University of Twente Student Theses
Identifying unknown variables influencing contract costs at MST
Straus, J.V.D. (2025) Identifying unknown variables influencing contract costs at MST.
|
PDF
2MB |
| Abstract: | We identify unknown variables influencing contract costs within the Contracts & Process Management subdepartment at Medisch Spectrum Twente (MST). MST, a top clinical hospital in the Netherlands, manages several external contracts critical to operational efficiency and cost control. The subdepartment experiences significant cost fluctuations, complicating resource allocation. The aim of this research is to uncover unknown variables that cause these fluctuations using data-driven methods, particularly Machine Learning models, to ultimately improve forecasting accuracy and contract monitoring. This research adopts the CRISP-ML, using anomaly detection and regression techniques to identify cause-and-effect relationships between variables. Due to their financial impact, we focus on two contracts—clinical chemistry and laundry. Anomaly detection and deletion were performed on normalised costs to exclude anomalies in contract costs before performing several regression methods. Linear regression, multiple linear regression, and statistical two-way analysis of variance were used to identify relationships between variables. Shapley additive explanations, Shapley additive explanations with an adjustment factor, and random forest Regression were used to quantify variables’ impacts on contract costs. Random forest regression was used to determine the accuracy when incorporating the statistically significant variables into the model. The random forest model predicted the costs of the laundry contract with a promising result (R² = 0.5926). In contrast, the prediction for the clinical chemistry contract showed a negative R² value (R² = -0.2955), highlighting a need for significant improvement in the model's performance for this context. Integrating the found variables into forecasting models can help MST optimise resource allocation and improve operational efficiency. Incorporating periodic anomaly detection into contract management practices will further enhance monitoring capabilities. This methodology offers a replicable approach for analysing other contracts at MST. The methodology can be shared with other healthcare institutions to drive improvements across the Dutch healthcare system. By utilising these insights, MST can address cost fluctuations, improve decision-making, and achieve greater control over contract management. |
| Item Type: | Essay (Bachelor) |
| Clients: | MST, Enschede, Netherlands |
| Faculty: | BMS: Behavioural, Management and Social Sciences |
| Subject: | 54 computer science, 58 process technology, 83 economics |
| Programme: | Industrial Engineering and Management BSc (56994) |
| Link to this item: | https://purl.utwente.nl/essays/104894 |
| Export this item as: | BibTeX EndNote HTML Citation Reference Manager |
Show download statistics for this publication
Show download statistics for this publicationRepository Staff Only: item control page