20. Addressing Endogeneity in Healthcare Operations through Copula-based and Machine Learning Methods
Invited abstract in session TB-13: AI and Machine learning in healthcare, stream OR in Healthcare (ORAHS).
Tuesday, 10:30-12:00Room: Clarendon SR 1.01
Authors (first author is the speaker)
| 1. | Meera Simon
|
| Quantitative Methods and Operations Management, Indian Institute of Management Kozhikode | |
| 2. | Shovan Chowdhury
|
| Quantitative Methods & Operation Management , Indian Institute of Management Kozhikode |
Abstract
Measuring cost at the patient level is important for the measurement and improvement of value. Healthcare providers must therefore be able to provide coordinated care with minimal expenditure which requires accurate estimation of cost. Understanding the importance of individual-level factors that determine expenditure, researchers have developed models that consider several customer-centric factors like socioeconomic and psychological factors and used regression models to analyse the effect of customer profiles and other factors on expenditure. It is, however, established that endogeneity is an important problem that arises while using regression, and many impactful academic journals require that the issue of endogeneity be addressed. Though till 2015, the issue of endogeneity was not of immense importance in operations management, the note from the editors of the Journal of Operations Management published in 2018 emphasised the importance of addressing the endogeneity, encouraging models that could incorporate endogeneity. In our study, we identify two endogenous variables that commonly occur in healthcare (namely, Length of Stay (LOS) and total expenditure ) and develop a model that can address endogeneity through bivariate copula which is an instrument-free approach that can effectively be used to model variables with suspected endogeneity. To accommodate other variables that are linked directly to the variables under study, we also use machine learning techniques.
Keywords
- Health Care
- Analytics and Data Science
- Service Operations
Status: accepted
Back to the list of papers