We have made a study about the inclusion of older cancer patients in clinical trials and we tried to assess the evolution of the inclusion of elderly patients in these trials before and after SIOG was founded. We analysed all clinical trials, that is to say phase I, phase II and phase III randomised controlled trials, that assessed therapies for haematological and solid tumours for patients older than 60 years old. We also identified all phase III clinical trials in order to retrieve subgroup analyses dedicated to older cancer patients. Then we compared the evolution of these clinical trials between two time periods, the time period that was the pre-SIOG period running from 2001 to 2004 and the post-SIOG period which ran from 2011 to 2014. What we saw in this comparison is that there was a tremendous increase in the absolute number of clinical trials which came from 128 in the first period to 415 in the second period and that this increase was mainly explained by subgroup analysis made in phase III randomised controlled trials.
We also found that the use of endpoints, because we analysed the clinical trials generally but also in detail, and we found that the use of endpoints was relevant for overall survival and patient reported outcome with an increase over time but we also found that other elderly specific endpoints such as functional status remained seldom used and that is something that we think should be encouraged in the years to come.
Do you have any further thoughts on quality of life and extending life treatments?
It’s what I said in my conclusion because we made some recommendations and I think we should use as primary endpoints endpoints that evaluate at the same time quantity and quality of life. In order to do so this is what we discussed during the question session. There are three main ways to do so: we can either use composite endpoints or co-primary endpoints or quality adjusted life year endpoints. In my opinion I would prefer to use co-primary endpoints because it evaluates independently the two endpoints. But some people like to evaluate composite endpoints but when we evaluate composite endpoints it can be very tricky to interpret when there is a difference in the evolution of endpoint. That is to say if we evaluate an aggressive therapy that improves survival but on the other hand decreases quality of life we can’t conclude on the composite endpoint so that’s why I prefer to use co-primary endpoints even though it can be more difficult for statisticians but it’s more meaningful.
Could you talk about your recommendations on subgroup analysis?
We made a logistic regression analysis to know what were the factors associated with an increase of subgroup analysis between the two time periods. One of the factors explaining this increase was the journals in which the clinical trials were published. Journals with a high impact factor, more than 20, tend to publish more subgroup analyses so that is something we said during the question session is that we should ask reviewers to ask for subgroup analyses for older patients for efficacy but also for toxicity because in the majority of cases the subgroup analyses are only about efficacy and not about toxicity.
There is some impact but we still have some work to do and the problem with when we improve our work is that the conclusions, we only get them ten years later so we have to wait. But there has been some progress and maybe something that we didn’t evaluate in the evolution of clinical trials is that we didn’t assess geriatric intervention trials. It’s another thing that we should recommend for the future is to design more trials on geriatric interventions.