Bringing Healthcare

to India's Last Mile Community

How a tech start-up is harnessing Artificial Intelligence to bridge India’s healthcare divide.

Growing up in remote villages in West Bengal, India, Sujay Santra observed firsthand how a lack of access to medical services could affect a community.

Learning from personal experience with a family member’s ailment and armed with the belief that technology could bridge the demand and supply gap, the 42-year-old entrepreneur launched iKure in 2010.

iKure enables access to affordable and accessible health services and technology-enabled integrated care, all while empowering community health workers. By the end of 2020, the Kolkata-based start-up, with the mission of “creating zero mortality in primary healthcare”, was operating clinics in seven Indian states. The nearest clinic to Tabageria, a village in West Bengal, was previously 50KM away. Tabageria now has a 17,000 square foot facility offering services such as a general physician, mother-and-child healthcare, an eye clinic, and a small pathology lab.

Community workers equipped with portable devices to measure blood pressure, hemoglobin levels, or heart rates carry out home visits to patients – such as pregnant women – who are unable to make the trip to the clinic. When the Covid-19 pandemic swept across India, iKure turned to telemedicine to continue providing healthcare.

The network covers a segment of the population that has been left particularly vulnerable due to the lack of access to medical services. In India, there is only one cardiologist for every 6,000 patients. There are usually no cardiologists in India's villages, which host 70% of the country's population. Overall, some 840 million people don't have access to the healthcare they need.
1 cardiologist/
There are usually none in India's villages, which host 70% of the country's population. Overall, some 840 million people don't have access to the healthcare they need.
people don't have access to

Because of this lack of healthcare, many diseases go undiagnosed. Nearly 2.5 million people die every year in India of cardiovascular diseases, making it the leading cause of death. There are also an estimated 73 million cases of diabetes, which, if undiagnosed, can lead to blindness.

When Bablu Shil, a 50-year-old farmer who toils every day under the hot sun and has a family of 10 to feed, suddenly developed severe chest pain, nausea, and shortness of breath, he immediately headed to Tabageria's iKure clinic. Recognizing the telltale signs of angina pectoris, a doctor at the clinic conducted an electrocardiogram (ECG) using a portable device and sent him to a hospital to see a specialist. The early detection of his condition enabled Shil to avoid surgery, saving him as much as RS 300,000 (USD 4,100) and added 20 years to his lifespan.

Since the pandemic, iKure has implemented measures to detect and monitor non-communicable diseases early.

As iKure's community workers and doctors interact with patients, measure their vitals, and carry out diagnostic tests, they collect a trove of data about the conditions they suffer from and their overall health. They also capture data about patients’ demographics and pre-existing conditions. To unlock the value of this precious data, iKure teamed up with IBM in December 2018 to develop an AI-powered risk and analysis model with a cloud-based interface.

Try your hand at this sample risk and analysis model
Step 1 of 3

Collect Data

The key to patient monitoring is iKure's wearable Braveheart patch, which captures several data points such as electrocardiogram (ECG), body movement, skin reactions and others.

The platform, built by a team of data scientists and AI experts from the United States, Singapore, and India, enables frontline health workers to load the data they capture on the ground into a smartphone equipped with iKure's app.

The data is then aggregated into 152 health parameters that include blood pressure, sugar levels, or heart rate.

Each of the 152 health parameters captured on the ground is then given a score based on severity. The system then analyzes to generate a list of the top 10 patients who need care immediately because they are at the highest risk of suffering from a heart attack.

Some at-risk patients are even fitted with a wearable Braveheart patch, which captures their heart rate, body movements, and skin reactions, instantly transmitting the data to routers.

The data analysis allows doctors to see the most urgent cases first, ultimately saving lives. It also enables them to optimize their workload and free up time to focus on providing care to them. They no longer have to manually process thousands of patient records to detect heart conditions.

The purpose is not to replace the doctors but to help them make the right decisions.
— Sujay Santra

Respecting the privacy of patients is paramount to that mission, especially where sensitive medical data is involved.

To ensure this, when iKure's algorithm analyzes the patients' parameters to generate insights, all the data is kept within the platform, which acts as a secure, enclosed system, removing the risk of leaks.

The platform also has a built-in bias monitor to ensure that when a physician chooses to prioritize treating patient A over patient B, the decision was made fairly and accurately. For instance, the bias monitor had once alerted Santra that men were being flagged as having a higher rate of cardiovascular disease than women.

He set out to investigate the potential bias by comparing iKure's patient datasets with the datasets used by other risk models with a proven track record. He found out that the bias was present in both datasets, meaning it was not a mistake but a reflection of men's higher susceptibility to cardiovascular disease. He was able to carry out the whole operation within the platform in a matter of minutes, without having to move data around or filter it.

Since 2010, iKure has treated over 1 million patients, covering a 10.2 million population, and aims to reach 50 million people within the next three years. iKure hopes to treat 200,000 patients by the end of 2021 using IBM's intelligent platform.

Ultimately, we hope to expand the predictive model to other diseases such as diabetes or hypertension. Without a platform like the one we worked on with IBM, we would never have been able to treat patients beyond a basic level.
— Sujay Santra
Find out more