Interpreting significance in an antidepressant drug trial

In a clinical trial to compare two antidepressants, mianserin and dothiepin, the Last Visit Carried Forward (LVCF) analysis was used for the statistical tests. By this method, no significant differences were demonstrated between the antidepressant response of each drug, as measured by the Hamilton Depression Scale. Statistical significance (or lack of it) does not necessarily correlate with clinical significance and, since antidepressant drugs usually take 10-14 days to exert any effect, a further analysis was undertaken excluding patients who completed less than 2 weeks of the trial. This further analysis then demonstrated a statistically significant difference in favour of mianserin (p less than 0.05). In designing statistical analysis for clinical drug trials, it is important to take into account the therapeutic characteristics of the drugs under trial.