How did we streamline the check-in process at ophthalmology clinics for the visually impaired?
Timeline:
July - September 2023 (3 months)
Team:
1 Research Coordinator, 1 UX Designer (me), 1 Subject Matter Expert
Tools:
Figma, Zoom
0->1 Design Project. Delivered a development-ready prototype
UX Research. Tested the Oculus Kiosk proof-of-concept with visually-impaired individuals; results provided content for research paper
Research Paper. Co-authored a research paper on ontology-based kiosks for ophthalmology clinics for conferences
Disclaimer
All of the concept, research, and original drawings created for the Oculus Kiosk is owned by Camille Dorset (MS, Applied Health Informatics at Fordham University). The Oculus Kiosk was a collaboration project between Camille Dorset (owner of the Oculus Kiosk), John Chelsom/Seven Informatics Ltd., and myself (a contract UX Designer).
Background
In the summer of 2023, I set out to enhance my UX design skills by seeking real-world projects that required creative problem-solving. Leveraging connections established through my school faculty, I reached out to explore potential opportunities.
Luckily enough, John Chelsom, CEO of Seven Informatics Ltd., with whom I had maintained contact over the past two years, knew of a project I could work on for Camille Dorset, one of his Master's students at Fordham University.
Following an initial email exchange, we scheduled a Zoom meeting to discuss the project in detail. That conversation marked the beginning of a rewarding collaboration.
As a health informatics undergrad student at the University of Victoria, my faculty is partnered with Seven Informatics Limited, an Oxford-based startup, where UVic health informatics co-ops frequently work for this start-up. Seven Informatics Ltd. is a health informatics research consultancy that works on a product called cityEHR, their own XML-based electronic health record (EHR) which they implement into clinics and hospitals primarily within the NHS.
As a research consultancy, Seven Informatics Ltd. has various research projects they work on. One of the recent projects was looking into testing the application of the cityEHR within a triaging context for ophthalmology clinics. One of his student researchers had been working on this project had previous proof-of-concept sketches done. These sketches were created based on the care pathway of ophthalmology clinics reported by the student researcher, who is also a SME in this areas as they worked in an ophthalmology clinic as a medical office assistant (MOA).
I was tasked with transforming these proof-of-concept sketches into a prototype that took into consideration UX principles, accessibility guidelines, and UX validation.
To further understand and frame the problem at hand, I set out to explore other American ophthalmology clinic's intake forms to see overlaps and differences between the information they provided. I would review these similarities and differences with the SME via call and verify the rationale behind including certain information and determining whether certain information was mandatory, optional, or unnecessary. A competitive analysis of 11 different intake forms was completed.
When starting this project, one of the biggest things that excited me was the medium I was designing on: a kiosk. I have had no experience designing for kiosk interfaces at this point. I went out to explore and do field work to study the UI layout of different types of self-service kiosks (SSKs) in the real world. These included kiosks in fast food chains (ie. McDonald's), self check-in kiosks in hotels, etc.
After playing around with some self-service kiosks in a real-world context, I was able to take note of a few thing that were common in many of these kiosks that could be taken into consideration when designing the Oculus Kiosk. Key takeaways regarding UI for navigation, touchpoint sizes, and element placement considerations were summarized.
I had various calls with the SME on the team regarding the check-in flow for a new and returning patient at the ophthalmology clinic they worked at, how patients' conditions are scored for triage, how insurance works in American ophthalmology clinics, along with general struggles patients faced when checking into their clinic. A few insights from these interviews are shown below:
Through SME consultations and understanding the context of American ophthalmology clinical check-in flows better, a persona emerged: the patient.
The SME and I discussed the patient intake process at an ophthalmology clinic. The main bottleneck occurred during the 15-20 minute intake process, exacerbated by the paper-based form system, which wasn’t fully accessible for all patients. If help was needed, clinic staff had to assist, involving an additional step. One key anecdote shared by the SME significantly influenced the design decisions for the Oculus Kiosk.
“There was a patient who came in to our clinic this week who tried to complete our new patient packet, but they had left rushed out of the clinic [crying] as they were frustrated they were unable to fill out the form. The patient happened to be fingerless.” - Ophthalmology clinic MOA
In addition to usability testing with participants with visual impairments, I used a low vision simulator in Figma to evaluate design decisions like larger target sizes, ensuring they were effective across conditions such as blurred vision, hemianopia, blind spots, retinal detachment, and central or peripheral vision loss.
Due to conference paper submission deadlines, usability testing was delayed until the 4th iteration, with much of the earlier testing conducted via lower-cost accessibility audits and SME consultations.
Blurred Vision
Retinal Detachment
Central Vision Loss
Furthermore, I used a WCAG contrast checker to verify if body text, header text, and UI components such as buttons and tiles were WCAG-compliant.
I recruited 5 subjects for usability testing based on Nielsen's Law of Small Numbers. There were three main flows they were tasked to complete:
Checking into a scheduled appointment.
Checking into a walk-in appointment as a new patient.
Checking into a walk-in appointment as a returning patient.
Recruitment aimed to represent the most common eye conditions in the general population. This happened to be refractive errors: myopia (~27%), hyperopia (~10-20%), astigmatism (~33%), and presbyopia (~100% over age 40).
Furthermore, refractive errors account for the largest portion of ophthalmic visits for various regions: 21.1% in Germany, 22% in Saudi Arabia, 27.5% in Southern Ethiopia, and 55.56% in a Northeast India study. Recruitment aimed to ensure a diverse range of eye conditions to reflect the general population. Below is the distribution of eye conditions among the test subjects.
After usability testing, participants completed a System Usability Score (SUS) survey, achieving a final SUS score of 87.5 in the fourth iteration, reflecting exceptional user-friendliness and effortless task completion. Despite a high SUS score, verbal feedback and think-aloud observations provided key insights for implementing new features.
TESTING & VALIDATION
This a summary of the feedback and comments made by test participants:
Strengths:
Effective usage of icons and visuals to represent conditions
Large target sizes helped users interact with the UI better
Easy-to-find entry points for each check-in situation
Weaknesses:
Contains jargon, though all users stated they would not need technical assistance while using the app, they said they would need assistance understanding terminology
Lengthy onboarding process for identification; especially for returning patients
Confusing form layout
New Changes:
Card scanner for health insurance card for streamlining onboarding/identification processes
Higher contrast borders around inputs, dividers, and tiles
Optional steps for steps that contain information that is often no ready at hand (ex. prescription)
Removal of inputs that can be auto-filled using health insurance number
Clearer disclaimers and consent forms
Confirmation pages
DESIGN
Referencing the SME’s anecdote of the patient without any fingers, larger target sizes were a priority in the design of the Oculus Kiosk. The digitization of the intake process at ophthalmology clinics would take away the struggle of holding a pencil for physically or cognitively-impaired patients. Pairing the digital touch-based interface with larger target sizes of buttons and UI elements, patients with a stubbed limb would be able to click buttons on the kiosk. Referencing the WCAG 2.1 guidelines many of the UI elements exceed the recommended minimum of 44 x 44 pixels.
Readability
Sans-Serif. I used two sans-serif fonts: Cal Sans for titles, buttons, and tile text, and Manrope for labels, placeholders, and body text. Sans-serif fonts reduce additional strokes that could impair readability.
Text Size. For better glanceable reading, I used bolded, larger text and wider line spacing, enhancing lexical decision-making (Nielsen Norman Group). The APH ConnectCenter states a common guideline for accommodating text for users with low vision is to have text be at least 18 points (24px).
Sentence Case. The decision to use sentence case improves legibility for users with visual impairments. It capitalizes the first letter of each sentence, enhancing word legibility and lexical decisions. By varying word shapes (e.g., upper vs. lower case), sentence case offers more visual context than title case, which appears as a block of text, reducing readability for those with blurred vision or farsightedness. Design systems like Google's Material Design support sentence case to improve accessibility.
In order to accommodate visually-impaired patients, I conducted usability tests with a colour-blind peer in order to get feedback on the difficulties regarding the current colour palette used for the Oculus Kiosk. Furthermore, WCAG 2.1 guidelines for Colour Contrast were referenced in order to align brand colours with accessibility needs. The colours used for the Oculus Kiosk aimed to have high contrast for all users regardless of if they had high visual acuity, visual impairment, colour blindness, etc.
During my interview with a subject matter expert (SME), a medical office assistant (MOA) at a New York City ophthalmology clinic, they highlighted the importance of clear visuals and simplifying complex ophthalmological jargon. Most patients are unfamiliar with clinical terminology, so using large images and icons helps visually impaired patients recognize information more easily. To enhance understanding, visuals were created for specific ocular and medical conditions, and condition names were simplified to aid quick identification.
Notably, 80% of usability testing participants viewed the icons as helpful secondary aids, using them to confirm their choice after first referring to the text. These icons were paired with a dark yellow (#FFB81C) background, known to be an accessible color for colorblind patients.
Ergonomic Hardware Experience
40% of usability test participants expressed a preference for scanning their health insurance card over manually entering personal and insurance details. This feature also improves accessibility for physically impaired patients, such as amputees. Implementing this functionality would require integrating a card scanner into the kiosk or tablet hardware. To complement the new hardware features, an intuitive onboarding experience was designed to guide users through the card scanning process.
The instructional graphics should be self-explanatory and provide users with system feedback and progress of their card scan. There are 3 distinct phases with 1 edge case for errors:
Enter your card
Scanning your card
Remove your card
Card error
Submitted to Springer for publication, this paper was also accepted for the HIMSS24 Conference. Check again shortly to download the paper once published.
DOWNLOAD AVAILABLE SOON!
The progression of the designs can be seen below:
Sketches
Sprint I Prototype
The design is nowhere near perfect or complete but due to timelines, we have left it as is and submitted the research paper work to conferences. Next steps for this project would include:
Development. Implementing this into a functional coded application integrated with the cityEHR. This will allow real-life clinical testing to be achieved sooner.
Testing with Blind Users. Evaluating accessibility for blind users and those with severe visual impairments is crucial, as it requires advanced accessibility accommodations and consideration of complex edge cases, such as screen reader integration, keyboard navigation, and form accessibility.
Testing with Users with Lower Tech Literacy. Placeholder.
Implementation into Clinic Setting. Once development is finished, using this as a real-life application in one of Seven Informatics' partnered clinic would allow more bottlenecks to be identified in everyday usage.
Internationalization. The cityEHR is an internationalized system, currently supporting Mandarin and Spanish with features that go beyond language translation, adapting the UI to align with EHR conventions specific to these regions.
Whoever is in charge of this project in the future should also further iterate the designs based on user validation via testing with various groups of patients who would use this application.
REFLECTION
Key Takeaways
Accessibility
Accessibility was a key consideration for all design decisions for this application. Most accessibility considerations were geared toward visually-impaired patients and issues such as color contrast and target size were emphasized in this project. This project has also taught me that accessible design is better design for everyone.
Considering Hardware Elements
The kiosk in nature is a hardware device. If a tablet is used, it can be connected to other hardware devices such as printers or card scanners. When designing a kiosk software, it is just as important to consider how the UX can be improved using the associated hardware devices.
Environmental Considerations
It was important to consider environmental elements to the UX of the Oculus Kiosk. For example, if a user is checking in and prompted to fill out their substance use habits, they would likely prefer to do this in a more private setting. One way to address this would be to host the Oculus Kiosk on a clinic tablet that they can take to their seat rather than a large kiosk machine that many people in the queue may be able to view.
2025
William Lee