Enhancing Healthcare Data Accessibility: A Web-Based Laboratory Test Result Analysis System

Abstract

Medical laboratory test results serve an important role in healthcare, helping in diagnosing, treating, monitoring, and preventing illnesses. However, its intricate nature can be confusing for a non-professional person, possibly causing misinterpretation, which can be a primary cause of emotional distress and struggles in making informed choices. Hence, interpreting medical laboratory test results may require assistance from a physician, thereby causing inconvenience and additional expenses.

This study implemented an experimental research design to determine the feasibility of the hypothesis. Meanwhile, the building phase took different stages: installation of the necessary software and programs, integration of EasyOCR model and GPT-3 API, and the combination of the components through different software, including HTML, CSS, Python, and JavaScript. Subsequently, the researchers tested the website's efficacy and reliability through different tests, such as a comparative test between the result provided by the doctor and the system, doing repeated tests to determine consistency, recording the response time, and detecting text boxes for the assessment of the OCR.

The result showed that the website's NLP accuracy stood at 87.11%, with a consistency rating of 92.85%. Meanwhile, the average response time was 52.06 seconds, and the OCR accuracy was determined to be 98.84%.

The total findings statistically outperformed the former studies, demonstrating an excellent performance of the web-based laboratory test result analysis system that suggests potential benefits for healthcare delivery enhancement while still leaving room for improvement. Enhancements to the system, such as using different natural language processing and optical character recognition models, are recommended, as well as widening the parameters covered by the system.