Health Studies User Conference 2026: Keynote presentation

I'll introduce Sally, who is our keynote presenter. Really pleased that she's here to give the keynote today. So, Professor Sally McManus is Professor of Social Epidemiology at City, St. George's University of London, and director of the Violence and Society Centre. She's widely recognized for her work on measuring mental health and understanding how social factors, such as inequality, violence, and adversity shape well-being across the population. Sally has played a leading role in England's National Mental Health Surveys, helping to build one of the most important evidence bases on population mental health anywhere in the world. Her research studies key issues, including self-harm and suicide, with a strong focus on trends over time and inequalities. Alongside her academic work, she has a long-standing role at the National Centre of Social Research, ensuring her research has direct relevance to policy and practice. So, over to you, Sally. Really looking forward to this. >> All right. Brilliant. I um delighted to be here, but also um yeah, can everyone hear me okay? Cool. Um really totally get it, but so sad not to be in a lovely room with you all, because this is one of the loveliest events in the annual calendar where we all come together. And what um what I like, what I really like about this event, is that we come together from so many different places. And that means that as a group, we collectively have this enormous reach. So, my kind of route to here is um I started out in a market research agency back in the day. That was BMRB International, no longer called that. And then I went to a non-profit uh social research institute. Back in the day, that was called SCPR, and some of you will remember it from that, and many of you will have passed through its doors as the National Centre for Social Research. Did a little bit of time in government, so um that was then the Strategy Unit, um Cabinet Office, and then the last few years been in a university. But what I think is really important about us as a collective is that many of us have worked across many these different sectors, and many of us have also many of you will have also worked, for example, in third sector organizations, in campaign groups, and using survey data in a range of different ways. We also, although we have this common kind of understanding of a focus on health, what are the consequences from poor health, and what what leads us into poor health, we also we come at it from so many different angles. So we're looking at this in terms of families and what goes on in families, socioeconomic factors, in housing, in neighborhoods, in schools, in workplaces, violence and crime in a range of different settings. So that's something which is also makes us a very diverse group. And then finally, I think and this is a really key sort of difference about us as a as a collective is that some of us know absolutely everything about one thing, particularly if you're coming from an academic uh environment, and others, I tend to put myself in this box, like to sort of spread across wide range of different topics. So we really come at this from many many different perspectives, but what I think >> [clears throat] >> is probably the thing that actually unites all of us on a very hot and sunny day to be thinking about uh surveys is that there's something, I get quite emotional, there's something about surveys which is incredibly democratic and beautiful in a way that it gives voice to whole societies. And I don't know of any other real way where people's own voice, and yes, we do constrain the options available. We do say this or this, and occasionally there's an open text box, but basically, we're treating everybody in the similar way, giving but different voices are coming out in that. And gosh, what I would love for us to come away from this event day is that we know the value of this, and this infrastructure, this way of gathering insight from lots and lots of different people only exists because there is support, infrastructure, investment, and all of us articulating a need for it. And sometimes we're the only ones who actually understand the real value of these surveys. So, what I wanted to do today was just a little bit like we can all come away with that elevated pitch, and when you're in the lift with Andy Burnham or whoever, that you're able to say why actually in a resource constrained environment this is actually worth fighting for. Um generally, we talk about respondents. I would like us to be talking about It doesn't just have to be the quality to talk about participants. There is agency involved if we recognize and talk about the people who are taking part in these studies are participants, and when we're talking about findings, we're talking about people. We're talking about children and adults and families. So, ensuring that this has a reach in everything that we say and do and we're articulating the need for this. Now, some people out there some people out there say surveys cost a lot and they're expensive. And I think that we're the ones, nobody else, it's not their fault, they don't know, but we're the ones who actually know that surveys cost well good surveys that's the key thing. Good surveys cost a lot. But just looking at what comes out of an event like today, they are an absolute bargain in terms of the mileage, the reach, the impact that can come out of them. So, what we have to be aware of is that of course my taxpayers' money how do we reach both the funders and the participants in ways that explain the value of what we're doing. And of course they'll come back and say, "Well, look, I know that there are nearly 4 million people last year who experienced domestic abuse in England and Wales and I know that I and a third of people can't get a GP appointment when they need one and I know that one in four parents aren't able to take their families out for a day out together and have the resource to be able to do that and that a fifth of the properties uh in the private rental sector fail decent home standards. Now of course you all know that the only reason we know any of these things is because of surveys and that there aren't really other mechanisms for finding out most of these things other than by going out, talking and asking people. I was going to have a picture here of John Cleese saying "And what what have the Romans done for us?" But actually UK has a very wisely cautious about the use of images and the copyright protection around them. So, you're just going to have to imagine John Cleese in his But, what what what are the surveys done for what other than be our best and only source on understanding the extent of social phenomena? So, how many, what proportion, who is affected and who isn't affected? It's our really our best way of knowing things are going up or things are going down. It's our best way of understanding gaps, what gets missed and inequalities. Who gets missed? Who doesn't get the services? Cuz we're never going to get that from service data in isolation. And is this nuanced way that we collect so much different information on individual people that enables us to profile a really complex and rich context environments and impact. Now, I know many of you are absolutely brilliant on the longitudinal and the birth cohorts and that's where the impact is really rich in terms of understanding individual trajectories. But, I just want to also make a case for those repeated cross-sectional surveys. So, many of you will have seen this slide before. We all have our favorite survey. My favorite survey is the badly named adult psychiatric morbidity survey. Favorite survey, least favorite name, but we just call it the mental health surveys. It's funded by NHS England by Department of Health and Social Care. Has been run by ONS in the past, National Centre for Social Research now with uh Leicester throughout. Um rich rich program of research. Actually, the longest-running mental health survey series in the world to have used the same measures over time with fresh cross-sections. Other places in the world have tended to change their questions and their measures. So, this is something that we just need to nurture and make the case for because if we lose this, we're losing 30 years of of of trend history. Um So, just focusing on those adult surveys that started back in 1993. Now, you think of the UK Data Service is where you get your data. It's much more than that. It is a source of historical record. This is the questionnaire in the 1993 adult psychiatric morbidity survey. Someone with a biro has written >> [clears throat] >> all of the codes, all of the different bits. You can see it all scanned in. It's It's It's a historical document of still where this survey is going. It was paper-based at the start and and has evolved over time. We don't even have the report from the '93 survey online because there isn't a digital copy available. So, there's the link down there. You can get the 1993 survey. We use it all the time. It was from another era, but it still matters. Now, some of you will know Howard Meltzer. Um he and a whole team of people at ONS got together with the epidemiologists and the psychiatrists and the clinicians and the civil servants, and they designed this study together. But, just to say I miss Howard and he and many of us do and and he was a very uh extraordinary person. And Sarah Morris at the National Centre for Social Research is carrying on and done an absolutely superb job with the 2024 survey. So, it continues and and new generations of leadership come through. Just wanted to flag up. So, from the 1993 survey, the highest income threshold was 30K or more. So, I think many of us are now in the highest income threshold group. And just the the the the the quality of um uh uh photocopying has definitely gone up uh over the years. But what I would just wanted to flag is that what's been key is that the way in which we've introduced the survey, everything many things have changed, but the way in which the survey is introduced is so similar. It's about the stresses and strains, it's about health and well-being as a whole, and that's really stayed consistent over time. We do have laptops now, so there have been have been changes. Okay. So, repeated cross-sectional, comparable methods, random fresh sample each time, right up to age 16, so we can look at that adolescent transition. Some in the data set over 100, though to you they're 95+. That's our disclosure. Um detailed interviews in-home, face-to-face, human being to a human being, verified, not an AI bot. We know that they're a genuine person there. And there's something about that rapport. Talking to interviewers, they know about the emotional investment it takes to be part of a study like this. They sit down at a table, and sometimes there's tears, and sometimes there's cups of tea, and there's laughter, and there's something incredibly human. It's an intervention in and of itself. Two-phase design and assessment of range of different mental disorders to a more detailed level than on many studies. And it's about that specificity that I was going to focus a bit on today. And this is perhaps another reason why face-to-face is also particularly important for a study of this nature, um particularly which is covering conditions such as uh autism and psychotic disorder as well. Now, what I just wanted to do, and what I may not be speed up a that. Is four short stories that just tell us things that we couldn't get from, I think, any source other than a survey. So, looking at the APMS data from '93 through to 2024, what are we seeing? We're seeing that broad pattern. Now, we use the clinical interview schedule revised, which is over 130 items. It's really, really detailed. 14 different types of neurotic symptoms, very complex algorithm. This pattern we're seeing from other data sets as well, which use simpler measures. But, >> [snorts] >> with a much more specific approach, we can start to see quite different patterns as well underlying those wider trends in mental health. For example, generalized anxiety disorder, very rare around the turn of the century, and has increased substantially over time. Another way in which the survey looks at mental health in a much more specific way is asking both about things like suicidal thoughts, where we are seeing an increase over time, but not a dramatic one, compared with the change we're seeing over time in terms of people's reporting of non-suicidal self-harm. So, just to compare those two trends, if we're looking at suicidality and self-harm, it's really important that we're we're really quite sort of precise and clear as to what statistics that we're looking at. And that means we need to have space for real specificity of measurement. What this chart shows is amongst people who reported that they'd made a suicide attempt, what sort of support did they get afterwards? Now, we see that women tend to be more likely than men to get support afterwards from whatever source they're going to, informal or formal. But, just to focus there on hospital emergency service A&E. If studies are dependent on looking at who presents to A&E after having self-harmed, A will be underestimating men in particular, but B will be not capturing 3/4 to 2/3 of those who have made a suicide attempt. So, it's really important that we look more widely. And it's very important where we're using health service data, but we need to use the survey data to put that in context. Consumption. So, lots of studies ask about substance consumption. This very, very messy chart has got lots of different age groups, and this is those reporting 16 or more on the audit. But if we just pick out two age groups, oh, my age group, we're a right mess, we are. Um 2% were had patterns of alcohol that were not good for you. So, we're we're talking here 16 plus, so this is uh, you know, high risk of alcohol harm, possible dependence. So, about 2% at turn of the century, we've doubled. So, you know, how are we going to age later on? There's going to be stuff going on. But do we not need to worry about the kids? So, um 16-24 year olds, sorry if you're there and I called you a kid, but you guys are doing great on the alcohol. Uh we're seeing that plummeting. But what do we know about drugs? So, for some reason, we take our drug data from crime surveys. Now, crime surveys are A victimization surveys. They ask about what illegal things have been done to you. So, they're not necessarily the best place to find out what illegal things you've been doing. Whereas in the context of a study that's about health and well-being, it's a bit less punitive, and we tend to find higher reporting. So, both in general, there's slightly different measures included, but also when we're looking at those specific individual drugs, we tend to find on the health survey slightly higher reporting than we get on the crime survey. But, whatever whatever measure we use, we tend to find the proportion in the population um just, you know, for the last sort of decade, not huge amount of change in prevalence of drug use. So, is that okay, then? Well, how do we explain then that the proportion of the population as a whole with signs of dependence, particularly on cannabis, has almost doubled this century, this decade only? So, something is happening there. Now, when we ask about dependence, we're asking about things like daily use, sense of dependence, if you're finding you're struggling with abstaining, you know, you're needing to use more, or you're experiencing withdrawal symptoms. So, just coming back to this chart here, proportion of people using stayed the same, proportion with signs of dependence is doubling. Now, actually, this is consistent with what we're seeing in police seizure of cannabis. Police seizure of cannabis is finding that cannabis is not the same drug. It's far more potent than it was in the past. It's far more likely to be cut with impurities, including the synthetic cannabinoids, and we've got changes in the ways in which people take cannabis. So, it's easier to take it at more regular periods of time with vaping devices and the like. So, we shouldn't just stop at consumption. We need to also be asking about things like dependence, and not many studies do that. So, it's really important that we keep the studies that do. We've also added in new items on the most recent study, and if you're interested in the opioid crisis in America, I'd highly recommend this book. It's an absolute fabulous read. Um but what we're finding is about 3% of adults reporting use of prescription opioids that were not prescribed for them. This is something we need to monitor. And unless we ask everybody in the population, we're not going to get a tick from The Americans only knew about the opioid crisis that was happening when people were dying and it was through their investigation of death statistics. Let's get in there before we see patterns like that. I don't think we are going to see patterns like that for many reasons, but we haven't got time to get into the map, but it's another reason. Just to flag up. If you ask people with signs of drug dependence, have you ever used drug dependence services? Most never have. We see the same thing for drinking. We see the same thing for uh problem uh gambling. And very unlikely to be found uh the majority of people or if you're looking at service data. And where they are using services, it tends to be things like self-exclusion blocking software. So, not the sort of thing that you would be picking up these people uh if you were looking at uh clinical data. Am I doing okay for time? Oh, I'll speed up a little bit. This is on on on mental health on the margin. So, another thing to say is that we need studies that both look at general distress, so GHQ and uh short form and things like that. But what we also need to be looking at are other sorts of conditions, those that are sometimes called SMI, so severe mental illness. It's a problematic name because actually uh major depressive episode is a severe illness. But what the survey data enables us to capture is those extraordinary socioeconomic inequalities in context to people with very particular conditions. And being able to look across and the patterns do vary by condition. So, what this shows is the proportion of people who using a detailed examination carried out in a clinical setting living in the most deprived areas compared with the least deprived areas. And we see an extraordinary and very strong pattern of association. Also, a survey like this, which I think is far braver than many, many in terms of asking about all rich range of exposure to adversities, we see that amongst people who've ever been in prison more than one in 10 tested positive for a psychotic disorder in the past year, which really tells us something about the nature of need amongst the prison population. And we also see that amongst those with who are currently claiming well, currently. This was a report that was done for DWP using data from 10 years ago. Claimants on employment support allowance, a very large proportion of those had psychotic disorder, which is an extremely impairing condition. Eight in 10 people with a psychotic disorder are in receipt of treatment much more so than for any of the other conditions that we looked at, but that means that one in five of them weren't in contact with mental health treatment. And that's particularly medication or therapy. And also we found that one in three didn't report that they thought they had psychosis. Now, there'll be many reasons for that and partly that's because it is quite possibly the most stigmatized of conditions which could fit reporting. But finally, I just wanted to flag up some really important much more nuanced work that Paul Bebbington led with this survey program. And what he really did was he used the survey program to move the field beyond genetic explanations to articulating a social contribution to the onset of psychosis. In particular, his work using the data in the surveys on victimization and child sexual abuse and the role that that played as well. So finally, around risk factors and context and violence and abuse in particular, which is an area that I've been looking at. Just wanted to talk through this paper which was done by absolutely brilliant guy at King's Fish Babsa. So, he's a psychiatrist and and a researcher and and that. And what he did was he used the APMS data to show to describe people who use violence. So, often we look at victims of violence, but actually the data also allows us to talk about people who use violence. And people who use violence against others do have worse mental health. When you adjust for their childhood experiences of violence, that explained so much of that association. So, it really was about showing the importance of early intervention can reduce violence potentially later. This is obviously a cross-sectional study. But what I thought was the most important thing and really hopeful thing about his analysis was that he looked, "Okay, they've got poor mental health. Let's compare them with other people with poor mental health." And he found that those people were no less likely People who use violence are no less likely to be using mental health services, which means mental health services is this amazingly non-punitive point of contact with people who are using violence and who may need help not to be using violence. And I thought that was a very powerful finding from the study. Now, there've been loads of analysis. Shaun Arum, Youth Thing, lots of people have been doing brilliant stuff with the data on on violence and abuse. And we've changed the questions in the later survey on that. So, there'll be there'll be more work. And if you're interested in some of that, we're just coming to the end of a five-year glorious five-year consortium grant from UKRI. There's a link there if you're interested. And it's a program of work where we've been pulling together data from police and crime surveys and the homicide reviews, health care and health surveys, data from specialist sexual domestic and violence services, and other sources in order to really explore and understand both the measurement and the interventions around violence and abuse. The data can be used if you're interested in the whole range of different types of conditions and change in the prevalence of different types of conditions, and can be used to explore demographics, age, ethnicity, sexual orientation, and other other comparisons as well as a range of different adversities. So, there's been some lovely papers focusing, for example, Natalie Chollerton on homeless people or on those who've been in prison or a range of of others. And sorry if I've not given you a shout-out, but there are so many. Um If you're interested in working with the data, I would first go and have a look at the report which Sarah Morris was absolutely amazing at bringing together Marion and others. Um And that's available on the NHS England website. You can just go via Google. Um UK Data Service all all of the surveys all of the surveys are available there. So you can go right back to 1993. The latest um 2024 data set is is just just coming in. So um that will be available soon. And then finally I think I've got a just a just wanted to say It's not a given that the survey series will go on. We haven't been commissioned to do another APMS yet. I just do that the full bit. We haven't been commissioned to do another APMS yet. Um it's been 10 years since the last mental health of children and young people survey. That's not happened. Um we have to use every opportunity that we get into that lift with Andy. Don't worry about Kier now, but I don't know where's who's taking over from where's there's a new one. Um whoever we need to and and this is where it comes back to that reach because we span we span government, we span market research, social research, academia, we span think tanks and a whole range of different places. Let's make that case for maintaining the infrastructure. We need UK Data Service and yeah. And do get in touch if you've got any questions if you're working with the data. Um also with Sarah or Marie or others we're all really keen to support people to work with the data.