OBJECTIVE To introduce the confirmation of research issues related to hospital-associated infections, data cleaning and selection of univariate analysis method.
METHODS The data of hospital-associated infections were characterized by complexity and variability, the research subjects were determined through the framework of 'population-intervention- comparison-outcome-study design' (PICOS). The data cleaning serves as the foundational processing for statistical analysis, and its core modes included missing value processing, repeated value processing and abnormal value processing. An appropriate method for statistical analysis of medical data plays an crucial role in drawing a reliable, valuable and significant research conclusion. The previous common errors were sorted out by reviewing literatures, which may help to avoid the repeated occurrence of problems regarding the data handling and data analysis.
RESULTS The principles for confirmation of research subjects for hospital-acquired infection, collection and cleaning of related data, and selection of method for univariate analysis of different types of data were introduced. The data cleaning methods and the selection of methods for univariate analysis based on the type of data and research objective, including descriptive analysis, univariate measurement data analysis, univariate enumeration data analysis and univariate ranked data analysis, were emphatically discussed. A flow chart of univariate statistical analysis was established.
CONCLUSION A definite research subject, standardized data cleaning and appropriate method for univariate statistical analysis may facilitate the improvement of scientificity and reliability of hospital-acquired infection-related clinical researches.
DownLoad: