Over the past few decades, the strengthening of the public health system and the Integrated Child Development Services (ICDS) as well as school nutrition through the mid-day meals has led to a gross reduction in the presentation of infectious diseases that used to affect children through lack of immunisation and nutritional disorders. We, in India, recently eliminated trachoma as a public health problem and it has become a rarity to see eye conditions secondary to vitamin A deficiency.
The National Programme for Control of Blindness has been among the most effective public–private-partnership programmes in increasing cataract surgeries. However, with the World Health Organization (WHO) looking at effective cataract coverage, there is a need to relook cut-offs for vision in surgery selection from 6/18 to 6/12 in the community.
India is home to a large ageing population. Age-related macular degeneration, wherein the sensitive part of the retina gets affected, is today creating hurdles for the aged, who require early and repetitive interventions, with drugs being injected into the eye to retain their residual vision. The explosion of lifestyle disease is also reflecting in diabetic and hypertensive retinal changes.
Rising awareness and changes in school education have increased the number of children identified with myopia (short-sightedness) and we see a sense of panic amongst parents worried about progressing myopia, as there is a heightened risk of associated eye conditions with every dioptre (largely to do with retinal changes) increase potentially later in life.
Early detection
Without doubt then, early detection of these conditions becomes imperative to effectively manage and possibly limit the impact on vision. India has around 25,000 ophthalmologists, which is only around 15 per million population. There is no second opinion that the country needs to add more eye specialists in order to reach all its people.
The need is indeed great. Corneal blindness, a major cause of vision loss in India, is on the rise, particularly in rural areas owing to limited healthcare access and donor shortfall. Largely caused by infections, eye injuries and vitamin A deficiency, corneal blindness afflicts around 1.2 million people in the country, often leading to irreversible blindness if not treated in time.
Conditions such as amblyopia (lazy eye) are also detected late among children, which could impact their functioning if not detected early and treated.
Every person after the age of 40 would require glasses to read small letters and doing near work (presbyopia), and again, a majority are unaware and continue to function with subnormal vision. They may, at best, procure glasses from opticians where other eye conditions may not be screened and identified.
Technology to the rescue
So, where is technology in all this? Exponential technology, especially Artificial Intelligence (AI) and machine learning, is expected to enhance accuracy in recognition and diagnoses of diseases, enabling greater personalised treatments in healthcare globally. AI is also aiding in quick analysis of vast amounts of clinical documentation and data to quickly identify disease markers. Early detection and better prediction of outcomes, smarter and more efficient care are possibilities now.
One of the early AI systems was IBM’s Watson, which, back in 2011, was geared to enhance better interpretation of communication in healthcare. Cut to 2024 and 2025, tech giants such as Apple, Microsoft, Google and Amazon too, are increasingly developing AI technologies for health.
This AI boom is seeing great application potential in ophthalmology as there is extensive imaging that is inherently undertaken as part of eye examinations.
How AI helps in ophthalmology
This AI evolution aids in a variety of screening platforms for a multitude of eye conditions. In the hospital, its use has been demonstrated from queue prioritisation in large institutions, to enabling inventory management and reorder levels, to converting real-time doctor patient conversations into an electronic health record, to diagnosis support systems and even patient engagement.
AI is also increasingly being deployed especially in retinal diseases. Deep learning (DL) algorithms have sharpened screening for diabetic retinopathy (DR), and age-related macular degeneration, better predicted the progression of myopia and outcomes of cataract surgeries. Machine learning algorithms are being trained to recognise glaucomatous changes too, from disc damage and nerve fibre layer defects, to wide-angle optical coherence tomography (OCT) images. AI tools could soon become mainstream in ophthalmic practice and practitioners and patients would need to embrace the same.
We have worked with partners on AI-based screening solutions for identifying diabetic retinopathy, keratoconus (a corneal disease), amblyopia screening (lazy eye) and cataract screening. At Sankara, specialists have built a GenAI-based CataractBot to engage with patients planning to undergo cataract surgery. There is also great potential in deploying AI to analyse surgical videos that helps specialists in training and monitoring outcomes across hospital networks in India. A voice-enabled feedback system also permits the exploration of emotions apart from just text-based or ratings-based feedback.
Key learnings
The key learnings from these deployments are many. Data that is used to annotate and build AI models must be varied and robust and there is a need to have ethnicity-specific data for these models to work effectively across India.
Undoubtedly, collaboration is critical. As clinicians, we have access to eye data and the concerns and methodologies that need to be prioritised. However, our engineering capabilities can be built, cultivated and deployed only through partnerships. While in India, the Indian Institutes of Technology (IITs) have partnered in eye care, globally, as well as in India, Microsoft Research and the University of Bonn are contributing a great deal in this area.
There is however, a need to set limits and carve out a clear scope for the models being built. These solutions work very well as a screening modality; with time they could evolve as diagnostic support. However, waiting for this evolution could delay possible utilisation in identifying millions through screening programmes.
AI is a tool and we need to not only look at the convenience or support to the healthcare provider, but also keep in view patients’ and their caregivers’ perspective. Patients are most comfortable with a ‘doctor in the loop’ model. The need for human confirmation and conversation is an imperative even in the age of AI.
While this is so, AI and exponential technology afford an opportunity to enhance the quality of care by identifying conditions early and assisting healthcare professionals technologically. AI enhances can be deployed at scale and thus, there is a grand opportunity to optimise delivery costs by blending tele-ophthalmology and allied health personnel in place of ophthalmologists in screening.
(Dr. Kaushik Murali is president, medical administration, quality & education, Sankara Eye Foundation India. kaushik@sankaraeye.com)
Published – December 10, 2024 02:28 pm IST
