University of Southern California Dissertations and Theses

Comparison of participant and study partner predictions of cognitive impairment in the Alzheimer's disease neuroimaging initiative 3 study

(USC Thesis Other)

Comparison of participant and study partner predictions of cognitive impairment in the Alzheimer's disease neuroimaging initiative 3 study

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content

Comparison of Participant and Study Partner Predictions of Cognitive Impairment in the
Alzheimer's Disease Neuroimaging Initiative 3 Study

by

Fangqing Liu

A Thesis Presented to the
FACULTY OF THE USC KECK SCHOOL OF MEDICINE
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
MASTER OF SCIENCE
(APPLIED BIOSTATISTICS AND EPIDEMIOLOGY)

May 2023

Copyright 2023 Fangqing Liu
ii
Acknowledgements
Data used in preparation of this article were obtained from the Alzheimer’s Disease
Neuroimaging Initiative (ADNI) database (adni.loni.usc.edu). As such, the investigators within
the ADNI contributed to the design and implementation of ADNI and/or provided data but did
not participate in analysis or writing of this report. A complete listing of ADNI investigators can
be found at:http://adni.loni.usc.edu/wp-
content/uploads/how_to_apply/ADNI_Acknowledgement_List.pdf
Data collection and sharing for this project was funded by the Alzheimer's Disease
Neuroimaging Initiative (ADNI) (National Institutes of Health Grant U01 AG024904) and DOD
ADNI (Department of Defense award number W81XWH-12-2-0012). ADNI is funded by the
National Institute on Aging, the National Institute of Biomedical Imaging and Bioengineering,
and through generous contributions from the following: AbbVie, Alzheimer’s Association;
Alzheimer’s Drug Discovery Foundation; Araclon Biotech; BioClinica, Inc.; Biogen; Bristol-
Myers Squibb Company; CereSpir, Inc.; Cogstate; Eisai Inc.; Elan Pharmaceuticals, Inc.; Eli
Lilly and Company; EuroImmun; F. Hoffmann-La Roche Ltd and its affiliated company
Genentech, Inc.; Fujirebio; GE Healthcare; IXICO Ltd.; Janssen Alzheimer Immunotherapy
Research & Development, LLC.; Johnson & Johnson Pharmaceutical Research & Development
LLC.; Lumosity; Lundbeck; Merck & Co., Inc.; Meso Scale Diagnostics, LLC.; NeuroRx
Research; Neurotrack Technologies; Novartis Pharmaceuticals Corporation; Pfizer Inc.; Piramal
Imaging; Servier; Takeda Pharmaceutical Company; and Transition Therapeutics. The Canadian
Institutes of Health Research is providing funds to support ADNI clinical sites in Canada. Private
sector contributions are facilitated by the Foundation for the National Institutes of Health
(www.fnih.org). The grantee organization is the Northern California Institute for Research and
iii
Education, and the study is coordinated by the Alzheimer’s Therapeutic Research Institute at the
University of Southern California. ADNI data are disseminated by the Laboratory for Neuro
Imaging at the University of Southern California.

iv
TABLES OF CONTENTS
Acknowledgements…………………………………………………………………………..……ii
List of Tables………………………………………………………………………………..……v
List of Figures……………………………………………………………………...………..……vi
Abstract……………………………….……………………………………………………...…vii
Chapter 1: Introduction………………………….……………………………………………..1
Chapter 2: Methods………………………….………………………………………………...…4
Data source………………….………………………………………………...…...4
Participants……………….………………………………………………...….......6
Statistical methods……………….………………………………………………..7
Chapter 3: Results………………………….……………………………………………...…..9
Chapter 4: Discussions………………………….……………………………………………...19
Limitations……………….………………………………………………...….....20
References………………………….……………………………………………………...……..22
Appendices………………………….………………………………………………………..…..25

v
List of Tables
Table 1: Baseline characteristics of participants and study partners who had complete
information and missing information in ADNI 3 study, stratified by study
partner type ……………..…………………………………………………….…….……………12
Table 2: The distribution of baseline diagnosis of participants within each APOE4 category…..18
Table 3. The distribution of study partner type of participants within each APOE4 category……18
Table 4: The coefficients and 95% confidence intervals of linear regression models at each
time point, stratified by study partner type……………………….…………………….…….…...25

vi
List of Figures
Figure 1: STROBE diagram: The number of included and excluded participants in
this study ………………..………………………………………………………………………..10
Figure 2: Variable importance and 95% uncertainty bound of participant and study partner
ECog in predicting ADAS 13 by study partner type………..……………….……………………15
Figure 3: Variable importance and 95% uncertainty bound of participant and study partner
ECog in predicting ADAS 13 by baseline diagnosis ………………...…………………………...16
Figure 4: Variable importance and 95% uncertainty bound of participant and study partner
ECog in predicting ADAS 13 by whether had APOE4 alleles or not……………………………18
Figure 5: Sensitivity analysis of including participants without APOE4 information:
variable importance and 95% uncertainty bound of participant and study partner ECog
in predicting ADAS 13 score by study partner type ………………………………………...……26
Figure 6: Sensitivity analysis of including participants without APOE4 information:
variable importance and 95% uncertainty bound of participant and study partner ECog
in predicting ADAS 13 score by baseline diagnosis …………………………………...…………26

vii
Abstract
Background
Alzheimer’s disease trials require that a study partner (SP) enrolls along with the participant, and
that the SP has sufficient interaction with the participant and can attend study visits. Therefore,
SPs play an essential role in the successful completion of trials. One role of SPs in AD trials is to
provide information about the participant's cognitive performance. Our goal was to investigate
whether participants or SPs are better at predicting the participant's cognitive performance.
Methods
We used data from the Alzheimer’s Disease Neuroimaging Initiative. Everyday Cognition (ECog)
was used to assess participant and SP evaluations of the participant’s cognitive performance. The
Alzheimer's Disease Assessment Scale 13-item cognitive subscale was used to measure objective
cognitive performance. Random forest models and variable importance were used to compare the
performance of participant and SP ECog, stratified by dyad. As part of exploratory analyses, we
investigated whether the results differed by baseline diagnosis and APOE4 status.
Results
Regardless of dyad type, SP ECog was better at predicting cognitive performance. When
stratifying by baseline cognitive status, cognitively normal participants were similar or better at
predicting cognitive performance. Among patients with mild cognitive impairment or dementia,
participants performed similarly or worse compared to their SPs. When stratifying by APOE4
status, SP performed better than the participants regardless of APOE4 status.
Conclusions
Overall, SP had similar or better predictions of participants' cognitive performance compared with
participants. SP provides valuable information in AD studies.
viii
Keywords: Alzheimer’s disease, Study partners, ADNI, Cognitive performance

1
Chapter 1: Introduction
Alzheimer's disease (AD) is the most common type of dementia. Approximately 6.5
million people in the United States over the age of 65 have AD and the total cost of caring for AD
patients in 2022 was $321 billion. There is currently no cure for AD, and it is predicted that by
2060, 13.8 million people in the US will have AD if there is no significant breakthrough in
treatment or prevention (2022 Alzheimer's Disease Facts and Figures, 2022). It is therefore critical
that we continue to advance research in this field.
Because AD trials require that participants enroll along with a study partner and that the
study partner also attends study visits, the success of these trials depends on both the study partner
and the participant. Study partners have several important roles in AD studies, and these often
differ by the clinical stage of participants. In the preclinical AD stage, the study partner can help
participants overcome distress and reduce anxiety to avoid extreme behaviors, such as suicide
(Grill & Karlawish, 2017). As the disease progresses and the participant's cognitive performance
worsens, the study partner can provide valuable information about the participant's daily function
and cognitive performance (Ready et al., 2004), which may assist in evaluating the efficacy of
treatment. Study partners have also been found to improve participant retention in trials (Amariglio
et al., 2015). In AD dementia studies, study partners provide surrogate consent for participants,
ensure trial compliance, and provide integral information about the participants’ cognitive function
(Black et al., 2014; Ready et al., 2004).
While study partners play a crucial role in trial success, previous work has found that the
study partner requirement is one barrier to the recruitment of AD trials. People who may otherwise
be eligible may not have a suitable study partner to enroll with (Largent et al., 2018), resulting in
slower accrual and delaying the investigation of new therapies for disease treatment and prevention
2
(Grill & Karlawish, 2017). Because the study partner requirement can be a barrier to recruitment,
it is important that we carefully understand the benefits and drawbacks of the study partner
requirement in AD studies.
Study partners are often a spouse, an adult child, a friend, or a caregiver. Previous studies
have shown conflicting results regarding whether there are differences by study partner type in the
dropout rate and in the quality of the information provided by the study partner. In a meta-analysis
of six NIA-sponsored trials, Grill et al. found that participants who enrolled with a spouse had a
lower dropout rate compared to participants who enrolled with an adult child or "other" study
partner type (Grill et al., 2013). Another study found that dropout rates did not differ by study
partner type after adjusting for participant’s age, as age was a confounder in the relationship
between dropout rate and study partner type (Bernstein et al., 2021).
Other work has investigated whether or not study partners provide additional information
than that provided by the participant. Ryan et al. (2019) found that study partners may be able to
better predict cognitive performance compared to participants (Ryan et al., 2019). Another study
suggests that study partner performance differs by study partner type. Among spousal dyads, the
study partner was able to better predict cognitive performance compared to the participant, while
among non-spousal dyads, the participant was able to better predict their own cognitive
performance (Nuño et al., 2019). Later work suggests that child study partners tend to report less
cognitive impairment compared to the spouse in the Aging, Demographics, and Memory Study
(ADAMS) that used the Blessed Dementia Rating Scale (BDRS) to measure cognitive
performance; however, the authors state that these results may have been due to differences in
patient and partner characteristics between the two dyad types in this study (Stites et al., 2022).
3
It is crucial that we understand whether there are differences between study partner types
so that we can work to mitigate these differences. As previously stated, the study partner
requirement can be a barrier to recruitment. Therefore, in our study, we aim to investigate whether
the participant-study partner difference in evaluating participants’ cognitive performance varies
by study partner type (spouse vs. non-spouse).

4
Chapter 2: Methods
Data source
Alzheimer’s Disease Neuroimaging Initiative (ADNI) data were used for this analysis
(adni.loni.usc.edu). The ADNI was launched in 2003 as a public-private partnership, led by
Principal Investigator Michael W. Weiner, MD. The primary goal of ADNI has been to test
whether serial magnetic resonance imaging (MRI), positron emission tomography (PET), other
biological markers, and clinical and neuropsychological assessment can be combined to measure
the progression of mild cognitive impairment (MCI) and early AD. For up-to-date information,
see www.adni-info.org (ADNI Data Sharing and Publication Policy).
The ADNI study has four phases: ADNI 1, ADNI GO, ADNI 2, and ADNI 3. The aim of
the ADNI 1 study was to obtain longitudinal MRI and PET data of participants with MCI and AD
dementia and explore biomarkers as outcome indicators in trials (Weiner et al., 2005). The ADNI
GO study acquired F18 amyloid imaging and blood and cerebrospinal fluid (CSF) samples. The
blood and CSF biomarkers were examined in the early stage of the disease to help AD diagnosis
early in time and predict cognitive decline. ADNI GO findings were validated in the ADNI 3 study.
(Weiner et al., 2010). The ADNI 2 study further investigated biomarkers to develop effective
clinical trials by enabling the refinement of eligibility criteria to enroll patients with AD or those
at higher risk for AD. Investigators performed neuropathological examinations by conducting
autopsies for participants from previous phases to verify the previous clinical diagnoses (Weiner
et al., 2010). The ADNI 3 study developed tau PET and functional imaging techniques for clinical
trials (Weiner et al., 2016). ADNI has shared data through the Laboratory of Neuroimaging (LONI)
at the University of Southern California, and the LONI Image & Data Archive (IDA). The R data
package of ADNIMERGE integrates coded clinical and biomarker data and documentation from
5
all ADNI studies into one table. The R package (https://ida.loni.usc.edu/login.jsp) was used and
downloaded from IDA on November 10, 2022. This package includes select tables from the ADNI
study.
The primary aim of this study was to investigate whether participants or their study partners
better predict participants’ ADAS 13 and whether the ability to predict cognitive status differs by
study partner type (spouse vs. non-spouse). As part of exploratory analyses, we also investigated
whether predictive ability differed by baseline diagnosis and by whether or not participants have
APOE4 alleles.
Participant and study partner demographic information along with participant medical
history were collected at baseline. Participant and study partner Everyday Cognition (ECog), and
participant Alzheimer's Disease Assessment Scale 13-item cognitive subscale (ADAS 13), Clinical
Dementia Rating Scale (CDR), Mini-Mental State Examination (MMSE), genetic and biomarker
samples, and brain imaging were collected annually through in-clinic visits.
Cognitive Assessment. Study partner and participant versions of baseline ECog were used
in our analysis. The ECog was developed as a strict standard to evaluate the daily functions of
older people and predict the level of dementia. The ECog questionnaire has 39 items assessing six
dimensions: Memory, Language, Visual-Spatial, Planning, Organization, and Divided Attention.
Participants and study partners were asked questions regarding each of these dimensions and how
their current function compares to ten years ago. For example, one question asks, “Compared to
10 years ago, has there been any change in remembering a few shopping items without a list”.
Each question is scored using a 4-point scale where 1 represents "better or no change compared to
10 years earlier", and 4 represents "consistently much worse" (Farias et al., 2008). In our study,
ADAS 13 was used as an objective measurement of cognitive performance. (Mohs et al., 1997;
6
Kueper et al., 2018). The Alzheimer's Disease Assessment Scale-Cognitive Subscale (ADAS-Cog)
was designed to evaluate the severity of cognitive dysfunction, and is one of the most commonly
used scales to evaluate the cognitive performance of patients with pre-dementia or dementia.
ADAS 13 scores range from 0 to 85, with higher scores indicating worse cognitive performance.
Compared with the original 11-item cognitive subscale, ADAS 13 improves response capability
and can be more representative of cognitive performance (Cho et al., 2021).
Participants
All participants included in this analysis were from ADNI 3 since this is the only phase
that collected study partner type. Upon enrolling in ADNI 3, participants were given the option to
also participate in the ADNI-Brain Health Registry (ADNI-BHR), which required the completion
of additional questionnaires and included the study partner’s relationship to the participant. The
ADNI 3 study enrolled 1105 participants. There were 441 rollover participants from ADNI 1,
ADNI GO, and ADNI 2, and 664 participants who were newly enrolled in ADNI 3. To be eligible
for ADNI 3, participants had to be between 55 and 90 years of age, be fluent in either English or
Spanish, and must have had a study partner who had frequent contact with the participant and
would attend clinic visits with the participant periodically. Participants who were cognitively
normal (CN) or had MCI or AD dementia were all eligible for ADNI 3. For more detailed
eligibility criteria, refer to the report by Weiner, et al., 2016. (Weiner et al., 2016). Written consent
was provided by each participant and their study partner.
For our study, participants who were missing study partner relationship, baseline ADAS
13, baseline ECog, study partner baseline ECog, APOE4 information, participants' age, MMSE
score, baseline diagnosis, or race were excluded from the study. If the original study partner
7
changed shortly after the start of the study and the study partner information available was that of
the original study partner, the participant’s information was not included in this analysis.
Statistical methods
Demographic information of participants and study partners is described using the mean
and standard deviation for continuous variables, and frequency and percentage for categorical
variables. We used random forests to compare participant and study partner ability to predict
participant cognitive performance in this study. Random forests were fit using the GRF package
in the R statistical software language (R version 4.1.2). Random forest is a machine learning
algorithm that uses classification and regression trees (CART) to build models. Trees are built by
splitting the outcome observations into groups based on the predictors in the model. At each split,
the algorithm randomly samples predictors among those available in the model and determines the
"best" split using the CART algorithm. By randomly sampling predictors, the algorithm induces
randomness in each tree. Many trees are built in this way to create a random forest (Athey et al.,
2019). The variable importance measure in the GRF package is based on a weighted sum of the
splitting frequencies, where variables used higher in the tree are given more weight. Each random
forest provides one variable importance measure for each variable. Due to the randomness of the
forests, we repeated this procedure 1,000 times, resulting in 1000 variable importance measures
for each variable. The estimated mean variable importance (eMVI) measure presented is the
average of these 1000 values for each variable. We also present the 2.5
th
and the 97.5
th
percentiles
based on these measures and present these as the 95% confidence bounds for the variable
importance. Random forests were used for this analysis because they inherently allow for
interactions and flexibility in the functional form of the predictors.
8
The primary analysis was based on the random forest models. Baseline participant and
study partner ECog were used to model ADAS 13 at baseline, year one, and year two. We built a
separate model for each timepoint and compared the eMVI corresponding to baseline participant
and study partner ECog for each model. The variable with the higher eMVI was determined to be
better at predicting ADAS 13 at that given time point. To examine whether the participant and
study partner performance differed by dyad type, we stratified models by study partner type
defined as spousal vs. non-spousal. We also fit linear regression models for each of the scenarios
to quantify the association between ECog and ADAS 13. These models were used to describe the
association between ECog and ADAS 13. All models were adjusted for participant’s baseline
diagnosis, age, gender, education, race, APOE4, baseline CDRSB, and baseline MMSE.
Coefficient estimates and the corresponding 95% confidence intervals are presented in Table 4 of
the appendix.
As exploratory analyses, we also stratified our analysis by baseline diagnosis and by
whether or not participants had at least one APOE4 allele. Baseline diagnosis was defined as:
cognitively normal (CN) and significant memory concern (SMC), early mild cognitive impairment
(EMCI) and late mild cognitive impairment (LMCI) as the mild cognitive impairment (MCI) group,
and AD dementia group. Baseline diagnosis, age, gender, education, race, APOE4, CDRSB, and
MMSE were included in all random forest models.

9
Chapter 3: Results
At baseline, study partner information was available for 664 participants who were newly
enrolled in the ADNI 3 study protocol and completed the ADNI-Brain Health Registry (ADNI-
BHR) questionnaire. After removing participants with missing information, there were 461 (69.4%
of 664) participants: 323 (70.0% of 461) participants enrolled with their spouse, 51 (11.1% of 461)
participants enrolled with an adult child or child-in-law, and 87 (18.9% of 461) participants
enrolled with other study partners such as friends, companions, and other relatives. Details
regarding the number of participants with missing information are provided in the STROBE
diagram (Fig. 1). The baseline characteristics of excluded participants and study partners are
presented in Table 1 of the appendix. Most of the excluded participants were missing APOE4
information (n=176). The excluded participants had similar characteristics to those with complete
data, except that the excluded participants (41.9%) had more MCI patients than included
participants (24.1%). Because most of the missingness was due to APOE4, as part of sensitivity
analyses we included participants who were missing APOE4 information and fit random forest
models without the APOE4 variable. Results regarding the performance of the study partner and
participant were similar. The plots of this sensitivity analysis are presented in Figure 5 and Figure
6 of the appendix.

10

Fig. 1 STROBE diagram: The number of included and excluded participants in this study
Demographic information of participants stratified by study partner type is presented in
Table 1 of the appendix. Participants across all study partner types had similar levels of education,
MMSE scores, CDR sum of boxes, and participant ECog total. Participants who enrolled with an
adult child were older, on average, (mean: 73.2, sd: 8.3) compared to participants who enrolled
with a spouse (mean: 71.2, sd: 6.6) or with a different type of study partner (mean: 70.9, sd: 8.1).
Participants who enrolled with an “other” study partner type had lower ADAS 13 scores (mean:
Assessed for
eligibility (n=664)
Excluded (n=203)
• Missing ADAS 13 score (n=8)
• Missing participant ECog (n=9)
• Missing study partner ECog (n=29)
• Missing APOE4 information (n=176)
• Missing participants’ age (n=4)
• Missing MMSE score (n=1)
• Missing baseline diagnosis (n=24)
• Unknown race (n=8)
*Some participants were missing more than one variable.
Total
eligible(n=461)
• Spouse (n=323)
• Non-spouse (n=138)
¨ Adult children (n=51)
¨ Others (n=87)

11
10.5, sd: 6.7) compared to participants who enrolled with an adult child (mean: 12.5, sd: 8.5) and
those who enrolled with a spouse (mean: 13.4, sd: 9.2). Mean ECog scores for participants were
similar for spousal and adult child study partners (Spouse: mean Study Partner ECog: 1.6, sd: 0.7;
mean Participant ECog: 1.6, sd: 0.5. Adult child: mean Study Partner ECog: 1.6, sd: 0.7; mean
Participant ECog: 1.6, sd: 0.6) to participants. However, study partner ECog scores were higher
among the “other” study partner dyads (mean Study Partner ECog: 1.2, sd: 0.5; mean Participant
ECog: 1.6, sd: 0.6) than participants’ ECog. The paired mean difference in ECog was 0.05 (sd:
0.8; 95% CI: -1.5, 1.6) between the participant and spousal study partner, 0.3 (sd: 0.6; 95% CI: -
0.9, 1.5) between the participant and adult child study partner, and -0.03 (sd: 0.7; 95% CI: -1.4,
1.3) between the participant and other study partner.

12

Variables Participants with complete data Participants with missing data
Study partner types Study partner types
Spouse Adult
child
Other Overall Spouse Adult
child
Other Overall Total
sample
size
Participants characteristics
Total, No. (%) 323
(70.1)
51
(11.1)
87
(18.9)
461
(100.0)
121
(59.6)
31
(15.3)
51
(25.1)
203
(100.0)
203
Baseline
DX, No.
(%)
Cognitively
normal (CN)
and Significant
memory
concern (SMC)
171
(52.9)
35
(68.6)
63
(72.4)
269
(58.4)
42
(38.2)
12
(44.4)
29
(69.0)
83
(46.4)
179
Mild cognitive
impairment
(MCI)
108
(33.4)
12
(23.5)
21
(24.1)
141
(30.6)
50
(45.5)
14
(51.9)
11
(26.2)
75
(41.9)
Alzheimer’s
disease (AD)
44
(13.6)
4 (7.8) 3 (3.5) 51
(11.1)
18
(16.4)
1 (3.7) 2 (4.8) 21
(11.7)
Age, mean (SD) 71.2
(6.6)
73.2
(8.3)
70.9
(8.1)
71.3
(7.1)
71.0
(6.8)
70.7
(9.0)
65.6
(7.8)
69.6
(7.8)
199
Female sex, No. (%) 141
(43.7)
43
(84.3)
63
(72.4)
247
(53.6)
47
(38.8)
29
(93.5)
37
(72.5)
113
(55.7)
203
Years of education, mean
(SD)
16.5
(2.3)
16.1
(2.4)
16.7
(2.4)
16.5
(2.3)
16.4
(2.4)
16.2
(2.6)
16.2
(2.2)
16.3
(2.4)
203
Race, No.
(%)
Am
Indian/Alaskan
0 (0) 1 (2.0) 0 (0) 1 (0.2) 1 (0.8) 0 (0) 0 (0) 1 (0.5) 195
Asian 8 (2.5) 4 (7.8) 1 (1.2) 13
(2.8)
7 (5.8) 2 (6.5) 5 (9.8) 14
(7.2)
Black 7 (2.2) 3 (5.9) 18
(20.7)
28
(6.1)
21
(17.4)
14
(45.2)
28
(54.9)
63
(32.3)
White 303
(93.8)
43
(84.3)
66
(75.9)
412
(89.4)
85
(70.2)
13
(41.9)
16
(31.4)
114
(58.5)
More than one 5 (1.6) 0 (0) 2 (2.3) 7 (1.5) 1 (0.8) 1 (3.2) 1 (2.0) 3 (1.5)
Unknown 0 (0) 0 (0) 0 (0) 0 (0) 6 (5.0) 1 (3.2) 1 (2.0) 8 (4.1)
Ethnicity,
No. (%)
Not
Hispanic/Latino
309
(95.7)
49
(96.1)
83
(95.4)
441
(95.7)
99
(81.8)
26
(83.9)
44
(86.3)
169
(83.3)
203
Hispanic/Latino 13
(4.0)
2 (3.9) 4 (4.6) 19
(4.1)
22
(18.2)
5
(16.1)
7 (4.6) 34
(13.7)
Unknown 1 (0.3) 0 (0) 0 (0) 1 (0.2) 0 (0) 0 (0) 0 (0) 0 (0)
Divorced 0 (0) 14
(27.5)
34
(39.1)
48
(10.4)
0 (0) 10
(33.3)
16
(31.4)
26
(12.9)
202
13
Table 1. Baseline characteristics of all participants and study partners available in our dataset,
stratified into groups based on those who had missing information and those with including in our
analysis.
Marital
status,
No. (%)
Married 323
(100)
19
(37.3)
21
(24.1)
363
(78.7)
119
(98.3)
6
(20.0)
10
(19.6)
135
(66.8)
Never married 0 (0) 1 (2.0) 12
(13.8)
13
(2.8)
1 (0.8) 2 (6.7) 17
(33.3)
20
(9.9)
Widowed 0 (0) 17
(33.3)
20
(23.0)
37
(8.0)
0 (0) 12
(40.0)
7
(13.7)
19
(9.4)
Unknown 0 (0) 0 (0) 0 (0) 0 (0) 1 (0.8) 0 (0) 1 (2.0) 2 (1.0)
APOE e4
carrier,
No. (%)
0 184
(57.0)
34
(66.7)
50
(57.5)
268
(58.1)
10
(58.8)
2
(40.0)
4
(80.0)
16
(59.3)
27
1 111
(34.4)
14
(27.5)
30
(34.5)
155
(33.6)
7
(41.2)
1
(20.0)
1
(20.0)
9
(33.3)
2 28
(8.7)
3 (5.9) 7 (8.1) 38
(8.2)
0 (0) 2
(40.0)
0 (0) 2 (7.4)
ADAS 13, mean (SD) 13.4
(9.2)
12.5
(8.5)
10.5
(6.7)
12.7
(8.7)
16.0
(9.8)
10.4
(6.5)
11.8
(8.2)
14.4
(9.3)
195
MMSE score, mean (SD) 27.8
(2.7)
28.1
(2.4)
28.8
(2.0)
28.0
(2.5)
27.3
(2.7)
28.3
(2.0)
27.9
(2.6)
27.6
(2.6)
202
CDR sum of boxes, mean
(SD)
1.1
(1.7)
0.9
(1.6)
0.4
(1.0)
0.9
(1.6)
1.6
(2.1)
0.9
(1.0)
0.5
(0.9)
1.2
(1.8)
203
Participant ECog total
assessed by themselves,
mean (SD)
1.6
(0.5)
1.6
(0.6)
1.6
(0.6)
1.6
(0.6)
1.7
(0.6)
1.6
(0.5)
1.5
(0.5)
1.6
(0.6)
194
Participant ECog total
assessed by their study
partners, mean (SD)
1.6
(0.7)
1.6
(0.7)
1.2
(0.5)
1.5
(0.7)
1.8
(0.8)
1.5
(0.7)
1.3
(0.5)
1.6
(0.7)
174
The difference between
Participant ECog and Study
partner ECog, mean (SD)
0.05
(0.8)
0.3
(0.6)
-0.03
(0.7)
0.04
(0.7)
-0.03
(0.7)
0.2
(0.5)
0.2
(0.4)
0.06
(0.6)
169
Study partner characteristics
Study partner age, mean
(SD)
69.3
(7.2)
46.0
(10.5)
67.8
(10.1)
66.4
(11.0)
68.7
(9.1)
40.3
(10.4)
62.8
(9.2)
62.9
(13.6)
197
Study partner female sex,
No. (%)
182
(56.4)
40
(78.4)
65
(74.7)
287
(62.3)
73
(60.3)
21
(67.7)
34
(66.7)
128
(63.1)
203
Duration of study partners
spent with participants in-
person (hours/week), mean
(SD)
116.3
(45.1)
22.3
(35.7)
33.9
(45.6)
90.4
(59.5)
123.4
(46.2)
53.9
(62.2)
34.2
(48.1)
90.4
(63.8)
203
Duration of study partners
spent with participants other
than in-person
(hours/week), mean (SD)
7.3
(22.6)
8.7
(19.7)
8.3
(21.9)
7.7
(22.1)
16.1
(40.9)
15.3
(33.5)
7.0
(9.5)
13.7
(34.7)
192
Lives with participants, No.
(%)
321
(99.4)
11
(21.6)
20
(23.0)
352
(76.4)
121
(100.0)
15
(48.4)
11
(21.6)
147
(72.4)
203
14
The primary objectives were to evaluate the difference in the participant and study partners
ECog assessments (for the participant) in predicting the participant cognitive performance
(ADAS13) and to evaluate whether there were differences by study partner type (spouse vs. non-
spouse). Models were stratified by study partner type (spousal vs. non-spousal). The non-spousal
group included adult children, child-in-law, friends or companions, other relatives, and others.
Overall, the baseline study partner ECog (eMVI=0.18 at baseline, eMVI=0.10 at year one,
eMVI=0.06 at year two) had greater eMVI compared to the baseline participant ECog (eMVI=0.04
at baseline, eMVI=0.03 at year one, eMVI=0.03 at year two) for all time points (Fig. 2A). The
differences between the variable importance of the study partner and participant ECog decreased
over time. Among spousal dyads (n=323), the eMVI was higher for study partner ECog
(eMVI=0.21 at baseline, eMVI=0.12 at year one, eMVI=0.06 at year two) compared to participant
ECog (eMVI=0.04 at baseline, eMVI=0.03 at year one, eMVI=0.03 at year two), and the difference
between participant ECog eMVI and study partner ECog eMVI decreased over time (Fig.
2B). Among non-spousal dyads (n=138), the eMVI for study partner ECog (eMVI=0.19 at
baseline, eMVI=0.17 at year one, eMVI=0.16 at year two) was greater than participant eMVI
(eMVI=0.11 at baseline, eMVI=0.12 at year one, eMVI=0.12 at year two) at all three time points
(Fig. 2C).

15
Fig. 2
A.

B. C.

The estimated mean variable importance and 95% uncertainty bound of participant ECog and study
partner ECog in predicting ADAS 13 score by study partner type: (A) All Participants, (B) Spouse,
(C) Non-spouse.

In the linear regression models, at baseline and year one, spousal study partner ECog was
positively associated (coefficient estimate=2.3 at baseline, 95% CI: 0.9, 3.7; coefficient
estimate=2.6 at year one, 95% CI: 0.8, 4.4) with ADAS 13, but participant ECog was negatively
associated (coefficient estimate =-1.9 at year one, 95% CI: -3.6, -0.2) at both time points. In year
two, spousal study partners were negatively associated (coefficient estimate =-5.4, 95% CI: -8.4,
-2.3) with ADAS 13, but participant ECog had a positive association. The coefficients of
participant ECog were not significantly different from 0 for spousal dyads at baseline and year
two. For non-spousal groups, study partner ECog had a positive association with ADAS 13 at
baseline and year two and had a negative association in year one. The coefficients of participant
ECog were positive at all three time points. The coefficients corresponding to participant ECog
16
and study partner ECog were not significantly different from 0 for non-spousal dyads at all time
points. All of these coefficients had wide 95% confidence intervals due to data sparsity. The
detailed coefficient estimates and 95% confidence intervals are presented in Table 4 of the
appendix.
Because the performance of the participant may differ by baseline diagnosis, as part of
exploratory analyses, we also stratified by diagnosis using random forest models. For participants
who were cognitively normal and those with significant memory concern, the eMVI of participant
ECog was higher than the eMVI of study partner ECog at baseline and year one but was lower in
year two (Fig. 3A). The eMVI was similar for participants and study partners in years one and two.
Among participants with MCI, study partner ECog had a higher eMVI for all three time points
(Fig. 3B). Study partner and participant ECog performed similarly in year two, and their
uncertainty bounds overlapped. For participants who had AD dementia, the eMVI of participant
ECog was similar to the eMVI of study partner ECog at baseline and year one that the uncertainty
bounds overlapped at these two time points (Fig. 3C). We were not able to investigate whether the
participant or study partner ECog performed better in year two due to data sparsity.

Fig. 3
A. B.

C.
17

The estimated mean variable importance and 95% uncertainty bound of participant ECog and study
partner ECog in predicting ADAS 13 score by baseline diagnosis: (A) Cognitively Normal and
Significant Memory Concern, (B) Early Mild Cognitive Impairment and Late Mild Cognitive
Impairment, (C) Alzheimer’s Disease dementia

We also stratified our analyses by whether or not participants had at least one APOE4 allele
as part of exploratory analyses as the motivation for participation in these studies may differ by
whether or not there is a family history of AD. For participants with no APOE4 alleles, the eMVI
was higher for study partner ECog compared to participant ECog (Fig. 4A). However, for
participants with one or two APOE4 alleles, the eMVI of study partner ECog was similar to
participant ECog (Fig. 4B). The distributions of participants’ baseline diagnosis by APOE status
are presented in Table 2. There were more cognitively normal and MCI patients in the no APOE4
alleles group (63.9% for CN or SMC; 54.3% for MCI) compared to the group with 1 or 2 APOE4
alleles (36.1% for CN or SMC; 45.7% for MCI), and there were fewer participants with AD
dementia in the group with no APOE4 alleles (30%) than in the group with at least one APOE4
alleles (70%). We also investigated the distribution of participants’ study partner type by APOE
status (Table 3). The group with no APOE4 alleles (56.1% for spouse; 59.5% for non-spouse) had
more spousal and non-spousal study partners than the group with at least one APOE4 alleles (43.9%
for spouse; 40.5% for non-spouse).

18
Fig. 4
A. B.

The estimated mean variable importance and 95% uncertainty bound of participant ECog and study
partner ECog in predicting ADAS 13 score by whether had APOE4 alleles or not: (A) Participants
with no APOE4 alleles (B) Participants with one or two APOE4 alleles

CN or SMC MCI Dementia Total
No APOE4 alleles 163 (63.9%) 75 (54.3%) 15 (30%) 253 (57.2%)
1 or 2 APOE4 alleles 92 (36.1%) 63 (45.7%) 35 (70%) 190 (42.9%)
Total 255 138 50 443
Table 2. The distribution of baseline diagnosis of participants within each APOE4 category.

Spouse Non-spouse Total
No APOE4 alleles 175 (56.1%) 78 (59.5%) 253 (57.2%)
1 or 2 APOE4 alleles 137 (43.9%) 53 (40.5%) 190 (42.9%)
Total 312 131 443
Table 3. The distribution of study partner type of participants within each APOE4 category.

19
Chapter 4: Discussion
Overall, our results are similar to those of previous studies. Ryan et al. (2019) investigated
the performance of participant and study partner ECog in predicting ADAS 13 cross-sectionally
and longitudinally in ADNI data and found that study partners performed similarly or better on the
prediction of participant’s cognitive performance than the participants; however, study partner
type was not available for this analysis, and the analysis focused on cognitively normal participants
with biomarker evidence of AD (Ryan et al., 2019). Our study only investigated the performance
of baseline ECog in predicting ADAS 13 and included all participants who were cognitively
normal, those who had MCI, and participants with AD dementia. Because study partner type is
collected in participants who were newly enrolled in ADNI 3, we were able to incorporate this
information into our analyses. Our results indicate that study partners had better predictions of
participants' cognitive status compared to participants. This differs from previous results of Nuño
et al., in which study partners performed similarly or worse in predicting future cognitive
performance compared to participants. The previous work, however, only included participants in
the preclinical stage (Nuño et al., 2019). Future work will be needed to stratify by both dyad type
and baseline diagnosis to validate our results.
As part of exploratory analyses, we investigated whether the performance of the study
partner and participant ECog differs by baseline diagnosis group. Previous work found that study
partners may have better predictions of participants' cognitive performance than participants when
participants have MCI or AD dementia (Tierney et al., 2003). Our results suggest that study
partners had similar or better predictions than participants with MCI, but study partners and
participants with AD dementia performed similarly. It should be noted that our analysis is based
on a small sample size, and therefore additional work is required to confirm these results. Among
20
cognitively normal participants, the participants generally performed better on the prediction of
their cognitive status than their partners (Farias et al., 2017). In our results, cognitively normal
participants predicted their cognitive level better than their partners at baseline, but their cognitive
performance became worse gradually in years one and two and predicted similarly to their partners.
We also stratified participants by whether they have APOE4 alleles or not as part of
exploratory analyses. Study partners played a more important role in cognitive prediction in
participants with no APOE4 alleles than in participants with one or two APOE4 alleles. When we
stratified by study partner type within each APOE4 group, there were more spousal and non-
spousal study partners in the group with no APOE4 alleles group compared to the group with 1 or
2 APOE4 alleles. However, the distributions of baseline diagnosis of participants' cognitive status
were different in the no APOE4 alleles group and the 1 or 2 APOE4 alleles group. Participants
with CN or SMC and MCI were the most significant proportion in the group with no APOE4
alleles. There was a greater proportion of AD participants in the APOE4 alleles group compared
to the no group without APOE4 alleles. These differences in baseline diagnosis may explain the
differences between the two APOE groups.
Limitations
There were some limitations in this study. First, study partner information was not collected
until ADNI 3. There were 203 participants excluded from our study due to missing information.
Most of these patients were missing APOE4; however, we conducted sensitivity analyses that
included patients with missing APOE4, and the results were similar to the results of our primary
analysis. Due to data sparsity, we could not investigate differences in the performances of different
study partner types within the non-spousal dyad group. Moreover, we did not have information
regarding the amount of time the study partner spent with the participant and other factors that
21
may impact the study partner's performance. Study partner type was therefore used as a proxy for
some of these factors, but future work should focus on investigating the factors that lead to
differences in performance by study partner type. Additionally, little information was available
beyond two years from baseline and therefore we could only conduct these analyses at baseline,
year one, and year two.

22
References
Amariglio, R. E., Donohue, M. C., Marshall, G. A., Rentz, D. M., Salmon, D. P., Ferris, S.
H., . . . Alzheimers Dis Cooperative, S. (2015). Tracking Early Decline in Cognitive
Function in Older Individuals at Risk for Alzheimer Disease Dementia The Alzheimer's
Disease Cooperative Study Cognitive Function Instrument. Jama Neurology, 72(4), 446-
454. https://doi- org.libproxy2.usc.edu/10.1001/jamaneurol.2014.3375
Athey, S., Tibshirani, J., & Wager, S. (2019). GENERALIZED RANDOM FORESTS. Annals of
Statistics, 47(2), 1148-1178. https://doi-org.libproxy2.usc.edu/10.1214/18- aos1709
Bernstein, O. M., Grill, J. D., & Gillen, D. L. (2021). Recruitment and retention of participant
and study partner dyads in two multinational Alzheimer's disease registration trials.
Alzheimers Research & Therapy, 13(1), Article 16. https://doi-
org.libproxy2.usc.edu/10.1186/s13195-020-00762-8
Black, B. S., Taylor, H., Rabins, P. V., & Karlawish, J. (2014). Researchers' perspectives on the
role of study partners in dementia research. International Psychogeriatrics, 26(10), 1649-
1657. https://doi- org.libproxy2.usc.edu/10.1017/s1041610214001203
Cho, S. H., Woo, S., Kim, C., Kim, H. J., Jang, H., Kim, B. C., . . . Seo, S. W. (2021). Disease
progression modelling from preclinical Alzheimer's disease (AD) to AD dementia.
Scientific Reports, 11(1), Article 4168. https://doi- org.libproxy2.usc.edu/10.1038/s41598-
021-83585-3
Farias, S. T., Lau, K., Harvey, D., Denny, K. G., Barba, C., & Mefford, A. N. (2017). Early
Functional Limitations in Cognitively Normal Older Adults Predict Diagnostic Conversion
to Mild Cognitive Impairment. Journal of the American Geriatrics Society, 65(6), 1152-
1158. https://doi-org.libproxy2.usc.edu/10.1111/jgs.14835
Farias, S. T., Mungas, D., Cahn-Weiner, D., Baynes, K., Reed, B. R., Jagust, W., & DeCarli, C.
(2008). The measurement of Everyday Cognition (ECog): Scale development and
psychometric properties. Neuropsychology, 22(4), 531-544. https://doi-
org.libproxy2.usc.edu/10.1037/0894-4105.22.4.531
Gaugler, J., James, B., Johnson, T., Reimer, J., Solis, M., Weuve, J., . . . Hohman, T. J. (2022).
2022 Alzheimer's disease facts and figures. Alzheimers & Dementia, 18(4), 700-789.
https://doi-org.libproxy2.usc.edu/10.1002/alz.12638
Grill, J. D., & Karlawish, J. (2017). Study partners should be required in preclinical Alzheimer's
disease trials. Alzheimers Research & Therapy, 9, Article 93. https://doi-
org.libproxy2.usc.edu/10.1186/s13195-017-0327-x
Grill, J. D., Raman, R., Ernstrom, K., Aisen, P., & Karlawish, J. (2013). Effect of study partner
on the conduct of Alzheimer disease clinical trials. Neurology, 80(3), 282- 288. https://doi-
org.libproxy2.usc.edu/10.1212/WNL.0b013e31827debfe
23
Kueper, J. K., Speechley, M., & Montero-Odasso, M. (2018). The Alzheimer's Disease
Assessment Scale-Cognitive Subscale (ADAS-Cog): Modifications and Responsiveness in
Pre-Dementia Populations. A Narrative Review. Journal of Alzheimers Disease, 63(2),
423-444. https://doi- org.libproxy2.usc.edu/10.3233/jad-170991
Largent, E. A., Karlawish, J., & Grill, J. D. (2018). Study partners: essential collaborators in
discovering treatments for Alzheimer's disease. Alzheimers Research & Therapy, 10,
Article 101. https://doi- org.libproxy2.usc.edu/10.1186/s13195-018-0425-4
Mohs, R. C., Knopman, D., Petersen, R. C., Ferris, S. H., Ernesto, C., Grundman, M., . . . Thal,
L. J. (1997). Development of cognitive instruments for use in clinical trials of antidementia
drugs: Additions to the Alzheimer's disease assessment scale that broaden its scope.
Alzheimer Disease & Associated Disorders, 11, S13-S21.
Nuno, M. M., Gillen, D. L., Grill, J. D., & Alzheimers Dis Cooperative, S. (2019). Study partner
types and prediction of cognitive performance: implications to preclinical Alzheimer's
trials. Alzheimers Research & Therapy, 11(1), Article 92. https://doi-
org.libproxy2.usc.edu/10.1186/s13195-019-0544-6
Ready, R. E., Ott, B. R., & Grace, J. (2004). Validity of Informant Reports About AD and MCI
Patients' Memory. Alzheimer Disease and Associated Disorders, 18, 11-16. https://doi-
org.libproxy2.usc.edu/10.1097/00002093-200401000-00003
Ryan, M. M., Grill, J. D., Gillen, D. L., & Alzheimers Dis Neuroimaging, I. (2019). Participant
and study partner prediction and identification of cognitive impairment in preclinical
Alzheimer's disease: study partner vs. participant accuracy. Alzheimers Research &
Therapy, 11(1), Article 85. https://doi- org.libproxy2.usc.edu/10.1186/s13195-019-0539-3
Stites, S. D., Largent, E. A., Gill, J., Gurian, A., Harkins, K., & Karlawish, J. (2022). Predictors
of who Serves as an Alzheimer's Disease Research Participant's Study Partner and the
Impact of their Relationship on Study Partners' Reports on Participants. Research on
Aging, 44(9-10), 734-746, Article 01640275221075739. https://doi-
org.libproxy2.usc.edu/10.1177/01640275221075739
Tierney, M. C., Herrmann, N., Geslani, D. M., & Szalai, J. P. (2003). Contribution of informant
and patient ratings to the accuracy of the Mini-Mental State Examination in predicting
probable Alzheimer's disease. Journal of the American Geriatrics Society, 51(6), 813-818.
https://doi- org.libproxy2.usc.edu/10.1046/j.1365-2389.2003.51262.x
Weiner, M., Aisen, P., & Petersen, R. (2005, March 2). Alzheimer's disease Neuroimaging
Initiative 1 (ADNI1) protocol
Weiner, M., Aisen, P., & Petersen, R. (2010, February 10). Alzheimer's disease Neuroimaging
Initiative Grand Opportunity (ADNI-GO) protocol
24
Weiner, M., Aisen, P., & Petersen, R. (2010, September 17). Alzheimer's disease Neuroimaging
Initiative 2 (ADNI2) protocol
Weiner, M., Aisen, P., & Petersen, R. (2016, May 24). Alzheimer's disease Neuroimaging
Initiative 3 (ADNI3) protocol

25
Appendices
Baseline Year 1 Year 2
Spouse
(n=323)
Non-
spouse
(n=138)
Spouse
(n=137)
Non-
Spouse
(n=49)
Spouse
(n=125)
Non-
Spouse
(n=51)
Intercept 31.5 (19.0,
44.0)
35.6 (18.0,
53.2)
56.5 (41.6,
71.4)
42.1 (14.7,
69.5)
45.4 (28.1,
62.8)
30.5 (-4.1,
65.1)
Baseline
diagnosis of
MCI
4.0 (2.2,
5.7)
0.9 (-1.7,
3.4)
0.4 (-2.1,
2.9)
0.5 (-2.7,
3.7)
2.9 (-0.4,
6.3)
4.0 (-0.3,
8.2)
Baseline
diagnosis of
AD
10.4 (6.9,
13.8)
8.3 (0.6,
16.0)
-0.8 (-5.6,
3.9)
7.3 (-0.6,
15.2)
2.4 (-5.2,
10.0)
10.1 (-1.1,
21.2)
Age per 5 years 0.5 (0.05,
0.9)
0.6 (0.1,
1.1)
-0.1 (-0.7,
0.5)
1.0 (0.3,
1.7)
0.5 (-0.3,
1.4)
0.5 (-0.2,
1.3)
Gender of male 1.2 (-0.03,
2.3)
0.3 (-1.5,
2.2)
1.9 (-0.03,
3.9)
-3.4 (-6.6,
-0.2)
2.7 (0.3,
5.0)
0.3 (-3.0,
3.6)
Education per
5 years
-1.3 (-2.5, -
0.003)
-0.3 (-2.0,
1.4)
-0.7 (-2.6,
1.3)
-0.4 (-3.1,
2.3)
-3.0 (-5.5, -
0.6)
-0.3 (-3.0,
2.4)
Race of White 0.8 (-1.5,
3.1)
0.2 (-1.7,
2.0)
2.4 (-2.2,
7.1)
0.4 (-3.3,
4.1)
-3.5 (-8.4,
1.4)
-2.0 (-7.1,
3.2)
APOE4 of non-
zero alleles
0.8 (-0.4,
1.9)
-0.1 (-1.6,
1.4)
1.2 (-0.6,
3.0)
0.7 (-1.7,
3.1)
1.1 (-1.2,
3.3)
-1.4 (-4.1,
1.2)
CDRSB 0.1 (-0.6,
0.8)
1.2 (-0.4,
2.8)
1.1 (0.4,
1.7)
1.0 (-0.7,
2.7)
2.8 (2.0,
3.7)
1.4 (0.003,
2.7)
MMSE -1.0 (-1.4, -
0.7)
-1.3 (-1.8,
-0.7)
-1.6 (-2.0, -
1.3)
-1.6 (-2.5,
-0.7)
-1.0 (-1.3, -
0.6)
-1.0 (-2.2,
0.3)
Participant
baseline ECog
total
-0.4 (-1.7,
0.8)
1.3 (-0.2,
2.8)
-1.9 (-3.6, -
0.2)
1.8 (-0.3,
4.0)
0.5 (-1.8,
2.8)
0.6 (-2.9,
4.0)
Study partner
baseline ECog
total
2.3 (0.9,
3.7)
0.5 (-1.8,
2.9)
2.6 (0.8,
4.4)
-1.5 (-4.5,
1.5)
-5.4 (-8.4, -
2.3)
0.3 (-5.1,
5.6)
Table 4. The coefficients and 95% confidence intervals of linear regression models in predicting
ADAS 13 scores using participant ECog and study partner ECog at each time point, stratified by
study partner type

26
Fig. 5
A.

B. C.

Sensitivity analysis of including participants without APOE4 information: The estimated
mean variable importance and 95% uncertainty bound of participant ECog and study partner ECog
in predicting ADAS 13 score by study partner type: (A) All Participants, (B) Spouse, (C) Non-
spouse.

Fig. 6
A. B.

C.
27

Sensitivity analysis of including participants without APOE4 information: The estimated
mean variable importance and 95% uncertainty bound of participant ECog and study partner ECog
in predicting ADAS 13 score by baseline diagnosis: (A) Cognitively Normal and Significant
Memory Concern, (B) Early Mild Cognitive Impairment and Late Mild Cognitive Impairment, (C)
Alzheimer’s Disease dementia

Abstract (if available)

Abstract Background
Alzheimer’s disease trials require that a study partner (SP) enrolls along with the participant, and that the SP has sufficient interaction with the participant and can attend study visits. Therefore, SPs play an essential role in the successful completion of trials. One role of SPs in AD trials is to provide information about the participant's cognitive performance. Our goal was to investigate whether participants or SPs are better at predicting the participant's cognitive performance.
Methods
We used data from the Alzheimer’s Disease Neuroimaging Initiative. Everyday Cognition (ECog) was used to assess participant and SP evaluations of the participant’s cognitive performance. The Alzheimer's Disease Assessment Scale 13-item cognitive subscale was used to measure objective cognitive performance. Random forest models and variable importance were used to compare the performance of participant and SP ECog, stratified by dyad. As part of exploratory analyses, we investigated whether the results differed by baseline diagnosis and APOE4 status.
Results
Regardless of dyad type, SP ECog was better at predicting cognitive performance. When stratifying by baseline cognitive status, cognitively normal participants were similar or better at predicting cognitive performance. Among patients with mild cognitive impairment or dementia, participants performed similarly or worse compared to their SPs. When stratifying by APOE4 status, SP performed better than the participants regardless of APOE4 status.
Conclusions
Overall, SP had similar or better predictions of participants' cognitive performance compared with participants. SP provides valuable information in AD studies.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

The effects of hormone therapy on carotid artery intima-media thickness and serum lipids by ApoE4 genotype

PDF

Association of carotid artery stiffness with measures of cognition in older adults

PDF

Relationship of blood pressure and antihypertensive medications to cognitive change in the BVAIT, WISH, and ELITE clinical trials

PDF

Cross-sectional association of blood pressure, antihypertensive medications, MRI volumetric measures and cognitive function scores in an aging population

PDF

An analysis of disease-free survival and overall survival in inflammatory breast cancer

PDF

Using ribosome footprinting to detect translational efficiency of Mbnl1x2 KO muscle cells

PDF

twas_sim, a Python-based tool for simulation and power analysis of transcriptome-wide association analysis

PDF

Long-term blood pressure variability across the clinical and biomarker spectrum of Alzheimer’s disease

PDF

Factors associated with long-term-care facility residence of patients with Alzheimer's disease from Alzheimer's disease research centers of California

PDF

Associations of ApoE4 status and DHA supplementation on plasma and CSF lipid profiles and entorhinal cortex thickness

PDF

Analysis of factors associated with breast cancer using machine learning techniques

PDF

Visual acuity outcomes after cataract extraction in Chinese Americans: the Chinese American Eye Study (CHES)

PDF

Effects of post-menopausal hormone therapy on arterial stiffness in the ELITE trial

PDF

Evaluation of factors influencing Los Angeles Tiered-Dispatch System’s improvement on bystander CPR rate and inter reliability between electronic patient care report (ePCR) and 911 call review on...

PDF

Need for tissue plasminogen activator for central venous catheter malfunction and its association with occurrence of vVenous thromboembolism

PDF

Subclinical carotid atherosclerosis, psychosocial measures, and cognitive function in middle- to older-aged adults

PDF

Incidence and survival rates of the three major histologies of renal cell carcinoma

PDF

Predicting neonatal outcomes among women diagnosed with severe preeclampsia and HELLP syndrome: a comparison of models

PDF

The evaluation of the long-term effectiveness of zero/low fluoroscopy workflow in ablation procedures for the treatment of paroxysmal and persistent atrial fibrillation

PDF

The risk estimates of pneumoconiosis and its relevant complications: a systematic review and meta-analysis

Asset Metadata

Creator Liu, Fangqing (author)

Core Title Comparison of participant and study partner predictions of cognitive impairment in the Alzheimer's disease neuroimaging initiative 3 study

School Keck School of Medicine

Degree Master of Science

Degree Program Applied Biostatistics and Epidemiology

Degree Conferral Date 2023-05

Publication Date 05/01/2023

Defense Date 04/28/2023

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag ADNI,Alzheimer’s disease,cognitive performance,OAI-PMH Harvest,study partners

Format theses (aat)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Nuno, Michelle (committee chair), Mack, Wendy (committee member), Raman, Rema (committee member), Siegmund, Kimberly (committee member)

Creator Email fangqing@usc.edu,fangqingliu@yahoo.com

Permanent Link (DOI) https://doi.org/10.25549/usctheses-oUC113091131

Unique identifier UC113091131

Identifier etd-LiuFangqin-11743.pdf (filename)

Legacy Identifier etd-LiuFangqin-11743

Document Type Thesis

Format theses (aat)

Rights Liu, Fangqing

Internet Media Type application/pdf

Type texts

Source 20230501-usctheses-batch-1034 (batch), University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the author, as the original true and official version of the work, but does not grant the reader permission to use the work if the desired use is covered by copyright. It is the author, as rights holder, who must provide use permission if such use is covered by copyright.

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA

Repository Email cisadmin@lib.usc.edu