Although Dr Davis and Mr Walsh raise some interesting issues in their letter, they seem to have largely misinterpreted our data. Our reply will specifically address the questions on or criticisms of our data, and a more general response will address their criticisms of the purpose and conclusions of our study.
Our article describes our study finding that delirium reversibility is less likely when there are repeated episodes of delirium.1 This is consistent with the implication that repeated episodes are unlikely to be independent events. Our use of the χ2 and t tests to specifically compare episodes of delirium was restricted to those comparisons of duration, number of precipitating factors per episode, likelihood of precipitating factors (probable or possible), and the distribution of 228 precipitating factors (either with reversal or nonreversal) in the first episode per patient. For comparative purposes the use of data from a single episode per patient is more correct than using data from multiple episodes per patient, hence we selected only the first episode for the comparison of precipitating factors in reversed vs nonreversed first episodes of delirium. This is stated clearly in the title of our Table 3. The P values in Table 3 relate to a comparison of totals for each precipitating factor as a proportion of all precipitating factors (not number of episodes) in reversed and nonreversed first episodes. This is already stated in a footnote to Table 3. Hence the P values as quoted in Table 3 are correct.
Lawlor PG, Gagnon B, Mancini IL, Pereira JL, Hanson J, Suarez-Almazor ME, Bruera ED. Clinical and Ethical Questions Concerning Delirium Study on Patients With Advanced Cancer—Reply. Arch Intern Med. 2001;161(2):297-299. doi: