Improving Online Clinical Trial Search Efficiency Using Natural Language Processing and Biomedical Ontology Mapping Approach

Summary form only given. With the increasing demand for healthcare, computer technology and the internet are playing a more important role for patients, practitioners, and researchers. Oftentimes the process of seeking or providing care does not start in a waiting room or in a doctor´s office, but online. Because of this, special attention must be concentrated in increasing the efficiency of online search engines so that the rest of the care process may also run smoothly. This research proposes a solution to improve the search efficiency for patients using natural language processing and SNOMED mapping techniques. In this research, clinical trials are extracted from ClinicalTrials.gov and n-gram method is applied to process the clinical trial contents. The processed terms are then mapped to SNOMED terms and a covariance matrix is formulated with the Jaccard similarity coefficient measuring similarity between a pair of clinical trials. Based on the similarity measures, the most relevant clinical trials are extracted for the searcher´s needs. In the end, a comparative study is conducted to prove the enhancement of the search efficiency. In conclusion, the combination of n-gram model and SNOMED terminology mapping to process the clinical trial contents is proved to improve the efficiency for the online search of the clinical trials. Future research with clinical trials will use multiple methods such as ontological and statistical approaches to improve the precision and recall of the search results. Another interesting next step may be to explore clustering by analyzing the correlation structure of the clinical trial contents.