Project case: Sorgente

Sorgente is a Dutch company specialised in medical nutrition. They work with manufacturers of medical nutrition, healthcare professionals and health insurance companies to optimise medical nutrition care for patients. Not only do they supply medical nutrition, they also have a customer care service that both users and health care professionals can contact whenever they have questions about medical nutrition.

Being a specialist in medical nutrition, Sorgente possesses a large database with information about patients, prescriptions and orders. The aim of this project was to analyse the data and use the outcomes for a poster presentation at ESPEN (an annual European nutrition congress) as well as for a campaign launched by Sorgente. The goal of the campaign was to identify patient characteristics and order behaviour of compliant patients, following up on the previous campaign which was about getting to know the patients better and better each day. Another part of the new campaign was to share these insights with prescribers in order to stimulate patients’ compliance and thereby enhance their health benefits and wellbeing. However, while the database was there, the means and time to kick-off the project were lacking. To help Sorgente to structure and analyse their data, Louella Schoemacher, statistical programmer at OCS Life Sciences, started an assignment on site in June 2020.

Louella was a good match for this project for she was able to combine the knowledge she gained during her studies Nutrition and Dietetics with her experience from working as a statistical programmer during the past years. The assignment started with getting familiar with the data and systems used by Sorgente. The next step was to organise the data and to structure it in a way that it could be used for analysis.

As the relational database was optimised for inserting data it is less suitable for analysing data. Therefore restructuring the data was the most important and time consuming step in this project  with multiple challenges. First, the data had to be retrieved from multiple data sources, resulting in different types of data with multiple levels of information that all had to be combined. Furthermore, after structuring the data, some of the research questions and definitions had to be carefully reconsidered as, only after the structuring of the data, it became clear whether the research questions could actually be answered with the data.

The approach to this project was very interactive; with brainstorm sessions and re-evaluating decisions. The result was  a dataset containing data from over 2000 customers and over 10.000 prescriptions. The characteristics of these customers and prescriptions were succesfully presented on a poster during the ESPEN congress.

After the ESPEN congress had taken place it was time to really dive into the resulting dataset and to see what conclusions could be drawn with regards to patient order behaviour and factors that could be influenced by presribers to increase a patients’ compliance. The dataset provided great insights in prescription characteristics and order behaviour in different types of diagnoses.

For Sorgente this project led to more insight into their data structures. Where the systems might be suitable for day to day practice, the way data is stored behind the scenes may not suit all research purposes. Providing insight into corporate data structures and restructuring it into analysis-ready datasets in close cooperation with the client is a perfect example of how OCS can help to get the most out of your data now and in the future. For OCS Life Science this was another real life experience on how data preparation for analysis purposes can take a large chunk out of a data science project’s total resource requirements and how to deal with the challenges effectively.