Making healthcare accessible: A machine learning approach
INDIA’S GROWING HEALTHCARE CRISIS
India continues to struggle with providing basic healthcare for its fellow citizens even after decades of rapid economic growth. Accessibility and disparity still are the primary reasons behind the misery. It won’t require an epidemic for the public health system to crumble. The situation is worse in rural and suburban areas. India’s Infant Mortality Rate (IMR), an important health indicator, remains three times higher than China’s and almost six times higher than that of U.S. and Russia, Maternal Mortality Ratio (MMR) more than twelve times higher than that of U.S., U.K. or China.
To make matters worse, India suffers from a severe shortage of doctors, the estimated number of doctors for 1000 people being 0.76. This is a lot lesser than most developed countries as shown
The overall burden of managing patients has increased by leaps and bounds over the years with a surge in non-communicable diseases on top of already immanent communicable diseases. The
challenges we need to address include dealing with the uneven ratio of qualified doctors to patients, making doctors more proficient at their tasks, delivering personalized healthcare and high-quality healthcare to rural areas. A plausible way to tackle the needs of the plenty and the services of the few is to seek innovative solutions leveraging the power of machine intelligence and human-machine collaboration.
SAARTHI: Charioteer of Healthcare
Artificial Intelligence (AI) has the capability of delivering solutions to make healthcare accessible and affordable to the masses at the same quality level irrespective of place and time. One of the major
applications is the automation of medical diagnosis via “differential diagnosis” which consists of distinguishing a symptom as the potential cause of a patient’s illness through A process of association or elimination. Keeping that in mind, we built an automated differential diagnostics tool named
SAARTHI as a solution :
- Acute shortage of professional doctors and nurses especially in rural areas of India often forces people to consult quacks. A differential diagnostics tool like SAARTHI can help those patients make better decisions prior to consulting doctors.
- With additional emphasis on common and highly infectious diseases in rural areas, SAARTHI will be able to identify those diseases earlier with better accuracy.
We also aspire to achieve bigger in the longer term by:
- Identifying patients in the need for immediate attention and care while sifting through the
probable diseases prompted by our tool via triage based approach.
- Along the way, enabling SAARTHI to learn actively and provide more reliable and efficient automated detection of diseases.
SAARTHI: Standout in Indian Context
“Health data is being collected to inform clinical decisions and [to shape] personalized predictive medicine. But there really isn’t integration to improve clinical trials and inform better health
practices.”-Heather Zumpano of IMST Telehealth Consulting
As a consequence of increasing access to health-related data, quite a lot of which is riddled with inaccuracies and open to manipulation, identifying a reliable source has become difficult. Along with the above lines to exploit, lack of personalized healthcare in deprived areas of India motivated us to
go one step further and compete with the already existing diagnostic tools such as WebMD symptom checker, Mediktor, Symcat, ADA, Infermedica, Isabel and so on:
The aspects in which SAARTHI is exclusive and unique is the emphasis on diseases prevalent in India, especially in rural areas. (Data Source: https://vizhub.healthdata.org/gbd-compare/india)
An App(le) a Day, Keeps the Doctor Away?
Through the use of technology and scientifically proven information about health-data, the general population would be more likely to gain access to clinical services and care. The end users of the application can be either the common man or a healthcare provider. SAARTHI will essentially help patients by optimising frequency of visits to doctors in terms of relevance. Furthermore, patients who while seeking appointments are often confused about which department to visit can decide for themselves using SAARTHI which can guide them to the right healthcare professional via potential diagnosis.
Doctors, despite their plethora of experience, might enrich their decision-making ability from an automated tool like SAARTHI supplementing their knowledge and recommending better treatments for patients. It also might inspire the doctor to better rule out alternate diagnosis apart from the obvious ones. According to a paper published in JAMA Internal Medicine, physician diagnostic errors are as high as about 15 percent. With that in mind, we can propose SAARTHI as a tool augmenting physician capability with machine intelligence to close this gap.
AI-enabled diagnosis can potentially aid even less competent people in making difficult decisions. Also consulting highly professional doctors only when confidence level of such automated tools are low will help in utilizing their skills in more intricate cases.
Although it will be a slow process for an emerging country like India, all in all, the inclusion of AI-based solutions in day-to-day medical decision making will eventually enhance patient healthcare in an affordable and accessible manner.