| Category | SOFT | P29 | Cloud-based Data Preparation for Medical Observational |
| Research |
| Abstract | The future of healthcare lies in endeavors such as precision medicine |
| where treatment options, clinical decisions, and preventive techniques |
| are tailored to patients based on their exact characteristics. For these |
| advancements to be made possible, clinical data collected from the |
| health records of patients in medical facilities must be obtained for |
| analytics and observational research. |
| |
| Electronic health records (EHRs) are the optimal source of data for |
| medical observational studies, but, unfortunately, current practices for |
| retrieving relevant EHRs lack automation and efficiency. Clinical data |
| mining is a complex endeavor due to the lack of standardization and |
| structure in the natural text comprising health records. In addition, |
| stringent patient privacy requirements at various organizational levels |
| further complicate procedures for preparing data for observational |
| studies. |
| |
| An autonomous, decentralized, privacy-enabled architecture |
| incorporating a novel semantic similarity algorithm is created for |
| automated EHR retrieval in this project. The semantic similarity |
| algorithm has high values of precision and recall, indicating that clinical |
| records relevant to an observational study’s specifications are |
| retrieved. Furthermore, the addition and removal of data-contributing |
| medical facilities does not impact data retrieval at other nodes. |
| Sensitive information present in the EHRs such as a patient's social |
| security number and name is identified and removed before retrieval |
| via an ontology-guided approach. This project promises to reduce the |
| time required for data preparation for observational research from |
| months to minutes. |
| |
| Bibliography | P. Buitelaar, "Ontology-based Information Extraction and Integration |
| from Heterogeneous Data Sources," International Journal of Human- |
| Computer Studies, vol.66, no.11, pp. 759-788, 2008."Paxata Give |
| Analysts Valuable Time Back for Analytics." Ventana Research, n.d. |