Healthcare professionals are facing various challenges in the war against COVID-19. These include a shortage of essential supplies, staffs requirements per hospitals, optimising hospital facilities, capacity management, and keeping healthcare workers safe. During situations like a pandemic outbreak, all the stakeholders, including policy-makers, care sites and their staff, individuals and technologists, should bring in all the resources to fight it. And that is exactly what an AI startup is doing.

H2O.ai, a Chennai and US-based startup, is utilising the latest technological developments in data, artificial intelligence and machine learning to deal with the COVID 19 crisis. The company believes that informed organisation and action will be essential to contain COVID-19 as well as help prevent the spread, and ultimately eradicate the virus and return to normalcy. Thus the startup has developed multiple AI/ML platforms to save people’s lives and is currently working with one of the largest healthcare systems in the US. 

How AI predicts?

With the help of data and artificial intelligence, H2O.ai’s platform can predict situations and provide data on various areas of health care. These data are an integral part for various stakeholders and helps in understanding the hospital staff requirements, population risk segmentation, predicting the spread of the disease, hospital supply chain requirements, sepsis predictions and so on. 

Data from H2O.ai’s platform is helping various departments to perform better and coordinate and share information. H2O.ai’s AI and machine learning models are used to predict where and when resources are most urgently needed - which hospital in which city or community, to help people more efficiently. In the supply chain, managing medical supplies including antibiotics, ventilators, and basic supplies such as masks and gloves, has become even more critical. “Using predictive modelling and analytics dashboards for supply chain distribution and optimisation that are prevalent in retail and the manufacturing sector, H2O.ai can help local healthcare officials, to predict equipment needs and all hospital care supplies,” says Sudalai Rajkumar, Data Scientist at H2O.ai.  

From screening to ICU transfer

More data from each source will help screening and testing to become faster and efficient. “While confidential and private data will be safe, the anonymous data gleaned can help the next patient, the next case, and the overall health of the community and ultimately defeat this disease. Working with testing and screening providers, the resultant data can be used to help make the AI and predictive models even more accurate,” added Rajkumar. Once the patient is transferred to the ICU, data is prime. Unplanned or incorrect ICU transfers can lead to high morbidity and mortality in hospitals - this is, even more, a challenge when hospitals are busy.  

Telehealth technology solutions targeting remote populations will also help to examine and determine real-time regions of contagion, containment or mandatory self-quarantine. The resulting data sets will be used to pinpoint more accurately the extent and spread. 

Analytics will improve efficiency and roll-out of telehealth solutions, in-house testing, and treatments.  

Contact tracing is another area. “Data emerging from states that have implemented more aggressive shutdowns shows that shutdowns of businesses have resulted in a 75x difference in spread. This is precisely what Singapore did to control the contagion,” says a spokesperson of H2O.ai. 

Data is the key 

“We can help with personalising outcomes for patients based on age, comorbidities like diabetes, high blood pressure, and immune-compromised,” says the spokesperson.  This will help in identifying and preventing the likelihood of sepsis for at-risk patients who are coming in for COVID-19. 

AI will also help in drug delivery. Repurposing existing drugs will also be necessary during a pandemic outbreak. Matching patients to clinical trials will help with drug discovery. For instance, China is testing HIV drugs as a treatment protocol. To optimise these efforts, data screening will be critical for patient matching and cohort identification. India is using antibiotics from the 1950s as the primary protocol because the virus has not been exposed to those drug protocols. Combination of HIV and they are using antibiotics from the 50s because most people are resistant. For more severe cases, Viagra is also being considered given its efficacy in the reduction of pulmonary hypertension, which is one of the key causes of death of COVID-19.