b) An unexpectedly high number of patients were not evaluable for primary endpoint according to the requirements specified in and implied by the protocol. Although the original protocol focused on evaluable patients, it did not foresee additional patients to be accrued to account for non-evaluable patients.

c) There were many clusters with only one or two or even no patients. These are problematic because they do not allow to properly calculating the within-cluster correlation required for the interim and final analysis. In fact, this renders patients who are evaluable of limited value for the Inhibitors,research,lifescience,medical analysis. While the trial design allows for varying and even small cluster sizes it is not clear how these very small clusters will affect the final results. Moreover, while there are approximate formulae to adjust the final sample size for varying cluster sizes, these only apply to approximate calculations which were not used during the trial development. These problems revealed from the interim Inhibitors,research,lifescience,medical analysis led to the following conclusions which were documented in a note to file: 1) 160 patients will be required, distributed to four patients Inhibitors,research,lifescience,medical per cluster with twenty clusters per treatment arm. 2) An additional five

per cent, i.e. eight patients, will be accrued to account for varying cluster sizes. 3) That is, in total 168 evaluable patients will be accrued to meet the design specifications of the protocol

(alpha=5%, power=80%). Because of the problems described in b) and c) even 240 accrued patients (the maximum foreseen in the protocol although there only evaluable patients were Inhibitors,research,lifescience,medical mentioned) would not have been sufficient to obtain 168 evaluable patients. Hence, it was decided by the trial team members at the SAKK CC on April 20, 2011 to accrue an additional 10 percent, i.e. 24 patients, resulting Inhibitors,research,lifescience,medical in a total of 264 patients. Context Several studies assessed whether the provision of HRQL data to oncologists, using touch pad symptom assessment devices, [24,44] by using prompt sheets [45,46] or summaries of HRQL, [47] Dichloromethane dehalogenase improve communication between oncologist and patient and symptom control. The provision of a summary of HRQL (EORTC-QLQ-C30) to patients and oncologists in a randomized crossover trial resulted in more frequent discussion of HRQL issues and detection of unexpected psychosocial topics and symptoms [14]. Longitudinal symptom assessment by means of see more computers is feasible in the multicenter setting even for a long time span. In a recently published study patients were invited to report symptoms (Common Terminology Criteria for Adverse Events CTCAE) on an online platform. Of 125 invited, 105 participated for the mean length of one year and showed a high compliance and high satisfaction with the system, however there was only a marginal effect on communication [11].