Longitudinal Studies and Determination of Risk
Longitudinal and observational study designs are important methodologies to investigate potential associations that may not be amenable to RCTs. In many cases, they may be performed using existing data and are often cost-effective ways of addressing important questions. The major disadvantage of observational studies is the potential for bias. The absence of randomization means that one can never be certain that unknown confounders are present, and specific studies designs have their own inherent forms of bias. Careful study design may minimize bias. Establishing a casual association based on observational methods requires due consideration of the quality of the individual study and knowledge of its limitations.
Key wordsLongitudinal studies cohort study case–control study bias risk factors sample size estimate
- 11.11. Dawson, B., Trapp, R. G. (2004) Basic and Clinical Biostatistics. McGraw-Hill, Toronto.Google Scholar
- 12.12. Glantz, S. A. (2005) Primer of Biostatistics. McGraw-Hill, Toronto.Google Scholar
- 14.Dupont, W. D., Plummer, W. D. (2004) PS: Power and sample size calculation. Available at http://biostat.mc.vanderbilt.edu/twiki/bin/view/Main/PowerSampleSize, accessed June 25, 2007.
- 15.Schoenfeld, D. A. 2001 Find statistical considerations for a study where the outcome is time to failure. Available at http://hedwig.mgh.harvard.edu/sample_size/time_to_event/para_time.html, accessed June 25, 2007.
- 20.20. Cox, D. (1972). Regression models and life tables (with discussion). J Roy Stat Soc B 4, 187–220.Google Scholar