Pandemic Preparedness Using AI




The COVID-19 pandemic continues to wreak havoc in countries around the world. After the Spanish flu pandemic in 1918, this is one of the most widespread and pervasive health calamities that has befallen the world. As of mid-July 2021 there were 188,128,952 confirmed COVID-19 cases and 4,059,339 deaths in around 200 countries and territories and 26 cruise and naval ships. Not only is the health and human cost significant, but the economic cost of it is also far-reaching with some estimating it as high as $16 trillion.


The global response to the COVID-19 pandemic has been largely reactive rather than proactive, even as the second and third waves are hitting countries. Entities such as the World Health Organization (WHO) have been monitoring epidemics and pandemics around the globe, but their efforts and preparedness were nowhere near up to scale for a global onslaught such as this.

A look at pandemics with high death tolls in the past couple of centuries (Fig. 1) shows that their incidence is on the rise, which is to be expected given the increase in global trade and travel.


According to the World Economic Forum there are 1.7 million ‘undiscovered’ viruses in mammals and birds, of which humans are susceptible to 827,000. If one were to look only at coronaviruses as the causative agent, the SARS epidemic in 2002-2004, the MERS outbreak in 2012, followed closely by the ongoing pandemic are a strong indication that global threats to human health due to a virus or bacterium are a reality of our times and need to be addressed.



Figure 1. There have been a number of pandemics with high mortality rates within the last three millennia


Even if reactive, the global response to the COVID-19 pandemic has been notable. Governments, medical institutions, academia, tech companies, and international agencies have taken significant steps in trying to contain and dissipate the COVID-19 pandemic. Quarantines and lockdowns have been used to control the spread of disease in spite of the very high cost to economies and individual freedoms. Surveillance and diagnostics through the scientific community have seen an unprecedented level of local and international cooperation.


Vaccine development and deployment have been accomplished in record time. And of course, the health care workers around the world have risen to the occasion and gone above and beyond the call of duty. However, there is no getting away from the reality that this herculean effort is still after-the-fact. How much more could be done if we had a proactive approach to pandemic preparedness.



Prepare: Detect, Prevent, Respond


There are three main areas of focus that need to be addressed for taking on any future pandemics: detection, control, and preparedness.


Detection:

Nearly all known pandemics are caused by pathogens of animal origin. Such diseases are called “zoonoses” and, in the most impacted countries, surveillance and early warning systems, as well pandemic prevention plans, have been deployed with concerted effort.

There are two ways in which disease outbreaks can be detected. The first is syndromic surveillance. This relies on the presence of health indicators such as symptoms and patterns, before a laboratory-confirmed diagnosis is made. However, this is not always the most accurate way to understand infection spread, as it lacks a definitive disease diagnosis.


The more reliable method is laboratory-based surveillance. This incorporates a number of methods to detect and confirm pathogen presence, of which metagenomics is proving promising in its ability to detect the presence of most microorganisms through a single sequencing technique, without the need for culture. A study published in the Journal of Clinical Virology in October 2020 highlighted the validity of a metagenomics protocol to identify coronavirus by simulating novel virus discovery and so provide a potential tool for pandemic preparedness.



Prevention:

There’s good news according to the World Economic Forum: we can prevent future pandemics.

A group of leading scientists say it is possible if we make an effort to protect the environment and restore its natural defences. “There is no great mystery about the cause of the COVID-19 pandemic – or of any modern pandemic,” said Dr Peter Daszak, chair of the panel which was convened by the Intergovernmental Science-Policy Platform on Biodiversity and Ecosystem Services (IPBES).

COVID-19 emergence has been entirely driven by human activities. “The overwhelming scientific evidence points to a very positive conclusion,” said Daszak. “We have the increasing ability to prevent pandemics – but the way we are tackling them right now largely ignores that ability.”

According to recent scientific studies, preventing disease spillover from wildlife costs 100 times less than trying to respond to such a disease once it has spread.


Preparation:

Covid-19 is not the first global health emergency, and it will not be the last. A recent report highlighted how many governments and leaders struggled to take early decisive action based on science, evidence and best practice at the onset. According to the report, responsible leadership and good citizenship have been key determinants of COVID-19’s impact. But preparedness is the way forward.


With many countries committing to providing affordable healthcare for all, the report states that investments in preparedness would only cost US$ 5 per person annually. Ensuring that healthcare facilities and staff are equipped and updated, that their staff are trained and prepared for pandemics, and that medical, communication, and process protocols for serious health threats are developed.

Not only is preparation needed to save lives, but also to mitigate the far-reaching effects of a pandemic.


“We know that preparedness makes economic sense, and we have developed tools and models for multi-sectoral cooperation. Learning from the pandemic and building on the previous progress should guide our steps to strengthen Health Security and thus help achieve the Sustainable Development Goals,” said Päivi Sillanaukee, Ambassador for Health and Wellbeing, Republic of Finland. Artificial Intelligence (AI) has a role to play in all three aspects of preparing for a future pandemic and other threats to human health.


What is AI?

AI is when computer systems are able to carry out human-like tasks, such as recognition, reasoning, and learning. These systems have algorithms that are highly adept at picking up trends and patterns. They can quickly and accurately sift and sort through massive amounts of data from various data sources, identify patterns in the data, make predictions as to the likelihood of these events occurring again, and then suggest actions that will produce the most favorable outcome.



AI in Pandemic Detection


Experts around the world have reiterated the importance of a global warning system for emerging infectious diseases (EIDs) for the timely detection of and response to any possible infectious hazards. Organizations such as the WHO have been very active in this area and it’s Coronavirus (COVID-19) dashboard gives a highly comprehensive and current overview of the current pandemic (Figure 2).

Figure 2. WHO Covid-19 Dashboard


Rick Bright, a former Director of the US Biomedical Advanced Research and Development Authority (BARDA) and currently Senior VP of pandemic-prevention activities at the Rockefeller Foundation, wants a global genomic surveillance system for viruses, but as a non-governmental, non-political entity. He sees sequencing data as essential but thinks building the capacity to gather data from public health departments, private industry, and academic labs, analyzing it quickly to create impactful analyses as more important, since it can inform decision makers to take strategic action.


AI can play a significant role here. In fact, Metabiota Inc, a unique company that specializes in pandemic threat management, used it’s epidemic detection system to identify and flag COVID-19 as an emerging coronavirus capable of causing significant economic loss, five days ahead of WHO. It claims to have the most comprehensive infectious disease database which covers a century of human outbreaks and the Global Epidemic Monitoring and Modeling (GEMM) platform to provide a comprehensive view of global epidemic risk. The Metabiota Epidemic Tracker is worth a visit.



AI in Pandemic Prevention and Control


Data-driven insights and predictive analytics can play a significant role in efforts to control the spread of a disease. For example, contact tracing, a process of identifying, assessing, and managing people who have been exposed to a disease to prevent further transmission, can be systematically applied to break chains of transmission. The following use case is another way with which AI helped mitigate the impact of COVID-19.


Use Case 1


Vantage created a real time simulation of US hospital and critical care capacity by US county and made it available to hospitals, government departments of health, and any other organization that needs insight to determine how to allocate resources during the COVID-19 pandemic.

Business challenge to solve: How to control the spread of COVID-19 in a way that ensures hospitals and critical care facilities do not get overrun, which would result in a significantly larger health crisis.

Solution: They sourced detailed published data on cases, hospitalizations, ICU admissions, and patient demographics. This was combined with census data to build a model that is region and county specific to estimate hospitalization and critical admission case rates using logistic regression and scoring tools. Vantage then sourced data on all US hospitals, their utilization and capacity, and used geospatial tools to map counties to individual hospitals and modeled the current capacity at a facility and county level.


Things to note in this use case:

  1. Vantage was able to extract the sourced hospital data into a list of 6,700 institutions with their staffed capacity, bed capacity and utilization rates all with geo-coordinates. And convert time series files into an analytics ready format using AI tools in a time efficient manner.

  2. To build a model quickly, Vantage was able to use extracted data from cases in Italy which was one of the first countries to be hit by COVID-19, to determine the probability of hospitalization and ICU admission by gender and age. Vantage was also able to model the length of hospital stay required.


AI in Pandemic Preparedness


After the flu pandemic of 1918, COVID-19 is the first virus to have such a pervasive and deep impact around the globe. The lessons learned from it are being put into practice in real time with new strains of the virus emerging and unleashing fresh devastation.

Preparedness for an epidemic will require major rethinking and reorganization. This impacts on healthcare and pharmaceutical industries, trade and commercial industries, tourism and entertainment industries, and at a social and behavioral level.


An example of how AI can help preparedness by a commercial enterprise:


Employee absence during an epidemic is a large driver of related economic loss and enterprise risk management should include preparations for this potential loss. Metabiota Inc simulated worker absenteeism by the economic sector during influenza pandemics and studied implications for economic impact.

They constructed a statistical model to predict absenteeism rates using CDC monthly lab-confirmed flu incidence/mortality and national survey data on worker absence (BLS). Industry-specific data was then analyzed to estimate the relative contribution by sector using absenteeism estimation as a function of incidence, mortality and sentiment. This fitted model was applied to global outbreak simulations to estimate temporal dynamics of pandemic-induced absenteeism rates. Such models can help quantify the potential indirect economic costs of epidemics.



Figure 3. Simulated absenteeism by sector in a study conducted by Metabiota Inc.


In fact, a Risk Report generated by Meta biota Inc alerted Bizzabo, an event management company, to create a suite of online virtual events months before their competitors. Meta biota’s critical data and analytics gave Bizzabo a crucial six week advance notice, providing them enough time to remodel and redesign their product and make it not only ‘pandemic-proof’ but highly profitable.


AI Taking on Virus Research


The COVID-19 pandemic has brought fresh focus on virus research, as a greater understanding of it’s biology will help in controlling it. AI is helping here too. Chen et al found that the main scope of AI in COVID-19 research includes, disease detection and diagnosis, virology and pathogenesis, drug and vaccine development, and epidemic and transmission prediction. They collected 2471 online publications and resources related to COVID-19 from databases such as Nature and Google Scholar. They then filtered out 443 papers that explicitly used AI methods to come up with the focus areas of research and the AI technologies being used in them. Figure 4 provides a summarization of the same.



Figure 4. Main scope of AI technologies being used in combating COVID-19. Data sourced from https://arxiv.org/abs/2007.02202.


AI methodologies able to speed up diagnostic procedures, enhance monitoring and tracking capabilities, predict the evolutionary stages of the contagion as well as its effects on society, plus simulate the results of a containment strategy, a medical protocol or a new drug molecule, can represent a revolutionary milestone in the progress of the world in facing these dramatic events.

It is largely on the strength of AI and ML that Peter Daszak, President of the EcoHealthAlliance says this pandemic is the beginning of the end of pandemics.



SEDGE Helps You

SEDGE - the powerful, flexible AI solution that uses advanced machine learning and predictive analytics - brings the versatility and predictive advantage into the real world. No longer is AI confined to big brands and scientists, SEDGE gives you the edge. SEDGE has made Covid Herd Immunity predictions available to the public in countries all over the world through its Covid Calculator.



The information or views expressed in this article is authentic to the best of our knowledge, and as such, it is prone to errors and the absence of some key information, for more information please review our Terms of Use.