# A Big Data Optimization Approach for Estimating the Time-Dependent Effectiveness Profiles Against Hospitalization for Double- and Single-Dose Schemes: Study Case, COVID-19 in Elderly Mexicans

## TL;DR

This study uses big data to analyze how well single and double vaccine doses protect elderly Mexicans from hospitalization due to COVID-19 over time.

## Contribution

A new method is introduced to estimate continuous time-dependent vaccine effectiveness profiles against hospitalization.

## Key findings

- Two-dose vaccine schemes offer significantly better protection than single-dose schemes.
- Vaccine effectiveness declines after an initial period of high protection.
- The method allows for detailed comparison of vaccine efficacy over time.

## Abstract

Background: The COVID-19 pandemic and its handling have made evident the cardinal role of vaccines in controlling the spread of diseases, especially around developed cities. Therefore, precisely characterizing their response has taken a relevant role. Unfortunately, substantial evidence has proven the time dependence of their effectiveness, requiring new approaches that account not only for single value estimations but also for time changes in the effectiveness. Methodology: A strategy is proposed to estimate a continuous profile representing the time evolution of the effectiveness against hospitalization. Such a strategy is showcased by characterizing the hospitalization behavior of elderly Mexicans during the COVID-19 pandemic (more than 15 million individuals). Results: It is demonstrated that practically total protection against hospitalization can be reached during a noticeable period. However, a substantial depletion in effectiveness occurs after such a plateau. Our methodology provides a continuous profile instead of only a few discrete values, offering insights unattainable by traditional strategies. Furthermore, the obtained profile details allowed for decoupling the effects of each dose independently, enabling the estimation of the expected effectiveness profile for a single-dose scheme. Conclusions: The comparison between both schemes (one or two doses) demonstrated that the two-dose scheme is far superior, offering a better investment for public health authorities. Concerning the strategy, the description capabilities of the proposal highly outperform currently available methodologies, allowing for detailed profiles describing the evolution of efficacy to be obtained. This not only opens the opportunity for fair comparison among available vaccines but also creates a tool for researchers studying the immune responses of polydose vaccines.