These differences in frequency and amount of alcohol consumption were supported by the finding that the patients with nosebleeds were significantly more likely than the controls
to have drunk alcohol within 24 hours of admission (P < 0-01, x2 test of proportions).