Committed to ensuring its future, the hospital we collaborated with adopted a new strategic plan in 2021, based on a vision for 2035, with short, medium, and long-term developments. The strategic plan relies on eight pillars, the fifth of which concerns digital transformation. Indeed, the hospital group aims to be recognized for its digital transformation and intensive use of data.
In Belgium, all hospitals must submit the RHM (Résumer Hospitalier Minimum) files to the Belgian federal government every semester. The RHM is a system of 27 files containing anonymized records of administrative, medical, and nursing data. While this data is a goldmine of information about hospital stays and patient pathways, it is often underused. It was within this context of exploring the content of RHM files that our collaboration with the hospital began.
The hospital wanted to explore the content of these files and an array of related medical data (APR DRG, ICD-10 codes, billing data, mortality indicators, etc.) longitudinally (i.e., over several years) to develop Management Cockpits. These cockpits aim to help department heads and hospital management address their most critical questions.
In collaboration with the medical teams, we developed a set of dashboards for various departments (diabetology, pediatrics, emergency) to assist practitioners and department heads in their decision-making. A general exploration dashboard also allows for free navigation within the data space and is used in brainstorming sessions for the creation of new dashboards for teams.
The developed solution and data model facilitates the easy addition of new data (RHM, DRG, billing, etc.) to expand the analyses performed. The implemented tool answers numerous previously unanswered questions. It is also the first tool of its kind to enable longitudinal and comparative analysis of the hospital’s data. They were able to analyze that they treated more than 430’000 patients (for more than 1.3 million stays) over a 7-year period.
The dashboards are presented in PowerBI, making data usage and exploration relatively simple thanks to the use of numerous filters and intuitive visuals that allow for deep dives into specific topics.
Our team began by developing a thorough understanding of the client’s requirements and data landscape. As medical data is coded, one of the first steps was connecting it to dictionaries, allowing for data labeling and making the data comprehensible for analysis. Much documentation is available online to understand the data structure.
The second step involved cleaning the data and developing a robust and consistent model that allows for the addition of new sources over time. Once the model was developed, we worked closely with the hospital teams on two points:
We regularly engage with the client for feedback sessions and dashboard updates or the integration of new data.
Stay tuned !