Hypertensight

Hypertensight

Hypertensive Retinopathy (HR) is a rare eye condition in which high blood pressure (hypertension) damages the layer of tissue at the back of the retina; It can lead to irreversible blindness if not diagnosed and treated in time. 

The diagnosis of hypertensive retinopathy (HR) faces significant obstacles due to the lack of efficient and accurate diagnostic methods and tools, which impedes early detection and effective treatment. According to a study by the American Academy of Ophthalmology, even though the percentage varies by demographic groups, 2-17% people with hypertension and no comorbidities are expected to be diagnosed with hypertensive retinopathy. This translates to about 22.6 million to 198.88 million people worldwide. Simply put, out of all people in the world, approximately 0.283%-2.486% have HR. 

The manual detection of HR involves analyzing fundus images to identify subtle signs of the disease. This process is not only time-consuming but also susceptible to human error, especially in the early stages when symptoms are less visible and harder to detect; It is in fact proven that signs of HR develop at much later stages of the disease. Timely detection is critical for effective intervention, yet the limitations of human vision often result in missed detection of early-stage symptoms, leading to inaccurate and delayed diagnoses. By the time the disease is detected, it may have progressed significantly, reducing the effectiveness of treatment and increasing the risk of severe complications.

Additionally, the extensive time required for doctors to examine these images often necessitates consultations with other ophthalmologists and comparisons with other cases. These factors can create a significant backlog of patient images exacerbated by the sheer workload on doctors, further delaying diagnosis and treatment.

In low-resource areas, the problem is compounded by the lack of awareness that hypertension can impact the retina and the absence of specialist doctors available to help. This results in many cases going undiagnosed and untreated altogether, further contributing to the burden of HR. This underscores how there is a paucity of accessible resources for the screening of this disease in such settings.

Therefore, there is an urgent need for accessible, precise, and rapid diagnostic tools to detect hypertensive retinopathy. Such tools would enhance the accuracy, efficiency, and overall quality of care in diagnosing and managing the condition, ultimately improving patient outcomes and reducing ‘the burden’ of this sight-threatening condition, as emphasised by another study published in the Indian Journal of Clinical and Experimental Ophthalmology that suggests taking ‘adequate measures’ to do so. 

Our solution, Hypertensight, marks a significant advancement as the first autonomous machine learning system designed specifically for early detection and diagnosis of hypertensive retinopathy from retinal images. Tailored to be integrated into ophthalmological clinical practices, Hypertensight employs AI backed technologies to substantially elevate the efficiency and accuracy of HR detection, especially during pathological examinations when the disease might be in its nascent stages. By delivering rapid automated diagnostic reports, Hypertensight equips healthcare providers with timely insights, facilitating prompt intervention and management to potentially mitigate the risk of vision loss associated with this condition.

Hypertensight employs a sophisticated AI-driven methodology that enhances digital images of the retina through a meticulous multi-step process. This process encompasses image resizing, extraction of the green channel, and application of Contrast Limited Adaptive Histogram Equalization (CLAHE) to optimize contrast and prepare images for detailed analysis. Central to the algorithm is YOLO v8, a robust Convolutional Neural Network (CNN) that is trained on extensive datasets of over 5000 annotated retinal images of HR patients at different stages of the disease as well as patients without HR. YOLO v8 excels in identifying subtle indicators of HR with exceptional sensitivity and specificity, including early signs of clinical features like arteriolar narrowing, hemorrhages, cotton wool spots, and optic disc changes.

Healthcare providers utilizing Hypertensight receive immediate feedback on the presence and severity of HR indicators in analyzed images. The system generates automated reports that succinctly summarize findings and provides standardized severity assessments. This streamlined process not only enhances diagnostic precision but also empowers clinicians to make well-informed decisions for a plethora of patients promptly, thereby optimizing patient care and potentially improving outcomes for hypertensive individuals at risk of vision impairment due to hypertensive retinopathy.

______

Green Channel extraction: Green channel extraction involves isolating and using only the green component of an image, which in fundus images, enhances the visibility of important features like blood vessels and lesions due to its higher contrast and better detail.

CLAHE:  CLAHE (Contrast Limited Adaptive Histogram Equalization) further enhances contrast in fundus images by adjusting image intensity levels, improving visibility of subtle details such as retinal structures and abnormalities, crucial for accurate medical diagnosis.

The 5000 fundus images: These fundus images are sourced from Shanggong Medical Technology Co., Ltd. that has compiled data from different hospitals/medical centers in China.

Hypertensight serves all hypertensive patients at risk of developing or having undiagnosed hypertensive retinopathy (HR). It is designed for healthcare professionals, including ophthalmologists, optometrists, but especially other general physicians in low resource settings involved in managing hypertensive patients. By integrating Hypertensight into their clinical practices, these professionals can significantly enhance their diagnostic capabilities and improve outcomes.

Specifically, Hypertensight targets hypertensive patients over 60 years-old, as they are at a heightened risk for HR, with the condition statistically more prevalent in men. Further, studies have shown that patients with hypertension for over five years are at a greater risk of HR, with incidences reported at 77.03% and 63.64% in two separate studies. Therefore, we aim to make Hypertensight easily available to patients with prolonged hypertension.

We also focus on certain racial groups within which HR cases is seen to be concentrated. Africans and Indians, for instance, face higher prevalence rates of hypertension and, consequently, HR. For instance, a study by Kabedi et al. found that 83.6% of hypertensive patients in Africa had HR. In India, the Department of Ophthalmology at Jawaharlal Nehru Medical College found that 49.33% of hypertensive patients had HR, with states like West Bengal showing prevalence rates of 62.25%.

As we have discussed the scarcity of accessible tools, particularly in rural low-resource settings, it is also important to highlight the diabolical fact that individuals in these areas, such as blue-collar workers, suffer from higher rates of hypertension compared to those in urban settings. To address this, we aim to make our screening method more accessible to these populations. On a slightly related note, we soon plan to develop a portable, inexpensive, and non-mydriatic fundus camera specifically designed for use in such settings. 

Lastly, we place special attention on hypertensive patients likely to have grade-1 or grade-2 HR, as these are the most common stages and the hardest to detect manually. Studies indicate that 76.35% of HR patients have grade-1 or grade-2 HR, and since detecting HR at these stages is crucial to prevent irreversible blindness, Hypertensight aims to reach this demographic of patients to save them from potentially life-long blindness. 

Currently, all these individuals are potentially underserved due to the complexity and time-consuming nature of diagnosing hypertensive retinopathy. The diagnostic process is error-prone and lengthy, causing significant suffering for patients who may experience delays in receiving crucial interventions. Moreover, individuals in low-resource areas often lack access to the necessary tools and knowledge to even diagnose the issue, leaving them at a significantly more pronounced disadvantage.

Hypertensight addresses these challenges by enhancing diagnostic accuracy and efficiency for healthcare professionals. This proactive approach allows for timely interventions, including treatment protocols that are crucial for preventing the progression of HR to severe vision impairment. By streamlining the diagnostic process and reducing reliance on specialised resources and manual assessments prone to human error, Hypertensight aims to improve overall health outcomes and quality of life for hypertensive patients vulnerable to HR.

We are currently in the pilot stage of development. Our solution, Hypertensight, has been launched in multiple communities and is actively being tested and iterated upon. Here are the key highlights of our progress:

1. Partnership with Leading Eye Clinics and Hospitals: We have partnered with Krishna Netralaya, a renowned eye clinic where Dr. Rajesh Rastogi, the lead surgeon, has endorsed our solution. Dr. Rastogi has not only advocated for Hypertensight but has also implemented it for patient care, allowing us to gather crucial feedback and data to refine our system.

2. Collaboration with Eye Care Equality: We have formed a partnership with Eye Care Equality, an organization dedicated to providing eye care support to blue-collar workers in India. This demographic often lacks awareness of hypertensive retinopathy and is at higher risk due to occupational stress and lifestyle factors. We have signed a contract with Eye Care Equality, and our first awareness and screening session is scheduled for July 8th. We have arranged for a fundus camera via a local charitable clinic and are accompanied by an ophthalmologist who the patients through Hypertensight would be referred to for further treatment and prognosis. (However, as mentioned earlier, our future plans include creating our own inexpensive, portable, and non-mydriatic fundus camera. We have the plans ready and would be able to produce them soon if provided with the prize grant). This initiative will help us reach and educate a vulnerable population, further validating our solution.

3. Outreach to the Delhi Ophthalmological Society (DOS): We have submitted an article to the prestigious Delhi Ophthalmological Society, which is currently under review. Once published, this article will reach thousands of ophthalmologists, raising awareness about hypertensive retinopathy and promoting our innovative diagnostic tool. This exposure is expected to generate interest and support from the medical community, facilitating broader adoption of Hypertensight.

4. Funding and Revenue Streams: While we are currently bootstrapped, we are actively deliberating on sustainable revenue streams to support our operations in the future. We are exploring various funding options to ensure the long-term viability and scalability of our solution.

The use of advanced AI and image processing techniques is what sets Hypertensight apart as a truly innovative solution in its approach to addressing the challenges of hypertensive retinopathy detection 

Central to Hypertensight's innovation is its integration and combination of sophisticated image processing technologies, notably green channel extraction and Contrast Limited Adaptive Histogram Equalization (CLAHE). These techniques enhances image contrast, and by improving the clarity and detail of retinal images, Hypertensight enables healthcare providers to identify signs of the condition more effectively than traditional methods allow. YOLO v8, a robust CNN that employs a more efficient and accurate backbone architecture (CSPNet) than seen previously, then excels in identifying diverse indicators of HR with exceptional sensitivity and specificity. Unlike manual inspection, which may overlook or misinterpret subtle signs, Hypertensight's AI-driven approach ensures comprehensive evaluation of retinal health, detecting subtle abnormalities such as arteriolar narrowing, hemorrhages, cotton wool spots, and optic disc changes reliably and swiftly.

This automation alone represents a significant innovation, as it reduces dependency on human interpretation, minimizes variability in diagnosis, and accelerates the delivery of diagnostic reports. Moreover, there have been no substantial or practical attempts at automating the process for the detection and early diagnosis of HR. In 2022, when we first conceived this idea, there was a severe paucity of studies on the use of AI to detect HR. However, in recent times, a few attempts (around 2-3) have surfaced. Yet, none of these have yielded tangible solutions adopted by doctors. This is because they are not only just theoretical proposals without practical implementation in clinical workflows, but they also remain unviable due to the lack of crucial image optimization techniques, resulting in inaccurate diagnostics. The clinical features in the images in these attempts are not clear enough, leading to unreliable results. Additionally, these attempts have not addressed the imperative need of real-time feedback and comprehensive reporting, further limiting their practical utility in busy clinical environments where rapid screening for many hypertensive patients is required. 

Talking about real-time feedback, Hypertensight represents a leap forward in clinical practice by providing rapid, automated diagnostic reports. These reports summarize findings, offer standardized severity assessments and recommend further evaluation when warranted. This streamlined approach not only enhances diagnostic precision but also empowers healthcare providers to initiate timely interventions.

Impacting the market, Hypertensight not only advances HR detection but also paves the way for automating the diagnosis of other rare retinopathies and diseases conventionally dependent on manual inspection, enhancing overall healthcare efficiency. This will ultimately improve outcomes across a range of ocular diseases, such as the rare thyroid eye disease, which impacts about 19 per 100,000 people.

Furthermore, the HR healthcare industry could be transformed, and the need for specialist doctors with years of experience in detecting HR would be reduced. This democratization of diagnostic capabilities means that even in areas with limited access to specialized ophthalmic care, patients can still receive accurate diagnoses and appropriate care by doctors who are not as trained in the field of HR. 

We are applying to The Amgen Prize because we believe that our solution has the potential to make a significant impact on global health outcomes. However, we do recognize that there are still barriers we need to overcome to bring our solution to scale.

A critical barrier we face is our AI algorithm, which is currently held back by the limited size of our dataset due to the rare nature of the disease. At present, we have only about 5,000 fundus images, split between undiagnosed and diagnosed eyes with hypertensive retinopathy. Given that around 110 million people suffer from the condition, our dataset is simply too small to effectively train a model designed to serve millions.

This limited dataset severely restricts our ability to ensure the model's accuracy, including the prevention of cases of dangerous false negatives. In critical conditions like hypertensive retinopathy, which can lead to irreversible blindness, the stakes are incredibly high. Therefore, it is imperative to increase our dataset until the algorithm consistently delivers highly accurate results. 

Specifically, a challenge we are facing has to do with the new images that are being provided by doctors globally. Ophthalmologists are now using Scanning Laser Ophthalmoscopy (SLO) to capture fundus images, resulting in rectangular images instead of the typical circular ones. Consequently, our algorithm occasionally struggles to produce accurate results when encountering these due to the absence of SLO-exported images in our training dataset. 

Additionally, we have observed inaccuracies in the algorithm's results for some images, again due to the model's insufficient exposure to a diverse range of images. As high school students, we aim to utilise the funding in R&D from The Amgen Prize wisely; we plan to hire experienced data scientists and AI specialists to curate a strong algorithm that goes beyond industry standards so that it can cater to hypertensive retinopathy patients accurately regardless of image aspect ratio and other factors. Therefore, the funding would also go into sourcing thousands of SLO-exported and conventional images from research medical institutions and feeding them into the algorithm to train/develop it further.

More importantly, we will use the funding to create our very own inexpensive, handheld, and non-mydriatic fundus cameras. These will be sent to every low-resource rural clinic in various communities that lack specialized doctors. Since general physicians in these clinics might not be aware of the complexities of ophthalmology such as the dilation of the eye, we hope to make our camera and software available to these areas so that thousands of hypertensive patients can be screened in these areas for hypertensive retinopathy without specialized resources. 

Furthermore, the public validity and the funds could help us navigate through complex legalities; such as getting approvals from drug authorities like the FDA (United States) and CDSCO (India).

Our team is exceptionally well-positioned to deliver Hypertensight due to our diverse skill sets, personal commitment, and direct engagement with the healthcare community. Our journey began at the prestigious BITS Pilani's Young Entrepreneurs’ Bootcamp (YEB) in December 2022. Both co-founders were focused on ideating a solution for the healthcare industry at BITS Pilani, one of the leading technical institutions in India. That's when we came up with the idea of addressing hypertensive retinopathy, a rare yet widespread ocular disease. Our concept won the runner-up prize from the four biggest venture capitals judging the competition, highlighting the potential of our solution.

Initially, we did not solidify our idea into a tangible product. However, a significant encounter with a Chinese patient living in the States reignited our determination. This patient was misdiagnosed with HR after two tests, one of which included fluorescein angiography, an invasive process. After seeing no improvement with medication, further tests revealed he had central retinal vein occlusion, not HR. This experience underscored the critical need for accurate and time efficient diagnostic tools. If doctors could misdiagnose him as positive when he was negative, it is just as  likely that there are cases of false negative diagnoses as well.

Motivated by this, we started building a prototype and began seeking feedback eight months ago. After multiple iterations in design and refining the AI model, we formally began developing the solution two months ago. We have since administered Hypertensight to three clinics and eye hospitals in Delhi and Gurgaon and are looking to expand further, contingent on securing funding to increase the size of our dataset by hiring research specialists, contacting medical institutions, and hiring data scientists to achieve our goal of 100% accuracy, as well as manufacturing and administering multiple units of our non-mydriatic fundus camera. 

Our team comprises members with distinct but complementary perspectives and skill sets. Sarthak Ahuja, with his background in AI and image processing, focuses on precision, efficiency, and innovation in our diagnostic tools. Sarthak has completed an AI/ML course at the National University of Singapore (NUS), where he worked with various algorithms. Additionally, he interned with AWS, leading a team of students to create a healthcare chatbot using tools like Amazon Lex. This chatbot provides diagnoses based on symptoms and helps book appointments with the nearest doctors in Singapore. 

Suhana Grewal, with her strong background in business and finance, provides critical insights into the operational and strategic aspects of Hypertensight. She has completed a course called Financial Markets from Yale University, adding to her understanding of the field. Her focus on business development and finance is crucial for scaling our solution and forming strategic partnerships. Her emphasis on usability ensures our technology addresses practical challenges in clinical settings and remains accessible to healthcare providers with varying levels of technological proficiency.

We actively seek feedback from the communities we aim to serve, incorporating their inputs and ideas into our development process. This iterative engagement ensures our solution is technically sound and practically valuable.  

For the near future, our solution visualizes a dual approach: Our software (Hypertensight) being used everywhere, including clinics and hospitals, in various communities, helping a plethora of hypertensive patients get early diagnosis of hypertensive retinopathy to prevent blindness and vision loss. Simultaneously, we also plan to create our own inexpensive, handheld, and non-mydriatic handheld fundus camera based on indirect ophthalmology to be used in low-resource rural settings where fundus cameras are not readily available, and specialists who can perform complex ophthalmic tasks such as dilating the eye to capture images are also rare. 

Secondly, we aim to expand its adoption among healthcare providers by refining its diagnostic accuracy. We aim to integrate Hypertensight into the workflows of at least 20 ophthalmology practices and optometry clinics in India, ensuring its usability and acceptance through feedback mechanisms and making sure it positively impacts over 5000 patients with hypertensive retinopathy. Concurrently, we will enhance the software's diagnostic algorithms based on real-world performance metrics and clinical validation, aiming to improve sensitivity, specificity, and overall diagnostic reliability. To achieve these goals, we will conduct targeted outreach and training sessions, collaborate with key opinion leaders, and establish partnerships with healthcare networks. We will measure progress through specific indicators such as the number of doctors, user engagement metrics, feedback scores, and improvements in diagnostic performance parameters.

Looking ahead to the next five years, our goals for Hypertensight include widespread global adoption and measurable impacts on patient outcomes. We envision Hypertensight being integrated into the standard practice of at least 200 healthcare institutions worldwide (many in rural areas as well), supported by regulatory clearances and certifications in key markets. Through longitudinal studies and clinical outcome data, we seek to demonstrate significant improvements in patient outcomes related to hypertensive retinopathy, such as reduced progression to vision-threatening stages and better management of retinal complications. Numerically, we aim to transform the lives of at least 100,000 patients with hypertensive retinopathy in this time-frame. 

Simultaneously, we also plan to manufacture and administer more than 1500 cameras to various rural areas across the world, starting with our own country which also happens to be heading towards being claimed as the "Hypertension capital of the world". Achieving these goals will involve continuous investment in research and development to advance Hypertensight's capabilities, including expanding its dataset for training, integrating new AI models, and enhancing features for comprehensive retinal health assessment. We will measure success through metrics including adoption rates, clinical efficacy data, user feedback scores, and regulatory milestones, ensuring that Hypertensight not only meets but exceeds expectations in delivering impactful solutions for the early detection and management of hypertensive retinopathy globally. Moreover, an added advantage would've been achieved if Hypertensight would be able to spread awareness about the harmful affects of hypertension on the retina as well among blue-collar workers and those living in rural areas.

Hypertensight is poised to revolutionize the management of hypertensive retinopathy by transforming how early detection is conducted, making it more efficient, accessible and accurate. Our solution aims to drastically reduce the incidence of irreversible blindness caused due to hypertensive retinopathy by providing accurate and time-efficient diagnosis. 

Our automated approach has and will continue to enable doctors to receive quick and reliable diagnostic results without the need for extensive manual analysis. The rapid feedback provided by Hypertensight enables healthcare providers to initiate treatment plans sooner, significantly reducing the risk of vision loss. By delivering immediate and accurate diagnoses, Hypertensight ensures that patients receive timely and effective care, which is crucial for managing HR effectively. 

Hypertensight's ability to consistently produce high-quality diagnostic reports consistently helps to standardize the detection process, minimizing variability in diagnosis and ensuring that all patients receive a uniform standard of care. This consistency is particularly important in busy clinical settings where time and resources are often limited. By improving diagnostic accuracy and efficiency, Hypertensight not only enhances patient outcomes but also alleviates the workload on healthcare providers.

How our solution will impact the problem and the outcomes for our target population is quite similar to how Google’s Deepmind AI brought about a drastic change in the detection of diabetic retinopathy, which to an extent is similar to HR but is caused due to consistently high levels of blood glucose (diabetes). Diabetic retinopathy is a significantly more common ocular disease, wherein 12% of all cases of global blindness are caused by it. 

Since the process of detecting diabetic retinopathy is extremely time consuming and prone to human errors too, especially in early stages, Google stepped in with its software acknowledging that "India alone has 70 million diabetics who must be screened, and there just aren’t enough trained clinicians to review their retinal scans.’ and that ‘it’s not humanly possible to screen these 70 million.' Google also emphasized how ‘artificial intelligence could help make diagnosing diabetic retinopathy easier by accurately interpreting retinal scans, perhaps the eyesight of millions could be saved."

6 years after the release of Google’s Deepmind, there was a report by the National Center for Biotechnological Information highlighting how “the accuracy of Google’s AI diagnosis was better than that of ophthalmologists” and that “the work efficiency is much higher than that of human doctors”. This underscores the potential of Hypertensight in the detection of HR and validates what we aim to achieve in the near future with the power of AI and machine learning. 

Initial implementation data from pilot studies of our solution reinforce our vision. All current users, majorly ophthalmologists, attest to the tool's efficacy and its significant contribution to their clinical practice; Dr. Rajesh Rastogi, a senior eye surgeon, claimed that "The system will not only provide a second opinion to specialist doctors but can provide better diagnosis accuracy with time efficiency. This system can be used specially in emergency treatment by physicians and neurologists in low-resource settings, especially the charitable hospital I visit often".

A key reference that underscores the potential of our technology is a research paper published in the MDPI journal:“Enhanced Tree Species Classification Using CLAHE and YOLO v8”. This paper demonstrates the successful application of Contrast Limited Adaptive Histogram Equalization (CLAHE) and the YOLO v8 model working in tandem to accurately classify different tree species based on the textual features of their trunks. The study highlights how CLAHE and Yolo v8 have the potential to make a strong AI image classification system when working in synergy: CLAHE effectively enhances image contrast, making subtle features more distinguishable while YOLO v8 excels in real-time object detection and categorization. Similar to our system, the algorithm here first increases the contrast of images, allowing for greater visibility of features and textures of the trunk, and then classifies them using the convolutional neural network (YOLO v8). We were in fact inspired by these findings to integrate both into Hypertensight.

To further enhance our model’s accuracy, we have incorporated the use of the green channel in image processing. The green channel is particularly effective in retinal imaging as it provides higher contrast for blood vessels and other relevant structures compared to the red and blue channels. By combining CLAHE with green channel processing, we achieve even greater contrast enhancement, facilitating the detection of minute abnormalities indicative of HR.

Preliminary tests and pilot studies have demonstrated promising results. Our system has shown an accuracy of 94.5% in identifying early signs of hypertensive retinopathy, significantly out-performing traditional manual inspection methods. Additionally, the integration of YOLO v8 allows for rapid processing, making our solution not only accurate but also efficient, capable of handling large volumes of images without causing delays in diagnosis.

Currently, our team consists of two dedicated members: Sarthak Ahuja and Suhana Grewal. Both are managing full-time responsibilities and 12th-grade assignments while meticulously working to ensure our expansion and success.

The ideation of our concept began in December 2022, about 1.5 years ago. Initially, it was developed for a competition, but we put it on hold. The importance of our project became clear when we met a Chinese patient who had been falsely diagnosed with HR, which was actually a different eye disease. This experience highlighted the need for an accurate solution. Eight months ago, we began developing a prototype, undergoing numerous iterations of the model and design. After incorporating initial feedback, we launched the final working solution two months ago.

1. Non-Negotiables: Zero tolerance for racism, casteism, sexism, bigotry, and any form of discrimination or harassment. Our company policy will clearly outline penalties for sexual harassment, bullying, and racism, ensuring a safe and respectful workplace for all team members.

2. Grievance Redressal Cell: We will establish a grievance redressal cell dedicated to addressing any concerns or issues that may arise. This cell will ensure that all team members have a voice and that their concerns are addressed promptly and effectively.

3. Professional Development: We will allocate a budget and time for professional development activities and relevant training sessions with flexible timings. This will ensure that all team members have the opportunity to enhance their skills and advance their careers.

4. Inclusive Hiring Practices: Our hiring process will focus on removing barriers to opportunity by implementing blind recruitment techniques, partnering with diverse organizations, and actively seeking candidates from underrepresented groups. We will prioritize hiring individuals who demonstrate a commitment to our values and mission. 

5. Constructive Feedback Culture: We deeply value constructive feedback and criticism, as demonstrated by the usability trials and surveys we have conducted throughout our development process. This receptiveness to feedback will continue in our interactions with future team members, ensuring that everyone feels heard and respected.

Creating a welcoming and inclusive environment for all team members is a core priority. As we expand, we will extend our current policies for skill development and flexibility to future team members, ensuring they have the resources and support needed for professional growth, as well as make sure that the team members we bring onboard are skilled in their respective domains. Our dedication to fostering diversity, equity, and inclusion will guide us as we grow our team and continue to develop Hypertensight, ensuring that we not only meet our impact goals but also create a workplace where all team members thrive.

Ensuring that our team is diverse, minimizes barriers to opportunity, and provides a welcoming and inclusive environment is of utmost importance as we expand. Our co-founders embody a collaborative leadership style that prioritizes open communication, mutual respect, and inclusive decision-making. This approach fosters an environment where all team members feel valued and empowered to contribute their unique perspectives and skills.

Our current team of two has demonstrated a commitment to diversity and inclusivity through our practices. As we plan to grow to a team of 25-30 members, including data scientists, AI experts, and healthcare professionals, we will continue to uphold these values. We aim to hire individuals who are passionate about our mission, including freelancers from diverse backgrounds, and we are committed to seeking talent from unexpected places, ensuring equal opportunities for all candidates regardless of their background.

To promote diversity, equity, and inclusion, we will implement comprehensive policies:

Our business focuses on providing doctors with a highly accurate and efficient diagnostic tool for hypertensive retinopathy (HR), achieved through the intersection of cutting-edge technology and a strategic business model.

We offer early and accurate detection of hypertensive retinopathy which significantly improves patient prognosis by identifying the condition at its earliest. Our user value proposition emphasizes rapid image analysis, saving doctors valuable time and allowing them to focus more on patient care. With enhanced diagnostic confidence, doctors can depend on the technology’s consistent and reliable results to support informed clinical decisions, ensuring that those threatened by HR receive suitable treatment. Moreover, we reducing diagnostic costs for healthcare providers by minimizing the need for extensive manual labour and multiple consultations. This leads to increased throughput, enabling providers to handle a higher volume of patients efficiently and optimize resource use.

We believe that our success as a business is determined by our impact. We therefore measure our impact in terms of improved patient outcomes by tracking metrics on early detection rates and successful diagnostic outcomes. Healthcare provider satisfaction is also gathered and analysed through feedback on the tool’s impact on their practice efficiency. With further expansion, we plan to assess accessibility and reach by measuring the number of healthcare providers and clinics adopting the tool, particularly in underserved areas.

Moreover, our hybrid technology serves as both a product and a service. Our software can be used as a product by integrating it into the existing technology used by healthcare providers—laptops, computers, or tablets. Moreover, our software is supported by continuous updates, data analysis, and customer support, and operates on a subscription model for private institutions, allowing it to be utilized as a service as well.

Lastly, our business model categorizes our activities into two types: program activities linked to development and deployment, as well as our non-program activities associated with business sustainability, allow us to consistently build, enhance, and distribute our technology. Within research and development, we focus on refining AI algorithms for efficient and accurate image analysis, enhancing image processing through larger datasets, and analysing medical data to improve algorithm accuracy. Product testing, which involves clinical trials with healthcare institutions, ensures the effectiveness of the technology while also providing us with opportunities to network and grow. Software deployment activities encompass maintaining a cloud-based platform for remote access and scalability, integrating the tool with existing healthcare systems, and providing ongoing technical support to users. Moreover, our heavy focus on marketing and outreach efforts involves promoting the tool in healthcare institutions and developing partnerships with research and medical firms to ensure provision.

Therefore, Hypertensight provides significant value to doctors and healthcare providers by enhancing diagnostic accuracy and efficiency. Through strategic use of resources and well-defined activities, we ensure continuous improvement and scalability of our product, aiming for widespread adoption and a substantial impact on healthcare efficiency.

Below is our Business Plan, Please click HERE if you want to view it with greater clarity: 

Our plan for financial sustainability relies on a holistic approach including sustained donations and grants as well as potential revenue from service contracts and selling our diagnostic tool on a subscription basis.

Firstly, as a non-profit, we actively seek donations and grants from private donors, philanthropic foundations and government agencies. Our mission to provide early and accurate detection of hypertensive retinopathy aligns with the vision of several health-focused fund-providing institution. Moreover, in its pilot phase, Hypertensight has already received approval and appreciation from leading ophthalmologists in India such as Dr. Rajesh Rastogi, a senior MMBS Gold Medallist surgeon/ophthalmologist with over 30 years of experience, from Krishna Netralaya, a top ranked eye hospital since 5 consecutive years. This is likely to increase our visibility and credibility, thus opening up potential opportunities for funding in the future.

Secondly, we aim to leverage our diagnostic tool's value proposition to establish service contracts with larger, privatized healthcare providers and institutions. These contracts will allow us to generate revenue through a subscription basis. Currently, we offer 4 subscriptions plans including:

(1) a monthly subscription of $75 for 30 images

(2) a monthly subscription of $100 for unlimited images

(3) an annual subscription of $1000 for unlimited images

(4) a pay-per-diagnosis subscription of $5 per image

The revenue generated via these subscription models from private medical institutions and hospitals will allow us to invest in expanding our outreach, especially to undeserving communities, where our model will be provided to rural hospitals and clinics for free, allowing us to fulfil our vision of enhancing healthcare accessibility as a non-profit.

Moreover, we plan to collaborate with universities and research institutions on studies related to hypertensive retinopathy and AI diagnostics. R&D organizations, focused specifically on eye-care, such as the International Centre for Eye Health. These collaborations can attract funding from academic grants and research initiatives. Moreover, we also plan to explore partnerships with governmental health programs such as NPCB & VI. By integrating our diagnostic tool into national healthcare systems, we can secure service contracts that provide a steady revenue stream.

Since we recently progressed into pilot testing, our project is yet to receive funding. Nevertheless, in terms of evidence of success, our current marketing and outreach efforts have increased our visibility to medical institutions in India such as Fortis Hospitals, which can open up funding opportunities. Additionally, our tool's performance in clinical trials and feedback from usability testing have demonstrated its effectiveness, making it more attractive to potential partners.

Our long-term financial sustainability plan includes continuously seeking grants and donations while expanding our revenue streams through service contracts and subscriptions. By maintaining a focus on both funding and revenue generation, we aim to cover our expected expenses and ensure the continued development and deployment of our diagnostic tool.

Thus, Hypertensight is well-positioned for financial sustainability through a combination of grants, donations, and subscription models. Our proactive approach in securing funding, backed with the demonstrated effectiveness and value of our diagnostic tool, provides a solid foundation for long-term success.