Physical Activity Patterns and Cognitive Health among Older Adults in the United States

Abstract

We assessed the association between physical activity (PA) patterns and cognitive health. Using the 2011–2014 National Health and Nutrition Examination Survey data among older adults (≥60 years), we defined scoring below the 25th percentile in the average z-scores from 3 cognitive tests as having low cognitive performance. We used latent class analysis to categorize PA patterns and examined their association with cognitive performance using logistic regressions while adjusting for relevant covariates. We identified three PA groups: inactive (50.2%), moderate intensity leisure (34.5%), and high intensity multiple activities (15.3%). Compared to the inactive group, the moderate intensity leisure and high intensity multiple activities groups were less likely to have low cognitive performance (adjusted proportion ratio .85; 95% CI: .75, .94; and .76; 95% CI: .57, .96). The results highlight the need for improving cognitive health of a large proportion of physically inactive older adults by promoting multiple types of PA.

Keywords

exercise cognitive function dementia health promotion physical activity cognitive performance latent class analysis older adults

What this paper adds

• Our study expanded the existing literature by using reliable objective measures of physical activity (PA) intensity and subjective measures of PA purpose to identify “unobserved” PA patterns among older adults in the first analysis step and examined the association between the identified PA patterns and cognitive performance in the second analysis step.

• The use of PA patterns instead of individual PA measures may advance our understanding of the relationship between PA and cognitive health among older adults.

Applications of study findings

• To promote physical activity (PA) among older adults, it is important to take a “whole-person” view of their PA patterns. Uptake and maintenance of PA may require development of comprehensive measures like PA patterns to understand what PA older adults like to engage in and how to support their PA engagement.

• We found that half of older adults in the US were physically inactive, and they tended to perform more poorly in cognitive tests. While longitudinal studies are needed, PA interventions to promote multiple types of PA among a large proportion of physically inactive older adults merit consideration.

Introduction

In 2022, more than 6.5 million older adults in the United States (US) are living with dementia, and approximately every 1 in 3 older adults die with dementia (Alzheimer’s Association, 2022). There are limited pharmacological options for treating dementia, and no curative treatment is currently available (Alzheimer’s Association, 2022; Lee et al., 2019); both facts highlight the importance of prevention strategies that target modifiable risk factors of dementia (Kivipelto et al., 2018; Livingston et al., 2020). A recent Lancet Commission on dementia prevention and treatment points out that 12 modifiable risk factors account for around 40% of dementia cases worldwide, and being physically inactive is one of such modifiable risk factors to prevent cognitive decline and dementia in later life (Livingston et al., 2020).

Physical activity (PA) is a multidimensional behavior in terms of activity types, habits, and levels that encompass dimensions such as the activity’s purpose, context, and intensity level (Haga et al., 2018). Currently, most observational studies on the association between PA and cognitive health have relied on self-reported measures of PA, while only limited studies used objective measures of PA (Hamer & Chida, 2009; Stephen et al., 2017). Self-reported or subjective measures of PA are generally easy to collect in a large population even though they may be subject to recall bias. While objective measures of PA can overcome this limitation by providing more accurate and reliable data, the deployment of such an approach is generally restricted to small studies due to its associated cost (Ogonowska-Slodownik et al., 2021). Also, objective measures of PA alone do not provide information on incentives and purposes of PA, which is an important aspect to understand the effect of various purposes of PA on cognitive health. Given that PA is a complex construct to evaluate empirically, a comprehensive measure that fully captures PA patterns should encompass objectively and subjectively measured PA. Better characterization of PA patterns among older adults in the US is needed. The current lack of focus on comprehensive PA patterns may hinder the investigation of the impact of PA on cognitive health among older adults.

This study aimed to understand the relationship between PA patterns and cognitive health among older adults in the US. We used latent class analysis (LCA) to identify PA patterns among older adults from subjective and objective measures based on data from the National Health and Nutrition Examination Survey (NHANES). We then assess the association between the identified PA patterns and cognitive health. The findings may inform interventions aimed at improving cognitive health by PA engagement among older adults.

Methods

Data Source and Subject Selection

As a serial cross-sectional survey, NHANES provides continuous, publicly available, and nationally representative data to assess the health and nutritional status of the noninstitutionalized civilian US population since 1999 (National Center for Health Statistics). The survey consists of household interviews and direct standardized physical examinations conducted in a specially equipped mobile examination center (MEC). All procedures and protocols conducted in NHANES were approved by the National Center for Health Statistics (NCHS) Research Ethics Review Board. Informed consent was obtained from all participants. Further details on the design and implementation of the surveys have been described elsewhere (Chen et al., 2018; Johnson et al., 2014).

The most recent available cognitive assessment data and Physical Activity Monitor (PAM) data were collected during the examination in 2011–2012 and 2013–2014 cycles. During these two cycles, overall response rates were 71.8% for the interview and 68.8% for the MEC (Brody et al., 2019). For adults aged 60 or older, response rates were 58.4% and 55.1%, respectively (Brody et al., 2019). For this study, we aggregated data from the 2011–2014 cycles of NHANES for participants aged 60 years or older for whom data on cognitive assessment, responses to PA questionnaire, PAM, and other related covariates were available (Supplementary Figure S1).

Cognitive Assessment

The cognitive assessment administered by NHANES included three validated tests to assess the cognitive health among participants aged 60 or older: the Consortium to Establish a Registry for Alzheimer’s Disease (CERAD) Word Learning subset to evaluate immediate and delayed recall of new verbal information (memory domain); the Animal Fluency Test (AFT) to examine verbal semantic fluency (a component of executive function); and the Digit Symbol Substitution Test (DSST) to evaluate attention and processing speed (Brody et al., 2019).

CERAD included three consecutive immediate learning trials and one delayed trial. During each of the three learning trials, participants were asked to recall as many words as possible immediately after reading the 10 unrelated words aloud. In the delayed recall test, which occurred after AFT and DSST, participants were asked to recall as many words as possible from the same 10-word list without review. The immediate recall (CERAD-IR) score, calculated as the sum of the three immediate recalls, ranged from 0 to 30, and the delayed recall (CERAD-DR) score ranged from 0 to 10. Participants taking AFT were asked to name as many animals as possible in one minute, and one point was awarded for each named animal. The DSST was a module from the Wechsler Adult Intelligence Scale, Third Edition. The DSST was conducted using a paper form with a key at the top that contained 9 numbers paired with symbols. Participants were asked to copy in 2 minutes as many number-corresponding symbols as possible in 133 boxes, and a point was awarded for each correct match resulting in a total score ranging from 0 to 133. For all three tests a higher score indicates better cognitive function. The study population only included those with valid test scores. For each test, we separately calculated z-scores by subtracting that test’s mean score from each individual’s score and then dividing by that test’s standard deviation.

Criteria for Cognitive Performance

We categorized participants’ cognitive status (low vs. normal) by the average of the z-scores from their three individual cognitive tests. We classified participants as having low cognitive performance if their average z-score was in the lowest 25th percentile. This dichotomized global cognitive measure has been used previously in other studies with data from national surveys, including NHANES (Blaum et al., 2002; Brody et al., 2019). Those who scored in the lowest 25th percentile likely included some participants with cognitive impairment, either due to normative aging, dementia, or delirium, along with participants who would have been in the lowest 25th percentile throughout their lives (Brody et al., 2019). The 25th percentile cut-off point was estimated from the full analytic sample for the global cognitive measure, incorporating the complex survey design.

Subjective PA Measures

We categorized the subjective PA data into 6 items (Supplementary Table S1), based on the Global Physical Activity Questionnaire (GPAQ) available in NHANES. GPAQ included questions related to work-related activities (e.g., paid/unpaid domestic chores/tasks that are either voluntary or out of necessity), active transportation, leisure time activities (e.g., sports, fitness, and recreational activities), and sedentary activities. The GPAQ has been validated with moderate to substantial reliability (Keating et al., 2019). Metabolic equivalent of task (MET) values were assigned to each of the PA activities using the NHANES guideline. Based on the corresponding MET values, a composite score (MET in minutes (mins)) reflecting the total activity per week was created for each PA activity, by multiplying the assigned MET with dose (i.e., frequency and duration). According to the calculated composite score, the levels of participation in each type of subjective PA measure were categorized into low (0 MET mins/week), medium (below 600 MET mins/week), and high (above 600 MET mins/week). The use of 600 MET mins/week is aligned with the Center for Disease Control’s recommendation for physical activity (i.e., 150 mins of moderate intensity activity/week). Participants were also asked to report the average time spent on sedentary activities per day. The levels of participation in sedentary activities were categorized as high (8–16 hours/day of sedentary activities) or low (0–7 hours/day of sedentary activities).

Objective PA Measures

For the objectively measured PAM data, NHANES participants were asked to wear an ActiGraph GT3X+ (Actigraph) accelerometer on their non-dominant wrist for 7 full and two partial days when awake and asleep. This new measure improved participation and available validated data over hip-worn accelerometers used in earlier NHANES cycles (79% vs. 68%) (Belcher et al., 2021). We used Monitor-Independent Movement Summary (MIMS) units in the publicly available NHANES data, which was generated using a four-step process that comprised interpolation, extrapolation, bandpass filtering, and aggregation of data for each axis. MIMS-unit is a novel accelerometer summary metric designed to capture meaningful human movement that may impact health, while minimizing environmental and movement artifacts via signal filtering. An expanded discussion of the MIMS-unit algorithm, including validation and comparison to other raw accelerometer-derived metrics, was described in detail elsewhere (John et al., 2019). Briefly, methodological steps included raw signal harmonization to eliminate inter-device variability (interpolation and extrapolation), bandpass filtering to eliminate non-human movement, and signal aggregation to reduce data to simplify visualization and summarization, resulting in MIMS-units that correspond to the total amount of movement activity (Belcher et al., 2021). We categorized the participants into four groups based on the quartiles of daily MIMS-unit (<25th percentile, 25th percentile to median, median to 75th percentile, and >75th percentile) (Supplementary Table S1).

Covariates

Sociodemographic characteristics included age (60–69, 70–70, 80+), self-reported gender (male, female), race/ethnicity (Non-Hispanic White (NHW), Non-Hispanic Black (NHB), Hispanic, or other), educational attainment (below high school, high school, above high school), marital status (yes (married/cohabiting) and no (divorced, widowed, separated, or never married), and income status (ratio of family income to poverty (≤1, >1)) were collected. Ever smokers were defined as the participants who reported having smoked at least 100 cigarettes over their lifetime. Self-rated health status was categorized as good (excellent/very good/good) and poor (fair/poor). Comorbidity burden (0, 1–2, 3+) was calculated as the total number of self-reported comorbidities, including diabetes, heart diseases (any history of heart failure, coronary heart disease, angina pectoris, heart attack), stroke, hypertension, asthma, arthritis, and COPD.

Statistical Analysis

All analyses were conducted using survey analysis procedures in SAS software (version 9.4TS1m6) to account for the complex sampling design of NHANES. Characteristics of participants were summarized using frequency and proportion (all characteristics were categorical variables). Differences in characteristics across the cognitive performance (low/normal) were tested with the Rao–Scott Chi-square test for each categorical variable.

The main analysis was a two-step process. We used LCA to categorize participants with similar PA characteristics into discrete groups in the first analysis step and examine the association between the identified PA patterns and cognitive performance in the second analysis step. LCA is a probabilistic approach to identify “unobserved” latent patterns (i.e., PA patterns) based on participants’ responses to the observed variables (i.e., the 8 items of subjective and objective measures of PA from NHANES, Supplementary Table S1) (Kongsted & Nielsen, 2017; Mooney et al., 2015). LCA estimated the posterior probability of latent class membership (i.e., PA groups) and the class-specific probabilities of participation in each item of PA.

We tested LCA models with 2 to 10 latent classes using the PROC LCA procedure in SAS (Lanza et al., 2007). The final model with the optimal number of latent classes was selected based on recommended measures (i.e., the Bayesian Information Criteria, the relative size of classes in each model, and class interpretability; Supplementary Table S2) (Dean & Raftery, 2010; Kim, 2014; Weller et al., 2020). After the PA patterns were identified, we used multivariable logistic regression to assess the association between PA patterns and cognitive performance (low vs. normal) while adjusting for the aforementioned sociodemographic and clinical characteristics. We reported the marginally adjusted proportions and proportion ratios, which were calculated using the SAS NLMeans macro.

Results

Characteristics of the Study Participants by Cognitive Performance

The study sample included a total of 2336 participants, representing 43,928,767 US older adults (Table 1). About 56.3% of participants were aged between 60 to 69 years, 29.7% were aged between 70–79 years, and 14.0% were aged above 80 years. Approximately 54.6% of the participants were female, and the proportions of NHW, NHB, Hispanic, and other race/ethnicity were 81.0%, 8.1%, 6.6%, and 4.4%, respectively.

Table 1.

Characteristic of Study Participants by Cognitive Performance Status.

	Total	Normal cognitive performance	Low cognitive performance
Unweighted N (weighted %)	N = 2336 (100%)	N = 1499 (75.7%)	N = 837 (24.3%)
Variable	Unweighted N (weighted %)	Unweighted N (weighted %)	Unweighted N (weighted %)	p-value
Age
60–69	1260 (56.3)	915 (63.5)	345 (34.0)	<.001
70–79	699 (29.7)	423 (27.8)	276 (35.6)
80+	377 (14.0)	161 (8.7)	216 (30.4)
Gender
Male	1128 (45.4)	661 (43.9)	467 (49.9)	.367
Female	1208 (54.6)	838 (56.1)	370 (50.1)
Race/ethnicity ^a
NHW	1173 (81.0)	853 (85.4)	320 (67.0)	<.001
NHB	553 (8.1)	298 (5.9)	255 (15.1)
Hispanic	422 (6.6)	221 (4.5)	201 (12.9)
Other	188 (4.4)	127 (4.2)	61 (5.0)
Education
<High school	557 (15.3)	191 (9.5)	366 (33.2)	<.001
High school	558 (21.8)	336 (18.9)	222 (30.9)
>High school	1221 (62.9)	972 (71.6)	249 (35.9)
Marital status ^b
No	999 (35.5)	596 (32.3)	403 (45.2)	.001
Yes	1337 (64.5)	903 (67.7)	434 (54.8)
Number of comorbidities
0	341 (16.0)	245 (17.6)	96 (10.9)	<.001
1-2	1314 (56.8)	870 (58.2)	444 (52.5)
3+	681 (27.2)	384 (24.2)	297 (36.6)
Self-rated health
Good	1686 (80.3)	1201 (85.6)	485 (63.7)	<.001
Poor	650 (19.7)	298 (14.4)	352 (36.3)
Smoking status
Never	1146 (49.9)	756 (50.5)	390 (48.2)	.924
Ever	1190 (50.1)	743 (49.5)	447 (51.8)
Income to poverty ratio
≤1	400 (9.3)	185 (6.4)	215 (18.5)	<.001
>1	1936 (90.7)	1314 (93.6)	622 (81.5)

Abbreviations: NHW, Non-Hispanic White; NHB, Non-Hispanic Black. Notes: We define scoring below the 25th percentile in the average z-scores from 3 cognitive tests as having low cognitive performance.

^aOthers (including Asians, others, and mixed race/ethnicities).

^bYes, married/cohabiting; No, divorced/widowed/separated, never married.

Compared to individuals with normal cognitive performance, those with low cognitive performance (i.e., in the first quartile of the global cognitive measure) were more likely to be older (80+ years: 30.4% vs. 8.7%; p<.001; Table 1), not NHW (NHW: 67.0% vs. 85.4%; p<.001), less educated (<High school: 33.2% vs. 9.5%; p < .001), in lower income status (income to poverty ratio ≤ 1: 18.5% vs. 6.4%; p < .001), non-married/cohabiting (45.2% vs. 32.3%; p=.001), as well as with more comorbidities (3+: 36.6% vs. 24.2%; p < .001), and poor self-rated health (36.3% vs. 14.4%; p < .001).

PA Patterns of Physical Activities

Based on BIC scores, well-balanced sample size within each class, and interpretability, we identified three PA groups: inactive, moderate intensity leisure, and high intensity multiple activities (individuals who were actively engaged in multiple types of PA) groups (Figure 1; Supplementary Table S2). Approximately 50.2% of participants were in the inactive group, 34.5% were in the moderate intensity leisure group, and 15.3% were in the high intensity multiple activities group. The class membership was interpreted based on the class-specific probability of participating in different PA types (Figure 1). For example, participants in the inactive group had the highest probability of having high level of sedentary activity, and the lowest probability of having the highest level of daily MIMS-units. They also had the lowest probability of having high levels of participation in all types of PA. Participants in the moderate intensity leisure group had the highest probability of having high levels of participation in moderate and vigorous leisure activity (36.6% and 21.1%, respectively). Participants in the high intensity multiple activities group generally had high probability of having high levels of participation in most types of PA: moderate and vigorous work-related activity (88.4% and 51.6%), moderate leisure time activity (34.8%), and active transportation (23.4%).

Figure 1.

Physical activities patterns were classified into the inactive group, the moderate intensity leisure group, and the high intensity multiple activities group. Notes: The figure shows the latent class-specific probability for the highest level of physical activity or sedentary activity. Participation in vigorous or moderate work-related activity, active transportation, and vigorous or moderate leisure time activity was categorized into low, medium, and high levels. Participation in sedentary activity was categorized in low or high levels. Levels of daily MIMS-units were categorized into four groups (<25th percentile, 25th percentile to median, median to 75th percentile, and >75th percentile). The highest level of daily MIMS-units was the >75th percentile.

Characteristics of the Study Participants by PA Groups

The inactive group was older than the moderate intensity leisure and high intensity multiple activities group (80+ years: 21.2% vs. 8.0% and 3.7%, p = .005, Table 2). Also, participants in the inactive group had lower income status compared to those in the moderate intensity leisure and high intensity multiple activities groups (Income to poverty ratio≤1: 12.3% vs. 6.7% and 5.5%, p < .001). Participants in the inactive group were more likely to have low cognitive performance than those in the moderate intensity leisure and high intensity multiple activities groups (33.5% vs. 16.3% and 12.0%, p < .001). There is no significant difference in other characteristics among participants across the three PA groups.

Table 2.

Characteristics of Study Participants by the Three Physical Activity Groups Identified From the Latent Class Analysis.

	Total	Inactive group	Moderate intensity leisure group	High intensity multiple activities group
Unweighted N (weighted %)	N = 2336 (100%)	N = 1247 (50.2%)	N = 806 (34.5%)	N = 283 (15.3%)
Variable	Unweighted N (weighted %)	Unweighted N (weighted %)	Unweighted N (weighted %)	Unweighted N (weighted %)	p-value
Age
60–69	1260 (56.3)	545 (45.0)	511 (64.3)	204 (75.4)	.005
70–79	699 (29.7)	418 (33.7)	218 (27.8)	63 (20.9)
80+	377 (14.0)	284 (21.2)	77 (8.0)	16 (3.7)
Gender
Male	1128 (45.4)	601 (44.3)	359 (42.8)	168 (54.7)	.825
Female	1208 (54.6)	646 (55.7)	447 (57.2)	115 (45.3)
Race/ethnicity ^a
NHW	1173 (81.0)	660 (80.8)	361 (78.7)	152 (86.7)	.216
NHB	553 (8.1)	309 (9.2)	190 (8.0)	54 (4.9)
Hispanic	422 (6.6)	197 (6.3)	167 (7.4)	58 (5.6)
Other	188 (4.4)	81 (3.7)	88 (6.0)	19 (2.8)
Education
<High school	557 (15.3)	338 (19.8)	164 (11.1)	55 (9.7)	.564
High school	558 (21.8)	332 (25.3)	159 (17.7)	67 (19.6)
>High school	1221 (62.9)	577 (54.9)	483 (71.2)	161 (70.7)
Marital status ^b
No	999 (35.5)	575 (39.5)	330 (32.7)	94 (28.4)	.087
Yes	1337 (64.5)	672 (60.5)	476 (67.3)	189 (71.6)
Number of comorbidities
0	341 (16.0)	115 (9.9)	168 (21.8)	58 (22.8)	.228
1-2	1314 (56.8)	680 (54.5)	463 (59.3)	171 (58.9)
3+	681 (27.2)	452 (35.6)	175 (18.9)	54 (18.3)
Self-rated health status
Good	1686 (80.3)	793 (70.9)	669 (90.3)	224 (88.7)	.442
Poor	650 (19.7)	454 (29.1)	137 (9.7)	59 (11.3)
Smoking status
Never	1146 (49.9)	570 (46.2)	446 (56.0)	130 (48.5)	.845
Ever	1190 (50.1)	677 (53.8)	360 (44.0)	153 (51.5)
Income to poverty ratio
≤1	400 (9.3)	243 (12.3)	117 (6.7)	40 (5.5)	<.001
>1	1936 (90.7)	1004 (87.7)	689 (93.3)	243 (94.5)
Cognitive performance
low	837 (24.3)	552 (33.5)	219 (16.3)	66 (12.0)	<.001
Normal	1499 (75.7)	695 (66.5)	587 (83.7)	217 (88.0)

Abbreviations: NHW, Non-Hispanic White; NHB, Non-Hispanic Black.

^aOthers (including Asians, others, and mixed race/ethnicities).

^bYes, married/cohabiting; No, divorced/widowed/separated, never married.

Association between PA Groups and Cognitive Performance

Compared to the inactive group, the moderate intensity leisure group was less likely to have low cognitive performance after adjusting for relevant sociodemographic and clinical characteristics (adjusted proportion ratio (APR) .85; 95% CI: .75, .94; Figure 2). Similarly, the high intensity multiple activities group was less likely to have low cognitive performance than the inactive group (APR .76; 95% CI: .57, .96).

Figure 2.

The associations between identified physical activity groups and having low cognitive performance after adjusting for sociodemographic and clinical characteristics. Notes: The marginally adjusted proportion was the adjusted proportion of having low cognitive performance by each characteristic.

Discussion

Based on a large nationally representative sample, we identified three distinct PA groups among US older adults aged 60 years or older: the inactive group, the moderate intensity leisure group, and the high intensity multiple activities group. It is concerning that half of the US older adults were in the inactive group. They tended to perform more poorly in cognitive tests than their counterparts in the moderate intensity leisure and high intensity multiple activities groups. While the PA-cognitive health association revealed in our study aligns with the previous work (Endeshaw & Goldstein, 2021; Hamer & Chida, 2009; Loprinzi et al., 2018; Stephen et al., 2017), the current study adds to the literature with improved rigor by using both subjective and objective measures of multiple types of PA and the application of LCA in identifying the underlying PA patterns. In addition, we used a global measure of cognitive health based on three established cognitive tests. Our findings highlight the need for public health interventions to improve PA among older adults in the inactive group as a modifiable strategy to improve their cognitive health.

PA is a complex behavior of human movement that is influenced by multiple factors, including physiological, psychological, social, and environmental correlates (Pettee Gabriel et al., 2012). However, earlier studies primarily rely on a single self-reported approach to measuring PA, or based on largely subjective self-reports, which may not fully capture PA as a complex behavior (Hamer & Chida, 2009; Stephen et al., 2017). With the advent of wearable devices and recognition of the multifaceted nature of PA, studies began to collect data on multiple PA types using objective and reliable measures of PA (Daskalopoulou et al., 2017). Our study expanded the existing literature by using reliable objective measures of PA intensity and subjective measures of PA purpose to identify “unobserved” PA patterns. To promote PA among older adults, it is important to take a “whole-person” view and understand meaningful motivators in their current lives. Uptake and maintenance of PA may require the development of comprehensive measures such as PA patterns in order to understand what PA older adults like to engage in and how to support their PA engagement.

It is worth noting that we observed a stronger negative PA-cognitive health association in the high intensity multiple activities group than in the moderate intensity leisure group. One possible explanation is that a higher dose of PA (i.e., higher intensity or frequency) confers more cognitive health benefits (Gallardo-Gómez et al., 2022; Sanders et al., 2019). Another reason is that participation in multiple types of PA activities (e.g., aerobic exercise, resistance training, flexibility, coordination, and balance) is more effective in improving global cognition and the activities of daily living skills compared to a single form of PA (Begde et al., 2021; Huang et al., 2022; Rezola-Pardo et al., 2020). For example, a recent clinical trial conducted in Spain showed that a PA intervention with multiple types of PA conferred greater improvements in physical performance than a walking intervention and can therefore support a more independent life for older adults (Rezola-Pardo et al., 2020). There is also evidence that different types of PA may positively affect cognitive health through different mechanisms (Huang et al., 2022; Netz, 2019). For example, aerobic and resistance training may improve cognitive health through improvement in cardiovascular fitness, whereas flexibility, coordination, and balance training may affect cognition directly (Netz, 2019). Finally, the combination of high intensity and multiple types of PA may also carry a synergistic impact on cognitive health. Encouragingly, a previous review study suggests that interventions to promote multiple types of PA have more potential to be effective at increasing PA levels in older adults (Zubala et al., 2017).

The study findings should be interpreted in the context of several limitations. First, because the analysis is cross-sectional, the directionality between PA patterns and cognitive health is difficult to establish. The identified association needs to be further explored through longitudinal studies and randomized controlled trials. Also, cognitive tests in NHANES were chosen for ease of administration in surveys. Despite their demonstrated validity, they may not be sufficient to capture an individual’s entire cognitive profile. These tests were not comparable to other established cognitive instruments that contribute to clinical diagnosed cognitive impairment or dementia (e.g., Mini-Mental State Exam and Montreal Cognitive Assessment). Therefore, participants who were considered as having normal cognitive performance in this study may still have cognitive impairment that was not captured by the cognitive tests in NHANES. Although NHANES oversampled racial/ethnic minority respondents, the relatively small number of racial/ethnic minority respondents may limit the power in detecting differences in cognitive health by race/ethnicity. Lastly, there may be potential recall bias in covariate information caused by different levels of cognitive status among survey participants.

Our results have implications for future research methods and PA promotion programs designed to improve cognitive health among older adults. Methodologically, we advocate the use of PA patterns instead of individual PA measures to advance our understanding of the relationship between PA and cognitive health. Although prospective studies with longitudinal assessment of both PA and cognitive performance are needed, our study highlights the vulnerability in cognitive health among a large portion of older Americans who are physically inactive and could potentially benefit from PA interventions with multiple types of PA.

Supplemental Material

Supplemental Material - Physical Activity Patterns and Cognitive Health among Older Adults in the United States

Supplemental Material for Physical Activity Patterns and Cognitive Health among Older Adults in the United States by Weixin Li, Yan Li, Ming Wen, and Bian Liu in Journal of Applied Gerontology.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This study is in part funded by a grant from the National Institute on Minority Health and Health Disparities of the National Institutes of Health (award number: R01MD013886).

Research Ethics

This study is conducted in compliance with the ethical standards of the responsible committee on human experimentation, both institutional and national. No animals were involved in this research. This secondary data analysis used publicly available de-identified data from the National Health and Nutrition Examination Survey (NHANES). All procedures and protocols conducted in NHANES were approved by the National Center for Health Statistics (NCHS) Research Ethics Review Board. Informed consent was obtained from all participants.

ORCID iDs

Weixin Li

Yan Li

Supplemental Material

Supplemental material for this article is available online.

References

Alzheimer’s Association (2022). Alzheimer’s disease facts and figures. Retrieved April 5, 2022 from https://alz-journals.onlinelibrary.wiley.com/doi/epdf/10.1002/alz.12638#:∼:text=Anestimated6.5millionAmericans,prevent%2CsloworcureAD

Begde

Jain

Hogervorst

Wilcockson

(2021). Does physical exercise improve the capacity for independent living in people with dementia or mild cognitive impairment: An overview of systematic reviews and meta-analyses. Aging & Mental Health, 1–11. Online ahead of print. https://doi.org/10.1080/13607863.2021.2019192

Belcher

B. R.

Wolff-Hughes

D. L.

Dooley

E. E.

Staudenmayer

Berrigan

Eberhardt

M. S.

Troiano

R. P.

(2021). US population-referenced percentiles for wrist-worn accelerometer-derived activity. Medicine and Science in Sports and Exercise, 53(11), 2455–2464. https://doi.org/10.1249/mss.0000000000002726

Blaum

C. S.

Ofstedal

M. B.

Liang

(2002). Low cognitive performance, comorbid disease, and task-specific disability: Findings from a nationally representative survey. The Journals of Gerontology. Series A, Biological Sciences and Medical Sciences, 57(8), M523–M531. https://doi.org/10.1093/gerona/57.8.m523

Brody

D. J.

Kramarow

E. A.

Taylor

C. A.

McGuire

L. C.

(2019). Cognitive performance in adults aged 60 and over: National health and nutrition examination survey, 2011-2014. National Health Statistics Reports, (126), 1–23.

Chen

T. C.

Parker

J. D.

Clark

Shin

H. C.

Rammon

J. R.

Burt

V. L.

(2018). National health and nutrition examination survey: Estimation procedures, 2011-2014. Vital Health Stat, 2(177), 1–26.

Daskalopoulou

Stubbs

Kralj

Koukounari

Prince

Prina

A. M.

(2017). Physical activity and healthy ageing: A systematic review and meta-analysis of longitudinal cohort studies. Ageing Research Reviews, 38, 6–17. https://doi.org/10.1016/j.arr.2017.06.003

Dean

Raftery

A. E.

(2010). Latent class Analysis variable selection. Annals of the Institute of Statistical Mathematics, 62(1), 11–35. https://doi.org/10.1007/s10463-009-0258-9

Endeshaw

Goldstein

(2021). Association between physical exercise and cognitive function among community-dwelling older adults. J Appl Gerontol, 40(3), 300–309. https://doi.org/10.1177/0733464820952242

10.

Gabriel

P. K. K.

Morrow

J. R.

Jr. Woolsey

A. L.

(2012). Framework for physical activity as a complex and multidimensional behavior. Journal of Physical Activity & Health, 9(Suppl 1), S11–S18. https://doi.org/10.1123/jpah.9.s1.s11

11.

Gallardo-Gómez

del Pozo-Cruz

Noetel

Álvarez-Barbosa

Alfonso-Rosa

R. M.

del Pozo Cruz

(2022). Optimal dose and type of exercise to improve cognitive function in older adults: A systematic review and bayesian model-based network meta-analysis of RCTs. Ageing Research Reviews, 76, 101591. https://doi.org/10.1016/j.arr.2022.101591

12.

Haga

Vrotsou

Bredland

(2018). Visualizing physical activity patterns among community-dwelling older adults: A pilot study. Sports (Basel, Switzerland, 6(4), 135. https://doi.org/10.3390/sports6040135

13.

Hamer

Chida

(2009). Physical activity and risk of neurodegenerative disease: A systematic review of prospective evidence. Psychological Medicine, 39(1), 3–11. https://doi.org/10.1017/s0033291708003681

14.

Huang

Zhao

Cai

Zhang

Wan

(2022). Comparative efficacy of various exercise interventions on cognitive function in patients with mild cognitive impairment or dementia: A systematic review and network meta-analysis. Journal of Sport and Health Science, 11(2), 212–223. https://doi.org/10.1016/j.jshs.2021.05.003

15.

John

Tang

Albinali

Intille

(2019). An open-source monitor-independent movement summary for accelerometer data processing. J Meas Phys Behav, 2(4), 268–281. https://doi.org/10.1123/jmpb.2018-0068

16.

Johnson

C. L.

Dohrmann

S. M.

Burt

V. L.

Mohadjer

L. K.

(2014). National health and nutrition examination survey: Sample design, 2011-2014. Vital Health Stat, 2(162), 1–33.

17.

Keating

X. D.

Zhou

Liu

Hodges

Liu

Guan

Phelps

Castro-Piñero

(2019). Reliability and concurrent validity of global physical activity questionnaire (GPAQ): A systematic review. International Journal of Environmental Research and Public Health, 16(21), 4128. https://doi.org/10.3390/ijerph16214128

18.

Kim

S. Y.

(2014). Determining the number of latent classes in single- and multi-phase growth mixture models. Struct Equ Modeling, 21(2), 263–279. https://doi.org/10.1080/10705511.2014.882690

19.

Kivipelto

Mangialasche

Ngandu

(2018). Lifestyle interventions to prevent cognitive impairment, dementia and Alzheimer disease. Nature Reviews. Neurology, 14(11), 653–666. https://doi.org/10.1038/s41582-018-0070-3

20.

Kongsted

Nielsen

A. M.

(2017). Latent Class Analysis in health research. Journal of Physiotherapy, 63(1), 55–58. https://doi.org/10.1016/j.jphys.2016.05.018

21.

Lanza

S. T.

Collins

L. M.

Lemmon

D. R.

Schafer

J. L.

(2007). Proc LCA: A SAS procedure for latent class Analysis. Struct Equ Modeling, 14(4), 671–694. https://doi.org/10.1080/10705510701575602

22.

Lee

G. E.

Kim

J. Y.

Jung

J. H.

Kang

H. W.

Jung

I. C.

(2019). Non-pharmacological interventions for patients with dementia: A protocol for a systematic review and meta-analysis. Medicine, 98(38), Article e17279. https://doi.org/10.1097/md.0000000000017279

23.

Livingston

Huntley

Sommerlad

Ames

Ballard

Banerjee

Mukadam

(2020). Dementia prevention, intervention, and care: 2020 report of the Lancet commission. Lancet, 396(10248), 413–446. https://doi.org/10.1016/S0140-6736(20)30367-6

24.

Loprinzi

P. D.

Edwards

M. K.

Crush

Ikuta

Del Arco

(2018). Dose-response association between physical activity and cognitive function in a national sample of older adults. American Journal of Health Promotion: AJHP, 32(3), 554–560. https://doi.org/10.1177/0890117116689732

25.

Mooney

S. J.

Joshi

Cerdá

Quinn

J. W.

Beard

J. R.

Kennedy

G. J.

Benjamin

E. O.

Ompad

D. C.

Rundle

A. G.

(2015). Patterns of physical activity among older adults in New York city: A latent class Approach. American Journal of Preventive Medicine, 49(3), e13–e22. https://doi.org/10.1016/j.amepre.2015.02.015

26.

National Center for Health Statistics . National health and nutrition examination survey. https://www.cdc.gov/nchs/nhanes/index.htm

27.

Netz

(2019). Is there a preferred mode of exercise for cognition enhancement in older age?—a narrative review [review]. Frontiers in Medicine, 6, 57. https://doi.org/10.3389/fmed.2019.00057

28.

Ogonowska-Slodownik

Morgulec-Adamowicz

Geigle

P. R.

Kalbarczyk

Kosmol

(2021). Objective and self-reported assessment of physical activity of women over 60 Years old. Ageing International, 47, 307–320. https://doi.org/10.1007/s12126-021-09423-z

29.

Rezola-Pardo

Rodriguez-Larrad

Gomez-Diaz

Lozano-Real

Mugica-Errazquin

Patiño

M. J.

Bidaurrazaga-Letona

Irazusta

Gil

S. M.

(2020). Comparison between multicomponent exercise and walking interventions in long-term nursing homes: A randomized controlled trial. Gerontologist, 60(7), 1364–1373. https://doi.org/10.1093/geront/gnz177

30.

Sanders

L. M. J.

Hortobágyi

La Bastide-Van Gemert

Van Der Zee

E. A.

Van Heuvelen

M. J. G.

(2019). Dose-response relationship between exercise and cognitive function in older adults with and without cognitive impairment: A systematic review and meta-analysis. PLoS One, 14(1), Article e0210036. https://doi.org/10.1371/journal.pone.0210036

31.

Stephen

Hongisto

Solomon

Lönnroos

(2017). Physical activity and Alzheimer’s disease: A systematic review. The Journals of Gerontology: Series A, 72(6), 733–739. https://doi.org/10.1093/gerona/glw251

32.

Weller

B. E.

Bowen

N. K.

Faubert

S. J.

(2020). Latent class Analysis: A guide to best practice. Journal of Black Psychology, 46(4), 287–311. https://doi.org/10.1177/0095798420930932

33.

Zubala

MacGillivray

Frost

Kroll

Skelton

D. A.

Gavine

Morris

(2017). Promotion of physical activity interventions for community dwelling older adults: A systematic review of reviews. PLoS One, 12(7), Article e0180902. https://doi.org/10.1371/journal.pone.0180902

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.46 MB