Designing a scalable and multi-institutional deployable cardiovascular data intergration solution

  • W. De Mulder

Student thesis: Master's Thesis


Technological advancements in healthcare are continuously improving the quality of life. Despite all technological innovations, many clinical systems and departments still operate independently, making it challenging to process and exchange the diversity of registered clinical information unambiguously. The recent growth of digital health initiatives and the Covid-19 pandemic of 2020 strengthens the value of data interoperability in healthcare to connect diverse software applications and information systems across different healthcare providers. The lack of data interoperability in various healthcare environments implies that leading healthcare organizations have to put a lot of effort into taking full advantage of the data-intensive culture without losing the meaning of the information, ultimately resulting in loss-making integration projects. This research highlights the need for standardization in healthcare.
We aim to propose a scalable software solution applicable in a multi-institutional health-care environment, reducing implementation burden and improving integration project profitability. The developed solution focuses mainly on the extraction of cardiovascular in-formation from diverse Electronic Health Record (EHR) platforms across Belgian (Flanders) and Dutch healthcare organizations. To avoid losing affinity with healthcare’s realistic com-plexity, we designed a software architecture and a prototype in collaboration with a Belgian medical institution acting as a vendor-specific EHR reference site. The proposed extraction method relies on a multi-institutional extraction method using vendor-specific standard-ized clinical content. Subsequently, we applied an efficient transformation process on the returned standardized dataset to deliver an unambiguously defined clinical dataset.
The proposed solution’s evaluation relies on three different pillars whereby we validated the proposed solution on correctness, efficiency, and performance. The applied validation mechanism results in an accurate software solution that meets the demands supporting the cardiovascular workflow. The efficiency evaluation methodology relies on a cost model estimating the implementation cost required to implement the prototype in another med-ical setting and estimates an expected cost to scale up the prototype accommodating ad-ditional clinical concepts. Although the validation results do not fully reflect reality due to restrictive factors we had to consider during the implementation phase, the prototype satisfies the demand to offer an accurate, efficient, and performing solution for the unam-biguous representation of cardiovascular data into a multi-institutional setting.
Future research can contribute to scale up the proposed software solution to extract a broader range of cardiovascular concepts. Since the proposed solution focuses only on one EHR vendor, further investigation into similar multi-vendor EHR extraction methods could make the software solution broadly deployable. Moreover, additional research can give us better insights into semantic data mappings’ maintainability if we use dedicated terminology solutions in the software architecture.
Date of Award23 Apr 2021
Original languageEnglish
SupervisorClara Maathuis (Examiner), Arjen Hommersom (Co-assessor) & dr. P.H.M. America (External assessor)

Master's Degree

  • Master Software Engineering

Cite this