0005)” and that “The total tracheostomy time was reduced from 34 to 24 days, although this difference was not statistically significant (P = 0.13)”.The second study was also a historical cohort study including ICU patients not under the care of an ENT unit who were discharged selleckchem to the ward with a tracheostomy at St. Vincent’s Hospital, Melbourne, Australia. A total of 280 patients were included in the study of which 41 were the control group and 239 received the intervention over three years. The intervention in this study was an intensivist-led, multidisciplinary team consisting of an intensivist, an ICU liaison nurse, a physiotherapist, a speech pathologist and a dietitian. This team undertook twice weekly ward rounds to review patients and to plan and oversee an individualised tracheostomy weaning programme.

The intervention patients were compared with patients who received standard care by a physiotherapist and speech pathologist with ad hoc input from doctors prior to the introduction of the intervention. The standard care in this study appears to be multidisciplinary, but is not a dedicated service. The study looked at decannulation time from ICU discharge as its primary outcome and hospital length of stay, length of stay after ICU discharge and length of stay of less than 43 days (the upper trim point for the disease-related group code for tracheostomy) as secondary outcomes.The methods of this study were fairly well documented, but some methodological weaknesses may affect the conclusions of the study.

The study presents patient groups chosen from similar populations; however, the difference in the size of the control versus intervention group is substantial (n = 41 and n = 239, respectively) and the larger group may include a wider range of patient types. Both groups were reportedly treated the same with the exception of the patients in the third year (2006) of the intervention group whose results may have been affected by the introduction of a nurse liaison service. Despite the study reporting similar populations and treatment of patients between the control and intervention groups, the historical control study design introduces the potential for unknown confounders and historical factors that may have affected the results.It appears that measurement of outcomes were not standardised, valid or reliable.

There was no loss to follow up with all patients being included in the final analysis.The study had three findings. (1) The median hospital length of stay decreased over the study period from 42 (interquartile range (IQR): 29 to 73) days to 34.5 (IQR: 26 to 53) days (P = 0.06). (2) The median hospital stay after ICU discharge was reduced in 2006 compared with 2003, from 30 (IQR: 13 to 52) days to 19 (IQR: 10 to 34) days (P < 0.05). This data was provided for the comparison between 2006 and 2003 only. Statistical significance was Cilengitide not reported for other intervention years.