Undiagnosed Dementia Is Associated with Poor Physical Function in Older Adults

Abstract

Background:

Older adults with a cognitive impairment, including those not yet diagnosed, may have deficits in their physical function.

Objective:

We sought to determine the associations of cognitive impairment consistent with dementia (CICD) diagnosis status on handgrip strength, gait speed, and functional disability in older adults.

Methods:

The analytical sample included 8,383 adults aged ≥65-years without history of stroke, cancers, neurological conditions, or brain damage who participated in at least one-wave of the 2010–2016 waves of the Health and Retirement Study. A handgrip dynamometer measured handgrip strength. Men with handgrip strength <26 kg and women <16 kg were weak. Gait speed was timed across a 2.5-m course and those with slowness had gait speed <0.8 m/s. Participants with difficulty or an inability in completing any basic activities of daily living had a functional disability. The adapted Telephone Interview of Cognitive Status evaluated cognitive function. Persons with scores <7 had a CICD. Healthcare provider dementia-related diagnosis was self-reported. Undiagnosed CICD was defined as no reported dementia-related diagnosis but had CICD, while diagnosed CICD was classified as reporting a dementia-related diagnosis. Covariate-adjusted logistic models were used for the analyses.

Results:

Persons with undiagnosed CICD had 1.37 (95% confidence interval (CI): 1.04–1.80) greater odds for weakness and 2.02 (CI: 1.39–2.94) greater odds for slow gait speed. Older adults with diagnosed CICD had 2.29 (CI: 1.32–3.97) greater odds for slowness and 1.85 (CI: 1.19–2.90) greater odds for functional disability.

Conclusion:

Screening for CICD could be recommended when defects in physical function are observed in older adults.

INTRODUCTION

Older adults are a rapidly growing age demographic that frequently utilize healthcare services when such facilities are available [1]. Examinations and screenings are the primary services provided to older adults during general healthcare visits [2]. For example, screening for certain age-related morbidities that are highly prevalent in older adults, such as cardiovascular diseases, are recommended for routine healthcare assessments [3]. Alzheimer’s disease and related dementias are another significant morbidity that is estimated to impact approximately 14 million Americans by the year 2060 [4]. Although screening for dementia-related conditions is an important step in making referrals for comprehensive cognitive testing, consensus regarding dementia screenings does not recommend or oppose such screening during routine geriatric healthcare visits [5]. Accordingly, many older adults could be living with undetected dementia because of an absence of regular screening [6].

Musculoskeletal- and neuromuscular-related morbidities and disabilities are similarly observed in older populations [7]. Physical function is an umbrella term that includes different categories of the disabling process, and each category of disablement has a distinct level of severity [8]. For example, muscle function is often assessed with handgrip strength (HGS), and low HGS represents the initial stages of disablement [9]. Deficits in muscle function precede poor physical performance. Measures of physical performance include mobility tasks such as walking speed whereby slow gait speed signifies a more advanced stage in the disabling cascade [8]. Functional disability, which is usually determined by having difficulty or frank inability in completing a basic self-care task, is the end stage of the disabling process [8]. While the disabling process is generally associated with muscle-related diseases such as sarcopenia [10], there is growing evidence that suggests deficiencies in physical function are also linked to cognitive impairment [11–13].

The common cause hypothesis, which suggests that common factors are responsible for eroding non-cognitive and cognitive processes during aging, serves as the foundation for explaining how physical measures and cognitive functioning are connected [14, 15]. As such, physical measurements, including HGS and gait speed, are bidirectionally associated with cognitive impairment [16, 17]. Therefore, it is possible that persons with undiagnosed dementia are living with weakness, slowness, or a functional disability. Given that physical function assessments can be more routinely evaluated in standard geriatric visits than extensive cognitive testing [18], older adults identified as having physical function deficits could be considered for cognitive screening as needed to identify those at greatest risk for dementia. Assessing physical function could be particularly important for helping to mitigate the high prevalence of undetected dementia [6], which may differ by demographic factors such as sex and race [19, 20]. Our study sought to determine the associations of cognitive impairment consistent with dementia (CICD) diagnosis status on HGS, gait speed, and functional disability in older adults. We also sought to examine these associations by sex and race as a secondary objective.

MATERIALS AND METHODS

Participants

This investigation completed a secondary analysis of publicly available data from the 2010–2016 waves of the Health and Retirement Study (HRS). Individual HRS datafiles were joined with the RAND HRS dataset as needed [21]. The HRS utilizes a panel design to observe health and economic factors in adults as they age [22]. Participants in the HRS complete core interviews, usually by telephone, biennially until death. Beginning in the 2010 wave, the HRS started including additional self-report measures of healthcare provider diagnosed conditions. The HRS also collects physical measurements such as HGS and gait speed, at participant residences during alternating waves. Thus, we analyzed the last wave in which each participant had data for HGS, gait speed, and cognitive function. Overall interview response rates for the HRS have been consistently been >80% [23]. More details about the HRS are available elsewhere [24].

The analytic sample included 8,383 older adults aged at least 65 years without neurological and sensory conditions (i.e., multiple sclerosis, cerebral palsy, epilepsy, Parkinson’s disease, Lou Gehrig’s disease, seizures, neuropathy) or brain damage. Although the HRS uses a panel design, we analyzed the last available wave completed for each participant to best estimate reported dementia-related diagnosis and CICD (n = 1,151 for the 2010 wave; n = 1,265 for the 2012 wave; n = 3,154 for the 2014 wave; n = 2,813 for the 2016 wave). Each person provided written informed consent before participating in the HRS and the University’s Behavioral Sciences Committee Institutional Review Board approved study protocols.

Measures

Dementia-related diagnosis

Participants reported to interviewers if a doctor had ever diagnosed them with dementia, senility, or any other serious memory impairment. Persons indicating an affirmative diagnosis were classified as having been diagnosed with a dementia-related condition. Participants had the ability to dispute their previously collected self-reported dementia-related diagnosis response during HRS interviews. This single-item dementia-related diagnosis measure has been used in other similar investigations [20, 25–28] and has demonstrated a moderate level of agreement with the Hurd et al. [25, 29] models for dementia.

Cognitive impairment consistent with dementia

An adapted version of the Telephone Interview of Cognitive Status (TICS) was used to examine cognitive function. This 27-point version of the TICS is a well-validated cognitive function screening tool designed for population-based studies such as HRS [30]. Each person was asked to engage in a series of cognitive assessments including: 1) instant and delayed word recall from a list of 10 noun-free words to test general memory (0–20 points), 2) serial sevens subtraction test beginning with the number 100 for evaluating working memory (0–5 points), and 3) counting backward at maximum speed for 10 consecutive numbers starting at 20 to examine mental processing speed (0–2 points). Participants with scores <7 were considered as having a CICD [31, 32]. This well-utilized CICD cut-point correctly classifies approximately 80% of persons as having dementia or not [31, 32]. More information about how cognitive function was examined in the HRS is available elsewhere [33].

Cognitive impairment consistent with dementia diagnosis status

The responses from the healthcare provider dementia-related diagnosis item and TICS assessment were used to define undiagnosed and diagnosed CICD. Persons who were cognitively intact had both no reported dementia-related diagnosis and CICD. Older adults with undiagnosed CICD had no reported dementia-related diagnosis from a healthcare provider but had a CICD, whereas persons with diagnosed CICD reported a dementia-related diagnosis from a healthcare provider.

Handgrip strength

A Smedley spring-type handgrip dynamometer (Scandidact; Odder, Denmark) measured HGS. Briefly, trained interviewers fit the dynamometer to the hand size of participants and explained the testing procedures. Participants practiced gripping the dynamometer with their arm at their side and elbow flexed at 90°. Starting on the self-reported non-dominant hand, each person squeezed the dynamometer with maximal effort, and then released their grip. Participants completed two HGS trials on each hand, alternating between hands.

Those who had problems standing or positioning their arm while grasping the dynamometer were allowed to be seated and place their arm on a supporting object during testing. Persons experiencing a surgical procedure within six months, swelling, inflammation, severe pain, or an injury to their hands in the previous month before testing did not engage in HGS assessments. Additional details about the HGS procedures used in the HRS are available elsewhere [34]. The highest recorded HGS, regardless of hand was used for determining strength capacity [35]. Men with HGS <26 kg and women with HGS <16 kg were classified as weak [36].

Gait speed

Trained interviewers measured and created a walking course in an unobstructed and preferably non-carpeted area of participant residences. A piece of tape was placed on the floor to identify the starting and ending points of the walking course. Pre-test instructions were provided to participants. Interviewers signalled when to begin walking by stating, “ready, begin” and with their toes aligned on the start line of the walking course, participants walked at a normal pace across the 2.5-m course. The interviewer began timing when the participant’s foot was across the start line and completely touching the floor, and stopped when the foot was likewise touching the floor beyond the finish line. After completing the first walking trial, interviewers reset the stopwatch and asked participants to walk back to the other side of the walking course using the same protocol for the second trial. Walking aids were permitted if they were normally used. Participants without sufficient space to perform the test in their residence, and those with problems from a recent surgery, injury, or other health condition that prevented them from walking may not have been part of gait speed testing [34, 37, 38]. The average of the two trials was calculated and persons with a mean gait speed <0.8m/s were considered slow [39]. Additional details about the gait speed procedures in the HRS are available elsewhere [34].

Functional disability

Ability to complete six basic activities of daily living were self-reported: dressing, eating, transferring in-or-out of bed, toileting, bathing, and walking across a small room. Persons reporting having difficulty or an inability in completing an activity of daily living were classified as living with a functional disability.

Covariates

Age, sex, race, educational achievement, height, and body mass were self-reported. Body mass index was calculated as mass in kilograms divided by height in meters-squared and individuals with a body mass index ≥30 kg per meters-squared were obese. Perceived health was collected with a single-item wherein participants were asked to rate their health as either “excellent”, “very good”, “good”, “fair”, or “poor”. Respondents also told interviewers if they were currently smoking cigarettes or had ever smoked at least 100 cigarettes in their lifetime (former smoker).

The 8-item Center for the Epidemiologic Studies Depression scale examined depressive symptomology [40]. Participants reported if they experienced any positive or negative emotions during the week before the interview. Scores ranged from 0–8, with higher scores suggesting more depressive symptoms. Persons with scores ≥3 were classified as depressed [40]. The HRS also asks participants about the frequency and intensity of their physical activity participation. Those who reported engaging in moderate-to-vigorous physical activity at least “once a week” were classified as engaging in moderate-to-vigorous physical activity [41].

Statistical analysis

All analyses were performed with SAS 9.4 software (SAS Institute; Cary, NC, USA). The HRS analytic recommendations guided our analyses [42]. Our analyses accounted for the complex sampling design for generating population-based estimates that were nationally-representative. The unweighted descriptive characteristics of the participants were reported as mean ± standard deviation for continuous variables and frequency (percentage) for categorical variables. Means and 95% confidence intervals (CI) for the descriptive characteristics of the participants by CICD diagnosis status were also presented to make comparisons between CICD diagnosis status groups.

Separate crude, partially-adjusted, and fully adjusted logistic regression models determined the associations of undiagnosed and diagnosed CICD (reference group: cognitively intact) on: 1) weakness, 2) slowness, and 3) functional disability. The partially-adjusted models controlled for age, sex, race, educational achievement, obesity status, self-rated health, cigarette smoking status, depressive status, arthritis status, and diabetes status. Moderate to-vigorous physical activity participation status was then added to the partially-adjusted models for creating the fully-adjusted models. The covariates included in the logistic models were prespecified by the investigators because they were thought to have been influential for our associations based upon clinical experience.

As an additional analysis, participants reporting a dementia-related diagnosis but no CICD were excluded (n = 32). Thereafter, the same series of fully-adjusted logit models analyzed the associations between CICD diagnosis status and each of the physical functioning assessments. For adhering to guidelines from the National Institutes of Health to examine sex as a biological variable in our study [43], we ran sex-stratified fully-adjusted logistic regression models for the associations of undiagnosed and diagnosed CICD on weakness, slowness, and functional disability. We also conducted fully-adjusted logistic regression models for the associations of undiagnosed and diagnosed CICD on weakness, slowness, and functional disability stratified by identified race (Black, White, other).These additional analyses were presented as supplementary because these analyses were not principal for our investigation. An alpha level of 0.05 was used for the analyses.

RESULTS

Table 1 presents the descriptive characteristics of the participants. Overall, participants were aged 75.6±7.2 years and 4,840 (57.7%) were female. The means and 95% CI for the descriptive characteristics of the participants by CICD diagnosis status are shown in Table 2. A lower proportion of participants who were cognitively intact (11.7%; CI: 11.0–12.4) had weakness compared to older adults with undiagnosed CICD (22.0%; CI: 18.7–25.3) and diagnosed CICD (23.7%; CI: 16.9–30.4). Likewise, a lower proportion of cognitively intact older adults (59.0%; CI: 57.9–60.1) had slow gait speed relative to persons with undiagnosed CICD (84.7%; CI: 81.8–87.6) and diagnosed CICD (84.2%; CI: 78.4–90.0). While there was a greater proportion of participants with functional disability who had undiagnosed CICD (20.8%; CI: 17.6–24.1) compared to those who were cognitively intact (15.2%; CI: 14.4–16.0), the greatest proportion of participants with functional disability had a diagnosed CICD (36.8%; 29.2–44.5).

Table 1.

Unweighted descriptive characteristics of the participants

	Overall unweighted sample (n = 8,383)
Age (y)	75.6 ± 7.2
Female (n (%))	4,840 (57.7)
Obese (n (%))	2,544 (30.4)
Fair or poor self-rated health (n (%))	2,177 (26.0)
Black race (n (%))	1,114 (13.3)
White race (n (%))	6,880 (82.1)
Never smoked (n (%))	3,731 (44.5)
MVPA participation (n (%))	4,271 (51.0)
Depressed (n (%))	1,502 (17.9)
High school graduate or above (n (%))	6,518 (77.8)
Has arthritis (n (%))	5,879 (70.1)
Has diabetes (n (%))	2,232 (26.6)
Weakness (n (%))	1,060 (12.6)
Slow gait speed (n (%))	5,135 (61.3)
Functional disability (n (%))	1,337 (16.0)
Undiagnosed CICD (n (%))	596 (7.1)
Diagnosed CICD (n (%))	152(1.8)

Results are presented as mean ± standard deviation or frequency (percentage) as indicated. CICD, cognitive impairment consistent with dementia; MVPA, moderate-to-vigorous physical activity.

Table 2.

Means and 95% confidence intervals for the unweighted descriptive characteristics of the participants by cognitive impairment consistent with dementia diagnosis status

	Cognitively intact (n = 7,635)	Undiagnosed CICD (n = 596)	Diagnosed CICD (n = 152)
Age (y)	75.2 (75.1–75.4)	79.0 (78.4–79.7)	79.5 (78.2–80.9)
Female (%)	57.6 (56.5–58.7)	60.6 (56.7–64.5)	77.0 (70.3–83.7)
Obese (%)	31.0 (30.0–32.0)	24.3 (20.9–27.8)	23.0 (16.3–29.7)
Fair or poor self-rated health (%)	24.3 (23.4–25.3)	40.4 (36.5–44.4)	51.3 (43.4–59.3)
Black race (%)	12.1 (11.4–12.9)	28.5 (24.9–32.2)	12.5 (7.2–17.8)
White race (%)	83.6 (82.7–84.4)	62.9 (59.0–66.8)	82.2 (76.2–88.3)
Never smoked (%)	44.5 (43.4–45.7)	44.5 (40.5–48.5)	43.4 (35.5–51.3)
MVPA participation (%)	52.1 (51.0–53.2)	40.0 (35.8–43.7)	36.8 (29.2–44.5)
Depressed (%)	16.7 (15.9–17.6)	29.7 (26.0–33.4)	32.2 (24.8–39.7)
High school graduate or above (%)	80.6 (79.7–81.5)	44.1 (40.1–48.1)	67.8 (60.3–75.2)
Has arthritis (%)	70.1 (69.1–71.2)	39.3 (65.6–73.0)	73.7 (66.7–80.7)
Has diabetes (%)	25.9 (24.9–26.9)	35.9 (32.1–40.0)	27.0 (19.9–34.0)
Weakness (%)	11.7 (11.0–12.4)	22.0 (18.7–25.3)	23.7 (16.9–30.4)
Slow gait speed (%)	59.0 (57.9–60.1)	84.7 (81.8–87.6)	84.2 (78.4–90.0)
Functional disability (%)	15.2 (14.4–16.0)	20.8 (17.6–24.1)	36.8 (29.2–44.5)

CICD, cognitive impairment consistent with dementia; MVPA, moderate-to-vigorous physical activity.

The weighted results for the associations of CICD diagnosis status on weakness, slowness, and functional disability are presented in Table 3. The estimates for each of the crude models attenuated after adding covariates. Persons with undiagnosed CICD had 1.37 (CI: 1.04–1.80) greater odds for weakness; however, no significant association was observed between those with diagnosed CICD and weakness (odds ratio: 1.04; CI: 0.65–1.67). Older adults with undiagnosed CICD had 2.02 (CI: 1.39–2.94) greater odds for slow gait speed; persons with diagnosed CICD had 2.29 (CI: 1.32–3.97) greater odds. Further, individuals with diagnosed CICD had 1.85 (CI: 1.19–2.90) greater odds for functional disability, whereas no significant association was found between those with undiagnosed CICD and functional disability (odds ratio: 0.80; CI: 0.61–1.06).

Table 3.

Weighted results for the associations of diagnosed cognitive impairment consistent with dementia status on weakness, slowness, and functional disability

	Weakness		Slowness		Functional disability
	Odds ratio	95% CI	Odds ratio	95% CI	Odds ratio	95% CI
Crude models ¥
Undiagnosed CICD	2.54	2.00–3.23	3.87	2.91–5.15	1.50	1.17–1.92
Diagnosed CICD	2.31	1.50–3.58	4.24	2.50–7.20	3.48	2.38–5.10
Partially-adjusted models ¥
Undiagnosed CICD	1.39	1.06–1.83	2.01	1.40–2.88	0.81	0.61–1.07
Diagnosed CICD	1.05	0.65–1.69	2.43	1.41–4.20	1.86	1.18–2.94
Fully-adjusted models ¥
Undiagnosed CICD	1.37	1.04–1.80	2.02	1.39–2.94	0.80	0.61–1.06
Diagnosed CICD	1.04	0.65–1.67	2.29	1.32–3.97	1.85	1.19–2.90

^¥Reference group: cognitively intact. The partially-adjusted models included age, sex, race, educational achievement, obesity status, self-rated health, cigarette smoking status, depressive status, arthritis status, and diabetes status as covariates. The fully-adjusted models included age, sex, race, educational achievement, obesity status, self-rated health, cigarette smoking status, depressive status, arthritis status, diabetes status, and moderate-to-vigorous physical activity participation as covariates. CI, confidence interval; CICD, cognitive impairment consistent with dementia.

Supplementary Table 1 shows the weighted results for the associations of CICD diagnosis status on weakness, slowness, and functional disability with participants reporting a dementia-related diagnosis but no CICD excluded. The findings from these supplementary models were consistent with those from our primary analyses. Moreover, the fully-adjusted, weighted results for the associations of diagnosed CICD status on weakness, slowness, and functional disability by sex are shown in Supplementary Table 2, while the same weighted results stratified by identified race are presented in Supplementary Table 3. The sex- and race-stratified results revealed similar trends to the principal findings of our study. Females with undiagnosed CICD had 1.43 (CI: 1.03–1.99) greater odds for weakness, but no significant association between undiagnosed CICD and weakness was observed in males (odds ratio: 1.30; CI: 0.79–2.13).

DISCUSSION

The principal findings of this investigation revealed that older adults living with undiagnosed and diagnosed CICD had greater odds for a physical functioning deficit. Specifically, persons with undiagnosed CICD had 37% greater odds for weakness. Individuals with undiagnosed CICD had 102% greater odds for slow gait speed, while those with diagnosed CICD had 129% greater odds. Moreover, older adults with diagnosed CICD had 85% greater odds for functional disability. Our findings for how CICD diagnosis status was differentially associated with a deficit in individual characteristics of physical functioning parallels with the severity of each physical function disablement stage.

Muscle dysfunction, as often determined by low HGS, signifies the initial stages of disablement [8, 9]. Low HGS is also associated with cognitive impairment [44, 45]. The hand dexterity and coordination involved in completing a maximal voluntary grip force task is mediated by the same neural systems that factor into cognitive function [7]. Declines in HGS and cognitive function are bidirectionally associated [17], and could be explained by shared body systems. For example, the changes in oxidative stress and blood inflammatory markers observed when losses of muscle strength occur are related to onset cognitive dysfunction [46, 47], and the age-related deterioration of the nervous system that is related to cognitive declines can be detected with physical tasks which require motor control such as HGS [7]. Considering that muscle dysfunction represents the onset of the disabling process and is linked to cognitive impairment, this may help to elucidate why we found that persons with undiagnosed CICD had greater odds for weakness.

Similarly, the global function, verbal memory, and executive function that contributes to overall cognitive functioning is associated with slow gait speed [48]. Slow walking speed in relation to cognitive dysfunction is explained by several mechanisms such as atrophy to regions of the brain, and changes in white and gray brain matter volumes [12]. Motoric Cognitive Risk Syndrome [49], which includes reported cognitive complaints and slow gait speed, is linked to dementia, thereby suggesting that slow gait speed should be considered as a factor for examining cognitive function in certain cases. Poor physical performance, as determined by slow gait speed, also signifies a more advanced stage in the disabling process that may require increased healthcare utilization (e.g., physical therapy, primary care) [8, 50].

The burden of functional disability is heavy because it is considered the end-stage of the disabling process [8, 51]. Diminishing fine and gross motor abilities that are in part necessary for completing basic activities of daily living are predictive of executive dysfunction and dementia [52]. Persons living with a functional disability could be receiving daily healthcare services at-home or in assisted living facilities [53], which in turn, increases overall patient-provider exposure and subsequently reduces undiagnosed dementia risk. The higher level of healthcare that persons with a functional disability are likely to be receiving explains why we found only persons with diagnosed CICD had greater odds for functional disability, but no significant association for undiagnosed CICD.

Our sex- and race-stratified supplementary findings show similar trends to that of our principal results. Older women with lower performance on physical tasks such as HGS may benefit from additional cognitive testing as needed, particularly because women generally perform better on certain memory tests that could be used to assess cognitive function [54]. Lower sample sizes for not-White participants may explain some of the non-significant race stratified findings in our study. Persons identifying as Black tend to have more dementia risk factors than those identifying as White, despite a lower diagnosed dementia prevalence [55]. Equitable screening practices and comprehensive cognitive assessments remain important actions in reducing dementia related diagnosis and treatment. Moreover, moderate-to-vigorous physical activity participation may factor into the link between physical tasks and cognitive function because physical activity helps to preserve both physical and cognitive health [56]. Physical activity participation should continue to be encouraged for overall health in older adults.

There are many older adults living with an undiagnosed CICD but no related diagnosis [20]. The prevalence of such persons could be influenced by the frequency and type of healthcare utilization [27]. While older adults who are not routinely visiting their healthcare provider are at greater risk for undiagnosed health conditions including dementia [57], screening for Alzheimer’s disease and related dementia may still be lacking [5]. Therefore, when other health conditions that are bidirectionally associated with dementia are observed, such as low physical function, dementia screening could be considered for mitigating undiagnosed dementia, especially when weakness or slow gait speed are present.

This work has several strengths. The HRS is a population-based study that provides nationally-representative data. Our findings support the use of simple physical measures for helping to reduce undiagnosed CICD when appropriate. Undiagnosed dementia is a significant public health concern that may elevate as the population ages. Future research should more precisely examine the role of physical measures for additional cognitive screening in older adults. Other measures of physical function (e.g., timed-up-and-go) and instrumental activities of daily living may reveal insights. New muscle function measures with electronic handgrip dynamometry and accelerometry is also a promising future research implication for determining an association with cognitive dysfunction and reducing undiagnosed CICD [58].

Although our work has strengths, some limitations should be noted. The CICD designation from the TICS does not confirm that a person has dementia but instead implies that a comprehensive cognitive examination from a relevant healthcare provider is needed. Misclassification for how we defined undiagnosed and diagnosed CICD may have occurred because we utilized the modified TICS and self-reported dementia-related diagnosis measures from the HRS. Proxy respondents were not considered for our study because they did not have information for measures included in our study. Given that persons with a severe cognitive impairment likely had a physical function deficit, our results are underestimated. Residual confounding may have also influenced the results. Criteria used to diagnose dementia-related conditions in clinical settings and population-based studies may differ and this has implications for how cognitive impairment is operationalized [59]. We analysed the last available wave where in participants had information for our explanatory and response variables, so longitudinal trends were not considered. Specificity in race lacked for those categorized as other race in the HRS. Although self-report information from HRS participants is generally reliable and valid [60], self-report biases may have factored into our findings. While the concordance between the TICS used in the HRS and a detailed neuropsychiatric assessment is about 74% [31], other sensitivity, specificity, and accuracy metrics may differ [61].

Conclusions

This study found that older adults living with undiagnosed or diagnosed CICD were at greater odds for a physical functioning deficiency. Perhaps most importantly, persons with undiagnosed CICD had greater odds for weakness and slowness. Given that low HGS and slow gait speed are bidirectionally associated with cognitive impairment, dementia screening could be especially warranted when weakness and slowness are detected. Observing the presence of health conditions that are linked to, and are perhaps biomarkers of dementia, may prompt assessments of cognitive function and dementia screening during routine healthcare visits. Increasing dementia screening when possible will help to reduce undiagnosed conditions, allow for appropriate referrals, and provide more timely interventions. Such referrals will aid in dementia-related treatment, and in turn, help older adults live longer and more independently.