I took a random selection of 20 abstracts from ASCO 2016. None of them mentioned that quality of life had been researched in the study but looking up the detail held in clinicaltrials.gov it turns out that 16 of them did have a quality of life component in them. Whether that will be reported when the study is published is a different question. Experience suggests that it will be mentioned, but that for the most part the detailed QoL data will go unreported.
These are the days of ‘big data’ in healthcare. Increasingly we are seeing databases linked, common standards being adopted to allow similar fields to be aggregated, and increasingly comparisons can be made between key outcomes. The International Cancer Benchmarking Partnership is providing fascinating comparisons on survival outcomes from several different administrations – and it is getting into the detail of differences in the ways that national data is gathered.
The first QoL in cancer study appears to have been in the 1970s. The first paper I could find which looked at developing a methodology dates from the 1980s. We now have a proliferation of methodologies dominated by the EuroQoL EQ-5D which gives the kind of generic background view needed for health technology agencies (HTA) to assess the value of the treatment. Other methodologies relate solely to the study to which they are being applied. Unfortunately for cancer patients, the EQ-5D is too crude report the kind of detail that cancer patients need to know to make informed choices about treatment. For that reason too it can only ever be part of the picture for HTA appraisals.
What do cancer patients need from quality of life information? The full picture of the impact of a proposed treatment in the context of the treatment pathway being followed. We want side effects information, how side effects are treated, and want to know how patients feel when taking the treatment in their own words, not those of some detached academic. We want to know about long-term side effects and what comes after if this treatment fails.
The pathway is the key issue in terms of quality of life – it is not a moment in time, it is life. However there are no data covering the succession of treatments, assessing the interactions that can arise as side effects aggregate, and the challenge of describing treatment and quality of life as end-of-life approaches is avoided.
I looked around to see whether anyone is aggregating quality of life data in a way which could provide this picture. No-one. Why not, after all there is at least 30 years’ worth of data around? The quick answer is the usual big data problem, different methodologies, data fields do not match, each study looks at the treatment being tested not at the patient. The pathway is ignored.
“No-one is aggregating quality of life data in a way which could provide this picture”
We are entering the world of targeted therapies. Targeting is a challenge for randomised studies: they can be unethical, either randomising patients when there is good evidence to indicate one arm is more effective than another, or using cross-overs, which invalidate the randomisation. Phase 2 studies usually do not include quality of life because there is no control group baseline with which to compare the treated patient. A strong response rate (60%+) is enough to ensure entry into clinical practice. One result is that patients are offered new treatments without being able to see a full picture of the step they are taking. Possibly just as important, health technology agencies are denied the whole picture of value which they need to make decisions.
Let us leave aside any questions about whether patients will accept such treatments – these issues disappear when survival itself is your primary challenge. Yet, as trials for targeted treatments reach statistical significance faster and with fewer patients than trials for traditional drugs, patients face increasing uncertainty about the impact of therapies they are taking. We need something to mitigate that uncertainty.
A first step would be aggregating existing QoL data to derive a picture of each cancer pathway, modified as clinical standards evolve. It would provide a baseline for comparing new treatments.
A second step would be using one tool/methodology for all QoL in cancer research. Agreeing that step when there are so many competing methodologies is likely to be difficult.
The third step would be requiring a relevant patient group to be the determinant of the measures to be adopted in each QoL study, small or large. Not easy to arrange but do-able.
A fourth step would be mandating full analysis and reporting of QoL data when it is undertaken in a clinical study. That should not be impossible.
If the cancer research community does not take on this challenge we could affect one of the binding factors of cancer care, trust between patient and doctor. Targeted therapy is an amazing development but we are journeying together into an unknown future built on that trust alone, something we haven’t done for more than thirty years.
We welcome contributions to this blog. If you have a topic you would like to write about, please send your post to Corinne Hall – firstname.lastname@example.org You can find our guidelines here – How to write a blog