Background: Keratoconus is a progressive corneal disease with thinning and scarring of the cornea. Diagnostic and treatment options are usually evaluated in large prospective or retrospective trials. Big data and smart data provide the possibility to analyze routine data for clinical research. In this article we report the generation of a monocentric keratoconus registry by means of computerized data analysis of routine data. This demonstrates the potential of clinical research by means of routine data.
Methods: A "clinical data warehouse" was created from all available routine electronic data. At the time of first presentation, each eye was classified into one out of four categories: suspected, early disease, late disease and status postkeratoplasty. Through integration of multiple data sources the clinical course for each patient was documented in the registry.
Results: A total of 3681 eyes from 1841 patients were included. The median follow-up time was 0.54 years. Patient age was higher in the groups with more severe stages of keratoconus, the proportion of female patients was higher in the group of suspected keratoconus and patient age and male to female ratios showed statistically significant differences between the groups (p < 0.001).
Conclusion: We were able to create a "clinical data warehouse" by linking multiple data sources and normalizing the data. With this tool we established a novel, monocentric keratoconus registry. Only the grading of disease severity and the exclusion of false positive results were carried out manually. In our opinion establishing a structured clinical data warehouse has a huge potential for clinical and retrospective studies and proves the value of the Smart Data concept.
Keywords: Big data; Data; Database; Keratoplasty; Smart data.