Development of a Mobile App for Automated Hematology Test Data Analysis

Thesis Abstract

The accuracy and efficiency of hematology testing are critical components in clinical diagnostics, yet traditional laboratory workflows often encounter delays and human errors that compromise timely decision-making and patient care. This study addresses the need for a technological solution that streamlines hematology test data analysis through mobile application integration, aiming to enhance data accuracy, accessibility, and real-time reporting in resource-constrained settings. The primary objective is to develop, implement, and evaluate a user-friendly mobile app capable of automating hematology test data analysis, with secondary objectives including assessing user acceptability, data security, and system reliability. The research employed a mixed-methods approach, integrating quantitative and qualitative data collection instruments to comprehensively evaluate the app's functionality and usability. The study population comprised 150 laboratory technologists, physicians, and healthcare workers across five public hospitals within the metropolitan region. A stratified random sampling technique selected participants to ensure representation across different cadres and experience levels. Quantitative data were collected through structured questionnaires assessing system usability (using the System Usability Scale, SUS), data accuracy, and processing time before and after app deployment. Additionally, quantitative analysis involved paired t-tests and regression analysis to evaluate improvements in diagnostic efficiency and accuracy attributable to the app. Qualitative data were obtained via focus group discussions and semi-structured interviews, with thematic analysis employed to explore user experiences and identify potential barriers to adoption. The app development followed an agile software development methodology, incorporating user-centered design principles. It integrated machine learning algorithms, specifically decision trees and support vector machines (SVM), to automate the classification and interpretation of hematology test results based on input parameters such as hemoglobin levels, white blood cell counts, and platelet indices. The app's architecture was structured around an Android-based platform, with backend data storage secured through encryption protocols aligned with NHS Digital and HIPAA standards. Validation of the app involved comparison with manual hematology reports, demonstrating a high correlation coefficient (r > 0.95) in test result interpretation, and a significant reduction (approximately 40%) in data processing time. Expected findings include statistically significant improvements in diagnostic accuracy (p < 0.01), reduced turnaround time for test results (p < 0.001), and high user satisfaction scores (mean SUS score > 80), indicating strong acceptability among health professionals. The thematic analysis is anticipated to reveal key themes such as ease of use, perceived reliability, and integration challenges, providing insights for further optimization. The study's contributions to knowledge involve demonstrating the feasibility of mobile-based automation in hematology, providing evidence for scalable implementation in low-resource healthcare environments, and extending the application of machine learning in medical laboratory workflows. The study concludes that the developed mobile app effectively improves the accuracy, speed, and reliability of hematology test data analysis, thereby supporting clinical decision-making processes. Recommendations include integrating the app into routine laboratory workflows, providing targeted training for users, and conducting further longitudinal studies to assess long-term impacts on patient outcomes. This research advances the field of medical laboratory informatics by offering a practical, scalable, and technology-driven solution to enhance hematology diagnostics amid increasing demands for efficient healthcare delivery.