Machine learning to predict the net promotor score and improve patient experience
Stouthuysen, Kristof Willems, Emma Heyvaert, Carl-Erik Distelmans, Tineke
Stouthuysen, Kristof
Willems, Emma
Heyvaert, Carl-Erik
Distelmans, Tineke
Citations
Altmetric:
Publication Type
Editor
Supervisor
Publication Year
2021
Journal
Book
Publication Volume
Publication Issue
Publication Begin page
Publication End page
Publication Number of pages
6
Collections
Abstract
Net Promotor Score (NPS) is a simple yet prominent indicator for customer satisfaction that is widely used in organizations. An in-depth analysis of the NPS and its drivers can provide organizations with highly valuable insights into their current deficiencies and strengths. CFO’s, for example, can use this metric to develop a strategy map. This case study explores the use of machine learning to perform an NPS key driver analysis at IndyCare, an Indian hospital group. The aim is to develop a model that predicts whether patients are promotors, detractors or passive customers for the hospital, and to assess feature importance of the models. The participants will also learn how to address a specific problem encountered when using NPS status as a target variable. In particular, the difference in scaling between the features and the target variable causes a low statistical fit of the various models. This problem is addressed by adding an additional feature to the model which considers subgroups of patients with common scoring patterns. Next to the analysis on the level of the hospital group, the participants will also go more into detail in this case study. More specifically, they will examine whether different patient groups have other drivers for their NPS. For this purpose, the participants will also explore unsupervised learning by making use of clustering techniques.
Research Projects
Organizational Units
Journal Issue
Keywords
Net Promotor Score (NPS)