A new approach to the use of datasets is needed to bring the best evidence to patients and society

By Mary Rice, EORTC Communication Consultant

The debate around the relative importance of randomised clinical trials versus real world data is a false one. It is not a question of choice, or of one being better than the other, but rather of accepting their complementarity and using the data thus generated in the best interests of patients. This was the primary message to come out of a panel discussion on ‘The future of clinical trials in cancer and the role of real-world evidence’, which formed part of the World Cancer Series; Europe Virtual Week, organised by The Economist.  Participants were EORTC Director General Dr Denis Lacombe, Dr Jennifer Tursi, Breast Cancer Global Medicine Team Leader at Pfizer, and Stefan Gijssels, Chief Executive, Digestive Cancers Europe.

Opening the discussion Dr Lacombe said that, until recently, RCTs had been the only way of bringing robust clinical evidence to cancer patients. This was still not in dispute, but there were now other forms of research and the challenge was how to integrate them. « One of the things I find most striking today is that RCTs are often simply associated with regulation – putting drugs on the market.  But it’s a lot more than that. For example, some EORTC trials randomise patients between existing therapeutic practices in order to discover which gives a better outcome. »

Another important aspect is the eligibility criteria for trials, he said. « The more you select your patients for a given trial, the more you are in an artificial world. We need into perspective the type of question we are asking and the type of methodology, and look at how they are complementary.»

In order to make progress in clinical research today, there is a need for much more than just clinical data. This means that datasets are becoming more and more complex, and getting access to them in a structured way is extremely difficult, as is finding a way through all the information they contain. « There’s also a risk that, by working with all of these data, we dilute those that may be important. Collecting data for collection’s sake can also be extremely expensive and may end up as a fishing expedition where at the end of the day you haven’t found much, » said Dr Lacombe. « We need to proceed prudently and make sure we have the methodology right before we can start to make promises to the community. »

Recognising innovation as such and deciding to how use it are still major questions.  « For example, when immuno-oncology first came to the market for Stage 4 melanoma, it was innovatory, and there is no dispute about that. But, the next question is about the duration of treatment. There is biological evidence that a shorter treatment might be just as effective, and the obvious next step would be to do a randomised discontinuation trial where you would compare two treatment durations. But this trial has never been done, and independent sector cannot do it because of the cost of the drugs. Even today, ten years after the innovation, this is still a major question, » he said.

If the answer to the question were to be that a shorter treatment was just as good, there would be savings for healthcare systems, and such savings could mean that more patients would be able to have access to the therapy. There is real world data in existence on the treatment, and attempts have been to analyse it. « But currently all we are likely to get out of it is a collection of the therapeutic practices of individual doctors, » said Dr Lacombe « and this not valuable for informing patients and society. So we still don’t have the answer to this question. »

This constitutes a large gap in treatment optimisation and leaves payers facing situations where there are no appropriately-organised datasets to answer the question that is being asked. « The whole system of drug development needs re-engineering in order to be able to provide such datasets. We should all be prepared to move out of our comfort zones in order to achieve this, » said Dr Lacombe. « All the data needed to do this exists – it’s just chaotic and insufficiently organised. It wouldn’t take a huge amount of effort to reorganise this so that we are able to optimise all the knowledge that we have and give accurate information to patients and to society. »

However, the need to carry out this kind of reorganisation is insufficiently understood at European level at present. Even though those working in cancer know about wastage, inappropriate treatment and screening, and over and under diagnosis, this knowledge appears to have passed policymakers by.

« We have a wealth of opportunity out there if we can reorganise, define the questions we want to answer, and apply the best methodology to those questions, and to try to further develop our knowledge of how to use all the datasets we have, » Dr Lacombe concluded.  « And we should not forget that we are not only talking about drugs. Cancer patients are also treated by radiation oncology and by surgery, and we need to understand the optimal integration of all these disciplines. The issues are exactly the same with other technologies. »