An Exploration of the Association between Mild Cognitive Impairment Subgroups and Dementia Progression Time Open Access

Qiu, Jiayue (Spring 2020)

Permanent URL:


MCI patients have a wide range of clinical presentations. To understand the heterogeneity of MCI disorders and related diseases, researchers have used latent class analysis with non- cognitive features such as neuropsychiatric symptoms as indicators. Using the NationalAlzheimer’s Coordinating Center’s Uniform Data Set (UDS) with 6,034 participants, this research aims to establish more informative MCI subtypes with neuropsychiatric features and to show their distinct associations with progression time to dementia which might suggest distinct disease etiologies. Latent class analysis and a subsequent proportional hazards regression model were used to analyze this association. We used a weight-adjusted three-step approach to better account for the uncertainty issues of the latent class membership assignment. As a result, we found 4 latent classes with varied neuropsychiatric characteristics, including two classes characterized by either uniformly mild or more severe neuropsychiatric features, a cluster characterized by a combination of high depression, anxiety, and apathy and another cluster characterized by both high agitation and high irritability. The subsequent statistical results from the proportional hazards model provide estimates of different relationships between MCI subtypes and subsequent times to conversion to dementia. We found different hazard levels that associate certain neuropsychiatric features, such as irritability and agitation, with earlier risk of dementia compared to the others. We believe the statistical results from this research may aid in the early recognition of dementia in a clinical setting.

Table of Contents


List of Figures and Tables

1 Introduction

1.1 Opportunities and challenges in early recognition of symptomatic dementia

1.2 Mild cognitive impairment subtypes

1.3 Subgrouping and latent class analysis

1.4 Survival analysis and estimating the time of progression to dementia

1.5 3-step approach to relating external variable with latent classes

2 Methods

2.1 Participants

2.2 Measures

2.3 Analysis

2.3.1 Latent class analysis

2.3.2 Adjustment to variables

2.3.3 The cox proportional hazards regression model

2.3.4 Adjustment to the 3-step approach

3 Results

3.1 The 4 latent classes

3.2 Summary statistics for the predictor variables

3.3 Distribution of event time

3.4 Latent classes, age, MMSE score, and education are statistically significant in predicting dementia progression

3.5 Dementia progression patterns for the 4 latent classes

3.6 Comparing the results between the traditional and the adjusted three-step approachesDiscussion



About this Honors Thesis

Rights statement
  • Permission granted by the author to include this thesis or dissertation in this repository. All rights reserved by the author. Please contact the author for information regarding the reproduction and use of this thesis or dissertation.
  • English
Research Field
Committee Chair / Thesis Advisor
Committee Members
Last modified

Primary PDF

Supplemental Files