Skip to main content

Transforming the understanding
and treatment of mental illnesses.

Celebrating 75 Years! Learn More >>

The Center for Global Mental Health Research Webinar Series 2023: Finding the Right Treatment for the Right People at the Right Time – Promoting Stratification in Global Mental Health Research


LEONARDO CUBILLOS: Good morning, good afternoon, and good evening, everyone. Thank you for joining this webinar. This is the first webinar of the 2023 Webinar Series of the Center for Global Mental Health Research at the National Institute of Mental Health.

Some of you attended the 2022 Webinar Series which hosted seven webinars last year. They are uploaded and they provide an important overview of different mechanisms and processes to apply to some of the calls for applications that the NIMH puts forth both for the training awards as well as for the research awards. So I invite you to access those recordings that are now live.

Today we are kicking off the first episode or first webinar of our 2023 series, and I am thrilled to announce that this webinar has been co-organized by NIMH and our colleagues from the Wellcome Trust, and they will soon be invited to come to the screen.

The topic for today is Finding the Right Treatment for the Right People at the Right Time where I think global mental health has a very important challenge, which is to provide access to care to individuals that are in need. It is equally important to ensure that they receive good quality care, and part of that is ensuring that the services and treatments we provide have the highest likelihood of helping these individuals and communities and to optimize the care we provide. This is precisely one of the areas where stratification comes into play.

I would like to invite my colleague, Usman Hamdani from Wellcome Trust to turn on his camera. Today we have three speakers. As I mentioned, Wellcome Trust is the co-lead of this webinar. Usman Hamdani, who is now on camera, is the research lead for the mental health translation team at Wellcome Trust in England. We also have Wesley Horton from the Foundation for the National Institutes of Health, and Arthur Caye from Universidad Federal do Rio Grande do Sul, Brazil.

Without further ado I will turn it over to Dr. Usman Hamdani.

USMAN HAMDANI: Thank you so much, Leo, for co-hosting this webinar with us, and welcome everyone to this webinar. I would like to begin by thanking you all for your time, and especially Leo for helping us organize this.

What do we mean by stratification in global mental health research? You all know that the mental health diagnostic categories are imperfect. They rely on subjective markers, and, as a result, whenever we have a diagnostic category there are huge variations in the characteristics of people who are within the same diagnostic category. We need objective markers that can help us classify these people into subgroups so that we can intervene and provide the right treatment to the right person at the right time.

These markers would be biological, psychological, social, digital, mechanistic response, risk stratification and so on and so forth. They could be genetic, they can be biochemical. Anything could be there. But this is an area of research in global mental health which we think is of strategic importance to Wellcome Trust and the NIMH, and hence this webinar.

As part of our strategy on this specific promoting of stratification in global mental health research, this is actually the third webinar that we are hosting, and when I will be talking about my talk as a part of this webinar, I will allude to the other webinars and where you can find them. So, as a result, you will have a complete list of like how we are looking at stratification in global mental health research.

Our two speakers today, Arthur Caye and Wesley Horton, are going to be talking about specific dimensions of stratification. Wesley is going to be talking about biomarkers and Arthur is going to talk about predictive algorithms and how they can help us stratify.

Without further ado, I will pass it on to Dr. Wesley Horton who will be talking about the work of the Foundation for the NIH and biomarker regulation and how is it relevant to global mental health research. So over to you, Wesley.

WESLEY HORTON: Thank you so much for the invitation and for having the chance to speak here today. I am really excited to walk through a little bit around regulatory engagement as well as biomarker qualification and what the Foundation for the NIH is doing in this space and how we move forward drug development tools to full use.

And I think very much relevant to the idea of -- at the end of any of these projects that will be ongoing and tackling solutions for mental health, we will need to bring together the right minds and bring together the right tools that will then be successful in bringing forward to patients tools that will be usable, that will be meaningful and, finally, things that will make a difference in people’s individual lives.

I want to start off just kind of giving you an overview of the goals of my presentation. I will give you a quick introduction to what the Foundation for the NIH is doing more generally, what areas we’re working on in mental health specifically. I’ll give a little bit of an introduction around the definitions of what a biomarker is and the clinical outcome assessment, what qualification looks like and how to develop a context of use that can specifically lay out evidentiary criteria for final use of those tools.

Finally, I will give a little bit about our letter of intent process with the FDA and how we work with the FDA to bring biomarkers forward through a qualification process.

Starting off, the Foundation for the NIH is a 501(c)(3). We were founded by Congress to support the mission of the National Institutes of Health, being a convening place for private sector and public sector partners to work together and collaborate and advance biomedical discoveries and breakthroughs for treatments and patients.

We work through many different partnership models. I’m highlighting here some of our major areas. One of these is the Accelerating Medicines Partnership which is usually a one-to-one collaboration with the NIH and private sector partners and the NFNIH to develop and advance a project specifically in an area. This is very much focused on larger-scale initiatives, and one of our flagship programs that we recently launched in the last two years is our Accelerating Medicines Partnership in Schizophrenia, which is looking to define trajectories and develop algorithms in clinical high risk for psychosis.

I will speak more specifically today about the Biomarkers Consortium and what we have ongoing in this space. We have two projects that have relevance in the mental health space. One is focused on inflammatory markers of major depressive disorder where we’re looking at developing a biosignature of MDD as well as Alzheimer’s disease and the differences between those.

Another project that we have ongoing is an SV2A PET project focused on synaptic density. This is focused on Alzheimer’s disease but has a bio portion to look into schizophrenia as well.

The Biomarkers Consortium is really focused on developing specific projects. It is more developed with private sector partners as opposed to being a full partnership with the NIH and so a little bit different in its scope but definitely an area of active work within the FNIH partnership models.

So the Biomarkers Consortium is looking at developing meaningful measurements. We are looking to lead and create cross-sector efforts to validate and qualify these biomarkers or other drug development tools to accelerate better decision-making. That involves really engaging with regulators, engaging with drug developers to ensure that what we are developing has the biggest impact, that the validation that’s developed supports the final tool utility so we can get these over the finish line and used by patients for finding solutions.

We have a lot of tools that we have generated within our consortium and we have been doing this for over 15 years. Many therapeutics have been advanced by the tools that we generated here -- nine clinical tools currently being used in drug development, five FDA guidance documents supported by the work of the Biomarkers Consortium, and at this point we have one clinical safety biomarker that has been qualified.

We have worked with the FDA developing the Biomarker Evidence Criteria, which I will talk about a little more today. We have many publications and many member organizations that joined as part of our steering committees within our consortium.

One thing I will note is our work within autism which developed two LOIs that have been accepted in the Biomarker Qualification Program. One is on eye-tracking for socio-communicative deficits, and one is on an EEG marker in these areas looking at developing and advancing those as stratification tools. Those were the two first psychiatric biomarkers accepted into the Biomarker Qualification Program.

This is giving you an idea of how it’s governed. Benefits of projects within our consortium involve the ability to work within our steering committees, which have very focused efforts within the specific disease area and a lot of expertise in driving these forward. Embedded throughout our programs are engagement from the NIH as well as the FDA to really ensure that what we develop here has long-term utility in being impactful for drug development and finding therapies.

One other piece here is that we are not just interested in the drug development aspect of this. One aspect of it having the Centers of Medicare and Medicaid services on our executive committee is to get feedback on payer perspectives around these tools, as we all know that one of the biggest hurdles will be not just making sure these tools work but making sure that patients and individuals have access to these tools in the end that could then find them and match to the right therapies.

I want to start at a high level, just giving a quick introduction to some of the definitions. We find it really important to be very specific in the language that we use so that we can make sure everyone is on the same page, so I want to go over a few definitions here around biomarker, clinical outcome assessment and surrogate endpoint.

Biomarker is really looking at an objective measure of a process, whether that be biological in basis, pathogenic or response to an exposure or to a therapeutic intervention. A clinical outcome assessment is reflecting how a patient feels, functions or survives. A surrogate endpoint is used in clinical trials as a substitute for a direct measure of how a patient feels, functions or survives, so it is a surrogate of that major point.

It is important to use these tools in assessing safety and efficacy of drug development, but they differ very significantly so you want to be careful not to equivocate some of these different terms and make sure that whatever you are developing specifically fits into one of these categories so that you can get the right engagement from regulatory and drug developers.

There are three different ways to integrate biomarkers or clinical outcome assessments into drug development, ways to advance these through what we consider the final qualification. Each drug company will be submitting an IND for their specific drug development pathway, and they will need to support that with what evidence they have specifically on what biomarkers they will include in that. And so a high threshold of evidence will need to be provided to ensure that what they are using will be able to detect a signal and be helpful for advancing their therapeutic.

Another pathway that’s more open is to go through the biomarker qualification program within the FDA so it’s a formal pathway, which I will get to in a little bit, but it allows for open collaboration with the FDA and drug developers and private sector to really engage on what the evidence is and stake a claim on what we would like to do in this space and make sure that whatever that is is really clear and everyone feels strongly that they can prove and show that it works in that specific way.

There is another way to get something qualified which is scientific community consensus, so, over the years a bolus of information is collected, information is categorized in a larger consensus sense, and the community can come to a consensus on what is a qualified biomarker. There are a lot of examples where we didn’t go through a formal process to get something qualified, but it does in the end show that consensus shows that this is working.

So, why do we want to develop biomarkers? One of the important parts that I think was mentioned earlier is around improving diagnosis. As we know, there are challenges with the current diagnostic categories in being able to understand what they truly mean and being able to subtype individuals in specific ways that are outside of the current paradigms of diagnostic codes. So we want to find different ways of understanding and detecting these earlier and to find more effective treatments by doing so.

Another aspect of this is better monitoring of disease, so being able to help clinicians plan accordingly and understand whether things are progressing or when things are expected to progress.

Another aspect is personalized medicine, being able to identify specific subgroups that might respond to certain treatments differently. And more efficient clinical trials. It’s a burden to be within a clinical trial and it’s important that we show the likely benefit of a particular treatment and show that in a cost-effective and efficient way that reduces burden for patients.

Early interventions are increasingly important as we understand biological processes start earlier than we actually detect symptoms. And improved drug development. Knowing whether a drug is safe, knowing when it’s efficacious is really important, and these are helpful indicators of how this would actually work.

In the end, what we would like is qualified biomarkers, and to do so we would want to have a qualified biomarker go through the qualification process. What this allows for is evaluation of the usefulness of a biomarker for a specific context of use in drug development, ensuring that that would be useful for patient care and providing an endorsement of that tool to support decision-making for individuals.

One thing to note is that while in the short term it is important for drug development, it’s usually in this space that some of the more novel techniques will take place that will enable more and better therapeutics and diagnostics in the future.

The Biomarker Qualification Program has a few steps here, just to give you an idea of how this works with the FDA. There’s a letter of intent that is submitted that proposes the context of use and its use within drug development, a Qualification Plan which generates the necessary supportive data to qualify a biomarker, and a Full Qualification Package which will allow you to support your qualification with that proposed context of use with all the data that has been generated so far. In the end, the FDA would provide a qualification recommendation on whether this is qualified for that proposed context of use based on a review of that entire package.

There are multiple different steps to conducting qualification of a biomarker. Understanding what the need statement is, what the unmet medical need is, what the drug development need is or what the knowledge gap is. Defining specifically what your biomarker is. Is it going to be a stratification biomarker for inclusion in the clinical trial? Will it be detecting a pharmacodynamic response? What question is it specifically addressing?

So that's the aspect of drug development but think about it also to the patients. What is the benefit and risk ratio here? Does it show improved sensitivity, selectivity? Does it have a mechanism that is novel in a novel area for it to be understood that would be helpful for understanding a drug’s effect.

And then the risk. What are the consequences of a false positive or consequences of a false negative?

Once you have developed this kind of frame set you then work within your specific qualification tool to say what is the evidence that I need to show the relationship between this biomarker and a clinical outcome.  What rationale is there on the use of this biomarker? What type of data and study design would support this most well? Do I need to perform a study that’s prospective? Do I need to have some sort of therapeutic intervention that would then detect a change in this specific measure? And what independent datasets are available for use within a qualification?

These are all the basis of thinking through how you would show evidence that would support the final utility of this tool.

Another piece is really statistical methods as we think about stratification and the fact that this will be a multimodal approach to think through what statistical methods you are doing, and to train and then test your models is also really important to show detailed underlying algorithm development and process to be very clear about what the underlying underpinnings of that model might be.

I’m going to talk a little bit about what might be helpful around how to develop a context of use and how to know when you want to go through a qualification process.

When you want to go through a qualification process you want to look at what the feasibility of this biomarker is for your specific context of use, and so you want to look at the data supporting the causal connections of this disease as well as the clinical outcomes to really ensure that this is a prepared aspect.

And carefully select your context of use. Consider the unmet need, what would resolve a gap that currently exists. Keep that context of use very focused and narrow. It might feel like you want to develop a context of use that would be for all diseases, but you might want to use a model disease area that is of known interest, to first subtype in that and then show later on some gain in other areas as well.

You want to understand the industry perspective, so we want to find tools that can actually open up further therapeutic interventions, and so getting that insight on how a biomarker would be used in a drug development or clinical trial context, collaborating and understanding what that utility would be and what they are looking for in order to make this usable. And then ask if the context of use is feasible with the evidence to date.

What I will talk about next is a little bit on how to develop and select your context of use using the BEST criteria. The BEST criteria was developed by the NIH and the FDA -- Biomarkers, Endpoints and Other Tools classification for a range of different biomarker types. Listed here are a few different examples of biomarkers that are of interest, and these are some of the categories that you will need to think through. Whether this will be a susceptibility or risk biomarker, whether this will be diagnostic, prognostic, monitoring. Whether this is a predictive biomarker, pharmacodynamic/response biomarker, which also includes some surrogate endpoints, as well as safety biomarkers.

It's important to read through some of the descriptions of these, considering what tool you are developing and kind of mapping onto these definitions to see which is the most applicable to how you would like to develop your specific tool.

Biomarker approach will involve a validation effort that will be kind of testing out what these biomarkers would do, looking at what measurements you are going to utilize, getting an understanding of the analytical validation, how well does it detect or do what you think it’s doing. What kind of parameters like test/retest -- do you need to make sure the specific tool is working in a way that can be usable and trusted?

And then looking through what your clinical validation looks like. What are the outcomes of interest? If you have a gold standard, that’s amazing. If you don’t have a gold standard, how are you going to anchor to known clinical measurements to show that this does have internal validity around the specific measure you’re targeting?

And this comes from the Evidentiary Framework Guidance that the FDA issued. It’s a great resource to dig into these.

Just to summarize some of these ideas, evaluate that unmet need, define that context of use using that BEST criteria. You want to engage with industry and private sector interests, ensuring that you also detect what is meaningful to patients, and working with nonprofits and patient advocacy groups to define that better.

You want to contact that rigorous analytical and clinical validation to ensure it’s reliable and has clinical relevance. And important for any effort that would involve the FDA or the regulators, is to ensure that you foster public-private sector partnerships so that collaborations are more useful for everyone and that the data will be usable by the broader community and that what successes you have can be carried forward.

There are also opportunities for early engagement. You have an idea, you have a biomarker, you have some ideas of how you would like to use this. There are always pre-LOI submission meetings that can be organized through the Critical Path Institute or within our Biomarkers Consortium which can then assist with bringing forward a qualification process and increase the likelihood of acceptance of that biomarker.

I want to highlight one quick initiative that we are supporting and have the fantastic support of many private sectors, both nonprofit and industry. We are happy to have the Wellcome Trust as well as the NIH really leading the way on this partnership around AMP schizophrenia. What is listed here are the ways that this project will enable an advancement within a new area, a diagnostic area of clinical high risk for psychosis, looking at deep phenotyping from different modalities, looking at what clinical outcome measurements would be most helpful and indicative of changes or predicting changes over time.

And then defining what you have as clinical endpoints. In this case, convergence to psychosis, this would be kind of a primary clinical endpoint that you are very interested in, but then defining what those trajectories are, whether that’s someone that’s in remission to the CHR state or not converting or non-remission. And so being able to subtype and stratify individuals into these categories would then enable different types of drug development outputs, whether that be clinical outcome assessments, multimodal biomarker signatures or, as I mentioned, future endpoints that can be useful in drug development.

Another piece of this is standardization, being able to do large initiatives like the ones that many are doing in the Wellcome Trust and with the NIMH which allow us to standardize against multiple different sites and make sure that those measurements are standard across, as well as ensuring that when you include an international component that you take into consideration the different cultural differences that may be driving -- or changes that may be different depending on those contexts that could be driving or changing those biomarker signatures.

That is all I have to present. Thank you so much for your time. I am happy to take any questions. Otherwise, I think I am going to pass it now over to Arthur Caye.

ARTHUR CAYE: Thanks, Dr. Horton. Good morning, everyone. I am going to talk a little bit about a specific project we call the IDEA. It’s Identifying Depression Early in Adolescence, and I think it gathers some of the things Dr. Horton was speaking about, stratification. I am not going to talk about the process but more about what we have done in this project which was funded by both the NIH and the Wellcome Trust.

I want to begin the discussing with why do we want to predict things in the first place? I think two words come beforehand. The first one is we want to predict relevant outcomes in medicine and mental health, and we want to predict outcomes that will change what we will do next and change the prognosis. For instance, if you have a sunny day, you can bring sunglasses or otherwise you would bring an umbrella. That’s a very simple concept but I think it translates to the idea of how can we prevent outcomes better.

So we have three ways of targeting prevention. We can do universal prevention and target everyone in our intervention. We have selective and indicated in the promotion of health.

In the case of depression we understand from past randomized clinical trials that the universal distractions were not effective and it has been recommended that we should not pursue classroom-based prevention programs for those who are high risk.

So we have (audio drop) one is indicated, so those that have minimal but detectable signs or symptoms that foreshadow a future mental disorder. I think this has been the main approach, for instance, for schizophrenia in high-risk individuals that have sub-threshold symptoms. Or we can use predictive models to find those who are at risk even when they don’t have any symptoms.

The definition of risk is very hard in mental health, as you know, how to identify the needle in the hay. In the case of depression and many other disorders I think the most relevantly used marker of risk is family history, but there’s a lot of limitations with that. The first is that we are only limited to those individuals who have a family history in the first place, and the second is that we know that family risk gives us very low relative risk in general -- two, three to five maybe in the case of depression.

Why did we choose depression? Talking about relevant outcome, we know that depression is responsible for most of the disability-adjusted life years, so the years lost to disability through a lifetime among mental disorders. We understand that it’s a very relevant condition that begins very early in life and that is why we chose to tackle depression in adolescents in the IDEA project.

We know that most of these DALYs are lost in low and middle-income countries, and that is another discussion that we are going to have afterwards. Most of the predictive models that we have are set, developed and validated in high-income countries.

We have these examples from other areas of medicine like cardiovascular disease, for instance. This is the Framingham Risk Score in which you don’t consider just one risk factor at a time, but you can combine those risk factors to compose a dimensional risk. A dimensional risk score has been one of the ways that, in cardiovascular disease, we could set trials to understand, for instance, that the treatment of hypertension or dyslipidemia only works and is only cost-effective with a positive cost-benefit ratio in very high-risk individuals. So we thought that this was a model that we could pursue in mental health as well and to take then a dimensional approach instead of a categorical approach.

In the development of a risk score for depression we went to cohorts. This is a Brazilian cohort that we have that already had data collected from birth to early adulthood, and we used the available data both in the development here in Brazil but also in other cohorts around the world to validate this data. I think this is one useful message for those that are wanting to develop predictive models, that we have lots of data that are available in different countries for us to propose prospective analysis that could deliver effective predictive models.

We selected predictors and those predictors include biological sex, ethnicity, maltreatment, school failure, social isolation and the relationship between parents and the child, running away from home and fights. Those are all predictors, and this is on purpose, predictors that are easily collectible. They are sources of demographics and cheap and easy to collect and that we could collect with the adolescent himself or herself, for instance, in schools because we wanted this score to be replicable across low and middle-income countries, settings with low resources, and because we didn’t have lots of evidence to support other kinds of predictors.

What we learned by looking at the performance of each of those isolated predictors is that the score, the IDEA risk score that combined all of those together, was better than using each of those alone, which validated our intention to use the multivariable model.

Of course, we have to understand that sometimes prediction has always uncertainty and so we have a prediction that many times is going to work and many times is not going to work, so there is a lot of implications of that, and ethical implications, discussed with the patients. Even when we predict depression with high certainty, it might not happen, and the implications of that.

We published this first risk score generated in Brazil with an area under the curve of 77 percent. This is similar to what we have, for instance, with the Framingham and other scores in general medicine and in mental health. But one of the challenges that we have in this field is that we don’t have -- it’s very rare that we have valid external replication. In child and adolescent mental health, for instance, only 5 percent of the proposed models have been validated elsewhere, and so replication and replication and replication is like in real estate we have location, location, location. In the prediction model we must do external replication for us to be confident that we can apply those models outside our original study.

What we have been doing after we developed in Brazil, we went to several other cohorts around the world and applied the same model without any adjustments to new unseen data in the United Kingdom, New Zealand, Nepal, Nigeria, USA and the second cohort in Brazil. Sometimes we didn’t have the same predictors available. You can see that we went from 11 predictors in the original sample to seven, to 10 in the different cohorts, so we had to of course run the model again and this generated loss of performance. But one thing we found very interesting is that we could replicate that the model was significantly better than chance in all the settings that we tested across the five continents, seven samples. So we were very impressed by how we lost performance but still were beating chance even when we developed the model in Brazil in a very specific context.

What we wanted to do next was to use this cohort to identify those at high and low risk and gain insights about biomarkers using the separation between low and high risk. So we wanted to examine these individuals and understand more about them. So we began with one movie, Back to the Future, and we went to the cohorts and looked at the data collected many years ago to predict an outcome years later.

But then we wanted to change to the Avatar movie using this for collecting data of high and low-risk individuals in Brazil. So we went to schools and surveyed more than 7,000 adolescents at those schools using the variables from the score and then separated the groups into high risk for depression, low risk for depression and those that already had depression in the schools. We compared the probabilities calculated for depression.

This is using data on adolescents 15 years of age, predicting depression three years later. We saw that, of course, females had a higher risk for depression and we know that depression is more frequent among girls in adolescence, but the density of the predictions were similar, although in Porto Alegre, the new sample, they were higher because we had a different case mix in this new sample.

The frequency of those predictors was similar across as well, but we had more frequent presence of predictors in the new sample. Also, we compared the structure between -- this is a network analysis comparing the structure between the predictors and it was also similar between the samples.

If you look on the X-axis you have the PHQ, the indices for depression. These are the low-risk individuals. They had low symptoms of depression and low probability as calculated by the score. You have the high-risk group, those that have low scores of depression but already had a high risk of depression at baseline, and those that were already depressed.

Then with these three groups we did very thorough clinical assessments. We collected blood samples with genome-wide gene expression, cytokines and inflammatory markers. We had saliva samples and neuroimaging, both functional and structural.

There are a lot of things in this. The low-risk group and high-risk group were similar in terms of clinical issues. For instance, the CDRS, which is -- and MFQ -- are two scales of depression. Irritability, trauma. They had more trauma of course. Their IQ was similar.

And now what we are doing is to analyze those groups in terms of their brain features, both structural and functional. Their inflammatory markers as well. The IDEA Flame, as we call it. We are using Checkbox to collect real-time data on a daily basis and comparing some expressions of text between those groups.

The Chrono IDEA, we are measuring how they sleep and how much they move. And we are now finishing the third wave of data collection.

So the regional score was predicting depression in three years. We have reached that now, and to our surprise and happiness as well, this is an analysis of how many of those at high risk converted depression three years later to the low risk, and the score worked and we have a risk ratio of around four between those groups.

This is our website which we are developing for further information and some interactive data. This is our financial support, some national and international institutions but it is all public funding.

These are some pictures of the team. Thanks a lot for your attention. I will be happy to take questions.

USMAN HAMDANI: Thank you, Arthur. We will have a question-and-answer session at the end. Before I go with my talk, I would like to thank Wesley and Arthur both. The purpose of these two talks were to introduce to the audience by examples and the details of what do we mean by stratification, and Wesley very elaborately pointed out what are the various stages for biomarker validation.

We know that there are so many studies going on in mental health research and we are working on many different biomarkers, but we know that they haven’t actually made it to the clinical settings as yet. So the purpose of the talk by Wesley was to actually help you imagine the process and know the resources where you can go.

Similarly, the talk that was presented by Dr. Arthur Caye, actually is a very nice talk in the sense that it actually tells you how. So thank you, Arthur.

Moving on to my talk, I will be sharing my screen, and the unique part about this webinar is that we are not just going to be talking about the subject, but there is a funding opportunity that Wellcome Trust is leading and it is open. So you can actually look at these ideas, look at your work, and you can put together an application and get funded and help us take this validation of markers in global mental health forward.

This webinar was recorded by us a week ago, and Professor Pim Cuijpers and Dr. Mark van Ommeren from WHO were the panelists, along with our Wellcome Trust team. I will be just presenting the part that is about the funding goal so you can know what are we looking for in our funding goal and put together the application. The due date is the 6th of June, so it’s about in a month’s time. And you are more than welcome to visit the Wellcome Trust website, look at the full webinar where Dr. Pim Cuijpers and Mark van Ommeren presented the details on how they look at stratification and what are different approaches and methodologies.

I will just go onto share my screen and share my audio.

… enlightening in the way that Dr. Pim Cuijpers has very elaborately summarized the field of stratification right from the beginning of where we started, what has progressed, what are different methodological approaches and what are the strengths and limitations of different methodologies.

And then, very accurately, Mark has brought in a policy and a practice perspective that highlighted that we need valid measures which should be biological but also psychological, social, digital, which are valid, reliable, open access. And I think these are all, most importantly, hypothesis driven. We really need those sorts of measures that are based on empirical grounds and can actually help us identify what is the right treatment for the right person at the right time.

This is exactly what is the crux of this funding call, and this is why actually we requested Dr. Pim and Mark to join hands with us to bring methodological expertise and the policy and practice perspective so that you can actually frame an application, because this funding call is led by our mental health translational team, so we have a translational focus.

We will be looking into applications which can actually help us move validation of markers for stratification in mental health along the development line. We emphasize that there needs to be robust pilot data to support your marker validation. There needs to be an empirical hypothesis that you need to define. Markers could be diagnostic, prognostic, predictive, mechanistic, treatment response markers, and all the range of markers, but these are the most common attributes.

We have a 20-minute recorded webinar that we are going to play now that will actually help you understand the whole call, and we are very happy to respond to your individual questions, so you can type them in the Q&A. We will be typing the answers and we will also be responding to them live after this webinar.

Thank you, Mark, thank you, Pim, and thank you, Lynsey, for the recorded part.

(Video presentation follows)

CHRIS CHRISTOFI: We are the management team at the Wellcome Trust within the mental health team. We would like to welcome you to our informational webinar on our mental health award, Finding the Right Treatment for the Right People at the Right Time.

In terms of the agenda, we will kick off with a round of introductions. This will be followed by Wellcome’s Mental Health strategy, then we will go to what do we mean by lived experience, then introduce the funding call, collaborating with people with lived experience in your research, and then we will conclude with how to apply for this funding opportunity.

First I would like to hand it over to Margaret to introduce herself.

MARGARET OSOLO ODHIAMBO: Hi. My name is Margaret Odhiambo, lived experience advisor to Nairobi.

CATHERINE SMITH: I am Catherine Smith, projects officer in the mental health translation team in London.

MARY: (Inaudible)lived experience Nairobi.

LYNSEY BILSLAND: I am Lynsey Bilsland. I am head of the mental health translation team.

USMAN HAMDANI: I am Usmani Hamdani. I am research lead in the mental translation team.

KATE MARTIN: My name is Kate Martin, I am head of lived experience in Wellcome’s mental health team.

CHRIS CHRISTOFI: Thank you, everyone. I would like to kick off with Wellcome’s mental health strategy and I will pass it to Lynsey.

LYNSEY BILSLAND: Thank you. The Wellcome strategy, we fund discovery research into life, health and wellbeing, and we also support research into solutions to the three of the largest health challenges facing humanity: infectious disease, climate and health and mental health. And we promote diversity and inclusion and research culture in all the work that we fund and we do. Last year we realized that we have $16 billion to spend over the next ten years in order to advance our strategy.

We will then focus in more detail on the mental health vision and mission. Our vision is a world in which no one is held back by mental health problems, and our mission is to drive a step change in early intervention in anxiety, depression and psychosis.

We acknowledge the challenges that exist around the diagnostic categories in mental health, and so we use the terms anxiety, depression and psychosis in a very broad sense so it will include any anxiety or depressive disorder including obsessive compulsive disorder and post-traumatic stress disorder. And also all forms of psychotic disorder, so including schizophrenia, postpartum psychosis and bipolar disorder.

In order to achieve our vision and mission we are looking to build a more integrated field of mental health science that convenes different expertise, and this includes lived experience. We are going to fund a broad range of research to really advance understanding of how biological, psychological and social mechanisms interact and result in the development and resolution of mental health conditions.

And then we are also looking to drive the development of better ways to intervene early, and this includes both improving our ability to group people, so to identify subgroups of people with different mental health conditions or risk of developing these conditions. This will hopefully allow us to develop more targeted and personalized interventions.

On the right side it shows some of the open opportunities on our website. We’ve got two current contract opportunities open, and then in addition to the call that you will hear about very shortly we also have another funding call that will open later this summer focusing on anxiety.

We can fund people in teams doing research across a number of levels, so from the subcellular through to the social and across a wide range of disciplines relevant to mental health. This might include biologists, psychiatrists, chemists and economists, ethicists and historians. We can fund basic clinical research for potential new and improved early mental health interventions, and these can be pharmaceutical and non-pharmaceutical including digital, and for use both in healthcare and non-healthcare settings.

If you are studying anxiety and depression, we ask that you use our currently agreed core measures in your research unless there are exceptional reasons why this will not be possible. This is in addition to other measures that you may wish to use.

Just to flag that while the mental health team focuses on anxiety, depression and psychosis, we also have a discovery research team that accepts applications on any other mental health condition as well as anxiety, depression and psychosis. They run three open funding schemes that run three times a year and cover a number of different career stages as well, so please do have a look at our website for information on those.

I am going to hand it over to Margaret to talk about lived experience.

MARGARET OSOLO ODHIAMBO: Thank you. Lived experience is a unique form of knowledge, insight and expertise that comes from having experiences of mental health challenges. It can either be diagnosed or not diagnosed.

In terms of Lived experience, is central to all our work at Wellcome and it helps in shaping the day-to-day work, thinking, direction and decision-making of the mental health team. Projects we develop or fund need to have meaningful involvement with people with lived experience expertise. We also work to integrate lived experience expertise in the field of mental health science.

I will pass it to Usman.

USMAN HAMDANI: Thank you, Margaret. In the next few slides I will walk you through the details of this funding call. This is a mental health award on finding the right treatment for the right people at the right time for anxiety and depression. Our strategy slides identified that we focus on three conditions: anxiety, depression and psychosis, but for this mental health award we are focused on only two conditions: anxiety and depression.

Before we go into the details of the call I would like to highlight a few aspects related to the broader remit of this call, which is about whether we describe it as stratified psychiatry or precision psychiatry or personalized mental health. All the stakeholders related to the field of mental health, including funders, clinicians, mental health scientists, end users including people with lived experience of mental health problems are becoming more concerned to ensure that the right people get the right treatment at the right time for their mental health problems.

How can we do that? We can do that by leveraging the progress made in recent decades in different scientific disciplines and applying them to validate objective markers in mental health in the same way as the objective markers for risk, diagnosis, prognosis, prediction and understanding mechanisms in other fields of health sciences such as cardiology and oncology.

We also understand that the field of mental health science is complex and multidimensional. There are not just biological but also psychological and social determinants, and we need to apply robust methodologies including advances in digital technologies to validate potential biological, psychological, social and digital markers that might inform meaningful identification of different subgroups of individuals with mental health problems.

This is the aim of this call: to support validation of biological, psychological, social or digital markers to enable stratification in anxiety and/or depression as early as possible.

Now, there could be different methods that can be used to measure the unique characteristics of the subgroups using different markers and these could be biological such as genetics, biochemical imaging, clinical scores or social, such as sociodemographic characteristics or behavioral such as psychological assessments. The stratification will allow us early targeted treatment and ensure that the right people get the right treatment at the right time.

Over to Margaret.

MARGARET OSOLO ODHIAMBO: In the context of this call, we hope to prioritize engagement with people with lived experience of mental health problems to help in validating these markers for early identification and intervention of anxiety and depression and to make sure the research and its findings are applicable and also applicable to the end users.

USMAN HAMDANI: Thanks, Margaret. So why do we need stratification in mental health research? We know that current mental health categories are imperfect, they rely on subjective measures resulting in significant heterogeneity of people within each diagnostic category, which in term impacts the development and provision of effective early interventions.

Stratified medicine aims to identify subgroups of individuals within a heterogenous disease population based on the unique characteristics of each subgroup such as underlying mechanisms risk factors, course of disease or treatment responses.

As we discussed, there could be a number of different markers that can be utilized to measure the unique characteristics of the subgroups. These are including but not limited to biological markers, psychological markers, social markers and digital markers. And we all know that there are a number of markers that are already reported in literature in terms of biological markers such as genetics, biochemical, imaging, clinical scores or psychological assessments that can be used for stratification. Or social markers such as the social demographic characteristics, environmental characteristics and so on and so forth.

And last but not least, digital markers and making sue of the progress that has been made in the field of digital science such as ecological momentary assessments, GPS tracking and so on and so forth.

So this is not a prescriptive list. Please work on the markers that you believe are the markers that can actually bring a change in the mental health science field. These are just meant to help you think about what are we asking for.

What we expect is that the use of stratification methods will enable early identification and targeted treatment in mental health. And we will be able to identify subgroups of individuals that will benefit most from a targeted pharmacological or non-pharmacological treatment and that will enable early intervention with the potential to alter trajectories of these conditions and have maximum impact on people’s lives.

We also believe that stratification can be used to improve understanding of disease pathophysiology, identify new targets for treatments, develop objective markers for disease risk, diagnosis, prognosis, response to treatment, and allow treatments to be developed, tested and applied to the most appropriate patient groups.

With this, I will hand it over to Lynsey to describe the details of the scheme.

LYNSEY BILSLAND: Thanks, Usman. At a glance, in the scheme the projects should focus on early identification and targeted intervention in anxiety and depression. As you mentioned earlier, these are broadly defined and include OCD, PTSD and bipolar disorder.

In terms of funding, we can fund academics and companies up to a total of five million pounds, and up to a duration of five years. However, if you need anything outside these limits please do get in touch with us before you apply.

There is no recommended amount or duration to apply for. We recommend that you apply for what you need and just be realistic.

We can fund globally except for mainland China and sanctioned territories. And if research is occurring in more than one location in the world we ask that a co-applicant is based in each of the countries in which the research will be taking place.

We are also looking to encourage equitable partnerships between high, middle and low-income countries, as per this call.

We ask that the lead applicant has appropriate and necessary expertise in order to drive and lead a research program and that it has been enforced, and we want evidence of this to be included in the application form.

And we are looking for applications from teams ideally, and these can be multidisciplinary teams and across diverse settings and across different career stages as well.

So what are we looking for? I am not going to take you through all of these, but we are looking for applications that focus on the analytical or clinical validation of markers, and they have to be for use in stratification for anxiety and depression.

The markers, as Usman said earlier, they can be biological, psychological, social, or digital, and they can be used either alone or in combination with observable or behavioral characteristics.; But they have to be able to enable stratification in anxiety and depression, and that might be according to risk and susceptibility, or diagnosis or prognosis, prediction of treatment response, or monitoring of disease progress.

We are looking for proposals to embed lived experience expertise, and this is at multiple stages in the proposal, so across planning, design, and the delivery of the research. We're looking for markers that are scalable, ideally, and that can be applied in a number of different settings and, ideally, through this call, would be tested in a number of different settings, as this will ensure generalizability of findings.

And then we would also like to consider the uptake of markers and from the outset, and there are some ideas around how you might do this on our website.

This next slide just shows the assessment criteria for the full applications. For full applications, we'll review the application against four weighted criteria. The first is the research question, the proposed methodology, and the potential for impact. That's a 40 percent weighting. The second criteria is the suitability and the expertise of the team, 20 percent weighting. The third criteria is lived experience involvement, again a 20 percent weighting. And the final criteria is around suitability of research location, research environment, and research culture. And that's 20 percent.

At the shortlisting stage, when we are shortlisting preliminary applications, we will use a simplified rubric.

I'm now going to hand over to Veronica for more on lived experience.

VERONICA: All applications will need to demonstrate clear plans for collaboration of lived experience at multiple levels, which we'll discuss next. It's particularly crucial to this call that we do that because that ensures that the findings and research will be meaningful and relevant to end users and reflect the priorities of people with lived experience.

Lived experience experts should be paid appropriately for the work that they are doing. We expect you to prove that you have budgeted and costed for collaboration lived experience throughout your proposal.It's important that you remember that as a research community we are all still learning about the best approaches to this kind of work, so we don't expect any sort of perfection at this stage, but Wellcome we're really interested in learning with other researchers and other organizations, and to that end we'll be having workshops with the lived experience team that we can share what we think are some of the best practices when it comes to this work.

And I'll hand over to Kate.

KATE MARTIN: Brilliant, thank you. Just a bit more about what we're looking for in proposals and applications. As we said, we don't have a set method or approach to collaborating with lived experience experts that we expect. The main thing is we really want this, the methods you're choosing to collaborate with lived experience experts, to be relevant, the most appropriate, the most influential, for your particular research and to really fit in with your research designs.

What we are looking out for when we're reviewing applications are that people with lived experience, lived experience experts, are involved in multiple stages and in multiple levels of the research.So looking at the governance, the oversight of the research, the planning, the design, the delivery, the dissemination, et cetera. And what we know is that for each of those stages you'll need to think about different methods, different models, approaches, to make sure that it's the most meaningful, the most impactful.

So we're looking, first off, are people involved in multiple stages? And then secondly, we're looking at the range of roles that people might take on. So are the range of other roles appropriate for the kind of research that you're doing?

Again, these are just sort of a list of examples, this is not exhaustive, and we don't expect all of these roles in every project, but just to give an example that some people might take on a role of expert advisors, some could be co-researchers, some may sit on advisory groups, et cetera. There may well be, of course, many other roles that people can take on. But to summarize, we're looking at people to be involved in the most meaningful ways at multiple stages of the research and taking on multiple different roles that are relevant, to have the most impact.

Over to Catherine.

CATHERINE SMITH: Thanks, Kate. Just now a little bit of information about how to apply for this funding opportunity. Before you do apply and write your application, we strongly recommend that you consult the additional guidance on our website. We know it is quite detailed and lengthy, but I hope it will give you a good sense of what we're looking for in applications, as well as what is considered out of scope.

For example, and I'm not going to go into this in detail but do consult the website on this. If you do not consider lived experience in your applications or if you're focusing solely on things like implementation science or epidemiology, it may be considered out of scope. So please do check that section in the website if you are under any doubt. We also have additional information on our policies and guidance on involving people with lived experience which you might find helpful.

I want to highlight in particular some material in the useful documents section. This is at the bottom of the webpage. Things like the MRC framework for the development, design, and analysis of stratified medicine research. So as I say, we do recommend you check the website for all of these resources and additional information on what we do and don't cover and the costs, et cetera.

Next, we ask that you submit your preliminary application on Grant Tracker. Again, you can find this on the webpage, by 5 p.m. British summer time on Wednesday the 7th of June, 2023.

Just to give you an overview of our timelines, after the submission of preliminary applications on the 7th of June, we will complete a shortlisting in July 2023. For those that are invited to submit a full application, the deadline will be 7 September 2023, with interviews in November, which will take place online for shortlisted applicants.

And just a little note on contacting us. You don't need to contact us before you write and submit your application, and if you do have a question on how to complete the application form, please contact our grants information advisors, and if you have broader questions about eligibility, what we offer, or our funding remit more broadly, please contact the mental health team at

Just a quick note that because we're seeking to be equitable in our response to potential applicants, we're unable to discuss the specific content of applications, but please do find our email address if you have broader questions relating to the form or eligibility.

We also share information on LinkedIn and Twitter so please do check us out there. With that, I'll thank you all for listening, and we look forward to receiving your applications and reading them in due course.

USMAN HAMDANI: Thank you so much for listening to it. I hope the context of the two talks. Over to you, Leo. Thank you.

LEONARDO CUBILLOS: Thank you so much, Usman, for this session. I would like to invite Arthur, Wesley, and Usman to turn on their cameras.

This was certainly a delight to hear what the Foundations of NIH are doing, very biological approach. Very interesting, Arthur, what you were doing in to observe how Brazil as a middle-income country is actually doing research on stratification, research that is top notch, but that can also -- and is also applicable to the reality of its own health systems. So that is very interesting.

Thank you, Usman, for chairing the official recording of the call.

It's now in my time 11:08.  We have 22 minutes to the end of the webinar.  Why don't we spend at least 15 minutes doing some Q&A. Without further ado, I will read from the top. Have a question for Arthur. This can be a brief answer, from Alex Conway.

ARTHUR CAYE: The universal interventions were not effective or they were low quality of evidence. There are two papers interesting for you to read further. One of those is in BMJ in 2012, there's a discussion, it's one trial which failed, and then there's discussion and editorial which I think it's interesting to read. And then there's Cochrane Review, I think it's 2016, in which they review the evidence for this school-based interventions, and they separate between universal strategies, targeting everyone, or high-risk group when universal strategies were very marginal effects.

LEONARDO CUBILLOS: If you feel comfortable, Arthur -- thank you for your answer -- if you could share the link to those papers in the chat box for the Q&A so that everyone can see them.

Another question for you from another participant, and sorry if I'm mispronouncing names or last names. For you, Arthur, how predictable are these mental health illnesses, in your opinion?

ARTHUR CAYE: I am not sure I understood the question precisely, but in the case of depression, there's a lot of ways in measure prediction and how well your prediction is good or not good. One of those is the area under the curve which I presented, and we could with these 11 measures have a good prediction. A good prediction is not a perfect or excellent prediction. It's considered a good prediction when you have it between .7 and .8. Being .5 in this measure is equal to chance, is flipping a coin. 1.0 would be a perfect prediction.

We have 0.8, which is much better than chance but not perfect. But there's a lot of ways to measure predictions, there's calibration, a lot of different ways of measuring how well you can predict mental health outcomes. There are other scores that I know for schizophrenia and other mental health outcomes. I developed for ADHD, as well the persistence of ADHD. So it depends on the model and the disorder we are talking about.

LEONARDO CUBILLOS: And another question for you before I turn it over to Wesley and Usman.From Melanie Abas, do you think the markers you found for the IDEA high risk group might link with predicting better or worse response to treatment? And if so, how?

ARTHUR CAYE: I think that is a very interesting question. We didn't test that yet. And I think I would guess the obvious that those with the high-risk group would have a worse response to treatment, but that is not necessarily true. And we might look at that in some NIH-funded trials that are available, with data available; I don't know if they have the exact same measures of risk, but maybe they have. This is an interesting research question that could be pursued with existing data. I would be happy to try that.

LEONARDO CUBILLOS: Thank you, Arthur.

Usman, questions for you. From (NAME), there's a question. Is there an expectation that we do a multi-country study? There is a similar question from Kovat Konidanchat(ph.). Can an early career clinician-researcher apply to establish our research group with international collaboration?

USMAN HAMDANI: Thanks for the questions. We do not require a study to be multi-country, but we do require that the markers that you select have a translational potential. We do promote collaborations between high-income countries and low/middle-income countries, and obviously, the study design and settings that you choose should be suited to your research question. Definitely this is one of the criteria that we are looking at, that the markers should have a translational potential in low-resource settings as well. So that's what we have to say about it.

In terms of an early career researcher who will be applying to this funding call to establish a research group, our focus is on supporting research that will focus on validation of markers. How that supports you, how do you use that to build your network, collaborators, partners, I think that is entirely up to you. You can put it in the plans in terms of the translational potential and the sustainability plans, but that's not something that the intent of the call is, but obviously you can use the resources and the opportunity to establish yourself in the area.

LEONARDO CUBILLOS: Thank you, Usman. There is another question, if this call is for clinical research only.

USMAN HAMDANI: We do need robust pilot data to support certification, and because this is the translational nature of the call, so we are not focused on the basic science research, but it is translational research that we're really focused on. So definitely the focus is on translational research rather than the basic research.

LEONARDO CUBILLOS: Thank you so much. Does this funding call include the use of a mechanistic biomarker to test the efficacy of the therapy targeting the mechanism?

USMAN HAMDANI: I can't comment on the very specific details, but yes, validation of mechanistic markers is within the scope of the call.

LEONARDO CUBILLOS: Is there an option to submit letters of support with the preliminary application?

USMAN HAMDANI: Yes. So there is a -- when you go onto the Grant Tracker, you can fill in the application, and then there are other support forms that you can upload things in there that you think are relevant. For example, your proof of concept data, your letters of support, your prior data that is not published that you're using to build the hypothesis or support the hypothesis. So, yes, once you go on to the Grant Tracker, you'll find a place where you can upload additional information, which doesn't go into the specific sections that we have for the application.

LEONARDO CUBILLOS: This is kind of a longer question. I am going to read it. Could you please quickly summarize what is required for the preliminary stage in terms of co-applicant and financial estimation? For example, it says, quote, the lead applicant is responsible for inviting all other participants to participate in their application, end of quotes, and that I am supposed to invite them through the system to ask them to log on, fill some information, and confirm their participation. So far I have not found how to do so.

USMAN HAMDANI: You can go onto the Grant Tracker and you can start to fill in the preliminary application. It's very straightforward, and it's really small. There are just four pages. You will have to put in the names of the partners and collaborators, and if you have any questions or you feel any difficulty, please feel free to write to the email address that's given, and we'll be more than happy to respond to any specific queries or troubles that you may have.

LEONARDO CUBILLOS: This is a two-part question, or actually two questions put in one. I would like to ask if the funding opportunity can be for woman, sic, depressed during pregnancy, by using the score of EPDS. First part. And do we need to cooperate with researchers from the UK?

USMAN HAMDANI: As we said in the webinar, we are open to fund any research across the globe except mainland China and sanctioned territories. So, yes, the application can come from anywhere, and you can definitely focus on validation of markers in postnatally depressed women. So that's a target population, because that is within the depression category. So, yes.

LEONARDO CUBILLOS: Does Wellcome Trust have examples of previous grants, like it's written previous grants that won, previously awarded grants that can serve as an avatar?

USMAN HAMDANI: We do not have such examples online available, but maybe the two talks that Wesley and Arthur gave, they can give you the idea of the scope of the project.

You can also go on our website and look at this full webinar. I just played the part specific to the funding call. If you go and play it full webinar, Professor Pim Cuijpers has actually cited many papers, which are very relevant to this funding call and he has really laid down some really important kind of information there which might help you put together the application.

So we do not have a sample completed application for the funding call, but you can benefit from these webinars and the information that's available. Please also look at the MRC stratification guidance, which is there on the webpage as well. That does cite some projects, and you can also find many of the projects submitted which are ongoing and are related to stratification in mental health research, and I'm sure you'll find them helpful in thinking through the details of this call.

LEONARDO CUBILLOS: So we can find the webinar that you presented with WHO, following the link that you provided to us at the beginning of the webinar.

USMAN HAMDANI: Yes, so that's the funding page link, and on the funding page you can find the link to the webinars and support documents and everything in there.

LEONARDO CUBILLOS: The study population, another question for you, are we supposed to have only most vulnerable groups or along the lifespan development, lifespan meaning adolescence, adults, per the question.

USMAN HAMDANI: Our focus is on early intervention, and the definition of early intervention will vary depending upon the condition, the context, and the settings. So you can suit yourself according to the study population you're looking at, the country that you're looking at, and the study condition that you are focused on, but it has to be an early intervention and that needs to be clearly articulated how this is an early intervention.

LEONARDO CUBILLOS: Wesley, a question for you. Is the Foundation of NIH accepting applications and applicants coming from low- and middle-income countries?

WESLEY HORTON: Yes, we do on a case-by-case basis have discussions on what could be possible for collaborations within our consortium. So happy to have any discussion. Feel free to reach out to me if you would like to propose maybe collaborating in a space. It may map onto an existing initiative. It might be something that --

LEONARDO CUBILLOS: Usman, there was a question about can you confirm that small businesses are eligible to apply to this call?

USMAN HAMDANI: If you go on our website and look at the grant conditions, any organization that is willing to or can sign up to our grant conditions is eligible to apply. So just look at our grant conditions and if you think your organization can sign up to them, yes, you can then apply.

LEONARDO CUBILLOS: Question for Wesley, or Usman potentially. It's for any of you actually. Could you give more examples of what is meant by a digital marker?

WESLEY HORTON: I guess I can jump in there. There are already digital markers that are used explicitly, heart rate, that are used more widely. I think what's more interesting is remote digital monitoring than just digital. So the ability to be remotely capturing information from someone's day-to-day that wouldn't be able to capture if you were having to have them come in on a monthly basis.

So those are the innovative aspects I think of digital measurements that I think we're most interested is the remote monitoring, and Usman, you probably have thoughts directly related to what your specific call is here.

USMAN HAMDANI: There can be many digital markers, as you've identified, but there could be also ecological momentary assessments that when used to predict like, for example, a lot of data is collected and then that can help you predict when someone is going to have a remission or a response or relapse of an illness, and similarly, remote sensing as you've said can be used for a number of things. GPS tracking can help you look at the activity, and similarly, the gadgets, use of gadgets, interactions; similarly, speech, natural language processing, speech markers, there's a lot of research that is going on on speech markers for stratification.

So the list is quite exhaustive, and if you look into the PubMed, you'll find many, many good papers which are on the subject.

LEONARDO CUBILLOS: Usman, would you consider -- thanks for sharing the link again to the webinar that you presented earlier for everyone. The link is now in the chat box. Usman, would you consider involvement of clinicians working directly with patients' a lived experience component?

USMAN HAMDANI: So clinicians working with patients with mental health problems do not constitute lived experience involvement. Lived experience involvement means people who had the experience of mental health symptoms, firsthand or they could be like carers, for example, a sibling, a parent, a child, or people who have actually experience of managing the mental health problems themselves or as part of their family or social network. So the clinicians won't constitute lived experience involvement in mental health research as part of this call.

LEONARDO CUBILLOS: Thank you, Usman. Arthur, question for you. What have been the most challenging aspects of doing research in stratification in a middle-income country, and what have been the most rewarding aspects of doing such research in a middle income country?

ARTHUR CAYE: There's one thing that we think we are very thankful for to NIH and Wellcome Trust, because as you might know, there have been some discussions about this in Science and Nature that we had politically very hard times in Brazil in terms of national funding. We begin with national funding that disappeared after some time. We were very lucky to have been funded by certain institutions.

So I think one problem is funding. The other problem is learning that much of the data that -- of course you always want to learn about what people have done in the past in terms of risk factors and analysis in our previous efforts, and we learn that we should begin at some point, we should begin from scratch, because some things do not translate to our settings. For instance, being an ethnic minority or not is very different in the UK, India, Nepal, and Brazil. So I think one of the things is that we kind of, we're more lost in translation in these terms in middle income countries, because we -- it's a whole new thing in some ways.

The most rewarding thing is learning that we can do competitive research and meaningful research that translates -- so the other side, so it's very interesting for us to have developed something here that can be used in the United States or in the UK, of course with adaptations, and I think for us it's a double reward when we do that.

I think that's it.

LEONARDO CUBILLOS: I appreciate your answers. Certainly it is challenging but it is an open canvas full of opportunities. Definitely within a country, high income regions, compared with low-income regions, and in the world higher income countries compared with middle and low income countries. The knowledge and learning goes in so many directions if we keep our minds open, of course.

Wesley, a question for you. How do you envision that this science can transform mental health?

WESLEY HORTON: I think it is important to realize that there are new paradigms that need to be created that are centered on patients and things that are meaningful to them. We have measurements that can detect the later stages, but the sensitivity and the ability to detect the earliest stages and move therapeutics and solutions to patients, giving them options, just the paradigm hasn't changed yet, and we really do need to find those solutions that can get to the earliest stages and monitor those earliest progressions.

So the ability to find those measurements, to validate those in a way that is trusted by the community and trusted by regulators and by industry, I think we will find a solution -- those solutions will enable a paradigm change for therapeutic interventions.

LEONARDO CUBILLOS: Thank you so much. Usman, over to you before we begin to close this webinar. Any additional thoughts or any closing remarks from Wellcome Trust?

USMAN HAMDANI: No, I think we have tried our best to elaborate on the different aspects of stratification through these webinars with experts. Please have a look at the scheme page. I have shared the link in response to some questions. Have a look at the webinar that's there. Have a look at the support documents that we have made available so that you can understand the scope of the call. Please do look at the eligibility criteria and the assessment criteria. It's important that you have some pilot data, and your marker that you select is hypothesis-driven and there is an empirical basis of what you are selecting and proposing to validate.

We have provided in the links, you can write to us, you can seek clarifications of questions that you may have, and last but not the least, NIMH is hosting the Global Mental Health Conference, and there is a track on stratification in mental health, and the website is live.So over to you, Leo, about the NIMH Global Mental Health Conference and the abstract submission that's there. So that's still another opportunity to be part of this effort to promote stratification in global mental health.

I would like to thank Arthur and Wesley and Leo for your time and to all the participants for the very engaging discussion and for listening to all the talks and participating. Thank you.

LEONARDO CUBILLOS: Thank you, Usman. The Global Mental Health Conference will take place between October 30 and November 1, 2023. The website is As Usman mentioned, registrations are open and so are submission of abstracts for the panels, presentations, and posters as well.

Our next webinar will be in July 26. We will have NIMH's Office of Clinical Research walking us through important elements of clinical trials that are considered important for applications submitted to NIMH. We will provide more details of this July 26 webinar.

As I said in the beginning, September of this year, we will have a similar webinar on precision psychiatry global stratification and the panelists will be our colleagues from NIMH.

So thank you all for your time, and we have Wellcome Trust and the NIMH team and the people that are working behind the scenes to put together this webinar series. Thank you so much for your time and we wish you a great rest of your day, wherever you are.

This webinar will be uploaded entirely in the coming days.

Thank you so much.