anticoagulant
treatment, diabetes), the statistical analysis was
unable to demonstrate any differences with regard to the risk of
renewed bleeding episodes, which can be explained by the
smaller size of the subgroups.
As far as the dropouts were concerned, we looked at whether
or not the five excluded participants might have effected a
change in the results (Table 3). In the study group, one case was
excluded, who experienced new bleeding episodes. In the
control group, four cases were excluded, where two cases
experienced new bleeding episodes and two did not (Table 8).
This means that the overall odds ratio at n = 105 was 0Æ817
compared with 0Æ784 (Table 3). Thus, the exclusion of these
patients did not change our primary outcome.
Discussion
This study focused on establishing the feasibility of mobilising
hospital patients under treatment for primary nose
bleeding without increasing the risk of new bleeding episodes.
We were unable to demonstrate any difference in the number
of new bleeding episodes between mobilised and immobilised
patients in the trial, which lead us to conclude that mobilisation
did not increase the risk of renewed bleeding. In fact,
the odds ratio was rather close to 1 (0Æ784); judging from
the distribution shown in Table 3, the risk of new bleeding
seemed almost identical for the two groups. Conversely, it
follows that mobilisation did not result in fewer new bleeding
episodes. The overall risk of new bleeding episodes thus
seemed to bear no relation to this treatment regime. The
insufficient number of observations forced us to abandon
attempts to create a statistical model to estimate the number
of new bleeding episodes and to determine any particular
characteristics among participants, who experienced new
bleeding episodes.
Our study included 100 epistaxis patients allocated into
two treatment groups of 50 patients each. For practical and
clinical reasons, the number of participants had to be
restricted. Ideally, the overall sample size should have been
500 patients to detect differences between the two treatment
regimes, but this would require a study period of at least
10 years. Such a long period would have placed a prolonged,
extra burden on the staff to recruit patients and collect data;