Enabling healthcare professionals to analyze public healthcare datasets without technical expertise.
US medical professionals need published research for career advancement, but most lack the technical skills to analyze large healthcare datasets. Medical students and early-career doctors face an even bigger challenge—producing publishable research while managing clinical duties without technical support.
Softblues Solutions built Lumono.ai, an AI-powered platform that lets healthcare professionals conduct medical research using natural language. The system analyses complex public healthcare datasets and generates publication-ready statistical analyses and reports—no programming knowledge required.
The Database Exploration section allows users to browse and categorize available dataset variables (Demographics, Dietary, Examination, Laboratory, Questionnaire). Users can review descriptions and 'like' relevant variables for their study. Selected variables can then be used to generate research questions in subsequent steps.
The Research Question Generation section helps users define their study's focus based on selected variables. Users can either automatically generate a research question or manually input their own. This accommodates both guided and independent research approaches.
Users review approximately 10 automatically generated research questions created from their selected variables, each exploring different relationships between population, exposure, and outcome. Users can browse and compare the suggested questions to find one that best fits their study objective. The selected question then serves as the foundation for defining the study design and conducting further analysis.
After defining the research question, the system automatically categorizes selected variables into population, exposure, and outcome groups. Lumono assists in cleaning and validating data to ensure consistency and reliability before analysis. Users can customize data inputs, adjust details, and provide preferences for information processing and display in subsequent analytical steps.
The Primary Analyses section provides initial statistical evaluation using descriptive and univariable analyses. Users view summary statistics to understand data characteristics and distribution. Basic regression and univariable testing reveal potential associations between variables.
The Advanced Analyses section provides deeper statistical insights to test finding robustness and evaluate effect modification. Users analyze subgroup interactions to understand result variations across populations. Advanced methods confirm findings and strengthen scientific validity.
PLATFORM ARCHITECTURE
