Seminars

When studying treatment effects on patient quality-of-life (QOL) in the context of successive or recurrent life events, both the QOL scores and the inter-event (gap) times contain useful complementary information. To analyze one source of data without the other might overlook essential aspects of the disease process on quality-of-life, and in some cases might lead to contradictory conclusions. To reconcile these views, we propose a unitary and more comprehensive nonparametric analysis that combines the two separate sources of data by introducing the QOL-adjusted gap time (QAGT) concept. The translation of the event data onto the QOL time scale and the correlation between the underlying gap times create a two-level dependent censoring structure that needs to be corrected in order to draw statistically valid conclusions. Inverse probability of censoring weighted estimators of the QAGT joint and conditional distributions are proposed and are shown to be consistent and asymptotically normal. Continuing research goals are to use pseudo-observations in modeling QOL-adjusted successive event parameters in terms of risk factors. Results modeling a single QOL-adjusted time-to-event will be described as motivation for future work. For one or more event times, once the problem is fully formulated, standard generalized estimating equations software can be used. This is joint work with Susan Murray.