Optimal Vaccine Allocation Using Recommendation-as-a-Service

Given the global nature of the Covid-19 crisis, a high degree of coordination is needed among global bodies such as the World Health Organization, National Governments, and local governing bodies. Local governments have the most granular data about the infected individuals, community spread, and hospital capacity.

Let’s take the example of officials in the State Health Department of a medium-sized US state that is battling with an increase in Covid-19 cases. A vaccine has just been approved by the US government. In all likelihood, the vaccine will be in limited supply initially, and they’ll need an optimal allocation strategy for saving lives and bringing the virus under control.

How can mathematical models help with optimizing Covid-19 vaccine allocations?

Given the urgency and scale of the problem, healthcare professionals and decision-makers need a solution that is easy to implement and test different scenarios. Advisory groups set up in countries like the US have already put forward drafts of plans to vaccinate different groups, while the World Health Organization (WHO) has created a global vaccine plan called Covax to buy and fairly distribute the vaccine worldwide.

To solve the problem of optimal allocation at a local level, there has been some preliminary research work published by researchers from the University of Washington. In this work, they developed a mathematical model of SARS-CoV-2 transmission and used optimization algorithms to determine optimal vaccine allocation strategies. They chose four different metrics, i.e deaths, symptomatic infections, and maximum non-ICU and ICU hospitalizations that they looked to minimize.

Fig 1. The mathematical model developed using the traditional way

Using this model they simulated several scenarios they found that any vaccine with medium to high vaccine efficiency would be able to considerably slow the epidemic while keeping the burden on healthcare systems manageable, then figure out the best target group first.

How can a “recommendation-as-a-service” solution, like Caboom, help to optimize patient prioritization?

Research, as we showed above, is just the start and needs to go through many iterations and accommodate other external factors such as political, ethical and societal.

At Leapfrog, we incubated a highly flexible “Recommendation-as-a-service” tool: Caboom using our expertise gained over the years in data science and scientific research projects. Caboom helps to speed up such research work by providing the ability to create and quickly iterate through powerful AI and mathematical modeling capabilities. It comes with features such as managing model development into projects, dataset management, and modules for each step of the process such as data processing, exploration, modeling, visualization and deployment. This means they don’t have to manage their experiments manually, which is a big issue in large collaborative teams.

Risk model by Caboom

Fig 2. Mathematical model developed using Caboom

Because of Caboom, we now have the following advantages:

Using Caboom means that the scientists and researchers can work on what they know best without having to worry about manually keeping track of their models, research and better collaboration with their peers.

If you are looking to solve problems like this or build any prioritization and personalization solution, send us a message and we’ll have a chat!

*Caboom is an internal project of Leapfrog.

*Also published at caboom.ai.

AI Team

This blog is written by the AI team at Leapfrog.

More in Blogs

Using FAISS Advanced Concepts to Speed Up Similarity Search in Recommender Systems  (Part II) Artificial IntelligenceInsights

Using FAISS Advanced Concepts to Speed Up Similarity Search in Recommender Systems (Part II)

Continued from the last blog. Improving on Flat Indexes Beyond the flat indexes that perform exhaustive searches, FAISS also has

Read more
Using FAISS basics to Speed up similarity search in Recommender Systems (Part I) Artificial Intelligence

Using FAISS basics to Speed up similarity search in Recommender Systems (Part I)

Measuring the similarity between vectors or arrays is a calculation that we encounter often while developing recommendation systems or other

Read more
Why is a No-Code Recommendation System worth a try? Artificial IntelligenceInsights

Why is a No-Code Recommendation System worth a try?

Lately, in the tech world, no-code systems have been getting immense attention, and for good reasons. These types of systems

Read more