R, du Berger R. Modeling cumulative dose and exposure duration provided insights regarding the associations between benzodiazepines and injuries. Journal of Clinical Epidemiology 2006; 59(4):393–403.
3. Sylvestre M-P, Abrahamowicz M. Flexible modeling of the cumulative effects of time-dependent exposures on the hazard. Statistics in Medicine 2009; 28(27):3437-3453.
4. Xiao Y, Abrahamowicz M, Moodie EEM, Weber R, Young J. Flexible marginal structural models for estimating the cumulative effect of a time-dependent treatment on the hazard: reassessing the cardiovascular risks of didanosine treatment in the Swiss HIV cohort study. Journal of the American Statistical Association 2014; 109(506):455-464.
5. van Gaalen R, Abrahamowicz M, Buckeridge DL. The impact of exposure model misspecification on signal detection in prospective pharmacovigilance. Pharmacoepimiology and Drug Safety 2015; 24(5):456-467.
6. Dixon WG, Abrahamowicz M, Beauchamp M-E, Ray DW, et al. Immediate and delayed impact of oral glucocorticoid therapy on risk of serious infection in patients with rheumatoid arthritis: a nested case-control analysis. Annals of the Rheumatic Diseases 2012; 71(7):1128-1133.
7. Sylvestre M-P, Abrahamowicz M, Capek R, Tamblyn R. Assessing the cumulative effects of exposure to selected benzodiazepines on the risk of fall-related injuries in the elderly. International Psychogeriatrics 2012; 24(4):577-586.
8. Young J, Xiao Y, Moodie EEM, Abrahamowicz M, Klein M, et al. The effect of cumulating exposure to abacavir on the risk of cardiovascular disease events in patients from the Swiss HIV Cohort Study. Journal of Acquired Immune Deficiency Syndromes 2015: 69(4):413-21.