Online Program Home
My Program

Abstract Details

Activity Number: 616 - Multidisciplinary Advances in Computing
Type: Contributed
Date/Time: Thursday, August 1, 2019 : 8:30 AM to 10:20 AM
Sponsor: Section on Statistical Computing
Abstract #302954 Presentation
Title: A Most Informative Index of Severity of Mental Health
Author(s): Barbara Clothier* and Maureen Murdoch and Siamak Noorbaloochi
Companies: CCDOR-Mpls VAHCS and CCDOR-Mpls VAHCS and University of MN and CCDOR-Mpls VAHCS and University of MN
Keywords: Dimension Reduction; Principal Component Analysis; Unidimensional Scalar Index; Poisson-inverse Gaussian Family of Distributions

Assume, for each patient, we have: number of Mental Health (MH) clinic visits, MH emergency/urgent care visits, MH related hospital admittances, and whether there were any occurrences of suicide or self-harm, substance (other than alcohol) abuse disorder, and alcohol abuse disorder diagnoses recorded within a certain time frame. We may be interested in constructing a unidimensional index that measures the “severity” of patients’ MH to include with other variables. Assuming severe cases are rare, one can use sufficiency arguments to construct such index scores (Noorbaloochi et al. (2010, 2019)). Here, we will provide the code that produced a unidimensional scalar index for a six-variable situation, and most importantly representing MH severity as one continuous variable.  We will discuss the results and the comparisons to other dimension reduction methods such as principal component analysis.  Given time, we will also show and discuss results from using other sets of variables that represent other constructs to illustrate the versatility of this method.  We will also discuss potential issues and solutions.

Authors who are presenting talks have a * after their name.

Back to the full JSM 2019 program