take a breath
a “mindful breathing” iOS app for young adults with anxiety
"Take a Breath" was my graduate thesis project for my M.F.A. in Interactive Media and Games. My thesis committee, Marientina Gotsis and Dennis Wixon, were instrumental to the completion of the project and related study. It is thanks to their guidance that I was able to explore an avenue close to my heart.
It was an unusual project for my program. I was only the second person to do a research project as my thesis, and the first to actually complete the study prior to graduation. This presented some unique challenges, and the project changed drastically from my initial proposal to the final product, as long-term projects so often do. While it began as a suite of small activities, I chose to focus on a single one for the purposes of running a study: reducing the number of variables would hopefully make the findings more accurate and specific.
Over the course of two years, it taught me a lot about the shifting goal posts of design, the often very frustrating realities of performing research on human subjects, and the unfortunate gap between how you hope something will be used and how it is actually used.
initial abstract
my original (kind of overambitious) pitch for the project as a suite of small applications
"Take a Breath" will be a mobile application designed to help teens and young adults with anxiety without preventing them from going about their daily life. It will be a collection of small tasks and mini-games based on the psychological concept of "grounding," which espouses self-reminders and simple, low-stress tasks as a means of decreasing dissociation and discomfort associated with acute anxiety.
It is aimed to relieve anxiety in the short-term in situations where using long-term therapeutic solutions is not possible or convenient: for example, on the train, in a waiting room, or at work. "Take a Breath" will be able to be used discreetly anywhere cellphones are allowed and, importantly, without calling outside attention to the user’s predicament and without causing the user additional stress or embarrassment.
why develop for mobile? why make something interactive?
Interactive media has an unparalleled capability to reach people under the age of 25 because it has been reaching us our entire lives. I wanted to make something with a wide reach, and there is no better medium suited to this demographic than the interactive.
(Please understand that I wrote these bold proclamations when I, myself, was actually 25.)
what is grounding, and how does it help alleviate anxiety?
Grounding helps an anxious person to become more aware of being in their body, as opposed to getting stuck in their head. Grounding encompasses both cognitive grounding (i.e. describing elements of the environment, recalling personal details, discussing current events with others) and sensory grounding (i.e. touching objects close by, clenching fists, paying close attention to sensory input).
Many grounding methods can be described as targeted redirection of attention, tasks which are not necessarily intellectually difficult but which require redistribution of mental resources away from anxiety and toward the current moment.
My initial goal with this project was to provide a convenient guide and reference on grounding techniques to a wider segment of the population.
why make something specifically aimed at teens and young adults?
Young people are at very high risk for mental illness—when considering anxiety specifically, the nationally-representative study conducted from 2001-2003 provided a lifetime prevalence estimate for anxiety disorders at 28.8%, with a median age of onset of 11. Those who identify as LGBTQ are at even greater risk of developing mental illness. Young people also don't necessarily have the financial means or the freedom necessary to access mental health care; they're often dependent on their guardians, and there's no guarantee that their needs will be appropriately met.
I wanted to make something free and accessible, something that could be use discreetly and in a socially-acceptable way—and what is less unusual than a teenager staring at their phone? Basically, I wanted to make something that could potentially mitigate those very real barriers to care.
what is "mindful breathing" and is there evidence supporting its efficacy?
Mindfulness-based therapy (MBT) is a method of stress reduction and is predicated on the practice of paying attention in a particular way: "on purpose, in the present moment, and non-judgmentally." The results of a meta-analytic review on the efficacy of mindfulness-based therapy suggest that it is a promising intervention for sufferers of anxiety and depression, and it was specifically found to be moderately effective for improving anxiety.
Included under the umbrella term MBT is the basic meditation exercise “mindful breathing,” which involves the specific application of mindful attention to the physical sensations associated with breathing. In one study, patients who were directed to do mindful breathing reported more decentering than those directed to use the other methods, and had a weaker association between the amount of repetitive thinking and negative reactions to thoughts than with the other two methods, suggesting that mindful breathing may help to reduce reactivity to repetitive thoughts, a symptom associated with anxiety.
project goals
Broadly, the goal was to determine what aspects of grounding could be effectively translated into a digital, mobile format and then to use those insights to develop a helpful application for young(ish) sufferers of anxiety. Additional specific goals and design considerations:
Accomplish general goals of grounding: 1) Reconnect the person to the present, 2) Orient the person to the here and now, 3) Connect the person to their body and remind them that they have control over it.
Create something that can help people who are currently underserved. Anxiety is an incredibly widespread condition, yet there are few resources and little education provided to the general public about how to identify it or how it can be mitigated.
Must avoid increasing the user’s anxiety, if nothing else; subtlety and discretion are important, and potential sources for additional anxiety can and should be avoided (i.e. being observed while performing exercises, being asked about the reason for installing an “anxiety app”).
prototype
created using Processing, an open-source graphical library
Four activities were initially implemented, as can be seen in the gallery below:
The "self-reminder" task prompts the user to recall personal details and concentrate on the feel of the world around them.
The "mindful breathing" task provides the user with a way to visually concentrate on their breathing by expanding the circle on inhale and contracting on exhale.
The "color matching" task asks the user to repeatedly click on squares until they are all the same color.
The "alphabet" task is the most “difficult” task, requiring the most attention. The user is asked to tap the letters of the alphabet backward. (There is no penalty for failed attempts on this or any other task.)
UX design and considerations for the prototype
Activities should be able to be experienced in any order. Each requires very limited instruction, so navigation between activities should be easy and quick.
Each activity should be designed to take only a few minutes to complete, so that they can easily fit into a busy schedule. "Take a Breath" should be intentionally designed for interstitial use.
Overall, the app must be accessible to as many people as possible. User experience (and reducing user frustration) is paramount.
less is more
narrowing scope for better testing
When I decided that I wanted to run a study on the app, it made sense to restrict the scope to just one activity. That way, we would have better specificity for the results and would reduce the possible confounding factors. After discussion with my committee, I decided to focus on the mindful breathing task because it seemed like the most useful or applicable. The benefits of mindfulness in treating mental illness are well-documented, and there really wasn't anything on the market which had the same level of interactivity.
(There are now hundreds of similar concepts on the app store, and I somehow firmly missed the boat. Oops.)
Because our app was going to be tested on human subjects, I had to undergo training and go through the IRB approval process.
"second pass" UX design
The user experience design specifications for the app needed to be amended and clarified so that they could be communicated to the iOS developer. They also needed to take into account the narrowed scope and the data collection required for the study.
data anonymization
On startup, the user needs to be able to enter a code they will be provided by email, which all data will be associated with rather than with personal information.
instructions
gesture tuning
The circle should be able to grow and shrink fully with a single gesture. This should account for hand size variation, and not default to, for example, an average "male" thumb.
satisfying interaction
The response to interaction should be "satisfying"—the swiping gesture should provide pleasing visual feedback to the user.
the actual implementation of the main screens
This is what the main "breathing" screen and the anxiety inventory screen looked like on the actual iOS app we used in the study.
study design
the thrilling conclusion to my months-long, epic battle with iStar
recruitment
Recruitment was conducted largely via the internet to maximize our potential reach and appeal to our target demographic: tech-savvy teenagers and young adults. Because this study did not require physical proximity, social media and email were ideal channels for recruitment.
Participants were classified as either showing elevated signs of anxiety (in which case they were assigned to the experimental group) or not (in which case they were assigned to the control group) based on the cut-off scores for the Anxiety subscale of the Depression Anxiety Stress Scales; participants scoring in the range of 10 to 19 on were classified as showing elevated signs of anxiety.
criteria for participation
Inclusion Criteria
-
Aged 18-24
-
Fluent in English
-
Owned and were comfortable using an iOS device
Exclusion Criteria
-
Participant showed signs of having an anxiety disorder by either:
-
Exceeding the 320 point cut-off score in the Visual-Analogue Scale of Anxiety, or
-
Exceeding the 19 point cutoff score of the Anxiety sub-scale of the DASS
-
All told, we had 21 participants: 9 in the control group and 12 in the experimental group.
research questions and hypothesis
RQ1. How will usage patterns change over time?
-
We plotted frequency of usage day by day for each user.
RQ2. Will application usage influence anxiety scores?
-
Participant change scores were determined by comparing post DASS scores to the baseline.
-
A linear and quadratic regression were computed using the change scores and the frequency of use.
-
We visualized the VAS-Anxiety scores over time.
H1. The majority of participants (75%) will not experience a clinically significant increase in anxiety.
-
The National Institute of Mental Health criterion for a clinically significant change on anxiety is a shift of 30%
-
We computed the difference between pre and post DASS scores, then rank ordered these differences, and compared the results to that cutoff.
procedure
An online screening questionnaire was used to determine eligibility. If the applicant was qualified, they were provided with an information sheet, asked to provide consent online, and given the baseline questions. If that all proceeded as expected, they were enrolled in the study for a two week period and provided with access to the app, which collected the inventory responses and usage data (ex. frequency of use, duration of use) during the study.
During the 2 week period of participation, the user was prompted to answer a four-question anxiety inventory periodically, so that change in anxiety level could be measured. Participants were presented with the inventory within the application itself in order to raise the chances of compliance. Intermittent sampling was to gather data during the testing period because of its longer duration.
After a maximum of 2 weeks, some participants were sent the post-test survey on Qualtrics and some were asked to participate in a phone interview.
project challenges
or, every thesis project inevitably winds up being a comedy of errors
The process of going from my original pitch and prototype, then seeking IRB approval, then contracting out the actual implementation of the app, and then actually running the study was long, confusing, and certainly a challenge. Despite all of the madness, everything got finished on time, and learned a lot in the process.
doing research means making so, so many decisions
This was my first time doing larger-scale, quantitative research, much less on human subjects where IRB approval was necessary. The process of writing the study protocol and having to make countless decisions on minutiae I didn’t fully understand was petrifying. It never felt like I had enough information, and I was worried that every step I took violated some sacred research social norms and sensibilities that I just didn’t know. (In reality, that was only the case some of the time.)
As with many of my experiences in grad school, it was a (controlled, ultimately safe, yet terrifying) sink or swim situation, where I just had to hit the ground running and figure it out as I went. What I found out, or confirmed, is that I am most productive when I am so busy that I have no time to worry that I might fail. Someone make that into a motivational posted I can stick on a wall.
the IRB is simultaneously great conceptually and supremely frustrating in practice
I'm mostly kidding, but it sure was a whole thing trying to iron out exactly what I could or could not do given that I was working with human subjects in the context of mental health. Our app (more accurately, our "intervention," because it was theoretically intended to provide a health benefit) was relatively low-risk for participants, but I still had to make some compromises on my original vision for the app.
When I (ignorantly and so naively) first conceived of the study, I thought we would use it to determine the efficacy of the app. As our app was technically a healthcare intervention, we had to first determine whether it was safe for people to use—we had to make sure it didn't make anything worse before we could even think about whether it was useful at all. That was the perspective we had to shift our thinking toward.
if working with adult human subjects is hard, working with minors is borderline impossible
When I initially came up with the idea for this project, I specifically wanted it to be for "teens and young adults" for the reasons discussed above.
I stand by that. I think that was a meaningful goal. Unfortunately, despite a lot of back-and-forth negotiation with the IRB, it wasn't feasible for me to include minors in the study, due to additional stringent requirements and the difficulty in obtaining parental consent given that participation was "remote" rather than in-person.
That change was probably the hardest to swallow, even though I was still able to target the 18-24 age range. It felt like I was forced compromise on the spirit of the project, which was the last thing I wanted to do. Ultimately, though, I realized that being unable to target that audience with this specific project and this specific study didn't mean that I was precluded from ever trying something like that again. All projects require compromise, and I was still able to hit a good portion of the age range I had intended to reach.
contracting out the app development was a mess
On a much lighter note, the process of finding someone to program the app was kind of ridiculous. I initially intended to target Android, but I somehow couldn't find someone who could do that, so I caved and found someone who would develop for iOS. (This seemed like a totally reasonable decision, right up until it came time for me to actually test the builds and then show it at events... at which point I remembered that the reason I wanted to go with Android was that I didn't actually own any Apple devices. I spent a lot of time that year glomming onto friends' phones, which I am sure they very much appreciated.)
Working with the developer was also more of a challenge than I anticipated—he became increasingly hard to get a hold of as time went on, and then entirely flaked out. At that point, it was a mad dash to find someone to replace him, because I knew nothing about iOS development, so I couldn't just insert myself into the development process directly.
As stressful as that was, all of the necessary functionality was present in the final iteration, if not handled exactly the way I had hoped. The communication lag and general time constraints precluded a lot of revisions, because eventually I absolutely had to proceed with the study. To put a positive spin on it, I guess that aspect of it felt sort of like working on a "real" professional project!
and after all that, what was the result?
a lesson in humility, probably!
Just to recap before I take a hard left, the following are all true:
-
The app was successfully designed and implemented
-
The study was designed and made it through the IRB-approval process
-
Participants were recruited for and enrolled in the study and data was collected as expected
-
The results indicated that the app was, indeed, safe to use
And those are all accomplishments I'm proud of! Working on my thesis was a fantastic challenge, and a really valuable learning opportunity.
But you want to know what else is true? People kind of didn't use it.
The app was safe to use, but that didn't mean that people thought to use it in a moment of crisis. The feedback I got on the concept of the app, on the experience of using it, was solidly positive... but the overall frequency of use was low, and almost no one actually used as I had hoped they would.
This was probably the roughest lesson in user interaction design I've learned so far: you can build the most beautiful, functional thing in the world—which, just to be clear, I am absolutely not saying is what I did here—but it doesn't mean anything unless people actually use it.
That has to be the goal. That has to be the litmus test.