CHAIN and Sybil: AI Tools to Improve Vaccine Efforts and Patient Scheduling
Wed, April 21, 2021

CHAIN and Sybil: AI Tools to Improve Vaccine Efforts and Patient Scheduling

 

Vaccines are the most recommended drugs to gain immunity from certain diseases. Recently, a startup proposed its plans to use artificial intelligence (AI) to enhance vaccine delivery and patient scheduling to improve the global vaccine efforts.

The startup behind the AI-driven vaccine augmentation project is called macro-eyes. It is currently seeking to improve vaccination programs in different parts of the globe. Its project involves a vaccine forecasting tool that leverages a unique combination of real-time data. The tool has been called Connected Health AI Network or CHAIN. The US private research university Massachusetts Institute of Technology (MIT) is working with the said startup company.

The State of Vaccination in the World

According to Our World in Data, an online source of research data, the main beneficiaries of vaccines are infants and younger children. These drugs can give them immunity to various vaccine-preventable diseases, many of which are incurable and can cause lifelong disability. While most vaccines can induce immunity, some can only give resistance for a certain period. One example is the vaccine for influenza that needs to be obtained every year. Doctors still recommend the flu shot because any percentage of protection is better than nothing, especially if multiple influenza strains are roaming in communities.

In 2018, 89% of children aged 12 months have been immunized against tuberculosis, leaving 11% without protection from the deadly respiratory illness. It was followed by 86% of children immunized against measles at first dose, 86% of children immunized against diphtheria, tetanus, and pertussis, 85% of children immunized against polio, 84% of children immunized against hepatitis B, 72% immunized against Haemophilus influenzae type B, 69% of children immunized against measles at first and second doses, 47% of children immunized against pneumococcal infections, and 35% of children immunized against rotaviruses. The low percentage of children immunized against rotaviruses showed that about 65% of children could be at risk of diarrheal diseases, one of the leading causes of pediatric deaths.

Aside from providing immunity among people, the second main goal of vaccination programs is to raise herd immunity in communities. Some people cannot be vaccinated due to preexisting health conditions, which make their bodies prone to even weakened pathogens in vaccines. But herd immunity or being surrounded by lots of other people immune to diseases can shield them from vaccine-preventable illnesses.

According to publicly available data, measles has the highest herd immunity threshold in the world between 92% and 95%. The disease's basic reproduction number or R naught is between 12 and 18. Pertussis follows with a herd immunity threshold of between 92% and 94%, and R naught of 12 to 17. Diphtheria is at third place with a herd immunity threshold from 83% to 86%, and R naught of six to seven. Influenza has the least herd immunity threshold from 33% to 44%, but its R naught is only from 1.5 to 1.8. Its low threshold score can be correlated to the lack of a vaccine with a 100% protection rate.

 

 

AI Leverages Data Sources to Improve Vaccine Efforts

At MIT, the founders of macro-eyes revealed the ongoing research of an AI-driven project to boost immunization programs. Its main objective is to increase vaccination rates, reduce vaccine wastage, and protect as many people from vaccine-preventable diseases. The objective can be significantly helpful among medical frontliners, who are experiencing numerous challenges in immunizing most population groups. Even if some do not wish to be vaccinated for whatever reason, the project remains a crucial augmentation in bolstering herd immunity threshold.

"Health care is complex, and to be invited to the table, you need to deal with missing data. If your system needs age, gender, and weight to make predictions, but for one population you don't have weight or age, you can't just say, 'This system doesn't work.' Our feeling is it has to be able to work in any setting," said Benjamin Fels, Chief Executive Officer and a co-founder of macro-eyes.

MIT associate professor Suvrit Sra, also a co-founder of the startup, detected a problem in the healthcare system with Fels since they worked together. The problem was the domination of algorithms in financial markets. Upon comparison, they found that healthcare was the least algorithm-based sector in the world. This led them to develop machine-learning algorithms back in 2013. Those algorithms were designed to measure the common factors among patients to determine better treatment plans.

Multiple data sources, such as images, medication details, financial information, and hospital costs, were utilized by those algorithms. Only through a grant from the Bill and Melinda Gates Foundation were they able to use the same algorithms to be used in vaccine utilization models. After those models were created, they were tested in select regions for pilot runs. Mozambique and Tanzania were chosen for testing and the goal was to make vaccine supplies more proactive.

 

 

The pilot runs highlighted the potential of CHAIN to decrease the likelihood of vaccine wastage. As of late, the algorithms reduced the wastage by 96% in three regions of Tanzania. While working with CHAIN, the experts developed another machine-learning tool called Sibyl. Unlike CHAIN, Sibyl specifically forecasts the chances of patients to show up for appointments. The tool could determine how many people would require non-emergency medical services. This would properly tweak the busy schedule of healthcare workers, particularly in this year of COVID-19.

Sibyl, the scheduling platform, has been trained with more than 6 million records of hospital appointments. The AI identified and analyzed the behaviors of patients and measured how many would appear in health facilities, how many would reschedule, and how many would ultimately fail their schedules. The most recent result of its testing showed a decrease in wait times of over 75% at one of the largest heart hospitals in the US.

The World Health Organization of the United Nations predicted in May 2020 that at least 80 million children younger than 12 months are at risk of vaccine-preventable diseases. With public quarantine protocols and rising cases in COVID-19, the global healthcare sector had to divert its manpower to the pressing problem in clinics and hospitals. On top of that, severed supply lines impacted vaccine supplies in various regions. Any tools to enhance these problems would be helpful.

In addition to augmenting vaccination programs and fixing schedules of healthcare workers, the founders are looking into using Sybil in the battle against COVID-19. One of the possible uses is directing COVID-19 patients to health clinics or hospitals with sufficient patient capacity. This way, hospitals can avoid wasting the time of patients and limit their movements outside while searching for a capable facility.