[MUSIC] EBELE OKOLI: Welcome to The Ability Summit 2022 Session, supporting the biggest growth area in disability demographic: Mental Health. My name's Ebele Okoli. I delve in all things Mental Health and Disability Intersectionality for Microsoft Accessibility Team.
I have the pleasure of introducing all of you to the session presenters. Each of whom will walk us through the significance and impact of mental health globally, how we Microsoft, support our employees, and how we pursue ethical research in this space. You will hear from Theresa Nguyen, the Chief Program Officer for Mental Health America, Kristen Roby Dimlow, Corporate CVP for Microsoft Total Rewards Performance and HR Business Insights, and Mary Zirinsky, the lead of the Human Understanding and Empathy Group within Microsoft Research.
Without further ado, I would like to hand over to our first presenter. Theresa, over to you. THERESA NGUYEN: Thank you, Ebele. I'm going to go ahead and start my slides.
A little bit about MHA and myself, MHA is one of the first mental health advocacy organizations in the country. We were founded in 1909, but what I especially love about Mental Health America is that, our work is really rooted in the lived experience of people with mental health conditions. I myself have struggled with complex chronic mental health conditions as early as seven. I remember the first time I had thoughts that were very difficult to control, and certainly by the sixth grade, it was really impacting my schooling, and like so many people, did not get help until I was in my twenties, and found Mental Health America as a social worker, and as an advocate and a person living with mental health conditions, where it's a community of people who really embrace what it means to live with a mental health condition, but also see the strengths that we have, that we bring to the table, and how we understand the context of our mental illness in America today. Next slide. As part of my work here at MHA, I run an online screening program.
Historically, MHA has been fighting for ubiquitous mental health screens as just the first step for individuals to get access to mental health care. In 2014, I launched a website to allow young people, primarily, to come online and take a mental health test, and they might be looking something up on the Internet like, "am I stressed or depressed?" "Depression test." "What is depression?" They come to our site. Here are all the screens that we do offer for community, for free. Ninety percent of our users come organically via the Internet, and take a test for depression, anxiety, bipolar disorder, or even psychosis. Next slide. A lot of the data that I'm going to share with you today is from this program.
I'll say, since starting in 2014, we've grown quite a bit. We used to annually serve about a million a year, and this changed in 2020, which I will be sharing with you. But for context, what we see from our screening work is that it matches a lot of both mental health research and census data. Given the scope of the size of the data, currently annually serving about five to six million users every year, predominantly female, which is consistent with the mental health research, young, which is consistent with what we know is, when mental health conditions start. If you're a parent with teenagers, you're a parent with children who are entering puberty for the first time, we know this is when we're most concerned about finding young people, and otherwise, our data very much matches censuses state to state.
Then certainly during COVID, we saw similar trends with our CDC partners, when there were incidences of rise, for example, in the Asian-American community or in the Black community, which I'll go into. Next. I don't think anybody now questions that COVID was very difficult for all of us, and certainly if you're living with a chronic condition, and certainly one like a mental health condition, where the sensitivity for us around stressors, around our cognition, the way we perceive the world really has a radical impact on our ability to attend, and pay attention or be as productive as we would have liked.
What we saw during the pandemic, which was something that we had not seen before, since tracking our data, since 2014, is that, the severity of screens, the number of people who come in to take a test and score positive increased by about 10 percentage points. We saw that start to come up, especially in the summer months, in July, August, and then they've stayed at 76 to 78 percent across all screens since late summer, which may not be surprising. As school youth started entering back into school during a pandemic period, whereas the early days of COVID was really attentive just to the chaos of what it meant.
But for many of our young people, they only had a month or so before they ended up going into summer, where we know many of our young people are doing better at this point. Next slide. To share this in a different perspective, I wanted to share about our psychosis screen. If you don't know anything about psychosis, psychosis is a symptom that is related to diseases like schizophrenia. When you see or hear things that are not there or you have trouble using your words.
We're attentive to the rates of increase in psychosis, because while we saw a leap in the number of people who are taking a screen, that change could be tied to some of MHA's business, which we saw leap up in May. But again, what we did not see, historically, what we saw during COVID, is the line that you're watching, the percentage at risk for even young people who are developing schizophrenia for the first time in their life jumped from 66 percent to 76 percent over the course of time during COVID. This meant that, for young people who may not have experienced the stressors that could have prevented the onset of something like schizophrenia, maybe not during a COVID time, the stress that we all felt as a community, was so profound that it moved the needle for us, in terms of when a young person might experience what we consider their first break of schizophrenia. Which indicates the enormity of the stress that we're experiencing, and the challenges our young people are experiencing in their cognitive functioning. Next slide. In the same way we identified that suicidal ideation was on the rise.
Unlike psychosis, where cognition is a concern in your ability to think, suicidal ideation is a reflective isolation in the community, depression and the desperation that we feel to want to get out of our current situation. You can see that, in the historic years, between 2015 to 2019, we've seen stable rates of suicidal ideation across the board. But starting again in that summer months of COVID in 2020, that number of young people who were reporting thinking about self-harm or death, started to rise up and has stayed at post-pandemic rates since then.
We're keeping an eye on when this is going to come down. Trying to understand when will we, as the collective community, feel less stressed? Is a good sign that hopefully these signals will also reduce. Next slide. Certainly, and this is fairly consistent with the exception of the Asian-Pacific Islander experience, in our historic data across the board, individuals who identify as Native American or identify as more than one race or multiracial, have historically scored higher than any other subgroup that we've seen.
During COVID, we saw changes in the Asian- Pacific Islander community. For us, likely is a rise of some of the API experiences with Asian hate, the nuances for black communities also similarly. But I will show you a slide of some of the difference that we see here. But I think the important point here is that, we know from a disparity or equity lens, that our Native American and multi-race screeners have always had a significantly more profound difficult time, and I don't think enough attention has been placed to these two communities.
I just want to highlight that if you're working with Native American populations or people who identify as multiracial, which we don't hear a lot about, there may be a time to turn on your brain, and say, "Hey, there's something special to consider here." Next slide. During 2020, we asked people, when you think about coming to take a mental health test, what are the top three concerns that you experience? This is fairly consistent. We saw some changes about, but loneliness and isolation consistently is the number one concern for individuals, which if you think about this from a productivity or a connection standpoint, we are really at our best when we have someone to share something with. COVID was an interesting time where we were trapped in some ways with very few number of people. We see this with youth, we see this with our adults.
Our young people losing social networks. Our adults learning to navigate a COVID experience as parents with just one another, and the unique context that we're navigating together. From a mental health perspective, it's not surprising that past trauma and relationship problems are absolutely a concern for people across the board. I think for many people during COVID, those issues, the past trauma and the relationship problems really started to come out in full force in the way that we had to navigate our concerns with one another. Next slide. But what is interesting
is when you think about the main concerns across different races and ethnicities. These findings have been fairly consistent through 2020 and 2021, our Black African American communities are concerned about financial concerns and racism at the highest rate. This has been consistent as a factor of understanding historic discrimination in America, and certainly, what we can do for the Black community in bringing additional equity from a financial standpoint and job equity, I think is really important. Native Americans are reporting more past trauma than anybody else. API community have talked about concerns about COVID.
This can be either fear of getting sick or sometimes, we are seeing what they're saying is, what does it mean to be tied to something like COVID-19 as a community? Latino communities are experiencing loneliness, and our multiracial screeners are reporting grief. I will say, since tracking these numbers in 2021, we have seen some of the stressors overall reduce to some pre-2020 with the exception of grief. The concerns about grief have started to increase significantly, I think as all of our stressor has come down and trying to contend with what it meant to lose a family member or lose someone. Just in the very most recent last months of data, we're seeing grief start to take a rise. Next slide.
As a response of this, I just want to share that as a result of support from Microsoft, MHA is working with Northwestern University to provide a text-based protocol treatment module. I wouldn't call it treatment, but it's a text-based service that we're going to give at very low cost. It's a project that's done both with Northwestern University, where we're taking different psychological strategies. The way to push people via text messages to learn skills and build them up in a low cost way, that's asynchronous.
It means you can access this any time you want, but the machine learning component actually learns your habits and your behaviors, and the way that you prefer to receive this content as well as what kind of content you have. I hear on the agenda that David Moore from Northwestern, my counterpart, will be sharing with you additional details. I just wanted to give him a shout out for our awesome collaboration. A huge thank you to Microsoft, actually, for the wonderful support that they've given us for the last couple of years. Next slide. I just want to end with a quick call to action. If anybody is struggling with a mental health condition or if you're wondering about your child or your cousin, go online to mhascreening.org
where you could take a free mental health screen. For the nerds like myself, there are two links here. One is a report that we do annually to summarize national data from our federal friends that's collected, and how that pairs up with MHA screening data, which I showed you, which is the final slide, which is our state and county data dashboard, which is going to be for the next year or so, a real time update of all the data we see, and how this breaks down at a local and state level community, so some epidemiological work that we're doing. That's it for me. Thank you.
EBELE OKOLI: Thank you so much for that, Theresa. Thank you for all of the insights that you provided at the population level about mental health and health seeking behaviors and all of the insights, even into the disability intersectionality, which is something we are starting to talk a lot more about as you look at the implication of having a disability and being of a certain racial group, and that multiracial component also a new component first, also think about as well. That being said, we here at Microsoft have been extremely intentional about mental health and well-being through our partnership with you and how we even invest in our our employees.
I would like to transition over to Kristen now, who will talk a little bit more about what we've done to support our mental health wellbeing of our employees here. KRISTEN ROBY DIMLOW: Thank you so much, Evely. I am really thrilled to be here and thrilled to be part of the Ability Summit. If we could go to the next slide. I'm really excited today to tell you a little bit about what we're doing within Microsoft to promote wellbeing. Before I dive into mental and emotional wellbeing, I just wanted to share that Microsoft is thinking about holistic wellbeing.
We think about how physical wellbeing, mental, and emotional, and financial wellbeing all are important and need to be in balance together. We find out that if you're off in any one of these areas, it manifests in the others. For example, if you have financial strain, if you have financial worries, it translates into anxiety, which could translate into physical symptoms. I think big employers like Microsoft have always been pretty strong on the physical benefits perspective and then also financial benefits in terms of pay and benefits.
But we've only recently really been focusing on the importance of mental and emotional wellbeing. When we think about our holistic platform, we call it the Be Well platform. We really try to think about each of these components.
What are the benefits and services that we offer to our employees so that they can, say, keep balanced in each of these areas. If you could move on. We think about programs and services across a number of different dimensions. These are just a couple of examples of what we offer here at Microsoft. First of all, a medical plan. It not only supports people's physical wellness, but also also their mental and emotional wellness.
We have an employee assistance program called Microsoft Cares, and that's offered globally. It offers all sorts of support for employees and their family's independence, so that they can reach out if they have the need to talk to somebody, if they need support. We have our Be Well global platform, which is a global software platform that allows employees to access a portal of services, trainings, ideas, ways that they can promote their holistic wellbeing. Then we also have a lot of opportunities here at Microsoft. For example, we have employee resource groups.
The resource groups are fantastic for promoting wellbeing. Then finally, external resources. We bring in outside training, things like today, where we can offer people perspective from the outside on how to promote mental wellbeing. Next slide. One of the things that we did a few years ago was we realized the importance of mental well-being, and we started an internal campaign called It's Okay Not to Be Okay.
What we realized is we wanted people to really be as comfortable talking about their mental well-being as they were with their physical well-being. We were really stunned by the overwhelming positive response from our employees. We had a lot of employees who were saying, "thank you so much for helping to remind us that it's okay not to be okay." We also had a lot of leaders who stepped up and shared their own journeys with respect to mental well-being for themselves or perhaps, a family member. We created this incredible team internal to Microsoft, it consists of a variety of wellbeing experts from different disciplines in the company. We have my own team, which promotes global benefits and global rewards.
Then we also have Ebele, who is from our legal team and our Ability team. We have people from our product groups who are trying to think about how Microsoft products come together to support mental wellbeing. We have people from our diversity and inclusion team who give us great advice on how to think about different aspects. It was great to see Theresa's data because you really start to understand the importance of thinking about this through so many different angles. What I wanted to share today is some of the output of that cross-functional Microsoft virtual team is these two things. One is your wellbeing was well worth it.
We have morphed our campaign from It's Okay Not to Be Okay into internal resources that are available to employees to support their wellbeing. We've also done a little bit of internal marketing just to continue to destigmatize and to ensure that employees can find the resources that are available to them. Then on the right, we've put some investment our learning organization has helped us by providing fantastic content that's targeted to both managers and employees to help us continue our journey on mental and emotional wellbeing.
Some of that might be for an employee just to think about, "Hey, would you like to have a conversation with somebody? Do you think you need support? Is there somebody on your team that you think needs to support? What can you do to be an ally to that person?" Then for managers, also to think about being great managers, how do you promote psychological safety so that employees can feel confident sharing any needs that they have from a mental health or wellbeing perspective? We've been so excited about the uptake in training. I think that these resources have been fantastic and we're starting to realize, our V-Team is thinking a lot about, where do we go next? We have so much opportunity and we feel like we're really scratching the surface, and yet, we really know that we're on track. Theresa mentioned the importance of COVID.
I think the silver lining of the increased anxiety and uncertainty people have had as a result of COVID has continued to destigmatize the importance of mental wellbeing and the ability for people to say, I'm not okay right now. We are learning a lot through our V-Team. We are continuing to try to push ourselves and to push the bar for all of Microsoft. You could go on the next slide. One of the neat things we did recently was we announced Headspace for all employees. We have a number of different things that people can use, but we did a 90-day pilot with Headspace globally, and it was so well received that we've now rolled it out to all employees.
We also use something called Talkspace, which allows people to chat. I mentioned the employee assistance program. We try to offer a lot of different ways that employees can support their mental well-being. This was a fun one that has been really well received by our employees. Next slide.
Where do we go from here? I mentioned that we're continuing to learn and grow and try to think of the different needs of our different audiences. We have a globally diverse organization where we do business in 109 countries, and so there's a lot of work to do just to understand how we localize this work. But part of that is continuing to tell the story. When we think of holistic wellbeing, we set up a calendar for ourselves.
This is just an example of the different topics that we touch on throughout the year. May is mental health month, and so during the month of May, we'll be doing an internal campaign on mental health just to remind employees of all the things that we have available. In July, we have our Be Well Games Challenge and a Fall Wellbeing campaign. Then in October through December as we get into our open enrollment period at Microsoft we'll have physical. This is a year-round thing.
I think if you're a large employer if you're thinking about how to support people, one of the things to be conscious of is people need just in time support and so the campaign that you did three years ago may not be something that's top of mind for somebody or you have a lot of new employees who come in every year. It's important to continue to message to employees and to managers the ways that they can support. Certainly, we want employees to understand the variety of programs and services that are available to them.
Next slide. This is an example of some of the collateral that we send. We will send an employee email, something like this, and then provide links to the services and to the support. Then next slide.
Then we think of this as essential. Your well-being isn't extra, it's essential. We want it to be part of everything. We really want holistic wellbeing. We want to continue to destigmatize mental health and mental wellbeing. We're thrilled with the response from employees.
We're really seeing that we're going on the right path and certainly, Theresa's data would tell me we need to do even more. From my call to action is really just continue. If you're a manager or if you're somebody in a position of privilege and power, please promote the resources that are available to support your team's wellbeing. Then also each of us, whether we're a manager or an employee, we can show others the importance of attending to our own mental and emotional wellbeing. One of the things we did recently at Microsoft is we converted sick time to being both sick and mental well-being time.
This was our way of saying, "Hey, please, it's okay. Take care of yourself." Important for leaders to continue to show that and for all of us to be allies to each other. Thank you so much for your time today. EBELE OKOLI: Thank you. That was
wonderful, Kristen. From there you can see that Microsoft has been very intentional about the mental health and wellbeing of its employees, from the benefits to the resources to the investment of its steering committee and mental health working group constantly prioritizing and strategizing around mental health. It is very clear that we are committed to the mental health and well-being of our employees.
Then we take that a step further with our next session, where we are now dedicated to doing some research in mental health and wellbeing with our MSR research under the leadership of our very own Mary Czerwinski. Over to you, Mary. MARY CZERWINSKI: Thank you so much, Ebele. It's so awesome seeing the two previous presentations and all the services that are available to employees and just the entire population.
I love it. My group has been doing research on stress detection for quite some time now, and I would love to talk to you a little bit about a few studies we did last summer. Next slide, please. It takes a small village to do this kind of research.
We have a big V team. But what we've found is that -- and it's not just our research that's showing this, it's actually well known in the psychological literature -- that the workplace is the number one stressor for American adults, and it costs the US economy quite a bit of money you see here one statistic, $300 billion annually. We know that COVID has exacerbated these issues as Theresa has said. We also saw from previous research that constant multitasking always leads to more stress, which then decreases perceived productivity and lowers well-being.
I think working from home, this has actually caused a lot of people who are caretakers, in particular, some issues with stress because they're trying to take care of others in addition to doing their work, and their work-life boundaries have blurred. This is causing mental health problems. It is the number one disability affecting over one billion people worldwide.
With the global pandemic, nearly half of Americans have experienced increased stress and worsening mental health issues. We even know our blood pressure is going up. Something's got to give. When we talk to people in our research studies, we hear that they really would like to manage their mood and know more about how to become more resilient to stress.
They just don't know how. That led us to go ahead and build an affect sensing platform that will allow us to design emotionally sentient systems, services, and agents that can help in the detection of stress and hopefully deliver just in time interventions and help with focus at work. Next slide, please. Our Stress Project is around a holistic service that tracks, understands, and supports information workers so that they can maximize their emotional resilience and increase their wellbeing at work. The way we decided to do this was via passive sensing and using machine learning on physiological and behavioral signals of the user. For instance, we can use the camera to detect your facial gestures, even your heart rate, and your breathing rate.
Then we can do machine learning on these signals without any wearable sensor. The other thing that we have going for us is that all of these information workers who are in our research projects consent and allow us to go ahead and do this tracking. But we can actually see what they're doing on their computer.
For instance, we can see how busy their calendar is. We can see how many notifications or email they're getting. We can see how late at night possibly they're working or if they have breaks in their schedule even. If it's needed, we can detect stress and some personalized interventions that might help you right in your time of need. Next slide. We did two studies this last summer. One was particularly on stress sensing, so how good can we get at actually seeing how stressed out you are? Again, all of the information workers that we studied last summer opted in and we went through ethics review.
The second study looked at psychological interventions from cognitive behavioral therapy and dialectical behavioral therapy and positive psychology, and we looked at the timing of sending these kinds of interventions and could we make them short enough and usable and useful enough that people would want to do them at work? Next slide, please. Our sensing platform is pretty complicated. It did take us about a year and a half to build the system.
Some of the parts of it had existed previously. But we do use the camera, the microphone, as I said, to collect some of these signals. It's invasive, but we did hear from people that they don't really like the camera turned on per say, and so the good news is because we had so many sensors, whether we looked at the context of what you were working on in your computer or even what your calendar looked like, that we can now actually turn off the camera and the microphone and still get pretty good at sensing. But the real system, the system we built for last summer, had all of these different kinds of signals being collected on our users at the same time. Next slide, please.
This is a typical study design that we use in my group. We usually do at least a four week longitudinal study where we will do all these recordings of all these sensor streams that I've been talking about. We usually have 50-100 information workers that we track again, all with their consent, and we do survey them a lot. We'll do a survey at the beginning of the study where we'll ask them their personality type, how stressed they are in their life, what major life stressors they've been going through, and all kinds of demographic information. Then at the end of the study we'll ask them what the experience was like. During the study, we usually do a beginning of the day and at end of the day little survey asking people how things went that day at work, how stressed out they were.
In this particular study, we asked between five and eight times a day how they were doing, what the cognitive demands on them were, how many resources they felt like they had to throw at those cognitive demands, how stressed out they were, how pleasant they felt, how busy they were, etc. What we found in the stress modeling system, thank you, was that if you just look at a user for a couple days, you're not going to be very good at detecting their stress. But if you look at about three weeks of data for an information worker, you can start to get much more accurate at detecting their stress. Again, this was with the camera turned on and the microphone turned on. What we determined from this last summer study was we have to personally model stress for each individual. The way I react to stress is not the same as the way you will react to stress so it has to be personalized and individualized.
What we saw was that users, as I said before, were definitely more comfortable keeping these kinds of signals on their own machines. They did not want management seeing how stressed out they were. Also they were less comfortable with the camera and the microphone turned on. The great thing is we learn what people were and weren't comfortable with. Interestingly enough, they're quite comfortable with us collecting their contextual information from their computers and their keyboard activity.
So we're going to be moving forward using those kinds of signals. Next slide, please. The second study was looking at the type and timing of interventions to help once we do detect stress. Next slide. In this particular study, we actually had a clinical psychologist who was working with us, as we always do and we took real psychological interventions from various kinds of therapies and we developed a catalog of fun, I would say, visually interesting designed interventions. We built a little chat bot in Teams that when and if you were in the condition where you were getting just in time interventions, we would use our stress detector to send you an intervention right in the moment when you really needed that help.
We also had different categories of interventions that I'll just touch on briefly. The other condition, there were two conditions. One was just in time help, and the other condition was schedule help for yourself. Different people liked these two kinds of intervention timings differently.
Some people really liked scheduling, for instance, a meditation break later in the afternoon when they knew they were trying to wind down and break up, the workday from the home day. Other people really relished the just in time. In fact, the just in time was significantly more popular than the scheduled interventions.
But different people liked both. Again, it's all about individuals and personalization. Next slide. Again, same kind of study. We had 86 participants, 43 in the pre-scheduled condition and 43 in the just in time condition for four weeks.
Same kind of thing as I explained to you before, they were running the stress sensing software, and then they got the interventions in the timing manner depending on their condition. Next slide. What we found in that study was that medium effort interventions were the least effective. People liked them the least.
It was the low and high effort interventions, interestingly, that were very effective and that people liked the most. The low effort interventions were by far the most well-loved, and they all had equal effectiveness. Low and high effort interventions were equally effective. The high effort interventions for those of you that are psychologists in the group probably led to more behavior change. We actually had a behavior change model that we looked at with our users and those that really wanted to change their behavior, the high effort interventions actually helped them move along that behavior change model cycle better.
The low interventions are great for just in time. The high effort ones actually help you change your behavior long term. In conclusion, we found that personalized stress models are really important, as are the personalized interventions. When we can see that you're stressed out, it's really nice to have those just in time interventions, and there is a tradeoff between what people are comfortable with e.g. the camera being on or not, and how good these models can be. But the models can be pretty good even when they're less invasive, stress sensing is going on.
We think we're at the verge now to ship this out to a large number of people at Microsoft. I hope some of you will volunteer for us and try this out and hopefully this will work its way into Microsoft products some way down the road. Thank you very much for your attention.
EBELE OKOLI: Thank you so much for that, Mary. Now, we'd like to transition to our calls to action just to recap on everything that you've learned here today. We'll start from the top with our speakers. I'll turn to Theresa first and then we'll have Kristen to follow and then Mary. Over to you, Theresa. THERESA NGUYEN: Thank you. If anybody is
worried about your mental health or want to learn more about mental health, visit mhascreening.org or if you're interested in policy and research, visit our main website at mhanational.org. EBELE OKOLI: Over to you Kristen.
KRISTEN ROBY DIMLOW: I would just say, remember, it's okay not to be okay, and be aware that there are benefits and services to support you and your teams and your family members, and then share how you support your own mental and emotional well being with your colleagues. Thank you. EBELE OKOLI: Fantastic, and Mary. MARY CZERWINSKI: I would just like to underline that, take advantage of the services that Kristen's talking about, and the website that Theresa's mentioned, and also try to schedule some time in your calendar so if you are feeling stressed or you know you're going to have a stressful day or event, schedule some time around it to take a little break, do some meditation, stretch, take a walk in nature. There's lots of support for you out there.
EBELE OKOLI: Fabulous. Thank you to all of the presenters that came out today to speak to us, and I hope and I know and I trust that everyone who attended is inspired by what we've got going on here, and hopefully they'll take it to your respective organizations and really drive for impact in supporting those with mental health and well-being needs. Thank you.
2022-05-13