The Antidepressant Advisor - version 1
Research type
Research Study
Full title
The Antidepressant Advisor: A decision support system for UK primary care - a feasibility study
IRAS ID
227010
Contact name
Roland Zahn
Contact email
Sponsor organisation
South London Primary Care R&D for Lambeth CCG
Duration of Study in the UK
2 years, 0 months, 0 days
Research summary
Sumary of Research
Depression is particularly disabling because most patients do not recover after treatment with one standard antidepressant medication or talking therapy. The National Institute for Health and Care Excellence (NICE) recommends using a choice of three antidepressants in the first instance, all of which work in a similar way. NICE also recommends a choice of antidepressants to be used in the second and third instance if the first antidepressants fail to help. There is, however, a lack of clear strategy for how to use these antidepressants, and it is difficult to receive advice from psychiatrists who have experience prescribing these antidepressants. As a result, second- and third-line antidepressants are only prescribed by some
GPs, meaning patients’ access is highly variable, which increases their burden of depression. A computerised tool to help GPs make decisions about antidepressant prescribing has been shown to improve depression in a US study. For UK GPs, however, there is no available antidepressant decision support tool. Here, our aims are to 1) develop the first such tool through collaboration with EMIS, the leading provider of GP electronic health records, and 2) to find out what factors are important to consider for future larger studies, such as how many participants drop out of the study, what impact the tool has on patient care and how satisfied GPs are with the tool. In our proposed study, we will randomly allocate GPs into two
groups: one group will provide care as they normally would; the second group will use the decision support tool to advise on choice of antidepressant. The findings of this study are necessary for planning future studies, which will investigate whether this decision support tool improves symptoms of depression. Service users have strongly endorsed the importance of our approach and will act as steering group members.Summary of Results
Each week, two million people in the UK experience a depressive episode. Unfortunately, current treatment approaches for depression rely on trial-and-error. A large proportion of people suffering from depression neither recover after treatment with antidepressants nor with psychotherapy. Their GP may feel unable to provide advice to them, yet they are not at risk enough to be referred to a specialist and, as a result, are left struggling with their mental health in primary care.If those “non-responders” could be identified early, many months of delays of being offered more suitable treatments could be avoided. If we were able to identify markers of poor prognosis, we could develop personalised treatment algorithms and pathways. The Antidepressant Advisor Study (study 3) sought to investigate the potential of brain magnetic resonance imaging (MRI) in predicting and understanding what distinguishes people whose depression improves in primary care from those where this is not the case.
Depression is associated with a wide range of cognitive, behavioural, emotional, and physical symptoms. Some of these symptoms have been linked to specific patterns of brain activation, which can be captured using neuroimaging techniques, such as functional MRI (fMRI). In our study, we wanted to investigate whether these patterns of activation could be used to predict and understand change in depressive symptoms over four months in primary care.
To investigate this, we recruited 45 participants with current depression who had not responded to at least two antidepressant medications. We also recruited 22 control participants with no depression history for comparison. After the baseline session, participants were asked to attend an MRI session in which they completed tasks. After the session, participants with depression were encouraged to optimise their treatment, and followed-up after four months of treatment by their GP. We made specific predictions and planned our analysis, all of which we registered publicly before analysing our data.
One of the most striking findings of the study was that less than half of our participants showed an improvement in their depressive symptoms (42%). Most participants also did not change their treatment (58%), which emphasises the need to find ways to improve treatment. We were delighted to see that our functional MRI measures at the initial study visit indeed predicted a significant proportion (around one third) of the subsequent change in depressive symptoms over four months.
More specifically, we measured brain activation related to various aspects of depression. For example, people with depression and anxiety were shown to be more attentive to negative stimuli. This bias can be probed experimentally by showing people happy and sad faces – an almond-shaped part in the depth of our brain, called amygdala, tends to respond more strongly to sad faces relative to happy ones in people with depression. This has previously been found to be helpful in predicting how well people respond to antidepressants. Accordingly, participants in our study who reported a subsequent improvement in their symptoms, were more likely to have shown reduced brain activation to sad faces relative to happy ones four months before. In other words, their brain amygdala response was more responsive to positive stimuli even before their symptoms improved, which may have helped with subsequent improvement in symptoms.
Another aspect of depression often reported in patients, is that of excessive self-blaming emotions, such as feelings of guilt, shame, and disgust towards oneself. People with depression tend to blame themselves for things, even when it is not necessarily their wrongdoing. We can measure this “finger of blame” by showing people blame-evoking statements of hypothetical social scenarios in which they are either behaving negatively towards their friend or their friend towards them. Previous work has shown that a measure of brain connectedness between two brain areas associated with this task was predictive of risk of recurrence. One area is called the right anterior temporal lobe and sits just beneath our temple, while the other area, more widely studied in depression, is called the subgenual cortex and sits in the depth of the midline of our brain.A healthy pattern of connectedness between these brain areas is thought to be important for integrating information that helps us to come up with a nuanced interpretation of failure and protects us against overgeneralisations as commonly found in depression. This connectedness can be targeted using brain training, offering a potential treatment for depression. Our previous studies have shown that the right anterior temporal lobe is important for understanding the social meaning of situations and helps differentiate the interpretation of when something goes wrong in a social interaction. In contrast, the subgenual region is relevant to how much an individual tends to attribute blame to either themselves or others (finger of blame). Interestingly, our current study showed for the first time that connectedness between these two brain areas whilst feeling self-blame is associated with subsequent improvement after four months in primary care. This could either be because this brain signature is associated with a subtype of depression who tends to get better more easily or because it influences how people respond to primary care treatment.
One further aspect of depression which we investigated in our study, was the connectedness between brain regions when not engaged in a task, the so-called “resting-state”. The idea is that, in depression, some brain regions communicate too much and others too little even when not engaged in a task – and this disrupted connectedness has been linked to depressive symptoms. Previous studies have shown that this connectedness could successfully predict response to subsequent treatment, and we found something similar in our sample of difficult-to-treat primary care patients.
Even though we were delighted to identify these neural signatures of subsequent improvement in depression, it is important to ask the question whether they perform better than standard clinical measures, such as scores of depressive symptoms or anxiety measured initially. When using these standard clinical measures, only 3% of the differences between people regarding their change in depressive symptoms over four months could be predicted. In other words, these measures tell us very little about subsequent change in symptoms. Encouragingly, when adding our neural measures, we were able to explain 32% of the differences between people in how much their depression changed over four months – almost ten times better than the clinical measures. These results show the strong potential of our fMRI measures as relevant predictors of how likely people’s depression is to improve in primary care.
Having said this, our measures are in themselves insufficient to be used to make predictions for each individual. Furthermore, because one session of MRI costs several hundred pounds and there is a shortage of MRI scanners in the UK, it is unlikely that the NHS will fund these scans routinely. However, the Centre for Neuroimaging Sciences at the IoPPN with whom we collaborated on this study, led work to develop a cheaper head-only mobile MRI scanner which could solve this problem. We are also working on additional measures to identify who is likely to benefit from treatment based on virtual reality and these could be combined with mobile MRI in the future to arrive at more accurate individual level predictions, as well as more fine-grained predictions about, for example, whether specific classes of medications would be beneficial or not.
Independently of practical uses of our findings, which will take at least a decade and may never come to fruition, our study has direct and highly relevant implications for the understanding of the brain mechanisms of depression. The three different functional systems investigated in our study have so far never been studied in the same population. Our identification of the most relevant brain signatures will inform further brain training and neurostimulation studies, as well as specific cognitive training-based approaches.
Relevant links:
- Paper on treatment gaps and why this study is relevant:
https://eur03.safelinks.protection.outlook.com/?url=https%3A%2F%2Fu2790089.ct.sendgrid.net%2Fls%2Fclick%3Fupn%3DXv3JSvJ-2B3M71ppf7N9agbQne05VVe0UaW9iiRd6kqFroC8tTWd7NtKM1gAVoGNMfXR1pEF-2FtYEU8i0i0WhKf-2Bg-3D-3DAtVS_E1aO2-2BZlVOSJJV-2FajQqskegTd6IRomHYTi-2Fbt8SH3YJdHvKudpKX-2FNZUvp0CTe7vJ1djUzuP7oglH-2FogUOIqO-2FjNlYzh7-2FxZbQsvlKJ7M90zf-2F3w-2F37LZ-2Bi0unapCbP77LcLRJO3rn5Ah8-2FlF2-2B1wB-2Bm43NNtEuLWHCeL5mGK-2BSro32w-2BQaIFCUy2IidQX7I0F-2BuWMoPf21VFdlnyTSqpA-3D-3D&data=05%7C01%7Capprovals%40hra.nhs.uk%7Cd7809fa84d594588f9a908dac89a5071%7C8e1f0acad87d4f20939e36243d574267%7C0%7C0%7C638042862338682146%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=5ATrWmErwSHTS4%2FXXFSz9D%2FJ7nCAP3Y7Q%2BfnKjySuEA%3D&reserved=0
- Pre-registered predictions and analysis plan: https://eur03.safelinks.protection.outlook.com/?url=https%3A%2F%2Fu2790089.ct.sendgrid.net%2Fls%2Fclick%3Fupn%3DXv3JSvJ-2B3M71ppf7N9agba4yu73OCS9U-2BkKS40W1kfZLMhYNSL6uAQ3D-2BHysQkrDAeHzkHYROXmcdnmZ2c0VCQ-3D-3DSmJW_E1aO2-2BZlVOSJJV-2FajQqskegTd6IRomHYTi-2Fbt8SH3YJdHvKudpKX-2FNZUvp0CTe7vAvKh8var2JyQfaJRMZz28CXpyXNQI-2F09jKO8uoPLCMZK-2FxwX9oFvAZ7jM-2B9BWXw6ZgNznKztOL6wZ7sIKKH4WE9ay3vlIFSNjArKdOl6-2FIYkEAFeUMSJMm-2B81eAZjyXcpT40WZw7E-2FaR-2BWkcBfTcFA-3D-3D&data=05%7C01%7Capprovals%40hra.nhs.uk%7Cd7809fa84d594588f9a908dac89a5071%7C8e1f0acad87d4f20939e36243d574267%7C0%7C0%7C638042862338682146%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=KI0%2Byl%2BvAOX0kocy3hmLLAuW3YG2rzB7nAjufxg4S38%3D&reserved=0
- Blogpost about brain training:
https://eur03.safelinks.protection.outlook.com/?url=https%3A%2F%2Fu2790089.ct.sendgrid.net%2Fls%2Fclick%3Fupn%3DXv3JSvJ-2B3M71ppf7N9agbRBMJetqSizLPi-2FOOjkmFHohnzdKqKFg16fBIiO6G88VbLU0nw6RRST9nBSdb-2Bu3iMFNOxhbjgm4ne-2BV340nHJBJ1R4f4JioZ7DUPsaAASajamEO-2B5r-2B6n-2FEZfgGI9IGaDA7U4FHPUtE5G37tZTp6bs-3DRHCY_E1aO2-2BZlVOSJJV-2FajQqskegTd6IRomHYTi-2Fbt8SH3YJdHvKudpKX-2FNZUvp0CTe7viMrbESEa-2FznsiSK3Y3KJcjNrZUf7F-2B74q9ijyPvkGF9FM4YAjfsF5JniYw-2BAbol7kya1sRLhUm6LnxTXHXolCVgxCw0cI24s-2FOrPFLeoj64DdeBCiDxeyA3sbfbiZuZEvvXRu4E-2BDa5NIRBzGXlvmg-3D-3D&data=05%7C01%7Capprovals%40hra.nhs.uk%7Cd7809fa84d594588f9a908dac89a5071%7C8e1f0acad87d4f20939e36243d574267%7C0%7C0%7C638042862338682146%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=0NKZMRv%2FyH0FJXWpiPSQIpPVWP1u3wiVRiH%2FXeze3IY%3D&reserved=0
- Work on head-only mobile MRI scanner:
https://eur03.safelinks.protection.outlook.com/?url=https%3A%2F%2Fu2790089.ct.sendgrid.net%2Fls%2Fclick%3Fupn%3DXv3JSvJ-2B3M71ppf7N9agbTouuFx9eZO-2FktuwZlrmxggwUZcHJ3IuINOoTlX0jkP3VIzCjj-2FetHp8HCT1FY-2FmIte0wXhyGCZ8q-2FTVHftuiRDncsnVw6JY-2FE8vLx8oz4FetvGiJR6KOTj3ov51hgH-2F4s6OTSUyfjG-2Bqii5Kl8kO2E-3DhtJd_E1aO2-2BZlVOSJJV-2FajQqskegTd6IRomHYTi-2Fbt8SH3YJdHvKudpKX-2FNZUvp0CTe7vn4CHsG1S5ODdoA53pE-2BM-2FKoxDXJE4H31P-2FsndIWvOaw5ZP6s-2Btb-2FqioUCM21eK14vfM5-2FoN0VtyPO4WWYE6BQtwnvwvva1mPACP9kZ8U1WXpvpFjqwMzdDw-2FuU4p3hZHcYAqub3JLTMazS9XrIUzfQ-3D-3D&data=05%7C01%7Capprovals%40hra.nhs.uk%7Cd7809fa84d594588f9a908dac89a5071%7C8e1f0acad87d4f20939e36243d574267%7C0%7C0%7C638042862338682146%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=ahQZiGGEM%2B3nNVV3IAWVQOuNhvANl%2F8UMjAqBfq4kJA%3D&reserved=0
- Paper on virtual reality in depression: https://eur03.safelinks.protection.outlook.com/?url=https%3A%2F%2Fu2790089.ct.sendgrid.net%2Fls%2Fclick%3Fupn%3DXv3JSvJ-2B3M71ppf7N9agbVDknLRlk3r9Y0tfvaGiBUDbwiDpf99N0bvwWSj9y-2FJiHFLPORQeHOTu4j5qfcFDGukq3GETK5Acx1ahIDyuD7w-3DI6T9_E1aO2-2BZlVOSJJV-2FajQqskegTd6IRomHYTi-2Fbt8SH3YJdHvKudpKX-2FNZUvp0CTe7vhGwTJKoyuBQe7mhrNjvIys97MLQE7kw-2FkrNTLrK-2BqBvj62Awfs61ZKdVVElUxR72hBnMIfxpnzRl1-2FfizYkaC8B4OvuWZuptrDa-2BZcNSIHzylUc9PlaEN62rRqhdEMH-2FW7rCdrneXw-2BB5kPB-2FnipQQ-3D-3D&data=05%7C01%7Capprovals%40hra.nhs.uk%7Cd7809fa84d594588f9a908dac89a5071%7C8e1f0acad87d4f20939e36243d574267%7C0%7C0%7C638042862338682146%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=8rJmcT3rYFUTgUUbNE9iTbBpenJcp02o%2BAy%2BXffPMJo%3D&reserved=0REC name
London - Camberwell St Giles Research Ethics Committee
REC reference
17/LO/2074
Date of REC Opinion
2 Feb 2018
REC opinion
Further Information Favourable Opinion