How data analytics and artificial intelligence shaped the response to Covid-19
From tracking Covid progression across geographies to sourcing secondary lifesaving drugs to streamlining the supply chain of vaccination, data analytics and artificial intelligence have been at the forefront of the fight against Covid-19.
When Bluedot, a Canadian Disease Outbreak Risk platform, published the possibility of Covid transforming into an ugly form using their AI-driven algorithms, epidemiologists got into action using various forms of SIR models to quantify the possible progression of the virus. Scores of technocrats developed social media listening tools to plug in gaps present in the data published by Government sources. Further, these social media tools along with advanced AI helped in mining and structuring critical information related to life-saving drugs, hospitals, oxygen concentrators, etc. Organizations are leveraging data analytics and artificial intelligence to transform their production lines and supply chain in order to create an efficient delivery network for vaccines. Truly, the year 2020–2021 has been a watershed moment for governments and industries leveraging practical yet life-changing applications of data analytics and AI. The usefulness and importance of analytics and the AI-led possibilities of managing the pandemic are now convincingly established.
Orchestrating a cohesive response to Covid-19
Disease forecasting models, built on complex algorithms consistently aided in decision making which influenced public bodies in enhancing testing capabilities, creating quarantine zones, and imposing lockdowns to flatten the curve. The multiple disease waves predicted by these models have been largely proven to be accurate and gave forewarnings on the possible grim situations if there is any complacency in adhering to protocols and preparedness.
AI and ML models have played a critical role in the drug discovery process. We have seen very promising deployments of these advanced models in many aspects of the drug design process, including use cases in chemical synthesis, drug screening, polypharmacology, and drug repurposing. Even within the testing phases of these molecules, AI was able to identify the correct test subjects, analyze a vast amount of genetic and other data, and speed up the time to market. These enhancements have helped scientists and decision-makers significantly reduce the time for discovery, testing, and roll-out.
Manufacturing facilities are using technology to ensure workplace safety and leveraging Computer vision-based systems to identify non-compliance like missing PPE kits and flouting social distancing norms. Mathematical Optimization models are determining optimal labor teams, operational scheduling, raw material forecasting, etc. to maintain manufacturing efficiency in tough times.
Managing the vaccination supply chain is one area where Data Analytics and AI adoption has been low. Leveraging these tools can help overcome the disruptions to vaccine supply chains while providing better logistical planning and management with predictive and prescriptive analytics. Estimating vaccination demands at the hyper-local level, optimizing current logistic networks, and planning last-mile delivery are a few top priority areas where AI/ML tools could be implemented to improve the availability of vaccines while minimizing wastage.
Streamlining healthcare infrastructure
AI and ML have made a positive impact in the area of Health Information Systems. These models helped generate customized messaging, in multiple languages and formats, for different groups of stakeholders. AI and ML-powered a host of applications providing streamlined information to the public, including information pertaining to PPE kits, over-the-counter medications, emergency services, and hospital and vaccine availability, among others. Advanced AI/ML capabilities have helped policymakers spread important messages quickly and effectively. Citizens significantly benefitted from this speedy and customized guidance.
Hospital management is another area that has leveraged these advanced mathematical models. The last few months saw hospitals making critical decisions based on real-time information to optimize their resources. AI has helped reduce patient wait times despite massive demand in many cases. It has optimized time for diagnosis by clinicians, allowing for better management of care facilities, medical inventory, real-time patient condition monitoring, and more.
Expanding healthcare capabilities with new analytics models
Clear trends have emerged in terms of AI for healthcare during the pandemic and stakeholders continue to find new, effective use cases as we speak. Even at a broad systemic level, we have seen greater AI adoption during the pandemic with the open-sourcing of large amounts of data. AI-embedded tools have enabled a generation of novel datasets by scraping through web information and general social media platforms. Feature engineering of new and novel parameters provided higher accuracy and continue to build innovative models. As a result, a large number of crowd-sourced models have emerged, helping detect anomalies in official data and news. These datasets are used to create forecasting models across a host of critical use-cases, sometimes with greater accuracy than those previously in use.
Driving continuous improvement in post-pandemic healthcare
Technology has proven its strength in supporting critical life-saving systems and processes. Even after the pandemic is behind us, healthcare organizations will encourage the usage of external data and internal data to increase the reliability of prediction models. The increasingly higher acceptance of social listening will help track the next pandemic much faster. Smart NLP tools will be explored further to match patterns from the humongous archive of past research data which will accelerate preventative and prescriptive responses to disease management. With the emergence of health technology companies capturing the most granular patient data, conducive data sharing platforms needs to be provided, as this digital dust can be the next game-changer from the public health management perspective. This data can be mined to infer disease lifecycle, the population at risk, demographic influence, etc. Workplaces will install IoT sensors for monitoring employees’ general health and behavior to recognize any unwell staff. We believe that the potential of AI goes way beyond curbing and abating disasters. It will soon pave the way for robust and effective systems that prevent healthcare disasters.
