Competing risks in clinical research
Van Kruijsdijk, R.C.M.; Eijkemans, M.J.C.; Visseren, F.L.J.
Nederlands Tijdschrift Voor Geneeskunde 156(46): A5176
2012
ISSN/ISBN: 1876-8784 PMID: 23151330 Document Number: 663616
In clinical research, study populations are often followed up until the occurrence of a certain outcome, such as breast cancer or death. Survival analysis, used to analyse the time-to-event data, is usually focused on one particular event; however, study participants may experience other events that prevent the event of interest from occurring. This is a situation of competing risks. An example is the risk of death in a study investigating breast cancer recurrence, as women who have died without having recurrent disease are no longer at risk of this recurrent disease. Competing risks can lead to biased results if not taken into account. Competing risks are important in studies including patients with a high risk of varying clinical outcomes, and in those studies with a long follow-up period. If competing risks are present, the competing outcomes can be combined into one outcome measure or should be taken into account in adapted survival analysis.