Psychometric reliability, validity, and generalizability of 3MSE scores among American Indian adults: the Strong Heart Study

OBJECTIVE

Modified Mini-Mental State Examination (3MSE) is often used to screen for dementia, but little is known about psychometric validity in American Indians.

METHODS

We recruited 818 American Indians aged 65-95 for 3MSE examinations in 2010-2013; 403 returned for a repeat examination in 2017-2019. Analyses included standard psychometrics inferences for interpretation, generalizability, and extrapolation: factor analysis; internal consistency-reliability; test-retest score stability; multiple indicator multiple cause structural equation models.

RESULTS

This cohort was mean age 73, majority female, mean 12 years education, and majority bilingual. The 4-factor and 2nd-order models fit best, with subfactors for orientation and visuo-construction (OVC), language and executive functioning (LEF), psychomotor and working memory (PMWM), verbal and episodic memory (VEM). Factor structure was supported for both research and clinical interpretation, and factor loadings were moderate to high. Scores were generally consistent over mean 7 years. Younger participants performed better in overall scores, but not in individual factors. Males performed better on OVC and LEF, females better on PMWM. Those with more education performed better on LEF and worse on OVC; the converse was true for bilinguals. All differences were significant, but small.

CONCLUSION

These findings support use of 3MSE for individual interpretation in clinic and research among American Indians, with moderate consistency, stability, reliability over time. Observed extrapolations across age, sex, education, and bilingual groups suggest some important contextual differences may exist.

Psychometric Properties of Controlled Oral Word Association (COWA) Test and Associations With Education and Bilingualism in American Indian Adults: The Strong Heart Study

The Controlled Oral Word Association (COWA) test is used to assess phonemic fluency and executive function. Formal validation of test scores is important for accurate cognitive evaluation. However, there is a dearth of psychometric validation among American Indian adults. Given high burden of dementia risk and key contextual factors associated with cognitive assessments, this represents a critical oversight. In a large, longitudinal population-based cohort study of adult American Indians, we examined several validity inferences for COWA, including scoring, generalization, and extrapolation inferences, by investigation of factor structure, internal consistency, test-retest reliability, and differential test functioning. We found adequate unidimensional model fit, with high factor loadings. Internal consistency reliability and test-retest reliability were 0.88 and 0.77, respectively, for the full group. COWA scores were lowest among the oldest, lowest education, bilingual speakers; group effects for sex and bilingual status were small; age effect was medium; and education effect was largest. However, Wide Range Achievement Test (WRAT) score effect was stronger than education effect, suggesting better contextualization may be needed. These results support interpretation of total COWA score, including across sex, age, or language use strata.

Educational and Clinical Associations With Longitudinal Cognitive Function and Brain Imaging in American Indians: The Strong Heart Study

BACKGROUND AND OBJECTIVES

Little is known about incidence of vascular and Alzheimer dementias in American Indians.

METHODS

We conducted a large, heterogeneous, population-based, longitudinal cohort study of brain aging in community-dwelling American Indians aged 64-95 years from 11 tribes across 3 states, with neurologic examinations, 1.5T MRI, and extensive cognitive testing. Visit 1 in 2010-2013 (n = 817) and visit 2 in 2017-2019 (n = 403) included all willing, surviving participants. Standardized cognitive tests at both visits included Modified Mini-Mental Status Examination (MMSE), Wechsler Adult Intelligence Scale digit symbol coding (WAIS), Controlled Oral Word Association (COWA), and California Verbal Learning Test short form (CVLT). Test materials added at follow-up included Wide Range Achievement (reading) Test (WRAT) and National Alzheimer Coordinating Center Uniform Data Set cognitive battery (v3 form C2), including Montreal Cognitive Assessment (MoCA). MRI neuroradiologists coded infarcts, hemorrhages, white matter hyperintensities, sulcal atrophy, and ventricle enlargement.

RESULTS

The mean time between examinations was 6.7 years (SD 1.1, range 3.8-9.1 years). Years of formal education had modest correlation with WRAT reading score (r = 0.45). Prevalence and incidence (respectively) of infarcts were 32% and 12.8/1,000 person-years (PYs) hemorrhages 6% and 4.4/1000 PY worsening sulci 74% and 19.0/1000 PY worsening ventricle 79% and 30.1/1000 PY worsening leukoaraiosis 44% and 26.1/1000 PY. Linear losses per year in cognitive scores were 0.6% MMSE, 1.2% WAIS, 0.6% COWA, and 2.2% CVLT. The mean MoCA scores were 18.9 (SD 4.3).

DISCUSSION

These are the first data on longitudinal cognitive and imaging changes in American Indians and first reports of Alzheimer disease-related features. The mean scores in MoCA were similar or lower than standard cutoffs used to diagnose dementia in other racial/ethnic groups, suggesting that standardized cognitive tests may not perform well in this population. Test validation, adaptation, and score adjustment are warranted. Years of education were a poor proxy for premorbid function, suggesting novel methods for cognitive score contextualization is also needed in this population. Evaluation of selective survival suggests attrition from death, and frailty should be accounted for in causal analyses. Overall, these data represent a unique opportunity to examine neurology topics of critical importance to an understudied population.

Adaptations of an RCT during COVID: Case Study in one Rural American Indian Community

The COVID-19 pandemic is global in nature but especially threatens American Indian and Alaska Native (AI/AN) communities due to pre-existing conditions and social determinants of health. Because of the higher risk to AI/AN communities, many tribal nations have been proactive in their policies to keep the virus at bay, including travel restrictions and lockdowns. This affected tribal programs as well as collaborative research projects. One project impacted is the Native CHOICES project, an ongoing randomized controlled trial with an AI/AN community that is focused on the prevention of alcohol-exposed pregnancies. Originally designed to be conducted via in-person motivational interviewing sessions, COVID-19 restrictions precluded the intervention from being delivered in-person as it was designed. The study team received valuable input from the project’s Community Advisory Board (CAB) and community-based staff to establish a feasible and acceptable way of conducting the intervention while respecting tribally-enacted COVID-19 restrictions. The goal of this brief report is to outline not just the process to adapting to COVID-19 but also to provide recommendations for future public health programs, including the ongoing need to consider gaps in access affecting resource-poor settings.

Alcohol Consumption during COVID among Women with an Existing Alcohol-Use Disorder

Prior to the pandemic, our research team implemented a randomized controlled trial of an intervention to reduce risk for alcohol-exposed pregnancy (AEP) in American Indian women. When active recruitment for the in-person trial was paused due to COVID, the research team moved to conducting follow-up surveys with participants who had completed the intervention to better understand changes to their alcohol use during the pandemic. We collected surveys from 62 American Indian women who had completed the Native CHOICES intervention. Baseline data collected pre-COVID included demographics and scores on the Alcohol Use Disorders Identification Test (AUDIT). Follow-up surveys conducted during the active pandemic period included a self-reported questionnaire about changes in drinking patterns. At pre-COVID baseline, all participants were engaged in heavy or binge drinking. At follow-up during COVID, 24.2% reported drinking more, and over half had at least one binge drinking episode. Approximately half reported reduced drinking. We found that risky drinking remained an issue during the pandemic for many American Indian women who had engaged in this behavior pre-COVID, while others reported reducing their alcohol consumption. As the pandemic abates, concerted efforts must be made to reach those with identified alcohol use disorders to offer resources and intervention as needed.

A culturally tailored intervention to reduce risk of alcohol-exposed pregnancies in American Indian communities: Rationale, design, and methods

INTRODUCTION

Prenatal exposure to alcohol can cause lifelong physical and cognitive challenges in the form of fetal alcohol syndrome and other fetal alcohol spectrum disorders (FASDs). The prevention of prenatal alcohol exposure is thus a public health priority - and one that should account for the particular needs of subpopulations, including in American Indian/Alaska Native (AI/AN) communities. Prior to conception, alcohol-exposed pregnancy prevention is accomplished by encouraging the reduction or elimination of risky alcohol use and/or promoting effective contraceptive use among risky drinkers who could become pregnant. The current study builds on promising findings about the impact of the Centers for Disease Control and Prevention CHOICES intervention with AI/AN communities by implementing a randomized control trial of Native CHOICES, a cultural adaptation of CHOICES, with AI/AN women in a rural reservation community.

METHODS

AI/AN women aged 18-44 who are at-risk for an alcohol-exposed pregnancy are being recruited. Participants are randomized in 1:1 proportion to the intervention and a services-as-usual, waitlist control condition. The Native CHOICES intervention consists of 2 motivational interviewing (MI) sessions, an elective contraception counseling session, and electronic messaging to boost the effects of MI. Data are collected at baseline and at 6 weeks, 3 months, and 6 months post-baseline. Those assigned to the control group are eligible to enroll in Native CHOICES following the completion of the 6 months post-baseline data collection. In addition to testing intervention effectiveness, the study is designed to yield a comprehensive economic evaluation, which will provide important information regarding the financial feasibility and sustainability of Native CHOICES for healthcare systems serving AI/ANs.