Abstract
The aim of survival analysis is to study if and when some event occurs. With continuous-time analysis subjects are followed until the time they experience the event or drop out. In practice, subjects cannot always be followed continuously and event occurrence is measured in time periods. This type of survival analysis is known as discrete-time survival analysis and measuring subjects discretely rather than continuously results in a loss of information. The aim of this paper is to study the effects of discretizing survival times for randomized controlled trials by means of a simulation study. It is shown that parameter and standard error biases of both approaches are small and those of the discrete-time approach are only slightly larger than those of the continuous-time approach. The number of time periods has a negligible effect on bias. Power levels hardly differ across the two approaches, so following subjects continuously is not necessary, except when a detailed estimate of the underlying baseline hazard function is needed.
References
2010). Mask use, hand hygiene, and seasonal influenza-like illness among young adults: A randomized intervention trial. The Journal of Infectious Diseases, 201, 491–498. doi: 10.1086/650396
(1995). Survival analysis using the SAS system: A practical guide. Cary, NC: SAS Institute.
(2005). Generating survival timesto simulate Cox proportional hazards models. Statistics in Medicine, 24, 1713–1723. doi: 10.1002/sim.2059
(1995). Controlling the false discovery rate: A practical and powerful approach to multiple testing. Journal of the Royal Statistical Society Series B, 57, 289–300. Retrieved from onlinelibrary.wiley.com/journal/10.1111/(ISSN)1467-9868
(1972). Regression models and life tables. Journal of the Royal Statistical Society Series B, 34, 187–220. Retrieved from onlinelibrary.wiley.com/journal/10.1111/(ISSN)1467-9868
(2002). Guidelines for the design of clinical trials with longitudinal outcomes. Controlled Clinical Trials, 23, 257–273. doi: 10.1016/S0197-2456(02)00205-2
(2000). Applied logistic regression (2nd ed.). New York, NY: Wiley.
(2012). Power analysis for trials with discrete-time survival endpoints. Journal of Educational and Behavioral Statistics, 37, 630–654. doi: 10.3102/1076998611424876
(2005). Effectiveness of home visitation by public-health nurses in prevention of the recurrence of child physical abuse and neglect: A randomised controlled trial. Lancet, 365, 1786–1793. doi: 10.1016/S0140-6736(05)66388-X
(2008). Powerful and cost-efficient designs for longitudinal intervention studies with two treatment groups. Journal of Educational and Behavioral Statistics, 33, 41–61. doi: 10.3102/1076998607302630
(2002). How to use a Monte Carlo study to decide on sample size and determine power. Structural Equation Modeling, 9, 599–620. doi: 10.1207/S15328007SEM0904_8
(2001). Effects of study duration, frequency of observation and sample size on power in studies of group differences in polynomial change. Psychological Methods, 6, 387–401. doi: 10.1037/1082-989X.6.4.387
(1993). It’s about time: Using discrete-time survival analysis to study duration and the timing of events. Journal of Educational and Behavioral Statistics, 18, 155–195. Retrieved from www.sagepub.com/journalsProdDesc.nav?prodId=Journal201853
(2003). Applied longitudinal data analysis. Oxford, UK: Oxford University Press.
(2013). An introduction to R. The R Core Team. Retrieved from cran.r-project.org/doc/manuals/R-intro.pdf
. (