A small hospital in a remote corner of Assam is using AI to detect strokes in tea planters

• Baptist Christian Hospital has integrated Qure.ai’s artificial intelligence solution to read brain CT scans in a matter of seconds.
• About 150 people are affected by strokes in Assam every day and 54,890 on a yearly basis.
• Baptist Christian Hospital has been using the technology since March-April 2021 and has been diagnosing about 100 patients on a monthly basis.

Nestled in the foothills of Himalayas, the Baptist Christian Hospital, in the Tezpur town of Assam, decided to move away from its traditional radiology solution in order to provide faster treatment to its patients affected by strokes and concussions. The hospital, which largely caters to the tea plantation workers in and around the town, has integrated Qure.ai’s artificial intelligence solution since then to read the brain CT scan in a matter of seconds.


Not many are aware, but strokes are one of the major causes of death in the Assamese region. About 150 people are affected by strokes in Assam every day and 54,890 on a yearly basis, according to a study released by a local hospital GNRC in 2018. One of its previous reports published in 2014 highlighted that 48 people die of these strokes in the state.

The situation is definitely a tough one in the region and it is aggravated due to the lack of radiologists — medical practitioners who use medical imaging to diagnose and treat diseases on the scene. According to a Global Healthcare Insights report published in 2017, there are about 10,000 practicing radiologists in India and it is not enough.

Prashant Warrier, co-founder and chief executive of Qure.ai, tells Business Insider that India has about 15,000 radiologists as of 2021 and they are usually concentrated in the metro or Tier I cities. This leads to people from suburban areas to either travel long distances to go to a healthcare facility or hospitals to rely on third-party radiologists. Both these solutions are not only inconvenient, but also not efficient.

Dr. Jemin J Webster of the Baptist Christian Hospital highlights that they used to rely on tele-radiology services that took about two hours to get readings of routine scans and about half-an-hour to one hour to get reading on emergency cases. While the doctors too are self sufficient to handle such delays, but a quick reading can make a huge difference.

The matters related to strokes and concussions are highly sensitive, as millions of brain cells are dying per minute. Even a delay of a few minutes can cause body functions like walking or talking to shut down. Therefore, Qure.ai decided to integrate its qER technology in Baptist Christian Hospital, a 130-bed secondary care hospital by non-profit Emmanuel Hospital Association of New Delhi.

How does the qER actually work?

Warrier explained that there are two parts to their artificial intelligence (AI)-enabled deep learning solution — the gateway and the application programming interface (API). The gateway are the x-ray machines and the computed tomography (CT) scan machines that collect the inputs from the healthcare centres and transfer it to the cloud. The API, then, scans, processes and creates a radiology report based on their algorithm. Notably, the company anonymises all the data before sending it from gateway to API to maintain users’ privacy.

Baptist Christian Hospital has been using the technology since March-April 2021 and has been diagnosing about 100 patients on a monthly basis. Dr Webster noted that the technology has enabled them to get the test results much faster.

“The report comes through two systems. One our original PACS system, a system through which we see whatever CT images we have done within our hospitals. The analysed reports are uploaded into the system also, and one more good thing is that they have integrated with messaging app Telegram. So whenever a CT for a patient is done, it comes with an alert on Telegram. In that manner, it is like you are pushing a patient for CT scan and by the time they are back, you get an alert on your phone with an analysis on their condition.”


Dr. Pooja Rao, co-founder and head of research and development (R&D) at Qure.ai, emphasised that the data is accurate. The company has trained their deep learning algorithms with a large as well as diverse data bank sourced from research labs. The company also publishes its accuracy rate on a regular basis, peer review publications and keeps updating its technology often.

Qure.ai has received US’s Food and Drug Administration (FDA) and Europe’s CE certifications to validate this technology.

Qure.ai has been using the same technology, with few variations, to detect tuberculosis and lung infections as well. Even the World Health Organisation (WHO) has validated that Qure.AI’s tuberculosis diagnostic software can replace human specialists in its annual Integrated Tuberculosis Screening Guidelines. It also partnered with British-Swedish pharmaceutical firm AstraZeneca for early detection of lung cancer. The use cases of the technology can be further expanded to diagnose other health issues by training the algorithm, Dr Rao added.

AI can essentially aid medical staff

Over the last year, the company’s qXR technology is also being used to detect COVID-19 across the world and also carried out operations in remote parts of Maharashtra as well as Jharkhand. CEO Warrier added that Qure.ai’s lung X-Ray system, when coupled with other metrics oxygen saturation, temperature and more, can be used to diagnose COVID-19 in areas where RT-PCR tests are not available. He noted that AI can provide an alternative to a trained resource.


Qure.ai’s CEO Warrier noted that the value of such AI solutions is much higher in lower income geographies and countries. He noted that even In India, the quality of care is much higher in big hospitals like Apollo and others, as they have quality radiologists. The quality of healthcare gets even better in high income countries.

“The value [for such AI solutions] for a place like Assam is very different because they don't have a radiologist on duty and for them interpreting that scan is very challenging. You are basically asking somebody who's not a radiologist to interpret the scan and correspondingly act on that patient during surgery, administering a blood thinner or something like that. The value in Assam, or Africa or Southeast Asian countries is tremendous because we can help those non-radiologists interpret x-rays or CT scans without necessarily getting radiologist input, ” he added.

A team of researchers led by Professor Alastair Denniston from the University Hospitals Birmingham NHS Foundation Trust in the United Kingdom conducted a study in 2019, to determine whether AI can diagnose illnesses as effectively as healthcare professionals.

The study noted that AI correctly diagnoses diseases in 87% of the cases, whereas detection by healthcare professionals yielded an 86% accuracy rate. The specificity for deep learning algorithms was 93%, compared with humans’ at 91%.

Qure.ai has raised about $16 million to date from Sequoia Capital India and MassMutual Ventures Southeast Asia. It has marked its presence across 450 sites across 40 countries over the years. It also has its presence across Maharashtra, Bihar, Uttar Pradesh and few other Indian cities, but its head CT scan feature is only being used in Assam’s Baptist Christian Hospital at the moment.

*The quotes have been edited for clarity of language

SEE ALSO
A startup that earned ₹15 crore selling juices in India’s suburbs is now looking at international expansion
Failure to pivot, COVID-19 blues force Amazon-backed Shuttl to look for a buyer
Fogg, Byju’s and PharmEasy grabbed the biggest investments of June 2021— here’s more