University of Twente Student Theses
Battling the labour shortages in the mentally handicapped-care : a data-driven approach
Smeerdijk, Ilan (2024) Battling the labour shortages in the mentally handicapped-care : a data-driven approach.
|
PDF
1MB |
| Abstract: | In this research, we explore the applications of data-driven working in the mentally handicapped-care. By evaluating how previous research can be translated to the specific problem context of the problem owners, and by constructing new data-driven models derived from this problem context, we get an overview of how data science can support decision-making within organisations in the mentally handicapped-care, and within the field of healthcare in general. This research is conducted at Trajectum, which specialises in the treatment of people with a light mental handicap. As an example of what data-driven working can entail for organisations such as Trajectum, we construct a risk prediction model, which assesses the relative chance of an aggression-incident occurring for each shift within the clinics of Trajectum. This model is trained on data of previous incidents and work rosters, provided by the company. Following interviews with staff members of the clinics, and based on a literature review, we ideate conceptual models describing possible relations between independent variables such as the experience/employment status of the present members of staff and the dependent variable i.e. the number of aggression incidents. By conducting multiple linear regressions (MLR) to verify the strength and statistical significance (α=0.01) of these ideated relations in the data, we come to understand the underlying causes of aggression-incidents. Using these relations, we construct a model that calculates the fitted number of incidents per shift. The average years of experience at Trajectum of the staff members working the shift seems to have a significant (negative) relation with the number of aggression-incidents that occur during that shift. Significant relations are also found with variables that describe the time since the last aggression-incident inside of the clinic and the average frequency of incidents inside of each specific department. The risk prediction model that followed from these verified relations was tested against past data that were not part of the training set used to determine the strength and statistical significance of the relations. The model proved to be a good predictor of incidents, posing a strong relation between the predicted number of incidents and the realised number of incidents. The performance of the model was strong at predicting incidents between clinics as well as within clinics (t-value = 28.6 and 15.0, respectively). Using this model, Trajectum can make data-driven decisions regarding the utilisation of staff and the risk of incidents occurring. |
| Item Type: | Essay (Bachelor) |
| Clients: | Trajectum, Zwolle, Netherlands |
| Faculty: | BMS: Behavioural, Management and Social Sciences |
| Subject: | 50 technical science in general, 54 computer science |
| Programme: | Industrial Engineering and Management BSc (56994) |
| Link to this item: | https://purl.utwente.nl/essays/102969 |
| 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