Seminars @ Biostatistics, UCPH
Upcoming seminars
Monday, December 09, 15:15
A number of works on causal inference with counting process data have attempted to draw on well-established conditions such as consistency, exchangeability, and positivity for identifying causal effects. Despite multiple claimed or conjectured identification conditions and formulas in the literature, counting process analogues of these identifying conditions have not yet been established. There is consequently a lack of tools to formally reason about identification, and it remains unclear when, or if, existing proposed identification formulas can be derived. There are furthermore no existing definitions of potential outcomes for counting process data, and characterizations of dynamic treatment regimes are lacking. In this talk, I define dynamic regimes and associated potential outcomes for data described by marked point processes (MPPs), and introduce MPP analogues of the commonly used consistency, exchangeability, and positivity conditions, which prove to be sufficient for identifying effects in MPP data structures. The conditions are martingale-centered, showing the critical role of martingale theory in identifying causal effects for data described by stochastic processes. The definitions and conditions align with well-established discrete-time conditions in the special cases where alignment is expected, and the results thus bridge the vast literature on survival (event history) analysis with counting processes in continuous time and discrete-time causal inference. Having established a set of identification conditions, I derive and characterize marginal g-formulas, and obtain formulas that are generally quite different from those that have been posed in related works on causal inference with stochastic processes. I discuss the relation with related work, the classical survival literature and the discrete-time causal inference literature.
Friday, December 13, 10:00
In observational studies with time-to-event outcomes subject to
competing risks, the g-formula can be used to estimate a treatment
effect in the presence of confounding factors. The construction of valid
pointwise confidence intervals and time-simultaneous confidence bands
for the causal risk difference, however, is complicated. A convenient
solution is to approximate the asymptotic distribution of the
corresponding stochastic process by means of resampling
approaches.
In this talk, we consider three different resampling
methods, namely the classical nonparametric bootstrap, the influence
function equipped with a resampling approach as well as a
martingale-based bootstrap version, the so-called wild bootstrap. We
compare these approaches with regard to asymptotic properties and based
on simulation studies and demonstrate their usage in a data example.
Map of CSS
You can find CSS next to the Botanical Garden, 5 minutes from Nørreport station.
Meeting room 5.2.46 is the library of the Biostatistics section, located in building 5, 2nd floor, room 46. See the map below for directions inside CSS.
