Exam Sentinel: Ambev AI-Powered Health Monitoring in Brazil
- Customer
- Zerrenner Foundation
- Project manager on the customer side
-
- Year of project completion
- 2023
- Project timeline
- March, 2023 - December, 2023
- Project scope
- 1250 man-hours
- Goals
- This project consisted of developing an AI that monitors and analyzes all the laboratory exam results we receive and compares them with a reference table for each type of exam. When the AI identifies a variation in an indicator or set of indicators, it automatically triggers one of our doctors to monitor the beneficiary and direct him/her, in a preventive manner, to a follow-up with a specialist.
- Project Results
- We currently receive around 3,000 exams per month and in 8 months of using the tool we were able to identify and preventively direct almost 100 cases of beneficiaries who could be facing more serious problems and with our prior action we were able to anticipate treatments and prevent the worsening of cases. We believe that with this approach we almost eliminated the risk of hospitalization and unnecessary intervention of the partner network.
The uniqueness of the project
This project was developed in partnership with a large network of laboratories, with which we developed an integration using APIs to receive all the results of exams performed by our beneficiaries in the laboratory network. We created an API HUB that identifies who our beneficiaries are and sends us their exam results. In this way, as a Foundation whose main objective is to take care of the health of our beneficiaries in a preventive manner. This way, in addition to reducing healthcare costs, we are also able to direct beneficiaries to the best possible treatment at an early stage.- Used software
- We used AWS Cloud with API Rest to develop the integration between us and the Laboratories and C# Language to develep the AI engine and Document DB to store the datas.
- Difficulty of implementation
- The biggest challenge in developing this project was convincing our partner, a laboratory network, to share the test results of our beneficiaries. The partner's biggest concern was due to the general data protection law, implemented in Brazil in 2018. We held several meetings, including with our legal department, to demonstrate that we are doing everything within the law and that all security resources have been adopted to ensure that the data is encrypted and all security "guardrails" have been implemented.
- Project Description
-
Zerrenner Foundation is the controlling shareholder of two important groups in Brazil: Ambev that is Owner of the brands Budweiser, Stella Artois and Corona, Antarctica, Brahma, among other and Itaúsa that is the holding of Itaú Bank, among other companies. We are responsible for providing healthcare, education and well-being for all Ambev employees and their dependents.
There are approximately 67 thousand lives that we support. We also have 2 schools where we provide free education for more than 3,000 students in São Paulo and Minas Gerais. Regarding the project, Zerrenner Foundation's areas of main activity is precisely the area of Health, where through health operators we provide the best service to our beneficiaries (Ambev Employees). Furthermore, we have Family Doctors who work in Ambev's large factories, we also have partnerships with laboratories and hospitals to receive the results of exams in a structured way carried out by our beneficiaries, the objective of which is to act preventively through prevention programs.
Exam Sentinel is the name of our project that aims to preventively identify changes in laboratory exams of our beneficiaries. This project consists of developing an AI that monitors and analyzes all the laboratory exam results we receive and compares them with a reference table for each type of exam. When the AI identifies a variation in an indicator or set of indicators, it automatically triggers one of our doctors to monitor the beneficiaries and direct them, in a preventive manner, to a follow-up with a specialist. This project was developed in partnership with a large network of laboratories, with which we developed an integration using APIs to obtain all the results of exams performed by our beneficiaries in the laboratory network.
The project was born precisely because we have this structure, our own doctors and partnerships with hospitals and laboratories, so that we can act preventively when we receive the test results. We created an algorithm that monitors these tests and alerts our doctors when something is not in line with our references. With this, depending on the result and which indicators are altered, our doctors contact the beneficiary and begin the work of recovering these indicators, bringing the beneficiary to what we call the "Correct Health Path". At the end of the day, we are able to provide the best possible quality of treatment at the most appropriate cost, applying genuine health care to our beneficiaries. This approach removes our beneficiaries from the network of health insurance companies, most of whose providers only want to make money with unnecessary or inadequate treatments. We have had excellent results in monitoring our beneficiaries and consequently reduced health costs, such as visits to the emergency room, reduction in hospitalizations and visits to ICUs.
We created an API HUB that identifies who our beneficiaries are and sends us their exam results. In this way, as a Foundation whose main objective is to take care of the health of our beneficiaries in a preventive manner. With this, in addition to reducing healthcare costs, we are also able to direct beneficiaries to the best possible treatment at an early stage. We also had the support of our own doctors to indicate which exams and/or sets of exams we should evaluate and which reference tables we should use to compare the exam results. The project lasted 9 months, during which we had the connection stage (Idea), then the planning stage, where we defined all the parameters and standards that we would use, and then we entered the execution and testing phase before putting it into production in December 2023.This project was developed entirely by the Foundation's internal team and the laboratory network team.
- Project geography
- This project covers 2,200 cities in Brazil where Ambev's large factories and our beneficiaries are located.