Pioneering Tech For Uncertain Times – COVID-19 Accelerates AI Adoption In Healthcare

 

 

For healthcare, it’s resulted in the increased adoption and appreciation of artificial intelligence (AI) technologies.

In many instances, lack of historical data and fragmented data sources along with rapidly changing and unexpected scenarios have made it difficult to establish reliable and timely prognoses for patients, determine the most appropriate treatment methods, maintain accurate visibility across operations, and manage the pandemic’s overall spread. Fortunately, some recent AI technologies have been designed to help providers overcome many of these problems.

 

 

Trustworthy AI that implements human-like reasoning and applies proven medical data, established lab work, and expert human knowledge is helping to drive the most dynamic advancements within the healthcare tech industry.

As described in a 2017 Journal of Physics paper by AI researcher Jagreet Kaur and physics professor Dr. Kulwinder Singh Mann, “AI in healthcare is needed to bring real, actionable insights and individualized insights in real time for patients and doctors to support treatment decisions.” The paper goes on to emphasize “…the importance of applying AI-based predictive and prescriptive analytics techniques in the health sector. The system will be able to extract useful knowledge that helps in decision making and medical monitoring in real time through an intelligent process analysis and Big Data processing.”

 

The benefits of AI-powered predictive modeling tools have been increasingly evident throughout the pandemic. A few notable examples include Penn Medicine’s COVID-19 capacity planning tool, CHIME, Washington State’s predictive dashboard, another COVID-19 risk assessment tool, and our own Beyond Limits dynamic predictive model, which is based on the SIR model in predicting the progression of pandemics.

 

These forecasting models, created in partnership with renowned medical professionals working directly in the fight against COVID-19, analyze the impact of social mobility on infection and hospitalization rates to better inform hospitals and policymakers in their efforts to combat the pandemic. They can estimate resource allocation supply and demand, enabling medical facilities to better prepare for personal protective equipment (PPE) needs by analyzing the percentage of patients that may need ventilators and extracorporeal membrane oxygenation (ECMO) in the ICU and dialysis units.

 

These models also can factor in impacts from social distancing modifications to stay-at-home regulations through cell phone mobility data, and are designed to aid relevant leaders in calculating the effects of viruses on individuals, medical facilities, and regional recovery strategies while supporting planners in forming procedural tactics for counties, states, and nations.

 

Developing predictive models in the war against contagious diseases is challenging. The likelihood of determining outbreak occurrences, alongside unfolding infection case numbers and deaths, are multidimensional problems with high variability. The most valuable models are designed to operate with several data sources and categories simultaneously.

The data that goes into such models is heterogeneous and highly dimensional by nature, while vast volumes of relevant data are continuously being collected by various parties. Therefore, the ability to integrate a variety of sources of time-series and dynamic spatial data into models with learning capabilities – which get smarter over time – is critical for more accurate decision making.

 

To effectively utilize all this data and create dynamic models with learning capabilities that function in near real-time, most conventional data processing, classification, pattern recognition, modeling, and forecasting tasks must be automated and streamlined.

 

In addition to predictive forecasting models, other examples of AI helping in the battle against COVID-19 include chatbots and telehealth initiatives that use conversational AI, vaccine discovery tools, and medical insurance solutions that optimize claims management, fulfillment, and billing.

 

 

Two of the biggest challenges facing this era in healthcare are the acceptance of AI and the proper implementation of the technology.

 

A 2019 Future Healthcare Journal article by Thomas Davenport and Ravi Kalakota sets out to discuss “…both the potential that AI offers to automate aspects of care and some of the barriers to the rapid implementation of AI in healthcare.” The article concludes by stating, “It also seems increasingly clear that AI systems will not replace human clinicians on a large scale, but rather will augment their efforts to care for patients. Over time, human clinicians may move toward tasks and job designs that draw on uniquely human skills like empathy, persuasion, and big-picture integration. Perhaps the only healthcare providers who will lose their jobs over time may be those who refuse to work alongside artificial intelligence.”

 

AI must be implemented not in a way that discounts expert input, but rather to work in conjunction with medical experts to amplify their capabilities. This means leveraging explainable AI solutions that are both auditable and accountable, providing transparent recommendations that keep domain experts informed and in control. This is the key to trusted AI.

 

Hospitals and healthcare organizations can further support and successfully implement AI through close collaboration with the developers of the AI solution. AI engineers should be working in tandem with the healthcare leaders, professionals, and doctors who will be using the solutions. Regardless of how promising any one AI solution may seem, it won’t likely be adopted if the UI/UX is confusing or impractical for day-to-day purposes. At the same time, coordination between the developers and users helps to dispel mistrust among healthcare staff, ensuring personnel is properly trained to utilize their new tools to the fullest potential.

 

COVID-19 has forced rapid disruption and evolution in virtually every major industry around the globe. Healthcare in particular is bearing the brunt of this change, but the introduction of new AI models, analytics, and diagnostic solutions will be critical to overcoming the coronavirus pandemic. Consequently, as AI becomes more entrenched within our existing healthcare systems, we’ll be better prepared to combat future outbreaks.