Slovenian Research Agency
Multi-state Models: A Variety of Uses
Biostatistics Unit, UK
The use of multi-state models in the analysis of
longitudinal data has increased substantially in recent years. This has been
facilitated by the availability of computer software but also reflects their
usefulness in the specification of data structures and their flexibility.
The use of multi-state models for a variety of problems will be illustrated
to demonstrate these characteristics. These problems will involve
potentially informative observation patterns, the challenges of panel data,
time to event analyses for events defined only by prolonged observation and
correlated processes. The application of causal reasoning in the context of
multi-state models will also be briefly discussed.