# Data science applied to the assessment of biological variation estimates

## TL;DR

This paper explores how data science can improve the estimation of biological variation using real-world data in clinical labs.

## Contribution

The paper introduces the novel use of data science algorithms to estimate biological variation from real-world data.

## Key findings

- Current algorithms for biological variation estimates are reviewed and described.
- Real-world data offers potential to enhance understanding of biological variation.
- Direct methods for biological variation have known advantages and drawbacks.

## Abstract

Data science is an umbrella term encompassing a set of tools and processes that make it possible to extract new information from structured or unstructured databases. This scientific discipline is gaining relevance in healthcare. In the clinical laboratory, the multiple applications of data science include the development of algorithms for obtaining population-based reference intervals or biological variation (BV) estimates. These algorithms contribute to overcoming the drawbacks of traditional or direct methods.

A review was performed of the state-of-the-art in algorithm-based methods for obtaining BV estimates using Real-World Data (RWD) in the field of data science.

A description is provided of the structure of the algorithms currently available for obtaining BV estimates based on the scientific evidence available. An overview is provided of the advantages and drawbacks of direct methods.

The use of RWD to obtain BV estimates is a novel discipline with a considerable potential for improving our understanding of BV.