Semantic modelling for clinical decision support based on SNOMED CT

From Master Projects
Jump to: navigation, search


About Semantic modelling for clinical decision support based on SNOMED CT

  • This project has not yet been fulfilled.
  • This project fits in the following Bachelor programs: {{#arraymap:Lifestyle Informatics|, |xXx|bachelorproject within::xXx|,}}
  • This project fits in the following masterareas: {{#arraymap:Human Ambience, Knowledge Technology and Intelligent Internet Applications, Information Sciences|, |xXx|project within::xXx|,}}


Description

Introduction

NL Healthcare Clinics are a new group of specialty clinics that provide orthopedic, ophthalmic and dermatological care on 25 locations, serving over 300.000 patients per year. We strive to incorporate new data science techniques into our daily practice, in order to provide better online services, e-health solutions and clinical decision support.

Problem

We aim to upgrade our data analytics platform to include more clinical data and semantically enrich them. However, within the current practice of data modelling there is no encompassing standard or design which we can use. Although many different standards exist, most notably SNOMED CT, research and work is required to integrate them within the setting of specialty care in the Netherlands.

Research Questions

- To what extent can a record be modelled using LOINC and SNOMED CT?

- Which barriers exist in jointly applying LOINC and SNOMED CT?

- How do LOINC and SNOMED CT relate to health record standards (e.g., openEHR or ISO standards)?

The research questions are answered by applying an implementation-driven approach, i.e., through developing and extending a data/entity model that incorporates the following design concepts: - SNOMED CT and LOINC as the lingua franca and de facto standard of clinical terminology

- Extensible with respect to ongoing development of APIs based on clinical object models in the IT domain

- Entity modelling following the design principles of Data Vault by Hans Hultgren, that will be used as a blueprint for implementation of the new version of the data vault.

Results

A proof-of-concept using the NumFOCUS (http://numfocus.org/) stack. A working database, piece of code or the like in PostgreSQL, Python and/or a markup language is preferred. Answers to the research questions.

In collaboration with:

Daniel Kapitan, Healthcare Clinics, daniel.kapitan@nlhealthcareclinics.com 06-24097295 Ronald Cornet, AMC, dept. KIK, r.cornet@amc.uva.nl 020-5665188