Recruitment, follow-up and analysis times in clinical trials of cancer treatment: a case study

Abstract

Summary A study has been made of the way in which the number of events available for analysis in a clinical trial was dependent on the recruitment period, the maximum follow-up time on individual patients and the length of time between the start of the trial and its analysis. The events considered were deaths, local recurrences and late radiation effects on normal tissue in patients treated for cancer of the laryngo-pharynx by two different fractionation regimes. The The third time needing specification is that at which the results of the trial are to be analysed. While early interim analyses may be advisable to monitor the treatment effects, the use of any interim result to stop entry into the trial or to trigger early publication must take into account the effect of repeated analyses on apparent significance levels. This is a topic that has been discussed by many authors, e.g. Armitage et al. (1969), Haybittle (1971) We have recently studied the way in which events accumulated in a trial that began to be planned over 28 years ago. Our results show how, if we had, at the time of planning this trial, the requisite data concerning the time-course of occurrence of significant events, we could have saved ourselves and the trial participants considerable time and effort. Unfortunately there was a lack of suitable retrospective data on which to base such considerations and the trial design was very much influenced by the clinical concepts then current. Our experience may therefore be of value for others planning clinical trials of cancer treatment. Materials and methodsIn 1962 the planning of a multicentre trial began under the auspices of the British Institute of Radiology to compare two radiotherapy regimes in the treatment of cancer of the laryngopharynx. The two regimes differed in their fractionation schedules. One employed five fractions per week (one each weekday), which was common radiotherapy practice at that time. The other used three fractions per week (on Monday, Wednesday and Friday), which, provided any difference in its therapeutic effect was clinically unimportant, would be beneficial to the patient because of a reduced number of attendances, and would be more economical in the use of radiotherapy machines and associated staff. Because of the known radiobiological effects of changing fractionation, the total dose to patients treated with three fractions per week was set at 11 or 13% less than that given to patients treated with five fractions per week, for the longer or shorter schedules respectively. Recruitment started in 1966.The initial aim was to recruit about 900 patients, which would have given a 90% power of detecting a difference of about 10% in 5-year event-free rate, should such a difference really exist. As will be discussed later, this could have been too modest an ambition for a trial that was set up to show that one regime (three fractions/week) was no worse in therapeutic outcome than the other and could therefore be preferred on non-therapeutic grounds. However, the number of parti...