Degree

Doctor of Engineering (DEng)

Department

Mechanical and Industrial Engineering

Document Type

Dissertation

Abstract

Asthma is a chronic condition whose symptoms are managed/prevented using medication and interventions. The overarching objective of this study was to evaluate the impact of patients' demographics on case management enrollment and healthcare utilization, as well as to develop machine learning models to predict high-cost patients.

To accomplish these goals, the Man-Whiteness test, the chi-squares test, logistic regression and odds ratios, and machine learning models were implemented. The average cost of the non-enrolled CM group was significantly higher than the enrolled group (p-value .0001). In addition, the non-enrolled groups had considerably more visits to the emergency department than the other group (p .0001), but the enrolled group had significantly more visits to specialists (p = 0.00024). The racial variable had a significant impact on enrolment in CM. The present study did not find any association between age and gender with CM enrollment.

Based on machine learning models, our prediction system is 70% trained and 30% tested. Depending on the parameters included in the dataset, our machine learning models achieved prediction accuracy ranging from 66% to 93%. A feature significance study found that procedure number, ED visits, age, primary care visits, and hospital outpatient visits characteristics are the most significant elements in constructing a predictive model for HCAP, with an average accuracy 93.08%.

Age was associated with emergency department, inpatient and outpatient hospitalization, primary care, and specialist visits with odds ratios (p-values) of 1.03 (<.001), 1.01 (<.001), 1.01 (<.001), 0.96 (<.001), and 0.99 (<.001), respectively. Asthmatic males are more likely to seek primary care (OR = 1.05, p = 0.033) and less likely to need hospitalization (OR = 0.72, p <.001). African American asthmatic patients were more likely to visit the emergency department than white patients (OR = 2.08 and p <.001), but less likely to visit outpatient hospitalization, primary care, and specialist with odds ratios (p-values) of 0.89 (<.001), 0.66 (<.001), 0.67 (<.001), respectively. Hispanic asthmatic patients were related with reduced inpatient and primary care use than other patients 0.47 (0.015) and 0.75 (0.007), but they were more likely to attend outpatient hospitalization than white asthmatic patients 2.54 (<.001).

Date

3-2-2023

Committee Chair

Nahmens, Isabelina

DOI

10.31390/gradschool_dissertations.6054

Available for download on Wednesday, February 28, 2024

Share

COinS