Clara Faria: Welcome to the Mind the Kids podcast. I'm Clara Faria an ACAMH Young Person Ambassador, and in today's episode, I have the pleasure to talk to Professor Amit Baumel from the University of Haifa. Professor Baumel is an associate professor in the Department of Community Mental Health at the university. His research focuses on ways to increase engagement with digital mental health interventions and on the development of tools to evaluate the quality of those interventions. Although he has an emphasis on digital parent training programs for children's mental health, Professor Baumel has also developed interventions supporting women with perinatal depression, people following first episode psychosis and more. He is also the winner of the 2024 ACAMH Digital Innovation Award. If you are fan of our In Conversation series, please subscribe on your preferred streaming platform. Let us know how we did with a rating or review and share with friends and colleagues. So welcome Professor Baumel Thank you so much for being here today. Can you start with an introduction, giving a short overview of what you do and why you decided to research digital mental health interventions?
Prof Amit Baumel: Thank you for having me. It's a pleasure to be here. I'm a clinical psychologist and former tech entrepreneur. And currently, as you mentioned, I'm an associate professor at the University of Haifa Israel, where I direct the Digital Intervention Psychology Lab. And my main line of research involves these two passions that I have, clinical psychology and computer programs. And I'm very interested in learning how can we integrate knowledge from psychological science and health computer interaction design into building programs that are doing a good job at translating usage into clinical change, meaning helping people get therapeutic impact into their life when they use these kind of applications. I decided to research digital mental health interventions first and foremost because I very much like these two fields and I liked using computers as a teenager playing games and so on. And I also like treating people. There was also a profound experience that led me to dive into this field. I remember as a clinical psychology intern working with teenagers who needed much more than one weekly session of psychotherapy. And this was not doable because we were limited in the amount of sessions that we could give for each person. And some of them eventually were hospitalised and this was a very, very difficult experience for me as a clinician. It's a very difficult experience for any, to any clinician. And at the same time, I do remember feeling like a broken record, repeating myself, repeating the same therapeutic process and instructions with other people who have mild to moderate problems. And I thought to myself that it would be cool if we would have these digital programs that could help people with mild to moderate problems on things that are much more, I would say, protocol based, more simple to help them with. And then potentially we could invest most resources into people with severe problems where I thought that professional help is much more needed. And I thought that we got very good training as clinical psychologist. I got good training. I just, wasn't sure how well my professional decisions are coming to play in each of the treatments that I provide.
Clara Faria: That's a fascinating story. And it's really interesting you mentioned therapeutic impact because that leads to my next question, your paper, the impact of therapeutic persuasiveness on engagement and outcomes in unguided interventions shine the light on a fundamental concept to improve engagement and efficacy of digital mental health interventions, which is, as you've said in your previous answer, therapeutic impact and therapeutic persuasiveness and it's a very long concept and I wanted to ask you if you can explain this concept to our audience and walk us through the rationale behind your paper.
Prof Amit Baumel : Now, many digital products aim to optimize usage, right? I mean digital products aim that user will use them for long duration of time, sometimes to make more money out of these users. Therapeutic persuasiveness asks something a bit different. It asks how well the digital program is designed to help users make positive changes in their life. So it's not focused on at least directly maximizing program usage. It is more focused on how we can effectively translate use into beneficial outcomes. And it includes several design features that works towards this goal. And these design features, they come from an overlap, I would say, in conceptualisation between behavioural change techniques in the psychological world and persuasive design in human-computer interactions. For example, we want to have good call to actions in the program which means that we make the goals that people want to achieve salient in their life. And we want to give people monitoring to understand, to monitor their current state, to understand where they're at and provide feedback based on this and to adapt the program. So to give a concrete example, if we take a parent training program, like the one that was mentioned in the, that was tested actually in the paper, we'd like to ensure that parents have positive interactions with their child before moving on to teaching them skills to address child non-compliance. This is when it comes to a program that tries to treat child behaviour problems, and well, parents might receive an e-learning module that teach them about positive interactions, and this is great. But from a therapeutic persuasiveness perspective, learning is not enough. Parents often struggle to foster positive interactions before because they have competing activities or distractions. They have job concerns they have smartphones in their hand, it might be easier just to play with a smartphone than to play with their child. So having a TP equality means, for example, that we want to help parents have the desired activity, positive interaction, salient in their mind. So just before they come back to home from work, they're going to get automated text message with TP, reminder, a motivational note about positive interactions during this day. So they remember that this is how they are supposed to be doing. The system also strives to document what parents are doing each day. So parents are prompted to complete very, very short questionnaire, less than one minute to complete about, for example, in this case, positive interactions that they did this day. And we ask it in a script manner in a way that tries to help parents then embed this sort of thinking into their lives. So for example, one of the question would be, on your way home from work, did you think about which positive interactions you should have with your child? And then parents get to learn that these questions will be asked every evening. So they prepare and become more accountable for this act. So we help them change their behaviours. And being therapeutically persuasive also means that we are data-driven. We don't ask questions just to ask, just for the sake of it. So the reports, the parents' reports are used to acknowledge their success, suggest additional actions, advance them to the next phase, or try to solve problems that we identified during this process. So from a behavioural change perspective, simply building an e-learning course is not enough to help people translate learning into meaningful change. The idea is to use all available features that we can use in this platform to make relevant activities salient, create accountability with well-designed questions, solve problems, and adapt the program based on user progress.
Clara Faria: So it sounds like therapeutic persuasiveness comprises a group of design features that are especially important for people, parents who not necessarily have the time or resources to proactively engage with the digital mental health intervention. And I wanted to ask you to what extent these features are important for parents?

Prof Amit Baumel: Yeah, well, I think an important term that is not discussed enough in digital interventions is effort optimization. And it refers to the idea that people don't have much time or capacity to invest in digital interventions. As a matter of fact, people want to exert a lot of effort into their mental health treatment. They would prefer psychotherapy. But actually, those people want to self-manage their states for many different reasons they would want to just take a nap and use it, but they are not very much willing into investing a lot of effort. So we need to think how we effort-optimized the intervention. So it better fits parents who generally have only small pockets of available time. So for example, when we suggest very bite-sized therapeutics, like engage for something for five minutes in a day for 30 days, people are much more inclined to do this then if we would tell them, take this course online, which then will have a very poor retention patterns.
Clara Faria: Going back to your paper a little bit and also touching base on the fact you just mentioned that, know, sometimes people have limited amounts of time to engage with this kind of interventions. In the paper, you mentioned user engagement in the discussion as one of the main challenges in unguided mental health interventions. And I wanted to ask you, what other challenges do you consider are hindering the impact of the digital mental health interventions in the real world?
Prof Amit Baumel: One key challenge is identifying and deploying the right distribution channel at the right time to offer the interventions. So for example, if a parent is concerned about their child, it can take a lot of time for them to eventually go ahead and seek support, like meeting a general practitioner to discuss the problem and say, I think I have a problem at home. Now, often by the time a teacher raises an issue about the child or a general practitioner refers the parent to an intervention, it's already late in the process of illness development. At that point, it just doesn't make sense to suggest people to use digital interventions. And we see very low desirability for these kinds of solutions all across the world. When a general practitioner or someone, you you come to a person, you come and meet and say, I have a problem. You went to meet a person. So right now you need persons, you need people it would not make sense at that point of view to refer them to digital interventions and we see very low uptake. So we need to find better opportunities to identify who needs a problem, maybe much earlier during the illness development phase, the prodromal phase of illness, and to introduce these kinds of interventions. And one solution could be nationwide screening procedures. So what I mean by that is that we can identify different opportunities where the whole population goes and meet someone and at that point to try to screen if there is an issue and then suggest the intervention so like we have. I don't know how it how it goes in the UK, for example, but in Israel when children they go and start the first grade all the children are going for for an eye exam to see if they need glasses which makes sense right I mean this is the time where we want to check it and I do have collaborators in Finland that I work with that have a very successful screening procedure for the treatment of child behaviour problems. So all parents take their children in Finland, all parents take their children when they are at the age of four, to an annual medical checkup. So all these parents, get this screening procedure where they need to answer very simple questions about the child behaviour problems. And those that are screen positive, are then being offered with this national program, digital program for parents that helps them address their child behaviour problems. And we have very high enrolment rates for this kind of program. And the idea that we have high enrolment rates is not because we did not wait for parents to come and say, have a problem. We came to parent proactively, we met them, we offer them to complete a screening procedure, and then those who came out positively screened were referred to get the service. So I think this is one key issue that we need to address and we don't have enough like these kind of examples all over the world. We need more examples like that. I think the complementary challenge is the market. So we need to ensure that providers of digital mental health solutions can be reimbursed for the services implemented in a sensible way. So as an example, if the only way to get reimbursed is for general practitioner who had to place a code within the system. And otherwise, the person who provided the mental health intervention can get reimbursed for providing the services. We created the bottleneck. We actually reinforced the use of a GP as the person who actually refers to services. And we know that this actually does not work. So we need to come up with these kind of plans that eventually enables the government also to pay for these services because if there is a market failure and nobody is going to pay for the service, then nobody is going to offer it as well.
Clara Faria: Yeah, those are really fascinating points you mentioned, especially about the uptake, because you mentioned that usually when the app is offered at the point where the young person or their family are already at a primary care service, the reception is usually bad. And that's also the experience we have here and also, usually if you talk to a general practitioner about what do you think about a digital mental health app, most of them are wary or have reservations. I think it's because precisely of what you said, because at that moment, the person is already at a point where she needs urgent referral and we need to find better opportunities to kind of implement those. It's an implementation issue.
Prof Amit Baumel: Definitely. I think the rates are something like 20 % acceptance rates. So one out of every five people that is being offered with intervention from a practitioner is going to accept it, something like that. And I think this is because it failed when they opened the door. It was just not the right time.
Clara Faria: Yeah, I think you have a very valid point there. And also, I don't know how the primary care system works in Israel, but in the UK, we have the added challenge of the time constraints. So usually GP consultations, they are very short, they last 15 minutes at most. So lots of GPs are already overworked you know, and then when you come to them and say, look, we have this amazing app for you to offer to your patient, usually they're not very, very happy about it. But yeah, really good challenges you mentioned. So talking a little bit about scaling those interventions, the potential to scale on guided mental health intervention seems very promising if we're able to overcome these barriers. I think, especially considering one of the things you mentioned in your first answer about people wanting more psychotherapy sessions. And you know, it doesn't matter. You can be in the most well-staffed healthcare system in the world. There is a limit of number of sessions we can provide as psychologists and mental health practitioners. So are there any additional ways we should think of digital interventions for them to support people?
Prof Amit Baumel: I think to realize the full potential digital interventions, we need to be very creative in how we use them across different levels. From a conceptual standpoint, there is a difference between an app that helps someone support themselves than an app that helps someone support someone else. In parenting program, for example, we help the parents help themselves, but we also help the parents change their child behaviour. And I think there are many ways we can do more in terms of using technology to foster therapeutic connections. In one of my studies, for example, we use 7 Cups of Tea, which is a platform that trains volunteers from all over the world to provide emotional support. And we were able to train these volunteers to effectively support women with perinatal depression. And the women were very happy with the support they received. We also had another project with people who experience first episodes psychosis. Now the idea with around women with perinatal depression is that many of them often experience loneliness and lack of social support. So to me, it made much more sense to develop technology that trains people to support these women rather than just giving someone experiencing loneliness a self-help app. It just didn't make sense. But we can go even farther in this direction. For example, new mothers who suffer from postpartum depression, for instance, they could benefit from volunteers who cook food for them. And this way they get both social and practical support during a time when taking care of a baby is so exhausting on the concrete level, especially while experiencing depression. So this kind of support could be organized for social media actually easily open and so you could build these kind of groups you know community groups with based on locations where you have somebody some volunteers that leads you know the other volunteers and then whenever someone says you know there is this is a new mother she she experienced some emotional problems then they get the community group to go ahead and cook for these women now we do have these kind of community groups in Israel not for women with postpartum depression for any new mother because it's very difficult for newborns, right? It's exhausting at the first months to actually create some normal routine. So it's really helpful. But I do think that if we are able to actively build these kind of groups for women who experience mental health problems and then focus these resources on them all through the world. And this is an example of how we use the technology actually to solve the environmental issue here which is the lack of social support. And I think we can do much more in this kind of reality.
Clara Faria: So it's fascinating that in your answer, you just reversed the logic and you were actually posing, you know, technology as a force for good. And do you think, you know, you gave the example of support groups online just to help mothers with or a postnataldepression, do you think that that could be scalable to children and young people?
Prof Amit Baumel: Yeah, definitely. So there are programs even as of today that first of all, if we talk about text based messaging and these kind of stuff, there are programs that can actually help people write the right text. So when someone gives an answer, they actually suggest to them what to exactly say or how to change the answer in a way that it would be more therapeutic. And of course, these kind of platforms are also very good at alerting and alerting for malicious content or if somebody is not doing something correctly. But from a more general perspective, I think when we discuss task shifting a lot. So what is task shifting? Task shifting is the idea that we teach someone who is less professional to do something that before that only professional did. But we can think about task shifting in many layers. And we do want to have mental health literacy around the community. We do want people to know what to do if, for example, they hear that someone around them said that they want to commit suicide. And I think that teaching all these kind of instances does not necessarily mean that we pass the responsibility to these people. It just means that we added another layer of support. And if we think about us human beings, you know, 5,000 years ago, and I'm not an historian. I don't know history too much, but we can like fantasize how it looks. know, people they sat in tribes, they met each other and they learn how to support each other in very natural way. Today, we have these kind of close relationships within our family, but outside of our family, we need to build it in different ways and we don't live with the community 24 seven. So I think that the use of technology to train people based on what we know as of today, how to support one another, how to help one another is not going to reduce the need in professional work. It's just going to help people get maybe skills that once before they had in a very natural way. And they think that part of it is being very clear at what the person is responsible for, what they can do and what they shouldn't do. And based on at least my studies, I did not find when we build these kind of top-down projects where we design them ourselves, I did not find one time that we have malicious attempts or things that we should have done differently because of safety issues.
Clara Faria: Interesting. And Professor Baumel going back to your paper where the focus is all about parent training to improve child behavioural problems, can you talk a little bit about what the science says when it comes to the impact of parental mental health on children and young people? And also vice versa, by supporting parent-child relationships, is it possible to effectively impact parent mental health?
Prof Amit Baumel: Parental mental health impacts how parents feel and the energy they have to invest in parenting and support their kids. So it is connected in both ways. And we do know today that children of parents with untreated mental health issues may experience higher stress levels, difficulty regulating emotions, and in some cases develop similar conditions themselves later on. There's another connection here. Parents...who have a child with behaviour problems, for example, naturally are not happy with what's happening at home right now. Parenting distress is often tied to feeling unequipped to foster better relationship at home. Feeling like you're a lousy parent clearly doesn't help one's own mental health. And this is where the connection between parent-child relationship and parental mental health becomes so important. If we help parents with children who have behaviour problems build better relationships with their child, it does not just reduce the negative effects of parental mental health struggles directly. It can at some level also reduce parenting distress and even improve the parents on mental health because they have a better relationship at home with their child, which is so important for them. And we can think the same, we can conceptualize it in the same way if we think about child anxiety where you feel you're such a lousy parent if you can't help your child overcome their anxiety. And it really is a cause for your own depression. But if then you feel that you are much more effective in your parenting and that you are equipped and you are doing very well in helping your child overcome their own anxiety, then you feel better about yourself. You have more meaning in your life, so you experience less mental health issues.
Clara Faria: Really interesting to see how it works both ways. And to wrap up, what future research would you like to develop or see advancing in the field of unguided mental health interventions and how to make sure those are scalable and implemented successfully?
Prof Amit Baumel: I do think first that it would be great to keep focusing on everything we discussed here today, product quality optimization, because we always have new technologies and new opportunities, and improved distribution channels. I think it's very important and also creative uses of technology for therapeutic connections. In my view, these are the basics anyway. Beyond that, I think there are two concepts I like to see more work on. The first is understanding how very small doses of interventions, such as single session intervention, for example, can create change over time. How can we make one session meaningful and guide people towards the right therapeutic vector, if you will? And it depends on so many factors, including the context of the users, the goals of the intervention, and its quality. I think it's important because we need to become more skilled at turning usage into correlatives change over time. And by the way, I don't think it's also only relates to digital interventions, also to psychotherapy by itself. The second is exploring how to use artificial intelligence to deliver highly effective interventions. And it goes way beyond the idea of chatbots. It leans towards artificial intelligence systems that learn about us and provide just-in-time feedback and suggestions based on evidence-based guidelines, they could also refer us to use different apps and different programs. And I think there's a lot of promising things to do in this realm. generally, it's a very interesting topic because technology keeps advancing all the time. So I feel there's always a lot to be done.
Clara Faria: Do you think we're there with large language models, I know you said it's beyond chatbots, obviously they can store great amount of information. And if you tell a chatbot, even a simple chatbot, if you're using GPT-4 to give an answer based on cognitive behavioral therapy principles, it will be able to do a reasonable enough job. But do you think we will get there with larger language models?
Prof Amit Baumel: Yes and no. So concretely, yes, I think that we are very close to a point where the text that the large language models create are therapeutic or precise as you get the text, the same language from a clinical psychologist who is highly trained. We are not there yet, but I think we are very close to it. And I do have a seminar in my university where I teach the mental health of tomorrow. And for the past seven years, I'm making the argument that things are really going to change very, very fast. And in the first three years, it was very difficult to convince the students about it. And then I started reading text that ChatGPT3 created. It was before everybody knew about ChatGPT. And people couldn't distinguish between the text that was being created by the original author, and the text of ChatGPT that tried to mimic this author, this was like, I think three years ago, and we made so many advancements since then and when there will be a large language model who is trained based on psychotherapy recorded sessions and based on different models that are focused on psychotherapy, I'm pretty certain that they will be able to provide text at the same quality. But then you have two other issues. The first is, well, text is not the same. It's not as salient as meeting someone, right? So the artificial intelligence could also mimic the audio, right? And it's very good at it. And it will be very good at it. And it could also mimic the video. It could appear, like in therapies. So we also saw this one. But then we remain with the last part, which is where I do feel that we... will not be able to solve with artificial intelligence. And it's the idea that I am as a human being knows that I receive support from another human being. And they know that I receive support from a human being that chose to help me, even though they might get paid for it. The idea that they made the decision to put an effort into helping me should be very meaningful or therapeutically meaningful because human beings, have... limit on the amount of energy that we are able to accept. We are biological entities. So once we decide to invest effort in something, other people who see this, they interpret it very much differently than if we do something that does not involve any effort. So any effortful activity relates to our sense of meaning. And I think we will not change that. And there is this, I always use this example with the Little Prince by Antoine de Saint-Exupery. I think I pronounce, I hope I pronounce his name correctly. So over there, is the story about the little prince who meets the fox and learns about friendship. And when he learns about friendship, he then learns about his rose, his rose that he left in his planet that he took care of and then had to fight with. And because of that, he left the planet and came to Earth. And on Earth, he saw like thousands of roses. All of them look the same as his rose on the original planet. And he came to understand when he met the fox, that these roses do not mean anything to him, unlike the other rose, even though they are all identical. And the reason is because of the time invested in the rose, because of the meaning he created when he took care of the rose that made this rose very, very special to him. And I think this idea of we feel in human relationship that it's special because of that, because the time we invest in one another and the meaning that it creates. And I just don't think we will get it from AI and this will be something that we will still want to get from other human beings.
Clara Faria: On that note, thank you so much, Professor Baumel for sharing your research and so many precious insights with us today. For more details on Professor Baumel work, can visit the ACAMH website, which is www.acamh.org You can also check Professor Baumel LinkedIn profile, which will be linked in the podcast page and our Twitter at ACHAM. ACHAM is spelled A-C-A-M-H.