Autism EHR • Tilly
Quickly created an electronic health record for ABA therapists that streamlines their clinical workflows, without all of the visual and functional clutter of legacy systems.
I designed & launched a uniquely clean, user-friendly set of clinical tools for autism therapists from 0 → 1 in less than 9 months (while also building practice management tools – scheduling, dashboards, payroll, etc – in that same time frame).
Outcomes
😀 100% user retention in 2024
In 2024, none of our active behavior analysts have left Tilly.
✨ Cleared that path for AI
Once we owned our clinical data, we launched time-saving AI features that increased our NPS by 19 points.
💭 Quick iteration
Thanks to my close customer relationships, all 4 top-requested features were already in development when MVP feedback was received.
Contributions
Design leadership & strategy
Team
Product, engineering, & research leads; 1 midlevel designer
Dates
Sep 2022 - Jun 2023
Problem overview
Tilly provides products and services to pediatric autism therapists to run their own small practices. Autism therapy is very data-driven and is often 30+ hours or care per week, usually in home. Data is collected throughout these play-based home sessions, where there's often a lot happening — children might be melting down, parents or siblings might be interfering, etc.

When we began this project, both our analysts (who supervise a client’s care) and our technicians (who implement therapy in the child’s home) complained about using third party software (Catalyst) for clinical data. With another system comes more overhead and duplicated or out-of-sync information. Additionally, competitor systems are bloated with features and are very hard to learn how to use — which is especially problematic given high industry turnover.
From a business perspective, we had the opportunity to save the annual subscription costs plus own our own clinical data, unlocking a myriad of opportunities in AI, measuring clinical outcomes, family visibility into progress, and product optimizations.
Competitors: Autism ABA data is highly complex and as a growing clinical field, there are a variety of ways to achieve each clinical objective. The legacy competitors in the industry suffer from bloated feature sets and, as a result, a lack of usability for newly trained technicians.
My approach
Learned complex clinical workflows
Doing this job of analyzing clinical data requires years of training, and we needed to understand it quickly. I was fortunate to partner with an incredible research lead who led formative interviews and walkthroughs while I triangulated with competitive research where I understood how existing tools were used through training videos.
Nurtured customer relationships
As we evolved and iterated, I took full responsibility of user feedback and relationships. I ensured that our primary beta testers felt heard and rewarded for their time, so they would remain engaged in small one-off asks and clinical feedback on our iterations. At the same time, I looked for opportunities to include new or external voices.
Prioritized simplicity & craft always
A key principle throughout the development of this project was to avoid duplicating features in competitor just to reach parity. We wanted a clean, seamless experience that was easy for a technician or analyst to learn with minimal training, while still enabling an expert user to have maximal functionality to try the latest techniques in ABA therapy. We looked for flexible solutions that scaled without resulting in overly complex set up.
Coached and grew a designer
After I led design for the launch of the MVP, I coached a mid-level designer through iterations and helped her build domain expertise and independence, while overseeing her work and relationships with clinicians to ensure high quality launches.


Solution
The Electronic Health Record (EHR) is a critical piece of Tilly's overall software stack, described in the following video (EHR starts at 1:46):
Solution: Data collection on mobile app
Our technician app allows technicians to view their schedule, collect data, and write session notes. Collecting data is made very user friendly with large tappable buttons and easy filtering options.
Technician app: Technicians use an app to collect data between 20-50 goals per child. We optimized especially for tablet, because technicians are equipped with tablets to bring into the homes.




Solution: Data analysis on web platform
Our web platform allows analysts to easily manage the goals each client is working on and analyze the data coming back from therapy sessions.
Analyst platform: Analysts use a web platform to manage each child's numerous goals and view all data collected with various important visualizations of that data.


Post-launch enhancements
AI Progress Reports. Owning our own data gave us the opportunity to solve a key pain point with AI — biannual progress reports that require manually copy-pasting and reporting on 50+ data points. Each report previously took ~10 hours to complete and can now be done in under 2 hours, saving analysts hundreds of hours per year.

Goal Libraries. Each child receives a unique care plan, but many goals can be reused across many children, like learning the steps to brushing your teeth. Analysts requested a way to easily copy or reuse existing goals and we delivered a clever a way to create, manage, and add from goal libraries.

Offline Mode. Technicians are often in homes with spotty service or WiFi. Losing all data collected or having to collect data on paper is a huge problem. This was a highly technical project to better anticipate what engineering could do in a reasonable timeline to store and retrieve this data, and how we could best translate the functionality to the end user via writing and notifications.

Graph improvements. In order to meet rapid launch timelines, the biggest sacrifices had to be made around graphing and analysis functionality. We selected some key Since the lack of functionality here caused the most pain for our end users, we quickly launched a space for more mature data analysis features, initially focusing on comparing multiple goals against one another.

What's next?
✨ Further opportunities for AI. We currently have a tool under development that will use AI with the data collected in a session to write a narrative note summarizing a session. We also look forward to using AI to help summarize and synthesize all the data collected over hundreds of hours of therapy.
👁️ Bridge gap between analyst web platform & technician mobile app. The users of the web platform who analyze the data are certified behavior analysts who employ and manage the technicians who collect data on the ir mobile apps. As a result, these analysts lack some visibility into the tool they train their technicians to use, and can't always tell how the program will look for each child once fully configured. We'd like to better understand how to support mobile app training and resources.
📉 Robust graphing & analysis tools. There is a lot of opportunity around letting analysts slice, dice, and compare the rich set of data they have. We have built the basics required, but we could easily build out a roadmap around maturing this feature set and further empowering our therapists with data.
MORE PROJECTS