Establishing a sorting protocol for healthcare databases

Background: Health information records in many countries, especially developing countries, are still paper based. Compared to electronic systems, paper-based systems are disadvantageous in terms of data storage and data extraction. Given the importance of health records for epidemiological studies, guidelines for effective data cleaning and sorting are essential. They are, however, largely absent from the literature. The following paper discusses the process by which an algorithm was developed for the cleaning and sorting of a database generated from emergency department records in Lebanon. Design and methods: Demographic and health related information were extracted from the emergency department records of three hospitals in Beirut. Appropriate categories were selected for data categorization. For health information, disease categories and codes were selected according to the International Classification of Disease 10th Edition. Results: A total of 16,537 entries were collected. Demographic information was categorized into groups for future epidemiological studies. Analysis of the health information led to the creation of a sorting algorithm which was then used to categorize and code the health data. Several counts were then performed to represent and visualize the data numerically and graphically. Conclusions: The article describes the current state of health information records in Lebanon and the associated disadvantages of a paper-based system in terms of storage and data extraction. Furthermore, the article describes the algorithm by which health information was sorted and categorized to allow for future data analysis using paper records. Significance for public health In our protocol, we explain the process by which health information collected from paper-based records were analyzed to develop an algorithm to allocate appropriate ICD 10 disease codes and categories to each entry. The algorithm allows for the sequential analysis of the information present and serves to overcome the issue of incomplete records. This is especially relevant in developing countries where paper-based records are highly relied upon, noting that our protocol has already been applied to a study in the Ivory Coast. The protocol allows for a health information database to be cleaned and sorted for research in environmental, social and occupational health. Furthermore, since data cleaning and sorting protocols are sparse, it is crucial to establish a standardized tool to enhance the quality of the conducted research.