Breaking Down Social Bias in Artificial Intelligence Algorithms for Cardiovascular Risk Assessment

K. C. Sabreena Basheer 05 Jun, 2023 • 3 min read

Cardiovascular disease (CVD) prevention is crucial for identifying at-risk individuals and providing timely intervention. However, traditional risk assessment models like the Framingham Risk Score (FRS) have shown limitations, particularly in accurately estimating risk for socioeconomically disadvantaged populations. Hence, researchers are now exploring the potential of artificial intelligence (AI) and machine learning (ML) algorithms to improve risk assessment. A recent groundbreaking study shows the use of AI algorithms to enhance cardiovascular risk assessment. What makes it better than traditional models is that it addresses the inherent bias in the FRS.

Also Read: Decoding the Blueprint of Life: AI’s Geneformer

Artificial Intelligence (AI) & Machine Learning (ML) algorithms can now assess the risk of CVD better than the Framingham Risk Score (FRS).

The Framingham Risk Score: Assessing Cardiovascular Risk

The Framingham Risk Score is a well-known algorithm to estimate an individual’s 10-year cardiovascular risk. Originally developed based on data from the Framingham Heart Study, the FRS primarily focuses on estimating the risk of developing coronary heart disease. However, it has been observed that the FRS may not accurately assess risk for specific demographic groups, such as those facing socioeconomic disadvantages.

Potential of AI and ML in Risk Assessment

Artificial intelligence and machine learning technologies offer promise in addressing the equity gaps and biases found in traditional risk assessment models. By leveraging vast amounts of data and complex algorithms, AI has the potential to provide more accurate risk estimates and improve patient outcomes. Nevertheless, a critical appraisal of these algorithms is necessary before integrating AI-informed decision-making into clinical practice.

Also Read: AI Discovers Antibiotic to Combat Deadly Bacteria

Unveiling Bias: Analyzing ML Algorithms

This study aims to employ an equity lens to identify and understand sources of bias within ML algorithms that aim to improve cardiovascular risk assessment relative to the FRS. Researchers will focus on aspects like race/ethnicity, gender, and social stratum to assess the extent of bias within these algorithms. The goal is to uncover and address any unfair discrepancies that may arise in the risk assessment process.

Methodology: A Comprehensive Literature Search

A comprehensive literature search will be conducted to address whether AI algorithms designed to estimate CVD risk and compare performance with the FRS can effectively address bias. Databases such as MEDLINE, Embase, and IEEE will be scoured for relevant studies. The search will not impose any date filters, although it will apply English language filters.

Eligibility Criteria for Inclusion

The study will consider papers describing specific algorithms or ML approaches that provide risk assessment outputs for conditions such as coronary artery disease, stroke, cardiac arrhythmias (specifically atrial fibrillation), heart failure, or global CVD risk scores. To maintain the focus on prevention, the study will not include papers discussing algorithms solely for the diagnosis of CVD.

Also Read: Groundbreaking News: FDA Grants Approval to Elon Musk’s Neuralink for Human Trials

Analyzing and Reviewing the Literature

The included studies will go through a thorough, structured narrative review analysis. This analysis will clarify the potential biases present within the AI algorithms with regard to the FRS. It will shed light on the effectiveness of these algorithms in mitigating bias and improving cardiovascular risk assessment. The study has received an ethics exemption from the General Research Ethics Board at Queen’s University.

A peer-reviewed journal will accept the completed systematic review for publication. Additionally, key findings from the study will be presented at relevant conferences. Researchers aim to drive discussion, collaboration, and further advancements in AI algorithms for cardiovascular risk assessment by disseminating these results.

Also Read: Breaking Barriers: ChatGPT’s Radiology Exam Triumph and Limitations Unveiled!

Our Say

The study aims to address the social biases inherent in the Framingham Risk Score by exploring the potential of AI & ML algorithms in cardiovascular risk assessment. By critically analyzing existing literature and uncovering sources of bias, researchers hope to pave the way for fair and accurate risk assessment tools that can benefit all individuals, regardless of socioeconomic background. This study represents a significant step towards improving healthcare equity and enhancing cardiovascular disease prevention. It also opens new avenues for the use of artificial intelligence and machine learning in healthcare.