OBJECTIVE
To develop a methodology for automated sociological survey data processing and to create a prototype of innovative information-analytical system for presenting results in graphical form and formalized textual format to support decision-making in healthcare.
MATERIAL AND METHODS
The study was based on a demonstration dataset designed to assess preparedness of healthcare organizations for emergencies. Each record in dataset included demographic and professional characteristics of respondent, as well as answers to 25 questions covering key aspects of preparedness for emergencies (presence of plans for epidemics and disasters, resources, training, interaction with emergency services, etc.). Data preprocessing included removal of missing entries and duplicates, validation via a derived variable and outlier detection using interquartile range. Exploratory data analysis was conducted using descriptive statistics and visualization. For group comparisons, a suite of statistical tests was applied. Natural language processing (NLP) methods were applied for processing of open-ended text responses. Textual responses were converted into quantitative assessments or thematic categories using semantic modeling and classification (a pre-trained RuBERT model was used).
RESULTS
A prototype of information-analytical system (IAS) for automated survey data processing was developed. This prototype was tested on survey of medical professionals evaluating preparedness of healthcare organizations for emergencies. The system is built on open technology stack and includes modules for data storage, processing (cleaning and aggregation), analysis of textual responses, and visualization of results providing a complete cycle of transforming raw survey data into informative reports. The information-analytical system ensures comprehensive analysis of large-volume survey data, identification of hidden patterns, and creation of interactive dashboards, charts, summary tables, and heatmaps. This tool enables healthcare professionals to obtain significant and up-to-date information on preparedness of healthcare system, particularly for emergencies and global challenges.
CONCLUSION
The study demonstrated the feasibility of integrating classical statistical analysis methods and NLP technologies on a single platform to assess healthcare system preparedness for emergencies. Original IAS is effective as a decision-making support tool. This system allows for transforming disparate survey data into quantitative indicators, promptly identifying problem areas in healthcare system and tracking their changes over time. The proposed methodology can be scaled and adapted to other tasks contributing to accelerated digital transformation of healthcare system.
