Iris is a comprehensive rehabilitation monitoring system made for people with blindness/low-vision (BLV) and their occupational therapists.

We worked closely with a vision occupational therapist (OT) to develop routines that help patients re-acquire essential daily skills. The wearable Iris tracker logs patients' sessions at home so that they can continue developing the techniques they learn from their therapy appointments. Building on existing OT paradigms, the Iris system creates a positive reinforcement cycle and fosters synchronous and asynchronous feedback between therapists and patients.

Everything comes together with the Iris cloud platform to make transitions between office visits and at-home practice seamless. The experience begins with therapists evaluating a patient's visual acuity through questions and activities on-site. Information about living conditions, daily goals, and comfort with technology are recorded with the Iris OT exercise generation platform. Based on this information, the therapist can begin to customize suggested practice routines with the patient to carry out after their appointment. Practice routines are uploaded to the Iris cloud, where it is securely stored and available for download by the patient through their mobile device.

Routines are synchronized from the patient's handset to the Iris tracker. Lightweight and durable, the Iris tracker is compact enough to be comfortably worn and integrates advanced sensors, including a wide angle camera, a time-of-flight sensor, and a motion-sensing gyroscope. Combined, the inputs from these sensors are fed into computer vision algorithms that run completely on-device, enabling Iris to recognize people, objects, and environments. Patients can choose to wear Iris however they like thanks to a variety of interchangeable magnetic backings. A single button controls the Iris tracker – simply press once to begin an exercise.

Spoken guidance provides instructions for patients to begin their practice routine. Because Iris tracks task completion in real time, it can also provide feedback and spoken assistance. Distinct tonal cues notify patients of progress throughout their sessions.

After completing a routine, the screen reader-compatible app provides metrics to the patient. Recordings of sessions can also be shared with the occupational therapist, who can watch and review key moments in preparation for the next appointment.

Upon returning to the OT's office, the therapist and patient can discuss successes and potential areas of improvement. By seeing a first-person view of how patients work through the activities, therapists can suggest new assistive techniques and tools.

Iris extends the BLV occupational therapy experience beyond the doctor's office. By empowering patients to independently practice new skills in their homes, Iris makes adapting to new visual conditions a more considered and natural experience.

Background

Blindness and low vision (BLV) are long-term, incurable conditions that affect the everyday living of people. Since visual conditions vary between individuals and deteriorate overtime, BLV rehabilitation needs to be customized for individuals' needs. In order to address such complexity, BLV occupational therapists (OTs) work closely with their patients to provide training with Instrumental Activities of Daily Living (IADLs) and assistive technologies. This requires comprehensive expertise and in-depth involvement of OTs, which causes an imbalance between the scarce professional service and vast patient demands.

Inevitably, BLV occupational therapy is expensive. The cost for private occupational therapy ranges from $100 to $200 per hour. Healthcare plans like Medicare can only provide limited coverage with BLV rehabilitation ("up to 9-hours of covered vision rehabilitation services over a consecutive 90-day period"). This presents a great opportunity for technology to democratize the service system of BLV rehabilitation by increasing its efficiency.

Stakeholders

In our proposed systematic addition to BLV rehabilitation, occupational therapists and patients are core considerations. Through our design we explore new ways to connect the two stakeholders in therapy. We also advocate for peripheral stakeholders, including healthcare networks and technology companies developing computer vision technologies to work together to realize change in the visual impairment rehabilitation space.

Research

We observed occupational therapy sessions in order to map the full user journey. For new patients, therapy starts with a thorough evaluation of patients' performance of essential living skills such as reading and writing. Patients will also establish learning goals with their OTs for the following therapy sessions. This ensures that the ensuing skill training will be tailored to the patient's needs. However, we found that returning patients had some difficulty translating skills learned in a healthcare setting to their lives at home. This observation aligns with the interview findings from both an BLV patient (male, 70s) and a BLV assistive technology YouTuber (female, 30s), both of whom mentioned that therapists' recommendations and training with assistive products can be hit-or-miss because OTs have limited information about living situations and daily routines.

A BLV occupational therapist brought up similar concerns in our interviews with her. First, she believes that a thorough evaluation of patients' conditions and pain points is fundamental to a successful therapy plan, but clinical evaluations cannot reflect all of patients' conditions. Second, she echoed that simulating real-life scenarios is more effective in helping patients improve their skills compared to clinical environments. However, due to the space limitations of her office, she cannot fully replicate the situations that patients might encounter at home. Although some patients can get home therapists for real-life IADL training, the OT still would like to understand patients' performance at home so that she can further personalize the clinical training. Finally, she emphasized the importance of patients learning from mistakes. Closely monitored therapy sessions allow patients to perform exercises in a safe environment, and patients are encouraged to learn from their mistakes.

Based on these research findings, we realized that practice at home could be a new opportunity in BLV occupational therapy. OTs currently rely on patients' subjective descriptions for situations outside the office, so more information about patients' performance would be valuable for OTs to customize rehabilitative exercises.

Competitive Analysis

We examined three different products that address the issue of BLV users' adaptation to real-life environments. First, many large eye centers have constructed simulation spaces to provide patients a safe place to practice IADLs. However, such infrastructure is very costly and not flexible enough to reflect the complexity of real-life situations. Another example is the AIRA app, which allows visually impaired people to connect with professional "seers" who provide guidance through the camera stream of the person with blindness/low vision. This is a light-weight solution, but it is very expensive for the patients due to the recruitment of professional workers. It also introduces an inherent privacy risk since the "seers" can see everything the user's phone is pointed at.

Our final concept is inspired by the technology that enables Seeing AI, a Microsoft research project that narrates the visual world around users using machine learning and computer vision. Compared to AIRA, Seeing AI is more affordable; it is also much more adaptable to diverse environments compared to a simulation space. However, just like AIRA, Seeing AI fosters users' reliance on the service rather than training for independence. We aim to create an affordable, lightweight experience for the visually impaired to guide them towards a fully independent future. Most importantly, the large amount of data from environmental recognition generates a greater healthcare opportunity beyond simply providing a "second pair of eyes". It can provide detailed insights about patients' real-life behaviors which are exactly what occupational therapists need but haven't yet been able to obtain. Applying computer vision towards real-life essential skill training represents a novel and promising method to help BLV patients achieve full independence.

Process

We conducted two rounds of user testing to refine the digital interfaces. First, we shared our low-fi prototypes with the OT to validate pillar features. After we incorporated her feedback into our design, we tested new iterations with peer designers to improve usability in information structure and visual design.

Design

Wearable Iris tracker

The design of the wearable Iris tracker was continuously refined to be easy to use and adaptable to different wearing preferences. Swappable magnetic backs mean Iris can be worn on the chest, clipped to a hat or bag, or worn with a lanyard for added security, according to feedback from the occupational therapist we consulted. A single button controls starting/stopping an exercise, moving between tasks, and recording a voice memo during a session. Made of durable polycarbonate and silicone, Iris is lightweight, durable, and tactile.

Behind the Iris tracker's simple design is advanced technology: it incorporates a wide angle RGB camera, a time-of-flight sensor, and motion-sensing gyroscope. These three components work together to enable local computer vision. In other words, Iris uses artificial intelligence to recognize objects, environments, and people and accurately gauges spatial relationships. By using a wide-angle lens, head- and chest- mounting both provide OTs with a first-person perspective of what the patient is facing.

Companion Phone App

We realized early in the prototyping process that patients would benefit from access to their personal rehabilitation statistics and trends. Inspired by fitness tracker ecosystems, we created a phone app that allows users to begin exercises and receive performance feedback when they finish. The phone app provides self-planning and weekly exercise reports, giving patient's greater guardianship over their rehabilitation.

Privacy

Due to the types of sensors incorporated in the Iris wearable tracker and the fact that patients would be bringing Iris into their own homes, privacy is incredibly important. We have designed Iris to meet the most stringent industry standards in the following ways: Information sent via Iris cloud platform is encrypted in transit and storage according to medical privacy regulations. Recordings from the wearable tracker are initiated and ended by the patient, who also positively confirms before uploading.

Audible feedback

We saw that during clinical training, OTs framed mistakes as an opportunity for patients to reinforce skills through their own trial and error. We designed the audio interactions to be encouraging when signaling different phases during a practice session, encouraging users to try techniques until they are able to master them.

Reference

https://www.aota.org/About-Occupational-Therapy/Professionals/PA/Facts/low-vision.aspx

https://innovation.cms.gov/files/migrated-medicare-demonstration-x/lowvisdemo_summary.pdf

https://www.aao.org/assets/0b1836f5-9f53-4478-b322-0217d0fa8510/635543512770730000/vision-rehabilitation-coding-module-pdf?inline1

https://www.special-learning.com/article/cost_of_ot#:~:text=The%20cost%20for%20private%20occupational,look%20around%20and%20compare%20costs