Figure 2.

Daily administration of morphine, midazolam and propofol and percentages of patients needing these sedatives. Data are expressed as mean DD/MDD. When comparing the summed data of seven days before and after tracheotomy there was a significant difference in dosage and percentage of patients using these sedatives before and after tracheotomy (P < 0.01 with the Wilcoxon signed-rank test and the McNemar test). However, a repeated-measurements analysis of variance showed that, from day -7 to day -1, morphine dosage declined by 3.34 (95% confidence interval -1.61 to -6.24), midazolam dosage by 2.95 (-1.49 to -5.29) and propofol dosage by 1.05 (-0.41 to -2.01) DD/MDD (P < 0.01). The percentage of patients using sedatives also decreased before tracheotomy. After tracheotomy there was no further increase in decline rate, and the dosage remained stable.