To mark World Mental Health Day, I’m writing a blog that covers two separate, but related, things. Both things relate to a question a patient asked me back in about 2014 and which I have thought a lot about since then
“Why aren’t researchers churning out new treatments for mental health problems the way new treatments for cancer get churned out?”
Author: Dr David Smailes
The first thing I wanted to write about is something that I think is implicit in that patient’s question, but doesn’t exactly answer it. It’s about how well treatments for mental health problems work in comparison to other types of health problems.
How effective are treatments for mental health problems?
A relatively widely held belief is that interventions for mental health problems aren’t very effective. However, at least some of the data we have suggests that treatments in psychiatry (such as medications for things like obsessive-compulsive disorder, depression, and panic disorder) work, on average, about as well as treatments for other types of health problems (such as medications for heart failure, asthma, and COPD).
A number of caveats need to be added here, as the authors of the study – Leucht and colleagues – noted. For example, it is difficult to compare how well treatments for different illnesses or health problems work because the outcomes researchers in different fields look at are so different. That is, if a psychological/psychiatric treatment doubles the likelihood that a patient recovers from depression, is that treatment as effective as one which halves the risk of someone dying from a heart attack? That’s a hard judgement to make. We also need to take into account things like the side effects of treatments, too. All of this means, in short, that it’s very difficult to say how well treatments for mental health problems work in comparison to treatments for other types of health problems. But looking at the data from Leucht and colleagues’ analysis, treatments for mental health problems seem reasonably effective.
That being said, I think no one would disagree with the claim that we need to develop novel, more effective treatments for mental health problems. Disappointingly, over the past few decades, few – if any – novel, more effective interventions for mental health problems have been developed. There are many possible explanations for the lack of novel, more effective treatments – low levels of funding in mental health research in comparison to things like cancer research; mental health problems simply being more difficult to understand and develop interventions for; researchers focussing on the wrong kinds of questions. But the possible explanation I’m focussing on here is that too much of the research done into the causes of mental health problems can’t be replicated.
The problem of replication
When research can’t be replicated, this means that when other researchers try to repeat a study someone else has performed, they fail to find the same results. The results of the original study, therefore, may be untrustworthy. Over the past decade, many scientific fields have re-examined what proportion of the findings they generate are replicable, and in several areas of psychology it has been estimated that only 30-50% of findings are replicable. Mental health researchers have, however, not really examined how replicable (or trustworthy) findings in our field are. But I see no reason to expect that our field will be different to other areas of psychology. This means that it is quite likely that much of the evidence we have about what factors might cause mental health problems will be wrong. And this is important because our understanding of what causes mental health problems tends to shape the treatments we develop to treat mental health problems. So, if we aren’t doing replicable research into what causes mental health problems, then we have little chance of developing novel, more effective treatments for those mental health problems.
Happily, there are many reasonably straightforward solutions to these problems. We know, for example, that running studies that (a) use larger samples than we have typically achieved in the past, that (b) pay more attention to how well variables are being measured, and that (c) involve fully open reporting of how the study will be run and how its data will be analysed generates findings that are more replicable and trustworthy.
How is the Psychology Department at Northumbria addressing this issue?
Here in the Department of Psychology, we have been conducting studies in line with these solutions, to try to improve the replicability and trustworthiness of mental health research. These studies have tested how replicable some findings in hallucinations research are when large sample sizes are used and have shown how we can examine whether variables in research have been measured effectively. More broadly, Northumbria University has joined the UK Reproducibility Network, an organisation that aims to improve the replicability and reproducibility of all areas of research in the UK. This is a really important step and shows that the university is committed to generating really robust, trustworthy research.
As I said earlier, some the solutions to the problem of research findings that can’t be replicated and so seem untrustworthy are quite straightforward, but they are often time-consuming and costly. This has meant that the take-up of these solutions has been quite slow. But, at least to me, it does feel like mental health research is beginning to move in the right direction. And the quicker that can happen, the sooner we should be able to start developing novel, more effective treatments to help people struggling with their mental health.
Interested in hearing about our research in to mental health?
Head over to our health and wellbeing research blog