Artificial Intelligence (AI) Capabilities Will Be Most Effective When Applied in Full Collaboration Between IT Leaders and Medical, Epidemic, and Healthcare Specialists
“In the fight against COVID-19, AI offers an important arsenal of weapons,” said Erick Brethenoux, research vice president at Gartner. “It allows predictions to be made about the spread of the virus, helps diagnose cases more quickly and accurately, measures the effectiveness of countermeasures to slow the spread, and optimizes emergency resources, to name a few. The power of AI should not be ignored or only partially leveraged, so long as it is applied in ethically responsible ways.”
The five areas where using AI to combat COVID-19 will have the most impact are:
Early Detection and Epidemic Analysis
AI techniques are used to understand, analyze, and predict how and where the virus is spreading or slowing down.
Automated contact tracing, for example, is used to build detailed social interaction graphs by analyzing a myriad of citizen data such as mobile phone locations and public facial recognition and backtracking the movement of people to identify the likely virus source. Individuals who encountered the source can then be notified, tested or quarantined.
“Other AI applications that fall within this area include epidemic forecasting and monitoring the development of herd immunity. Such capabilities are obviously highly relevant in the short term as society tries to ‘flatten the curve’ and minimize the burden on our healthcare systems, but they are also important in the long term if new, hopefully, smaller, outbreaks reoccur,” said Mr. Brethenoux.
Containment
Considering the huge societal and economic impact of ‘one-size-fits-all’ measures such as lockdowns, collaboration with non-IT experts is paramount when applying AI to containment efforts.
“Behavior analytics derives new insights by accounting for the dynamics of human behavior, culture and individual thinking to answer questions around social distancing compliance or the emergence of unwanted group behaviors,” said Pieter den Hamer, senior research director at Gartner. “Law enforcement can predict where and when people may not adhere to stay-at-home orders or social distancing through predictive enforcement and dispatch enforcement units accordingly.”
Triage and Diagnosis
The use of AI-enabled self-triage has already gained popularity as telehealth practices, including virtual health assistants, were made available to help people identify if they are possibly infected and what the appropriate next steps are. Augmented medical diagnosis and triage are also key AI capabilities that help in this area.
“AI is known to improve the accuracy of certain diagnoses if augmented with human judgment, especially in more complex cases,” said Mr. den Hamer. “Prognostic modeling, or predicting how the disease will likely develop in patients, can also be used to improve treatment recommendations. The fact that AI has a role to play in assessing patient risk and prognosis is not something to overlook, especially when there is a possible shortage of medical professionals.”
Healthcare Operations
AI plays an important role in streamlining healthcare operations and optimizing scarce hospital resources during a pandemic. Healthcare CIOs and CDOs can use predictive staffing to improve personnel allocation by analyzing anticipated patient numbers and their individual prognosis, and cross-referencing them with the availability of qualified medical staff, materials and equipment.
“Remote patient monitoring and alerting with the use of AI also allows patients to stay at home, lowers the burden on hospitals, and enables a better understanding of how symptoms develop over time,” said Mr. den Hamer.
Vaccine Research & Development (R&D)
AI graphs and natural language processing (NLP) can enable medical researchers to scour through many thousands of relevant reports to draw connections between data at an unprecedented pace. Augmented vaccine R&D also identifies coronavirus countermeasures including those that have already been tested on humans.
“Healthcare CIOs and CDOs should explore every avenue of AI to fight COVID-19 using an ongoing and systematic process of AI application identification and prioritization. Technologists should not overestimate their ability to understand what makes sense from public health and medical perspective and then work with healthcare professionals to create and actively advertise an open marketplace that shares AI applications, models, and data transparently,” said Mr. den Hamer.
