Integrating and Validating Heart Health Calculators in Medical Platforms

Abstract

We believe the underutilization of clinically validated algorithms for heart health assessment in current medical platforms undermines patient confidence in using associated screening instruments. The importance of earlier detection and improved health outcomes in heart disease motivated the development of these screening instruments. Available risk assessment tools are stand-alone online apps or screening devices that typically do not consider individual risk profiles. Combining risk estimation with screening instruments like Medsensio allows for targeted heart health assessment, potentially leading to earlier detection, resource optimization, and increased patient engagement.

The primary objective of this master project was to develop, test, and apply a screening tool in a medical framework.

This study presents the researcher's methodology for developing and evaluating risk prediction algorithms for heart disease. We followed a thorough process in developing the Medsensio Screening tool. A literature analysis informed the tool's design, ensuring it aligned with recent studies on preventative healthcare. The technique then comprised detailed design and implementation, internal testing, established algorithm utilization, rigorous integration evaluation, and in-depth user/physician workshops for input.

The primary outcome of this thesis is a functional risk assessment tool embedded within a state-of-the-art medical system. According to the findings, the Medsensio Screening Tool achieved higher acceptance rates than its web-based counterpart. User Evaluation further demonstrates that individuals trust the screening tool's output. Moreover, the tool uses LLMs to provide consumers with customized recommendations. Our results show that the LLM's recommendations generated considerable curiosity and inspired user trust, indicating a successful user experience.

This thesis not only demonstrates the successful integration of a risk assessment algorithm into a medical platform but also holds promise for the future of heart disease prevention. The study found that users highly accepted the tool and trusted its results, indicating a positive shift in patient engagement. Additionally, the tool utilizes LLMs for personalized recommendations, a feature that could significantly enhance the tool's effectiveness. Future work will explore strategies to further enhance user experience, paving the way for a more proactive approach to heart health. Moreover, subsequent investigations should verify its effectiveness and evaluate its potential to positively influence medical results.