Mild Neurocognitive Disorder, Social Engagement and Falls Among Older Primary Care Patients

Abstract

Objectives:

1 To estimate the association between social engagement (SE) and falls; 2) To examine the relationship between mild neurocognitive disorder (MNCD) and falls by different levels of SE.

Design:

We performed a secondary data analysis using prospective cohort study design

Setting:

Primary Care

Participants:

425 older adult primary care patients at risk for mobility decline. As previously reported, at baseline, 42% of participants exhibit MNCD.

Main Outcome Measures:

The outcome variable was number of falls during two years of follow-up. Exposure variables at baseline included: 1) MNCD identified using a cut-off of 1.5 SD below the age-adjusted mean on at least two measures within a cognitive performance battery; and 2) SE, which was assessed using the social component of the Late-Life Function and Disability Instrument. High SE was defined as having a score ≥ median value (≥ 49 out of 100). All models were adjusted for age, sex, education, marital status, comorbidities, and pain status;

Results:

Over two years of follow-up, 48% of participants fell at least once. MNCD was associated with a higher rate of falls, adjusting for the covariates (Incidence Rate Ratio=1.6, 95% CI:1.1-2.3). There was no significant association between MNCD and the rate of falls among people with high SE. In participants with low SE (having a score less than 49.5 out 100), MNCD was associated with a higher rate of falls as compared to participants with no neurocognitive disorder (No-NCD).

Conclusions:

Among participants with low SE, MNCD was associated with a higher rate of falls, but not among participants with high SE. The findings suggest that high SE may be protective against falls among older primary care patients with MNCD.

Among adults 65 years or older, falls occur with high frequency and increase the risk for disability, hospitalization, premature nursing-home admission, and mortality.^1,2 Rates of falls are even higher among those identified as being at risk for mobility decline.³ In addition, cognitive impairment is linked to all aspect of mobility skills (such as gait speed, balance, sit down/standing up) and falls among older adults within cross-sectional studies^4,5. Mild Neurocognitive Disorder (MNCD) identifies individuals with noticeable decline in cognition that goes beyond normal aging as defined by DSM-5.⁶ MNCD has a relatively high prevalence among community-dwelling older adults (25.2%) and negatively impacts their well-being and functional independence.^7-9 MNCD is associated with functional limitations, leading to traumatic events, and MNCD can progress to dementia.¹⁰ Morley indicated that 50% of patients with MNCD would progress to dementia in five years, but 20-30% would revert to normal. Another study shows that 47% had returned to normal 7.5 years later.¹⁰ Comorbidities including hypertension, heart diseases, and diabetes, arterial stiffness, common genetic variants linked to cardiovascular diseases (CHD), could all be causing the cognitive decline via heart and brain mechanisms. ^{10,11 12} Reversion from MCI to normal cognition was associated with specific lifestyle activities such as doing fieldwork or gardening or attending community meetings, participating in sports, and reading books or newspapers.¹³ Therefore, MNCD is a critical period potentially amenable to treatment targeting the prevention of avoidable negative health outcomes such as falls in older adults.

Since pharmacotherapeutic therapies for MNCD do not exist, alternative therapeutic approaches to MNCD are a priority for geriatric care. Therapies addressing behavioral or lifestyle modifications may be potential avenues for preventing or slowing the downward spiral of deterioration caused by cognitive impairment and falls. Social engagement (SE) is a behavioral factor that refers to the participation of older adults in different group-related activities, such as getting together with friends and family members (other than a spouse/partner), volunteering, and attending religious services. ¹⁴ Some studies indicate that social engagement (SE) is associated with a potential reduce in the rate of cognitive decline¹⁵ and positive impact on well-being and health outcomes such as disability and mortality among older adults. ^14,16 A study by Tangalos et al. indicated that better outcomes among patients with MNCD were observed among those having spouses, good financial resources, and attentive families.¹⁰ Those studies did not specifically examine the relationship between SE and falls in older adults with MNCD. However, one previous study of secondary data derived from the Boston Rehabilitative Impairment Study of the Elderly (Boston RISE)¹⁷ conducted by our group showed that high SE is associated with a lower rate of falls based on cross-sectional data¹⁸. There is an absence of data evaluating the association between SE and falls prospectively.

Building on the findings from our previous study, we hypothesized that SE is associated with the number of falls and SE modifies the association between MNCD and the risk of falls prospectively. Therefore, the objectives of this study were 1) to examine the association between SE at baseline and falls after two years of follow-up; 2) to investigate the association between MNCD and falls based on SE status (high vs. low).

METHODS

We conducted a secondary analysis of data collected from December 2009 to January 2014 within the Boston Rehabilitative Impairment Study of the Elderly (Boston RISE, R01AG032052). The methods of Boston RISE have been published and previous studies addressed different scientific aims.¹⁷ Data were collected among primary care patients at risk for mobility decline such as self-reported difficulty with walking half a mile (6 blocks) or climbing one flight of stairs (10 steps) over two years of follow up and continued with certain outcomes, including falls for two additional years. The analytical sample consisted of 425 individuals with a mean age of 76.6 years, of which 68% were female. The inclusion criteria of this study were patients living in the community, aged 65 or older, with English proficiency and at risk of physical function decline (self-reported difficulty walking a half mile or climbing a flight of stairs). Exclusion criteria were patients with a terminal illness, severe cognitive impairment score (MMSE) <18, Short Physical Performance Battery score (SPPB) <4, and severe visual impairment. Other exclusion criteria included: amputation of a lower extremity, use of oxygen supplementation, uncontrolled hypertension (a primary care provider determined primarily before referring their patient to the study and also if the BP was 180 systolic or greater and/or 110 diastolic or greater), heart attack, major surgery in the last six months, planned major surgery, and a plan to move away from the Boston area within two years.

Primary Outcome

The outcome of this analysis was the number of falls per year. Falls were defined as unintentionally coming to rest on the ground or lower surface or floor.¹⁹ Falls were measured based on the question, “How many times have you fallen to the ground during the last 3 months?”. These data were collected every three months during two years of follow-up after a baseline visit, then converted into the rate of falls per year. This approach to the measurement of falls has proven to have good sensitivity and specificity which also included individuals with cognitive issue.^19,20 In the first and second years, 97.2% and 95.0% of fall responses were completed, respectively.⁴

Key Measures

Social Engagement (SE) as the independent variable was assessed using the social role domain of the disability component of the Late-Life Function and Disability Instrument (LLFDI).^21,22 SE consists of nine Likert scale questions such as “how often do you…”: (1) keep in touch with others; 2) visit friends and family in their homes; 3) provide care or assistance to others; 4) volunteer; 5) take part in active recreation; 6) travel out of town; 7) invite people into home; 8) go out with others to public places; and, 9) take part in organized social activities. The total score of SE was converted to a scale of 0 -- 100. High SE was defined as having a score of at least 49.5 as the median value of the SE total score vs. low SE (having a score less than 49.5 out 100).¹⁸

Cognition was measured using a comprehensive cognitive battery performed at baseline to evaluate attention, executive function, learning, and memory published in the previous study²³. Briefly, the cognitive tests included: 1) Trail Making Test (TMT) parts A (TMT-A) and B (TMT-B) were used to measure processing speed, sequencing, cognitive flexibility, and executive function.²⁴ Completion time in seconds was used to evaluate TMT. 2) Digit Symbol Substitution Test (DSST) from the Wechsler Adult Intelligence scale was used to measure processing speed, attention, and executive function ²⁵. The participants were asked to copy symbols to match the numbers, and the number of correct digit-symbol pairs was recorded, producing a score.²⁵ 3) Hopkins Verbal Learning Test (HVLT-R) was used to measure verbal memory and learning. The test consisted of 12 words of concrete objects that were read aloud by the examiner across three learning trials. The total score was based on three measures of interest, including Total Recall of words correctly recalled on trials 1-3, Delayed Recall, the number of words correctly recalled after 20-25 minutes, and Recognition Discrimination indices (Recognition), the number of true responses minus the number of false recognized on a subsequent recognition task ^26,27. The raw scores of these tests were converted into age-adjusted standardized scores (z-scores) using data from healthy, age-matched peers. ^23,28 We used a cut-off of 1.5 standard deviations (SD) below the age-adjusted mean of the cognitive performance score to identify impaired performance on a given cognitive measure.²³

Mild Neurocognitive Disorder (MNCD) is defined by a noticeable decline in cognition that goes beyond normal aging defined by DSM-5.⁶ Although MNCD cannot be diagnosed without a documented decline in performance,⁶ using cut-off points is advantageous in determining lower cognitive performance. For the purposes of this study, we defined MNCD using a previously published approach by impairment in at least two measures within the neuropsychological test battery (TMT-A, TMT-B, DSST, HVLT-R Total Recall, HVLT-R Delayed Recall, and HVLT-R Recognition Discrimination).²³ According to DSM-5, participants were classified as either cognitively normal (No-NCD) or MNCD.⁶

Covariates included age measured at baseline as a continuous variable, gender (male vs. female), race (white vs. non-white), and education measured in three categories: less than or equal to 12^th grade or General Educational Development certificate; undergraduate, vocational or technical degree; and graduate or professional degree. The total score of pain was measured as a continuous variable using the Brief Pain Inventory.²⁹ Comorbidity was recorded using the validated and well-established Self-administered Comorbidity Questionnaire.³⁰ The list of comorbidity variables included heart disease, high blood pressure, lung disease, diabetes, ulcer or stomach disease, kidney, liver, anemia or other blood diseases, cancer, depression, osteoarthritis or degenerative arthritis, back pain, and rheumatoid arthritis.³⁰

Statistical Analyses:

First, univariate analysis was conducted to examine the distribution of demographic characteristics of the sample (Table 1). A correlation matrix was generated to evaluate collinearity among potential covariates. A series of negative binomial regressions were conducted to test how both MNCD and SE were associated with the rate of falls. Hierarchical models were conducted to understand the changes in the estimates within iterative multivariate models. Model 1 adjusted for race, gender, education, marital status, and the number of comorbidities. Model 2 was model 1 plus the addition of pain as an adjustment variable. In addition, we ascertained how the association between MNCD and the rate of falls differed between low SE vs. high SE with thresholds of SE being defined by the median value. Stratified analyses by SE status were performed to investigate the relationship between MNCD and sub-types of MNCD and falls in each stratum of high SE and low SE. Incidence Rate Ratios (IRR) and 95% confidence intervals (CI) were presented. The significance of p-values was set at <.05, and all analyses were conducted with SAS (9.4).

Table 1:

Baseline Characteristics of Sample (n=425)

Characteristics	mean (SD) or n (%)
Age [mean (SD) in years]	76.6 (7.0)
Male [n (%)]	137 (32.2)
Living with spouse/partner (yes) [n (%)]	145 (34.1)
White and Non-Hispanic [n (%)]	350 (82.4)
Education [n (%)]
- <= 12th grade or General Educational Development certificate	180 (42.3)
- Undergraduate, vocational or technical school	140 (32.9)
- Graduate or professional school	105 (24.7)
Pain score (range: 0-10) [mean (SD)]	2.5 (1.9)
Pain (score ≥ 4) [n (%)]	97 (22.8)
Comorbidity (range: 0-23) [mean (SD)]	6.3 (3.3)
Comorbidity categories
<= 4	136 (32.0)
5-6	106 (25.0)
7-9	94 (22.0)
>= 10	89 (21.0)
MNCD (yes) [n (%)]	179 (42.1)
Low SE < 49.5 [n (%)]	207 (48.7)
Falls [n (%)] over two years
No fall	220 (51.8)
One fall	88 (20.7)
>=2 falls	117 (27.5)

Notes: MNCD: Mild Neurocognitive Disorder

No-NCD: without neurocognitive disorder

SE: Social Engagement

The Boston RISE study was approved by the Institutional Review Board at Partners and Spaulding Rehabilitation Network/Massachusetts General Hospital. All participants provided written informed consent.

RESULTS

The characteristics of the sample are presented in Table 1. The mean age of the participants was 76.6 years. About two-thirds were female, and a majority were non-Hispanic white (82.4%). Approximately one-third of the participants were living with their spouse or partner. More than half of the study participants had more than 12 grade/General Development Certificate or undergraduate or graduate or professional degree. The prevalence of MNCD at baseline was 42%. The median SE was 49.5 (range: 0-79) with low SE classified as below 49.5 points. More than one-fifth of the study participants suffered from pain. The average number of comorbid conditions among the study participants was 6.4 (range: 0-23) with 43% having at least seven chronic health conditions. Over 2 years of follow-up, 48.2% reported at least one fall.

SE was not associated with the rate of falls in unadjusted model (IRR= 0.75, (95% CI=0.55, 1.03) and adjusted models (Table 2). Those with MNCD had a 60% higher rate of falls than those without MNCD, adjusting for all covariates (IRR=1.60, 95% CI= 1.14, 2.25) (table 2).

Table 2:

The Association between Baseline MNCD or Social Engagement (SE) and Rate of Falls during Two Years of Follow-up

	Model 1			Model 2
	IRR	95% CI		IRR	95% CI
MNCD (Yes vs. No) (n=425)	1.71	1.20	2.42	1.60	1.14	2.25
SE (High vs. Low) (n=425)	0.78	0.56	1.08	0.83	0.60	1.15

Notes: MNCD: Mild Neurocognitive Disorder

SE: Social Engagement

Model 1: Adjusted for age, sex, race, education, marital status and comorbidities

Model 2: Model 1+ Pain status

IRR: Incidence Rate Ratios

CI: Confidence Interval

High SE ≥ 49.5 out of 100

The number of falls per 100 person-years stratified by MNCD and SE status are shown in Figure 1. Among participants with low SE and MNCD, the number of falls was much higher than those with low SE and No-NCD (89 vs. 61 falls per 100 person-years, p<.05. However, among participants with high SE, the difference in the number of falls between participants with MNCD vs. No-NCD was much smaller (61 vs. 55 falls per 100 person-years, p<.05) (Figure 1).

Figure 1: Number of Falls per 100 Person-years Stratified by MNCD and SE status.

Notes: MNCD: Mild Neurocognitive Disorder

No-NCD: without neurocognitive disorder

SE: Social Engagement

Low SE was defined as less than 49.5 score

A stratified analysis by SE status was conducted to examine the association between MNCD and the rate of falls (Table 3). The results indicated that MNCD was associated with a higher rate of falls compared to No-NCD among patients with low SE (IRR=2.27 (95% CI= 1.35-3.80). However, no association was observed among those with high SE (Table 3).

Table 3:

The Association between Baseline MNCD and Rate of Falls during Two Years of Follow-up, Stratified by Baseline SE Status

		Model 1			Model 2
		IRR	95% CI		IRR	95% CI
Among patients with low SE (n=207)
Association between MNCD status and the rate of falls		2.27	1.35	3.80	2.04	1.23	3.39
Among patients with high SE (n=218)
Association between MNCD status and the rate of falls		0.99	0.61	1.63	1.02	0.63	1.66

Notes: MNCD: Mild Neurocognitive Disorder

SE: Social Engagement

Model 1: Adjusted for age, sex, race, education, marital status and comorbiditie

Model 2: Model 1 + Pain status

IRR: Incidence Rate Ratios

CI: Confidence Interval

DISCUSSION

The main findings of our study were that 1) SE was not independently associated with a higher rate of falls among participants after two years of follow-up; and 2) the relationship between MNCD and falls differed by levels of SE, where among participants with low SE, MNCD status was associated with higher rates of fall but among participants with high SE at baseline, the association between MNCD and falls was mitigated.

This prospective study expanded on our prior cross-sectional analyses of Boston RISE data. The results of the study suggested that higher levels of SE may be protective against prospective fall risk for primary care patients with MNCD. While this has important implications for the benefits that social engagement may provide, future studies need to examine the change in SE over time and understand how that influences the risk of falls more remotely.

An important aspect of SE to consider includes the types of SE (the quantity and quality of social engagement), which may play a role in the relationship with falls. The previous studies conducted by Quach et al. and Trevisan et al. (which was conducted with community-dwelling older adults) showed that a larger quantity of social engagement was protective against falls regardless of the type of SE. ^31,32 Because of limitations in our data, we cannot examine different types of SE (quantity versus quality of SE). However, this is an important consideration for future studies of these relationships.

We found that a high SE level was protective for the association between MNCD and falls. The influence of SE on the association between MNCD and the rate of falls might go through a mechanism of increased physical activity caused by involvement in social engagement activities ³³. According to the Social Cognitive Model, people participating in social activities receive support and learn best practices and behaviors from good models. ^34-36 This finding can be explained by the fact that SE helps older adults, especially those with MNCD, access more social support and information, which may help them recognize their risk of falls. In addition, high SE is associated with well-being and may help people feel more confident in their own abilities.³⁷

Self-efficacy and a sense of control are recognized as helping older people cope with cognitive impairment, pain, and mobility decline.³⁸ Previous studies indicated that SE might reduce the rate of cognitive decline in older age.^15,39 In addition, Mendes de Leon et al. stated that SE could promote resilience via social support from an individual’s social network, especially when someone suffers from diseases that show severe progress over time.³⁸ Our findings also highlighted two important treatment concerns. First, the findings highlighted the potential adverse role that social isolation (i.e. with COVID-19) may have played in the lives of older adults, and second, the results highlighted the beneficial role that treatments targeting enrichment of social networks and activity might play. Boston RISE was not designed to address either concern, but the results pointed to fertile lines of inquiry for future research. More specifically, future studies should consider the pattern of social support, resilience, and exercise-related data among patients with MNCD and explore the benefits of interventions expanding social networks on falls in this population.

Study limitations

This study had several limitations. First, this was not a population-based study and the results may not generalize to primary care patients in other regions. Second, our measures of SE and falls were based on patients’ self-reports, which could be affected by recall bias. Third, we evaluated SE and MNCD at baseline and did not evaluate the impact of prospective changes in either factor. Because of the nature of the research protocol, causal statements about relationships cannot be offered. Last, the sample size of Boston RISE was relatively small and did not allow for more nuanced evaluation, such as evaluating the influence of SE on different sub-types of MNCD.

This study's major strength included its prospective design, its unique inclusion of variables addressing cognition, SE and falls, and its focus on a population of patients for whom the investigated relationships are highly relevant. Our MNCD and cognitive measurements were based on standardized cognitive batteries that have been validated in previous studies.¹⁷ Missing data and dropout in this study was less than 5%, limiting bias associated with sample attrition.

Conclusions

For a sample of older primary care patients at risk for functional decline and falls, this study demonstrated that higher levels of SE among those with MNCD was protective against falls risk over two years of follow up. These associations should be confirmed through prospective analyses of larger, representative data that are adequately powered to study these associations more closely.

Abbreviations

Boston RISE

Boston Rehabilitative Impairment Study of the Elderly

BPI

Brief Pain Inventory

CI

Confidence Intervals

DSST

Digit Symbol Substitution Test

GED

General Educational Development

GEE

Generalized Estimating Equation

HVLT-R

Hopkins Verbal Learning Test, Revised

ICF

The International Classification of Functioning, Disability and Health Model

LLFDI

Late Life Function and Disability Instrument

MNCD

Mild Neurocognitive Disorder

MMSE

Mini-Mental State Examination

RR

Rate Ratios

SD

Standard Deviation

SE

Social Engagement

PHQ-9

Patient Health Questionnaire −9

SPPB

Short Physical Performance Battery