We commend Isojärvi et al1 for their article, "Serum Lipid Levels During Carbamazepine Medication: A Prospective Study." This study should greatly contribute to the medical literature, but we are concerned that some of the conclusions are misleading because of their inappropriate statistical analyses. Since multiple comparisons were made using the t test (at baseline, after 2 months, and after 1 year),1 there is an increased risk of making a type I error.2 In other words, the chance of finding a significant difference when one is actually not present is increased. An analysis of variance for repeated measures with a posthoc test would have been more appropriate in analyzing statistical significance.3 With the lack of patient-specific data, we cannot reassess the results for significance using this method, but a Bonferroni correction can be used to decrease the chance of making a type I error when multiple comparisons
Wagner ML, Preblick R. The Effect of Carbamazepine on Serum Lipid Levels. Arch Neurol. 1994;51(8):748–749. doi:10.1001/archneur.1994.00540200022007
Customize your JAMA Network experience by selecting one or more topics from the list below.