 |
|
 | |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
|||||||
|
In reviewing the epidemiologic literature on TMD pain, Drangsholt and LeResche (1999) found lack of adequate sample size to be a weakness in the majority of the studies examined. Specifically, sample size was found to be inadequate in over 80% of 133 prevalence studies and in over 90% of 253 analytical studies (cross-sectional, cohort, and case-control studies aimed at investigating risk factors). If the sample is too small, differences in proportions can appear important when they are simply due to the play of chance, or a clinically meaningful difference may not be statistically significant. An additional concern when sample size is too small is that the sample often needs to be divided into subgroups by age and gender and/or by diagnostic subtypes, resulting in further reduction of the sample size. Investigators are increasingly interested in examining prevalence, risk factors, and longitudinal course of specific diagnostic subtypes of TMD (e.g. myofascial pain or painful joint problems). Sample size will need to be a significant consideration in the study design if investigations of this type are to yield meaningful findings. Information
on how to calculate appropriate sample size for specific study
designs can be found in Friedman
et al., 1998.
|
|||||||
