How Can We Design Clinical Trials for Older People? We Asked an Expert
There are significant challenges in the design and implementation of clinical trials for these patients.
Complete the form below to unlock access to ALL audio articles.
Geroscience is the study of the molecular mechanisms that underpin and drive the aging process and how these lead to disease. The field aims to develop clinical interventions to extend the period of optimal health as people age. Increasing the knowledge gained from geroscience clinical trials is more important than ever, with the number of older people projected to increase and to place significant pressure on global health systems.
Many older people develop multiple different conditions as they age (known as multimorbidity) and as a result, are often taking several drugs to manage these conditions (known as polypharmacy). This presents significant challenges for the design and implementation of clinical trials in these patients.
At ELRIG Drug Discovery 2024, Technology Networks sat down with Miles Witham, professor of trials for older people at Newcastle University, to learn more about these challenges and how novel approaches can be used to improve current research.
In your talk, you mentioned aging is hard to study – could you elaborate on why this is the case?
One of the problems is that aging involves multiple conditions and multiple organ systems. The way that we traditionally do clinical research is that we focus on a single disease or organ system and study it in isolation, so the old way of doing research is not a good fit to the challenges that aging brings.
Some of the challenges are because we haven't got out of that old mindset into a new way of looking at how we develop or test drugs in clinical trials. In the old model, you would select a target – one that was very specific to your condition – and measure a narrow range of things in trials that were relevant to that condition. For instance, you'd develop a drug for a lung condition, and a lot of the things you would measure would be about lung function, breathlessness or production of sputum. If you try and multiply that for aging, then you run into problems very quickly, because you're looking at a heterogeneous group of people with different combinations of conditions, and you can't look at a single condition in isolation. You can either measure lots of different things for a lot of different conditions, which becomes very burdensome and very expensive, or you can try and find different ways to do trials, to recruit people and to measure outcomes. That's important, because a good drug or a good intervention for aging needs to be of benefit across multiple organ systems. We need to demonstrate that benefit and that it doesn't make one organ system better while worsening other conditions; that is something that we're not very good at measuring in standard clinical trials.
One of the things we need to think about is who we will be treating with these interventions. It's worth considering two groups of people at either end of the life course because what makes a good intervention will vary.
For example, if we're treating people who are very old, have multiple conditions and are very frail, then the intervention needs to have very few side effects. Any side effects need to be very mild, because even a minor side effect could have a major impact on somebody who is very frail. There also shouldn't be too many contraindications to using it; for instance, there are many drugs that we can't use in people who have kidney disease, but most older people have kidney disease. You don't want an intervention that you can't give to anybody who's got kidney disease or that is likely to interact with the medication that they're already taking.
At the other end of the life course, if you're thinking about prevention of conditions related to ageing by treating young, healthy people, you're probably going to need to do it for decades in order to bring them into old age in good health. That group of people may tolerate your intervention much better because they're young and healthy, but will they keep taking it for decades if there are side effects? We also need to be sure that there are no long-term side effects; problems with cancer, problems with pregnancy or problems accelerating other conditions over the life course that perhaps wouldn't be apparent within a few years of taking it but might be a problem decades down the line.
Each of these groups will have different considerations when we think about what a good treatment is. Aging is so common and pervasive; it affects all of us. If we're going to make a difference at a population level, we must have treatments that are affordable for the whole population. The model of developing expensive targeted treatments for a very small number of people isn't going to work for a problem like this.
There are some advantages in using repurposed drugs as they do cut out a lot of the initial development and toxicology work that's necessary to bring a novel drug to the point of human testing. I think there are real advantages in speeding up the process, and there are also advantages in that even if the repurposed drugs aren't perfect, they might give us proof of concept and allow us to do geroscience trials and show that they are feasible. If we've got that proof of concept, it will give other people confidence to come into the market and develop novel entities for a particular indication.
The downside of repurposed drugs is that they're not going to be ideal for the particular indication that we're looking for because they weren't designed to actually deliver for that indication. Some of the effects we're looking for may be off-target effects or effects that are a consequence of the target but might not be the best way to treat that particular molecular pathway, but I do think that they've got some merit. And certainly, some of the work we've done in the past has been very much focused on repurposing drugs, because as academic institutions, that's a much easier path to give that proof of concept.
The MET-PREVENT trial was a Phase 2 trial of metformin versus placebo in older people who were very frail and who had sarcopenia – muscle weakness in old age. We deliberately designed the trial to make it much easier for older people with frailty and sarcopenia to take part. This is a group of patients who don’t walk very far and find it difficult to get out of a seat, so expecting these people to come to a hospital and walk half a mile down a corridor is hard for them. These patients also often get ill quite often because they have multiple conditions, and they have very high dropout rates in trials.
These have been real problems in doing sarcopenia trials or trials with older, frail people in the past. The MET-PREVENT trial used quite simple outcomes, and we were able to deliver those outcomes and do the study visits in people's own homes. We gave people a choice, though, and that flexibility was key. A lot of people wanted us to go and see them in their own homes, whereas a small proportion of people said absolutely not – so there was still the option that they could come into the research center for their visits if they wanted to. Because of that, we were able to retain almost everybody in the trial. Over 4 months, we lost 2 out of 72 people; 1 died and 1 withdrew from the study, giving us a 97% retention rate, which is much higher than we would usually expect for this group.
How can we promote inclusivity in geroscience trials to make sure that underrepresented groups take part?
This is an increasingly recognized problem as the people who go into clinical trials don't always look like the people who've got the conditions. This is true of whatever condition you look at; a good example is heart failure, where the average age of people with heart failure is around 85, but the average age of people in clinical trials is around 65. It's not just age, there are lots of people who don't get into trials: people who have multiple conditions, ethnic minority groups, people from poor neighborhoods and people who are socioeconomically deprived. Often clinical trials aren’t designed to make it easy for those groups to take part; these can be simple things such as, if you are from a socioeconomically deprived area, how are you going to get to the trial center? Often you need a taxi ride – many trials don't do that, and people don't want to be out of pocket to take part. There's an increasing recognition that you need to bring these underserved groups into clinical trials so that we get evidence that reflects the people who we are treating in practice.
You also can't assume that because clinical trials showed a benefit for people aged 60 with no other conditions that you're going to get the same benefit in people aged 80 with lots of different conditions. We as clinicians want that evidence, and patients do too. There is a trust problem now in that we as clinicians don't believe the evidence sometimes because we don't feel it's relevant to our patients. So, finding ways to include these groups in clinical trials is absolutely key, and there have been many recent reports including recommendations from the World Health Organization saying we have to do better. Within the National Institute for Health and Care Research (NIHR) we had a project called the INCLUDE Project that provided a road map for how people could go about designing better ways to include some of these underserved groups, and the solutions are different for every underserved group. We've generated some guidance for how to bring older people into trials, other people have created guidance on how to bring ethnic minority groups into trials better and others have done group work on how to bring people in socioeconomic deprived regions into trials. That practical guidance on how to better design and deliver trials will provide us, hopefully, with better evidence in the future.
What potential do you see for AI and machine learning to improve geroscience trials?
I think there are ways that we can leverage AI. The first thing to say is that AI doesn't exist in a vacuum; it's not all about the technology. It's about having the right people and processes as well as the right technology. If we can bring the three of those together, then we'll make progress.
There are a few different ways that AI can be helpful. One is helping us to find and recruit the right patients. We know patients want to take part in clinical trials, and sometimes it's quite hard to match patients and trials together. The other way is around outcomes, particularly outcomes in early phase trials and geroscience, where we are potentially measuring a lot of different outcomes across the many different organ systems. We need ways to integrate that high-dimensional biomarker data because a traditional clinical trial has a single outcome, and that doesn't provide a good fit with the need to measure a lot of different markers across many conditions.