AI-enhanced Diagnostic Software for Convergence Insufficiency

Convergence insufficiency (CI) is a prevalent binocular vision disorder characterized by the eyes' inability to maintain alignment during close tasks like reading, leading to symptoms such as eye strain, blurred or double vision, headaches, and difficulty focusing. This condition affects approximately 7.5% of the global population, disproportionately impacting children (5%) and one in six adults, yet it remains largely undiagnosed.

The primary challenge with CI is its underdiagnosis, often due to standard vision screenings' inability to assess binocular vision issues. These screenings typically overlook CI, mistaking its symptoms for other problems like learning disabilities, which can significantly affect educational and professional development.

Our solution addresses this by enhancing the detection of CI through advanced AI-enhanced diagnostic software. This tool aims to improve diagnosis rates, ensuring timely and appropriate interventions, including vision therapy and surgical options if necessary. By improving early detection and treatment, we aim to mitigate CI's impact on academic performance and long-term career prospects, enabling affected individuals to achieve their full potential.

The software would work as follows (the pseudocode, if you will):

1. Ask the user for any history of convergence insufficiency symptoms

2. Analyze the responses using AI algorithms to determine a relation to CI

3. Record the user performing binocular eye exercises (e.g a convergence test) from the user’s device camera

4. Use conventional computer vision or convolutional neural network (CNN) models to analyze the captured eye movement, assessing the relation to CI

5. Give an overall assessment of likelihood of CI

Young students who have convergence insufficiency experience difficulty with reading and school if left untreated, and this applies also to working adults who engage in similarly demanding vision tasks. Our software would be made accessible to any child, youth, or adult with a personal device, allowing students and working professionals to find out at an early stage if they have CI, enable them to seek care, and improve the productivity of their academic and working lives. 

Medical professionals in vision clinics can use the software to improve diagnostic accuracy and efficiency, providing a valuable tool for optometrists and ophthalmologists.

Finally, individuals in areas without access to specialized vision care can benefit from more readily available diagnostic tools, which can be used in a primary care setting or through remote consultations.

As a leader deeply rooted in the convergence insufficiency (CI) community, my personal journey underscores my commitment to this cause. From a young age, I struggled with reading and visual tasks, compensating with head turns without understanding the need for proper eye coordination. It wasn't until my freshman year at the University of Michigan, after years of academic challenges, that I was diagnosed with CI. This late diagnosis, a result of widespread information asymmetry and limited public awareness, could have been avoided.

The support and treatment I received at the University of Michigan Kellogg Eye Center were transformative, inspiring me to ensure others do not face similar delays. My firsthand experience with CI—from the initial symptoms to the diagnostic processes and treatment options—has equipped me with valuable insights into the everyday realities of those affected by this condition. This personal experience is complemented by the technical expertise of my teammate, Adarsh, a dedicated computer scientist driven by a passion to apply technology to solve prevalent health issues. I, too, am an engineer who wants to leverage technology to do that. 

Our team's technical aptitude and direct connection to the CI community and the Kellogg Eye Center position us uniquely to develop a solution that is not only technically sound but also deeply informed by community needs. We are committed to incorporating feedback directly from those affected by CI to ensure our solution genuinely addresses the challenges they face. This approach guarantees that our project is guided by the community’s needs, and is also an authentic representation of their experiences and aspirations.

Our team has only recently committed to the project, so it exists as a concept currently.

The initiative's commitment to harnessing technology for tackling significant social challenges resonates deeply with our mission. Beyond the financial support, the coaching and development programs will significantly enhance our team's leadership skills. Given the delicate nature of working with doctors and handling patient data, we are particularly keen on leveraging the initiative's and MIT's network to facilitate crucial medical and technological collaborations. 

Our solution integrates modern AI with traditional diagnostic methods, enhancing eye care by making diagnoses faster, more accurate, and more accessible using devices most people already own such as smartphones and computers. This approach contrasts with others like VR-based methods that require specialized equipment and that have not taken advantage of the capabilities of AI (efficiency, accuracy, and scalability). The widespread adoption of our AI tool could set new standards in eye care, enabling earlier detection and intervention for conditions like CI, mitigating learning or work difficulties. Additionally, the continuous application of AI diagnostics produces valuable data, improving our understanding and treatment of CI. This shift towards preventive care, emphasized by our software, could influence broader health sectors to integrate AI, significantly improving healthcare delivery and patient outcomes.

Our software would allow users to screen for convergence insufficiency from the comfort of their device, powered by the capabilities of cutting-edge AI. Our immediate goals are to make the software known and accessible to education institutions (K-12 schools and universities), companies, vision clinics, and individuals by strategic partnerships and media efforts; and thus, due to widespread accessibility, we expect to see higher rates of CI diagnoses than current statistics indicate. In the long-term, this enables people with CI – armed with the knowledge that they possess the condition – to start seeking appropriate treatment as early as possible; improves the quality of life of CI patients; increases the diagnostic accuracy of CI for doctors in vision clinics; and ultimately leads to a reduction in the underdiagnosed nature of CI (which is the biggest assumption in our theory of change).

1. Decrease the underdiagnosis of convergence insufficiency

• Indicators: increase in diagnosis rates, feedback from doctors on patient outcomes

2. Increase public awareness of convergence insufficiency.

• Indicators: the number of schools, companies, vision clinics, and individuals who use our software; the amount of media attention that the software receives.

3. Increase accessibility to CI diagnostic tools

• Indicators: user adoption rates, cost-effectiveness of using software

4. Improve quality of life of CI patients

• Indicators: changes in academic and occupation performance and performance otherwise in individuals with CI diagnosed via the software

Our diagnostic software uses computer vision technology to analyze eye alignment and movement through a device camera, identifying signs of convergence insufficiency. It employs AI and machine learning algorithms, trained on extensive datasets of eye movement, to differentiate between normal and impaired convergence.

Advanced data analytics process and interpret the diagnostic data, providing healthcare providers with detailed insights into a patient’s eye movements. In addition, our software would feature an inclusive UI/UX design, making it accessible and easy to use. This allows both patients to conduct self-tests and view their results, and clinicians to efficiently navigate diagnostic tools and manage patient data.

Myself + Adarsh Pettapa (Computer Science Student, University of Michigan)

In total: 2 people work on the solution

I also have a strong relationship with an Associate Research Scientist at U-M who helps me to brainstorm ideas, direct me along helpful paths, and offers his decades wisdom in engineering design. He does not produce any technical material for nor does he work directly on the solution. But he is a helpful resource. 

We have just started working on our solution, so the number of years is currently zero. 

Currently, on the leadership team there exists just Adarsh and I – two international students in the United States – with Adarsh hailing from India and I from New Zealand. We value diversity and will ensure that our hires going forward will represent the diversity we, MIT, and the wider community desire. We plan to learn more about DEI through engagement in workshops and continuous feedback mechanisms with our stakeholders to inform our efforts in promoting and celebrating diversity.

Our value proposition centers on an AI-enhanced diagnostic software for convergence insufficiency, known for its high accuracy and efficiency. This tool provides significant benefits for the following groups: 

1. Individuals: Early and precise diagnosis facilitates effective treatment plans, reducing symptoms and enhancing quality of life.

2. Healthcare Providers: It offers a dependable diagnostic tool that integrates seamlessly into existing workflows, boosts diagnostic capabilities, and promotes better patient outcomes.

3. Educational Institutions: The software identifies students who may benefit from vision therapy, potentially boosting their academic performance and enriching their educational experience.

Our primary customers include optometrists, ophthalmologists, and other eye care specialists. Secondary customers encompass schools, pediatricians, and general practitioners who might utilize the software for preliminary screenings. The main beneficiaries are individuals, both children and adults, who are suspected of having or are at risk of CI.

We would distribute the software directly to healthcare professionals, institutions, and individuals. Our marketing strategy includes online advertising, participation in medical conferences, and forming partnerships with educational, healthcare organizations, and possibly government.

Our revenue will be generated by the periodic subscription fees that we would charge organizations and individuals, which would bring in recurring revenue throughout our operations. Large sums of initial capital could be generated through the one-time licensing fees that we would charge large organizations, etc., who would have perpetual use of the software. 

Beyond that, revenue could be generated also through service contracts to government as part of large public health initiatives, especially in school health programs to monitor student vision health. In combination, forming partnerships with healthcare technology firms and education technology companies can open up additional revenue streams by integrating the software into existing platforms and systems, thereby extending our market reach.