A Decade of Decline in Serious Cognitive Problems Among Older Americans: A Population-Based Study of 5.4 Million Respondents

Abstract

Background:

Numerous studies suggest the prevalence of dementia has decreased over the past several decades in Western countries. Less is known about whether these trends differ by gender or age cohort, and if generational differences in educational attainment explain these trajectories.

Objective:

1) To detect temporal trends in the age-sex-race adjusted prevalence of serious cognitive problems among Americans aged 65+; 2) To establish if these temporal trends differ by gender and age cohort; 3) To examine if these temporal trends are attenuated by generational differences in educational attainment.

Methods:

Secondary analysis of 10 years of annual nationally representative data from the American Community Survey with 5.4 million community-dwelling and institutionalized older adults aged 65+. The question on serious cognitive problems was, “Because of a physical, mental, or emotional condition, does this person have serious difficulty concentrating, remembering, or making decisions?”

Results:

The prevalence of serious cognitive problems in the US population aged 65 and older declined from 12.2% to 10.0% between 2008 and 2017. Had the prevalence remained at the 2008 levels, there would have been an additional 1.13 million older Americans with serious cognitive problems in 2017. The decline in memory problems across the decade was higher for women (23%) than for men (13%). Adjusting for education substantially attenuated the decline.

Conclusion:

Between 2008 and 2017, the prevalence of serious cognitive impairment among older Americans declined significantly, although these declines were partially attributable to generational differences in educational attainment.

Keywords

Cognitive impairment dementia memory temporal trends

INTRODUCTION

More than 47 million individuals worldwide have dementia, and that number is projected to increase to 131.5 million in 2050 as the size of the older adult population grows [1]. However, a recent systematic review indicated a decline in the incidence and prevalence of dementia in developed countries [2]. If these trends are confirmed in other studies using large population-based samples and if these trends continue into the coming decades, future projections of the number of people living with dementia will need to be substantially modified.

Numerous population-based studies have already found significant declines in age-specific incidence of dementia. Between 1988 and 2008, Britain reported a 20% reduction in the incidence of dementia [3]. A similar trend has been observed in France, where a 30% reduction in the incidence of dementia occurred between 1988/1989 and 1999/2000 [4]. A study from the Netherlands found a 25% decrease in dementia incidence between 1990–2000, as well as a significant increase in total brain volume in study participants in the 2000 cohort in comparison to the 1990 cohort [5].

Community-based studies from the United States (US) have also shown significant declining trends in the age-adjusted incidence of dementia [6, 7]. An analysis based on the Framingham Heart Study in Massachusetts reported a 44% decrease in dementia incidence between 1977 and 2008 [6], while a study based in Rochester, Minnesota found a 30% decline in dementia incidence between 1985 and 1994 [7]. Two cohort studies based out of Manhattan and Indianapolis have shown declines in dementia incidence of 53% and 39% respectively (1992–1999; 1992–2001) [8, 9]. Further, a study investigating short-term trends in dementia over a 5-year period in Germany (2006–2010) found a 3.3% and 5% decline in dementia incidence for men and women, respectively [10]. Additionally, two provincially representative Canadian studies based on universal health insurance records showed significant declines of 7.4% in Ontario between 2002 and 2013, and 11.1% in Saskatchewan between 2005/2006 and 2012/2013 in age-specific incidence of dementia [11, 12].

In contrast, there are some geographically limited studies indicating a stable or increasing incidence of dementia. A population-based study of older adults in three Chicago neighborhoods found no significant change in the incident of Alzheimer’s disease (AD) over an 11-year period, suggesting that the risk of AD remains constant [13]. A recent study based in Amsterdam found an increase in incidence of persistent cognitive decline over a 23-year study period (1992–2015) [14].

Changes have also been observed in dementia prevalence [2]. Although trends in prevalence are more variable than incidence, a recent systematic review highlighted numerous population-based studies from the US that have shown a declining trend in dementia prevalence [2]. A nationally representative US study found a 42% relative decline in dementia prevalence over 17 years (1982–1999; 3% per annum) [15]. Further, a more recent study using the nationally representative Health and Retirement Study (HRS) found a 24% relative decline in dementia prevalence between 2000 and 2012 (2% per annum) [16]. An additional study using the same HRS dataset with different methodologies found a less rapid decline in age-specific dementia prevalence between 2000 and 2012, observing a decline of 12.6%, or 1% per year during the study period [17]. The latter study’s methodology included additional modeling for mortality, timing of interviews, and changes in the collection of proxy interviews that occurred during the study period in the HRS [17]. This estimated decline is similar to one observed in a population-based study in England, which found dementia prevalence declined from 8.3% in 1991 to 6.5% in 2011, corresponding to a 12% decline in the projected number of individuals with dementia per decade [18]. Another study using the HRS dataset found a 30% relative decline in dementia prevalence over a decade (3% per annum; 1993–2002) [7]. Additionally, a study of short-term dementia trends in the US found a 1.7% decline per year in the age-specific prevalence of dementia per year during their 5-year study period (2011–2015) [19]. Based on these studies, the reported decline in dementia prevalence in Western countries is between 12%–30% per decade.

However, a few cohort studies have reported a stable or increasing prevalence of dementia [2]. Two Swedish studies found that dementia prevalence has remained stable over time [20, 21]. One study found that age- and sex- specific dementia prevalence was virtually identical between two population birth cohorts; however, this study was restricted to an older population of individuals 75 years or older [20]. The other study found no significant difference in prevalence between birth cohorts; however, this study only looked at two ages (70 & 75 years) across the two points of time used in the study [21]. Other studies that reported increases in prevalence frequently relied on health records data, which means their findings may be influenced by increasing trends in diagnosis, rather than a true increase in prevalence [2].

The predominant trend of declining incidence and prevalence of dementia seen in Western countries is encouraging. The cause of this declining trend is yet to be fully explained; however, there are several factors that could be contributing to this decrease. Increasing levels of education over the past several decades has been cited as a primary explanation for the decline in dementia rates [16]. Between 1940 to 1980, the prevalence of the adult population with a high school diploma increased from 24.5% to 66.5% [22]. Higher levels of education are associated with a decreased risk of dementia, potentially due to the strengthening of an individual’s cognitive reserve, allowing them to better compensate for the impacts of dementia pathology [23]. An alternative explanation for the link between education and cognitive decline relates to the fact that education level is inversely associated with a number of potentially modifiable risk factors for dementia including hypertension, obesity, hearing loss, smoking, depression, physical inactivity, social isolation, and diabetes [24].

Many of the previous US studies that investigated the temporal trends of dementia have been restricted by geographic area or shorter study periods. The objectives of the present study were to use 10 consecutive years (2008–2017) of a very large series of cross-sectional nationally representative surveys of community dwelling and institutionalized American older adults:

To detect temporal trends in the age-sex-race adjusted prevalence of serious cognitive problems among Americans aged 65+;

To establish if these temporal trends in serious cognitive problems differ by gender and age cohort;

To examine if these temporal trends are attenuated by generational differences in cohort educational attainment.

METHODS

Sample

As has been described elsewhere [25], this study used data from 10 consecutive cross-sectional waves (2008–2017) of the American Community Survey (ACS). The ACS is an annual survey which replaced the long form of the decennial census. The ACS uses a representative sample of the US population including those living in the community and institutions, such as nursing homes. It is important to note that the ACS is a series of cross-sectional surveys rather than a longitudinal cohort study.

The annual response rates ranged from 89.9% to 98.0% among community dwellings older adults and from 94.7% to 98% among institutionalized older adults. The sample size of those aged 65 and older ranged from 467,736 in 2008 to 610,327 in 2017, resulting in a total sample size of 5,405,135 respondents.

Measures

The same questions were used in the ACS throughout all 10 years of data collection. The definition of serious cognitive problems was based upon the question, “Because of a physical, mental, or emotional condition, does this person have serious difficulty concentrating, remembering, or making decisions?” (yes/no). Year of data collection (2008–2017) was entered as a portion of a decade (e.g., 2008 was entered as 0.1, 2009 as 0.2 and 2017 as 1.0). Year of age was entered into the logistic regression analyses categorically, rather than as a continuous variable, due to the non-linear association between age and dementia. Age was top coded at 97 years. For the prevalence tables, age was presented in 5-year groups (65–69, 70–74, 75–79, 80–84, 85–89, 90–94, 95+) as well as for the whole sample (age 65+). Race/ethnicity was categorized into Hispanic (of any race), non-Hispanic white, non-Hispanic black, non-Hispanic American Indian Alaskan Native, non-Hispanic Asian-American, and Native Hawaiian Pacific Islander. Sex was based on self-report or proxy report of male or female. Education was entered categorically as no schooling, some schooling but less than grade 3 completed, then each year of education completed from grade 3 through grade 11, grade 12 but no diploma, grade 12 regular high school diploma, General Education Diploma (GED), some college but less than 1 year, 1 or more years of college but no diploma, associate’s degree, bachelor’s degree, master’s degree, professional degree, and doctorate degree.

Statistical analysis

Prevalence data were generated for both genders together and for each gender by age cohorts (all respondents [65 +] and 65–69, 70–74, 75–79, 80–84, 85–89, 90–94, 95+) for each year of data collected (e.g., 2008, 2009,. . . 2017). For both genders together, three separate logistic regression models were conducted for each age cohort (65+, 65–69, 70–74, 75–79, 80–84, 85–89, 90+). For each age cohort, three models were conducted to determine the odds of serious memory problem per decade. The models included the following variables in the analyses, respectively: 1) year of data collection (recorded as part of a decade, e.g., 0.1, 0.2 for 2008 and 2009, respectively); 2) year of data collection, age in years, race (and sex in the analyses which combined men and women); and 3) year of data collection, age in years, race (sex in the analyses which combined men and women), and education. Three separate logistic regression models were also conducted for each gender and age cohort and all age cohorts together that included the above variables. All variables were entered as categorical variables.

All data were weighted to adjust for non-response and differential selection probabilities. All missing data in the ACS were imputed in the US Census Bureau using a sophisticated hot-decking technique [26]. All sample sizes are represented in their unweighted form. This study used a weighting variable created by the US Census Bureau based on the inverses of selection probabilities [27]. All analyses were conducted using IBM SPSS 25.

RESULTS

Table 1 provides the total prevalence of serious cognitive decline among American older adults 65 + between 2008 and 2017, as well as separate analyses by gender and age cohort (65–69, 70–74, 75–79, 80–84, 85–89, 90–94, 95+). All ages and genders showed a substantial decline in the prevalence of serious cognitive problems over the decade, with the exception of men age 65–69 in which some fluctuation was observed. In the total population of older adults age 65 and over, the prevalence of serious cognitive problems declined from 12.2% in 2008 to 10.0% in 2017. The prevalence among men declined from 10.2% in 2008 to 8.8% in 2017, while the prevalence among women declined from 13.6% in 2008 to 10.9% in 2017.

Table 1

Temporal trends in the prevalence of cognitive problems by decade (2008–2017), gender and age cohort. Source: American Community Surveys annual data (n = 5,405,135)

		Age 65–69			Age 70–74			Age 75–79			Ages 80–84
	Unweighted Sample	Total	Males	Females	Total	Males	Females	Total	Males	Females	Total	Males	Females
2008	467736	5.6	5.7	5.5	7.2	7.0	7.3	10.9	10.2	11.4	16.9	15.2	18.0
2009	479970	5.4	5.7	5.1	7.0	6.8	7.2	10.5	9.8	11.0	16.3	14.7	17.3
2010	490811	5.4	5.6	5.3	6.6	6.4	6.8	10.2	9.6	10.6	15.5	14.0	16.6
2011	524112	5.4	5.5	5.3	6.7	6.5	6.8	10.2	9.2	11.0	15.5	14.3	16.3
2012	537548	5.4	5.6	5.2	6.7	6.5	6.8	10.1	9.6	10.6	15.1	13.7	16.2
2013	544427	5.5	5.8	5.2	6.6	6.5	6.6	9.7	9.1	10.1	15.3	13.6	16.4
2014	566003	5.5	5.9	5.2	6.5	6.5	6.4	9.6	9.1	10.1	15.1	13.7	16.0
2015	583383	5.4	5.9	5.0	6.2	6.4	6.1	9.5	8.9	10.0	14.7	13.8	15.4
2016	600818	5.5	6.0	5.1	6.4	6.8	6.1	9.1	8.4	9.7	14.7	13.7	15.4
2017	610327	5.5	5.8	5.2	6.3	6.3	6.3	9.0	8.4	9.4	14.2	12.4	15.4
Unweighted Totals	5405135	1715795	809255	906540	1288672	595559	693113	963874	427926	535948	718090	295966	422124
Linear by linear association		0.174	11.83	8.38	124.72	6.93	161.99	336.15	131.20	197.39	284.96	91.69	175.94
p		0.676	0.001	0.004	< 0.001	0.008	< 0.001	< 0.001	< 0.001	< 0.001	< 0.001	< 0.001	< 0.001
		Age 85–89			Age 90–94			95 and older			Total (age 65 and above)
		Total	Males	Females	Total	Males	Females	Total	Males	Females	Total	Males	Females
2008		26.1	22.2	28.2	39.5	31.6	42.4	44.1	43.9	44.1	12.2	10.2	13.6
2009		25.1	21.9	26.7	37.5	30.5	40.1	41.7	42.9	41.4	11.7	9.9	13.0
2010		24.2	21.1	25.8	36.1	29.5	38.6	35.5	31.7	36.7	11.2	9.5	12.4
2011		24.0	20.4	26.1	35.6	28.4	38.4	33.3	24.1	36.3	11.1	9.4	12.5
2012		23.5	21.0	25.0	34.4	28.3	37.0	40.2	38.3	40.9	10.8	9.4	11.9
2013		23.2	20.6	24.6	34.8	28.5	37.5	40.0	34.3	41.9	10.7	9.3	11.8
2014		22.9	20.1	24.6	34.4	28.6	36.8	38.9	28.7	41.8	10.5	9.2	11.5
2015		22.9	19.9	24.6	34.1	29.0	36.4	39.3	34.4	41.0	10.3	9.2	11.2
2016		22.8	20.1	24.5	34.2	28.9	36.5	36.6	30.0	39.1	10.2	9.2	11.1
2017		22.6	20.2	24.1	32.9	27.9	35.1	37.3	30.0	40.4	10.0	8.8	10.9
Unweighted Totals		449276	164993	284283	250799	72441	178358	18629	4869	13760	5405135	2371009	3034126
Linear by linear association		236.51	38.94	176.10	294.00	19.89	255.18	10.02	11.83	1.26	2154.26	311.81	1905.75
p		< 0.001	< 0.001	< 0.001	< 0.001	< 0.001	< 0.001	0.002	0.001	0.262	< 0.001	< 0.001	< 0.001

In the logistic regression analyses, there was a 20% decline in the crude (i.e., unadjusted) odds of serious cognitive problems over the decade when examining both genders and all age cohorts 65 and older together (OR = 0.80; 95% CI = 0.79, 0.81) (See Table 2, column 2). Men had a 13% decline in the odds of serious cognitive decline per decade (OR = 0.87; 95% CI = 0.86, 0.89), and women had a 23% decline per decade (OR = 0.77; 95% CI = 0.76, 0.77). In the 65–69 age cohort, there was no significant decline over the decade in serious cognitive problems when both genders were combined (OR = 1.01; 95% CI = 0.98, 1.03), but men showed a significant 6% increase over the decade (OR = 1.06; 95% CI = 1.03, 1.09), and a women showed a significant 5% decrease over the decade (OR = 0.95; 95% CI = 0.92, 0.99). In all other five-year age groups, the crude odds of serious cognitive declined significantly over the decade for men (ORs range from 0.81 to 0.95), for women (ORs range from 0.79 to 0.83), and for both genders together (ORs range from 0.80 to 0.87). Further adjustments for age and race (and gender in non-sex specific analyses) resulted in very little change to the ORs (see Table 2, Column 3). For example, the age-race-sex adjusted odds of cognitive impairment over the decade for those aged 65 and older declined 16% for all older Americans, with a 19% decline for women and a 10% decline for men.

Table 2

Temporal trends in the odds of cognitive problems per year). Odds and 95% Confidence Intervals (p < 0.001 unless otherwise indicated). Source: American Community Surveys annual data 2008–2017 (unweighted n = 5,405,135)

Age Cohort (weighted sample size)	Unadjusted odds of cognitive problems per decade (2008–2017) (95% CI)	Odds of cognitive problems per decade (2008–2017), adjusted for year and age, race (includes adjustment for sex in non-gender specific analyses) (95% CI)	Odds of cognitive problems per decade (2008–2017) (95% CI), adjustment for year of age, race, (includes adjustment for sex in non gender specific analyses) and education (95% CI)
Age 65–69
Both Genders 65–69 (n = 1738036)	1.01 (0.98, 1.03), p = 0.677	0.98 (0.96, 1.01), p = 0.170	1.19 (1.17, 1.22), p≤0.001
Men 65–69 (n = 817986)	1.06 (1.03, 1.09), p = 0.001	1.04 (1.01, 1.08), p = 0.019	1.24 (1.20, 1.28), p≤0.001
Female 65–69 (n = 920050)	0.95 (0.92, 0.99), p = 0.004	0.93 (0.90, 0.96), p≤0.001	1.15 (1.11, 1.19), p≤0.001
Age 70–74
Both Genders 70–74 (n = 1284442)	0.87 (0.85, 0.89), p≤0.001	0.86 (0.84, 0.89), p≤0.001	1.04 (1.01, 1.07), p = 0.003
Men 70–74 (n = 589949)	0.95 (0.92, 0.99), p = 0.009	0.95 (0.92, 0.98), p = 0.004	1.14 (1.10, 1.18), p≤0.001
Female 70–74 (n = 694492)	0.81 (0.78, 0.84), p≤0.001	0.80 (0.77, 0.83), p≤0.001	0.96 (0.93, 0.99), p = 0.020
Age 75–79
Both Genders 75–79 (n = 952391)	0.81 (0.79, 0.82), p≤0.001	0.80 (0.78, 0.81), p≤0.001	0.91 (0.88, 0.93), p≤0.001
Men 75–79 (n = 418072)	0.81 (0.78, 0.84), p≤0.001	0.80 (0.77, 0.83), p≤0.001	0.90 (0.87, 0.93), p≤0.001
Female 75–79 (n = 534319)	0.81 (0.78, 0.83), p≤0.001	0.79 (0.77, 0.82), p≤0.001	0.91 (0.88, 0.94), p≤0.001
Age 80–84
Both Genders 80–84 (n = 711985)	0.83 (0.81, 0.84), p≤0.001	0.81 (0.79, 0.82), p≤0.001	0.88 (0.86, 0.90), p≤0.001
Men 80–84 (n = 289169)	0.84 (0.81, 0.87), p≤0.001	0.82 (0.79, 0.85), p≤0.001	0.90 (0.86, 0.93), p≤0.001
Female 80–84 (n = 422817)	0.83 (0.80, 0.85), p≤0.001	0.80 (0.78, 0.82), p≤0.001	0.86 (0.84, 0.89), p≤0.001
Age 85–89
Both Genders 85–89 (n = 449772)	0.83 (0.81, 0.85), p≤0.001	0.81 (0.79, 0.83), p≤0.001	0.86 (0.84, 0.88), p≤0.001
Men 85–89 (n = 162260)	0.88 (0.84, 0.91), p≤0.001	0.86 (0.82, 0.89), p≤0.001	0.91 (0.87, 0.95), p≤0.001
Female 85–89 (n = 287512)	0.82 (0.80, 0.84), p≤0.001	0.79 (0.77, 0.81), p≤0.001	0.84 (0.82, 0.87), p≤0.001
Age 90+
Both Genders 90+ (n = 268508)	0.80 (0.78, 0.83), p≤0.001	0.81 (0.79, 0.84), p≤0.001	0.86 (0.83, 0.88), p≤0.001
Men 90+ (n = 77311)	0.89 (0.84, 0.94), p≤0.001	0.88 (0.83, 0.94), p≤0.001	0.94 (0.88, 0.99), p = 0.027
Female 90+ (n = 191198)	0.79 (0.77, 0.82), p≤0.001	0.79 (0.76, 0.82), p≤0.001	0.83 (0.80, 0.86), p≤0.001
Total (age 65+)
Both Genders 65+ (n = 5405135)	0.80 (0.79, 0.81), p≤0.001	0.84 (0.83, 0.85), p≤0.001	0.95 (0.94, 0.95), p≤0.001
Men 65+ (n = 2371009)	0.87 (0.86, 0.89), p≤0.001	0.90 (0.88, 0.91), p≤0.001	1.01 (0.99, 1.02), p = 0.321
Female 65+ (n = 3034126)	0.77 (0.76, 0.77), p≤0.001	0.81 (0.80, 0.82), p≤0.001	0.91 (0.90, 0.92), p≤0.001

However, further adjustments for education level resulted in substantial attenuation of most analyses (see Table 2, column 4). All cohorts age 75 and older, for both sexes individually and combined, had between 6% and 17% significant declines across the decade in the odds of cognitive impairment (ORs range from 0.83 to 0.94) after adjustment for education level, race, and age. In sharp contrast, for men aged 65–74 and women aged 65–69, once education was taken into account, the odds of serious cognitive problems significantly increased over the decade (ORs range from 1.14 to 1.24). These diverging trends with the addition of education into the analysis resulted in women, and both genders combined, age 65 and older, having a significant decline in the odds of serious dementia decline over the decade (OR = 0.91 and 0.95, respectively), in contrast to men 65 and older who showed no significant change across the decade (OR = 1.01, 95% CI = 0.99, 1.02).

DISCUSSION

Findings of this nationally representative study indicate the unadjusted odds of decline in memory problems across the decade was higher for women (23%) than for men (13%) aged 65 and older. When age and race were taken into account, the odds of cognitive impairment over the decade were only modestly attenuated to a 19% decline for women and a 10% decline for men. All gender and age cohorts aged 65 and older showed significant age-race adjusted declines in the odds of serious cognitive problems over the course of the decade, with the exception of men age 65–69, who experienced a 4% increase in age-race adjusted odds over the decade.

Based on analyses of the ACS, the US population of those aged 65 and older increased 30% from almost 39 million to more than 50 million between 2008 and 2017. However, the number of Americans with cognitive impairment only increased 7% during this period from 4.728 million in 2008 to 5.064 million in 2017, due to the substantial decline in the prevalence of cognitive problems over the decade. However, if the prevalence of serious cognitive impairment had remained constant at the 2008 prevalence of 12.2%, rather than declined to the 2017 prevalence of 10.0%, an additional 1.13 million older Americans would have had serious cognitive problems in 2017.

Logistic regression analyses indicated that education played the most significant role in attenuating the odds of decline, explaining more than half of the observed decline in the prevalence of serious cognitive problems across all older adults age 65 and over, and completely attenuating the relationship in the analysis restricted to men only. Several studies have suggested that increasing levels of education can explain the decrease in age-specific incidence and prevalence of dementia [4 , 19]. Higher levels of education are associated with a decreased risk of developing dementia [23 , 29], higher cognitive ability in normal aging [29 –32], and a delayed onset of cognitive decline [33 –35]. Education is a main component of cognitive reserve, which asserts that individual differences in neural networks may result in a specific brain pathology resulting in varying clinical manifestations [36, 37]. One study showed that among individuals with a specific neural pathology, those with higher educational levels had a lower risk of dementia, indicating that educational attainment may mitigate the clinical symptoms of dementia [23]. Additionally, a study using cognitive lifestyle scores, including educational level, occupational complexity, and social engagement, showed that individuals with more active cognitive lifestyles had significantly higher neural densities compared to those with less cognitively active lifestyles [38]. Since many of the known modifiable risk factors for dementia, such as hypertension, obesity, hearing loss, smoking, depression, physical inactivity, social isolation, and diabetes are correlated with lower levels of education, the strong inverse association we see between education level and cognitive problems may be partially attributed to these unmeasured factors [24].

However, even after adjusting for education, there remains an unexplained decline in the odds of serious cognitive problems of approximately 5% per decade (OR of both genders aged 65+ = 0.95), meaning approximately one-third of the observed age-race adjusted decline in odds remains unaccounted for after education was included in the analysis. For women in particular, the decline in the odds of serious cognitive impairment could not be entirely explained by educational attainment. This result has been shown in other studies, where after adjusting for the defined educational factor, there was still a persistent decline in the rate of dementia [4 , 19].

There is a need for further research to examine the role that other factors may have on the decline in dementia, such as the reduction in hypertension, smoking, and stroke among American adults. From 1970–1990 there was a 30% decrease in hypertension prevalence [39]. In addition, between the 1960s and 1980s there was a 37% and 16% decrease in smoking rates for men and women, respectively [40]. Finally, comparisons of stroke incidence showed a 46% decline between the 1950s and 1970s [41]. However, even after adjustment for cardiovascular risk factors, several studies have continued to observe a decrease in dementia risk, indicating the presence of additional factors contributing to the decline in dementia prevalence and incidence [4 , 42].

There may be other cohort specific effects that have contributed to the decrease in dementia prevalence, such as improvements in oral health over the last several decades. Considerable research has identified oral infections such as periodontitis to be associated with cognitive impairment and increased dementia incidence and prevalence [43 –45]. A 2018 study showed that the bacteria that causes periodontitis, Porphyromonas gingivalis and its corresponding antigen, gingipain, which has neurotoxic effects, were present in the brains of AD patients [46]. Mice with oral P. gingivalis infections had resulting brain colonization by the bacteria and increases in the production Aβ₁₄₂ which is an essential component of amyloid plaques [46].

Further improvements in health and nutrition, such as the addition of folic acid into flour may be affecting the prevalence of cognitive problems. In 1998, the US and Canada implemented the addition of folic acid to cereal grains nation-wide to reduce the development of neural tube defects (NTDs) during pregnancy [47]. Along with decreasing rates of NTDs, when given to older adults (55+) folic acid has the additional benefit of lowering levels of homocysteine, which is a risk factor for cardiovascular disease [48]. Moreover, higher levels of homocysteine are associated with cognitive decline, supporting the findings of this study and other research reporting decreases in cognitive problems among older Americans [49. 50]. However, European countries that have reported declines in dementia rates such as the United Kingdom (UK), France, Germany, and the Netherlands do not fortify cereal grains with folic acid, suggesting factors other than folic acid fortification must also be at play [47].

Technology use is another factor that may play a role in cognitive decline. Frequent computer use is associated with better cognitive performance [51]. The rate of technology use among older Americans has increased. In 2016, 67% of older Americans reported using the Internet for tasks such as email or general information searching [52]. This has increased markedly since 2000, when only 12% of older adults reported internet use [52]. Additionally, technology use among older adults is associated with lower levels of loneliness, which is a risk factor for dementia [53 –55]. However, other research indicates that rates of loneliness among older Americans have remained steady. Two nationally representative surveys performed by the American Association for Retired Persons (AARP) in 2010 and 2018 found comparable levels of loneliness in Americans over the age of 70 of 25% and 24% respectively [56, 57].

Environmental factors such as the declining rates of air pollutants may also contribute to the decline in dementia prevalence. Levels of common air pollutants such as nitrogen dioxide, carbon monoxide and fine particulate matter (PM_2.5) have decreased 51%, 65%, and 43%, respectively, between 2000 and 2019 [58]. A recent systematic review suggested that higher levels of nitrogen dioxide and carbon monoxide in the air are associated with elevated dementia risk [59]. More specifically, exposure to pollutants from traffic emissions such as PM_2.5 and nitrogen dioxide can have neurodegenerative effects on older adults [60]. A recent study using a cohort of older American women found that women living in areas with PM_2.5 levels higher than the approved standard experienced accelerated global cognitive decline and an increase in all-cause dementia by 81% and 92%, respectively [61]. This result was supported by a UK study which showed increasing dementia risk with increasing levels of PM_2.5 and nitrogen dioxide in the air [62]. These two findings are further corroborated by a large population-based Canadian study, which demonstrated that closer proximity to major roadways was associated with higher dementia incidence [63]. This evidence points to the possibility that decreasing levels of air pollutants may be contributing to the reported decline in cognitive problems among older Americans.

Another environmental factor which may be contributing to the decrease in dementia is the removal of lead from gasoline. The phase out of leaded gasoline began in the US in the 1970s, reducing the population’s exposure to environmental lead. The reduction of environmental lead exposure is important to consider, as lead is a known neurotoxin. High levels of lead exposure are associated with declines in cognitive functioning [64 –66]. The removal of lead from gasoline caused a dramatic lowering of blood- and bone-lead levels. A study comparing two waves of the National Health and Nutrition Examination Surveys (NHANES) found that between 1976 and 1988, blood-lead levels decreased 78% among the American population [67]. Additionally, the VA Normative Ageing survey, which measured bone-lead levels (a maker of lifetime lead exposure) during the 1990s in older men, found that those 75 years and older had bone lead levels 20% higher than the 65 to 69 cohort, while those 60 years and younger had half the level [65]. The youngest cohort in that study recently became the oldest in our current population, bringing with them a lower lifetime lead exposure and potentially fewer accompanying neurological effects. It is plausible that the decrease in lead exposure in the 1970s contributed to the observed decline in the prevalence of serious cognitive problems over the past several decades, but it is evident that more research is still needed.

Limitations

There are several limitations of this study that must be considered. The question of cognitive problems was based upon self or proxy report and not upon review of medical records or medical examination. The question “Because of a physical, mental, or emotional condition, does this person have serious difficulty concentrating, remembering, or making decisions?” is a marker of cognitive impairment rather than a medical diagnosis of dementia. Although this is not a diagnosis, this question has been determined to be reliable and valid by the US Census Bureau [68]. On a positive note, we anticipate that it would be unlikely that the interpretation of the question changed substantially over the decade between 2008 and 2018, but we could find no research that had investigated changes in interpretations of questions over time to support this belief. In contrast, researchers have noted that health records data and mortality data may be influenced by increasing trends in diagnosis, which makes it challenging to assess temporal changes in true prevalence [2]. Unfortunately, the ACS did not include data on many of the factors associated with dementia including health behaviors (e.g., smoking, obesity, physical inactivity, excessive alcohol consumption), health conditions (e.g., diabetes, hypertension, depression, traumatic brain injury), social isolation, and air pollution, and so we could not examine the degree to which temporal changes in these risk factors collectively may have contributed to the observed decline in serious cognitive problems [24, 69]. Future research would benefit from inclusion of these factors in prospective studies.

Despite these limitations, this study provides current information on temporal trends in cognitive problems over a 10-year period using a large nationally representative sample of approximately half a million community dwelling and institutionalized American older adults per year. Additionally, the sample size provided enough power to examine the prevalence and odds of cognitive problems per year across gender and 5-year age cohorts.

Even with the positive findings of the dramatic decline in the prevalence of cognitive problems among older Americans, it is important to consider that the unparalleled increase in the size of the older population will still lead to an increase in the absolute number of adults experiencing cognitive impairment. Furthermore, the youngest men studied (those aged 65–69) were the only group to show no decline in crude odds of serious cognitive problems over the decade, and men 65–74 actually had significantly elevated odds of serious cognitive problems when education was taken into account. The reason for this anomaly is unclear, but as large cohorts of baby boomers transition into the 65 + age group, it is essential to closely monitor trends to determine if the younger baby boomer men (those born after 1952) also show similar vulnerability. If that is the case, the salutogenic trend in decreasing prevalence of serious cognitive impairments among older Americans, observed in this and other previous research, may stall. Future research is needed to better understand how cohort and other group differences may influence vulnerability to serious cognitive decline.

Footnotes

ACKNOWLEDGMENTS

The authors would like to thank Andie MacNeil for her assistance with manuscript preparation.

Authors’ disclosures available online ().

References

Prince

, Comas-Herrera

, Knapp

, Guerchet

and Karagiannidou

(2016) World Alzheimer report 2016: Improving healthcare for people living with dementia: Coverage, quality and costs now and in the future. https://www.alzint.org/u/WorldAlzheimerReport2016.pdf

Stephan

BCM

, Birdi

, Tang

EYH

, Cosco

, Donini

, Licher

, Ikram

, Sievro

, Robinson

(2018) Secular trends in dementiaprevalence and incidence worldwide: A systematic review. JAlzheimers Dis 66, 653–680.

Matthews

, Stephan

BCM

, Robinson

, Jagger

, Barnes

, Arthur

, Brayne

(2016) A two decade dementia incidence comparison from the cognitive function and ageing studies I and II. Nat Commun 7, 11398.

Grasset

, Brayne

, Joly

, Jacqmin-Gadda

, Peres

, Foubert-Samier

, Dartigues

, Helmer

(2016) Trends in dementia incidence: Evolution over a 10-year period in France. Alzheimers Dement 12, 272–280.

Schrijvers

EMC

, Verhaaren

BFJ

, Koudstaal

, Hofman

, Ikran

, Breteler

MMB

(2012) Is dementia incidence declining? Trends in dementia incidence since 1990 in the Rotterdam study. Neurology 78, 1456–1463.

Satizabal

, Beiser

, Chouraki

, Chêne

, Dufouil

, Seshadri

(2016) Incidence of dementia over three decades in theFramingham Heart Study. N Engl J Med 374, 523–532.

Rocca

, Petersen

, Knopman

, Hebert

, Evans

, Hall

, Gao

, Unverzagt

, Langa

, Larson

, White

(2011) Trends in the incidence and prevalence of Alzheimer’s disease, dementia, and cognitive impairment in the United States. Alzheimers Dement 7, 80–93.

Noble

, Schupf

, Manly

, Andrews

, Tang

, Mayeux

(2017) Secular trends in the incidence of dementia in a multi-ethnic community. J Alzheimers Dis 60, 1065–1075.

Gao

, Ogunniyi

, Hall

, Baiyewu

, Unverzagt

, Lane

, Murrell

, Gureje

, Hake

, Hendrie

(2016) Dementia incidence declined in African-Americans but not in Yoruba. Alzheimers Dement 12, 244–251.

10.

Doblhammer

, Fink

, Zylla

, Willekens

(2015) Compression or expansion of dementia in Germany? An observational study of short-term trends in incidence and death rates of dementia between 2006/07 and 2009/10 based on German health insurance data. Alzheimers Res Ther 7, 66.

11.

Kosteniuk

, Morgan

, O’Connell

, Kirk

, Crossley

, Teare

, Stewart

, Bello-Haas

, McBain

, Mou

, Forbes

, Innes

, Quail

(2016) Simultaneous temporal trends in dementia incidence and prevalence, 2005-2013: A population-based retrospective cohort study in Saskatchewan, Canada. Int Psychogeriatr 28, 1643–1658.

12.

Sposato

, Kapral

, Fang

, Gill

, Hackam

, Cipriano

, Hachinski

(2015) Declining incidence of stroke and dementia: Coincidence or prevention opportunity? JAMA Neurol 72, 1529–1531.

13.

Hebert

, Bienias

, Aggarwal

, Wilson

, Bennaett

, Shah

, Evans

(2010) Change in risk of Alzheimer disease over time. Neurology 75, 786–791.

14.

van den Kommer

, Deeg

DJH

, van der Flier

WM.

, Comijs

HC.

(2018) Time trend in persistent cognitive decline: Results from thelongitudinal aging study. J Gerontol B Psychol SciSoc Sci 73, S57–S64.

15.

Manton

, Gu

, Ukraintseva

(2005) Declining prevalence ofdementia in the U.S. elderly population.. Adv Gerontol 16, 30–37.

16.

Langa

, Larson

, Crimmins

, Faul

, Levine

, Kabeto

, Weir

(2017) A comparison of the prevalence of dementia in theUnited States in 2000 and 2012. JAMA Intern Med 177, 51–58.

17.

Hudomiet

, Hurd

, Rohwedder

(2018) Dementia prevalence in theUnited States in 2000 and 2012: Estimates based on a nationallyrepresentative study. J Gerontol B Psychol Sci Soc Sci 73, S10–S19.

18.

Matthews

, Arthur

, Barnes

, Bond

, Jagger

, Robinson

, Brayne

(2013) A two-decade comparison of prevalence of dementia in individuals aged 65 years and older from three geographical areas of England: Results of the cognitive function and ageing study I and II. Lancet 382, 1405–1412.

19.

Freedman

, Kasper

, Spillman

, Plassman

Short-term changes in the prevalence of probable dementia: An analysis of the 2011-2015 National Health and Aging Trends Study. J Gerontol B Psychol Sci Soc Sci 73, S48–S56.

20.

Qiu

, von Strauss

, Backman

, Winblad

, Fratiglioni

, Twenty-year changes in dementia occurrence suggest decreasing incidence in central Stockholm, Sweden. Neurology 80, 1888–1894.

21.

Wiberg

, Waern

, Billstedt

, Ostling

, Skoog

(2013) Secular trends in the prevalence of dementia and depression in Swedish septuagenarians 1976-2006. Psychol Med 43, 2627–2634.

22.

U.S Census Bureau (2006) A half-century of learning: Historical statistics on educational attainment in the United States, 1940 to 2000 (census 2000 PHC-T-41).

23.

EClipSE Collaborative Members, Brayne

, Ince

, Keage

HAD

, McKeith

, Matthews

, Polvikoski

, Sulkava

(2010) Education, the brain and dementia: Neuroprotection or compensation? Brain 133, 2210–2216.

24.

Livingston

, Sommerlad

, Orgeta

, Costafreda

, Huntley

, Ames

, Ballard

, Banerjee

, Burns

, Cohen-Mansfield

, Cooper

, Fox

, Gitlin

, Howard

, Kales

, Larson

, Ritchie

, Rockwood

, Sampson

, Samus

, Schneider

, Selbæk

, Teri

, Mukadam

(2017) Dementia prevention, intervention, and care. Lancet 390, 2673–2734.

25.

Deng

, Fuller-Thomson

(2021) Temporal trends over a decade in serious vision impairment in a large, nationally representative population-based sample of older Americans: Gender, cohort and racial/ethnic differences. Ophthalmic Epidemiol, doi: 10.1080/09286586.2021.1889001.

26.

Slud

(2015) Impact of mode-based imputation on ACS estimates. American Community Survey Research and Evaluation Program. United States Census Bureau. https://www.census.gov/content/dam/Census/library/working-papers/2015/acs/2015_Slud_01.pdf

27.

National Research Council (2007) The weighting of ACS 1-year period estimates. In Using the American Community Survey: Benefits and Challenges. The National Academies Press, Washington DC, pp. 184–208.

28.

Sharp

, Gatz

(2011) Relationship between education and dementia: An updated systematic review. Alzheimer Dis Assoc Disord 25, 289–304.

29.

Contador

, Del Ser

, Llamas

, Villarejo

, Benito-Leon

, Bermejo-Pareja

(2017) Impact of literacy and years of education on the diagnosis of dementia: A population-based study. J Clin Exp Neuropsychol 39, 112–119.

30.

Glymour

, Kawachi

, Jencks

, Berkman

(2008) Does childhood schooling affect old age memory or mental status? Using state schooling laws as natural experiments. J Epidemiol Community Health 62, 532–537.

31.

Vemuri

, Lesnick

, Przybelski

, Machulda

, Knopman

, Mielke

, Roberts

, Geda

, Rocca

, Petersen

, Jack

Jr (2014) Association of lifetime intellectual enrichment with cognitive decline in the older population. JAMA Neurol 71, 1017–1024.

32.

Skoog

, Backman

, Ribbe

, Falk

, Gudmundsson

, Thorvaldsson

, Borjesson-Hanson

, Ostling

, Johansson

, Skoog (2017) A longitudinal study of the mini-mental state examination in late nonagenarians and its relationship with dementia, mortality, and education. J Am Geriatr Soc 65, 1296–1300.

33.

Reuser

, Willekens

, Bonneux

(2011) Higher education delays and shortens cognitive impairment: A multistate life table analysis of the US health and retirement study. Eur J Epidemiol 26, 395–403.

34.

Terrera

, Minett

, Brayne

, Matthews

(2014) Education associated with a delayed onset of terminal decline. Age Ageing 43, 26–31.

35.

Then

, Luck

, Villringer

, Angermeyer

, Riedel-Heller

(2015) Compression of dementia morbidity by higher education. Alzheimers Dement 11, P201.

36.

Stern

(2009) Cognitive reserve. Neuropsychologia 47, 2015–2028.

37.

Steffener

, Stern

(2012) Exploring the neural basis of cognitive reserve in aging. ,. Biochim Biophys Acta 1822, 467–473.

38.

Valenzuela

, Matthews

, Brayne

, Ince

, Halliday

, Kril

, Dalton

, Richardson

, Forster

, Sachdev

, Medical Research Council Cognitive Function and Ageing Study (2012) Multiple biological pathways link cognitive lifestyle to protection from dementia. Biol Psychiatry 7, 783–791.

39.

Burt

, Cutler

, Higgins

, Horan

, Labarthe

, Whelton

, Brown

, Rocella

(1995) Trends in the prevalence, awareness, treatment, and control of hypertension in the adult US population. Data from the health examination surveys, 1960 to 1991. Hypertension 26, 60–69.

40.

Garfinkel

, Silverberg

(1990) Lung cancer and smoking trends in the United States over the past 25 years. Ann N Y Acad Sci 609, 146–154.

41.

Broderick

, Phillips

, Whisnant

, O’Fallon

, Bergstralh

(1989) Incidence rates of stroke in the eighties: The end of the decline in stroke? Stroke 20, 577–582.

42.

Derby

, Katz

, Lipton

, Hall

(2017) Trends in dementia incidence in a birth cohort analysis of the Einstein Aging Study. JAMA Neurol 74, 1345–1351.

43.

Kaye

, Valencia

, Baba

, Spiro

, Dietrich

, Garcia

(2010) Tooth loss and periodontal disease predict poor cognitive function in older men. J Am Geriatr Soc 58, 713–718.

44.

Noble

, Borrell

, Papapanou

, Elkind

MSV

, Scarmeas

, Wright

(2008) Periodontitis is associated with cognitive impairment among older adults: Analysis of NHANES-III. J Neurol Neurosurg Psychiatry 80, 1206–1211.

45.

Stein

, Desrosiers

, Donegan

, Yepes

, Kryscio

(2007) Tooth loss, dementia and neuropathology in the nun study. J Am Dent Assoc 138, 1314–1322.

46.

Dominy

, Lynch

, Ermini

, Benedyk

, Marczyk

, Konradi

, Nguyen

, Haditsch

, Raha

, Griffin

, Holsinger

, Arastu-Kapur

, Kaba

, Lee

, Ryder

, Potempa

, Mydel

, Hellvard

, Adamowicz

, Hasturk

, Walker

, Reynolds

, Faull

RLM

, Curtis

, Dragunow

, Potempa

(2019) Porphyromonas gingivalis in Alzheimer’s Disease brains: Evidence for disease causation and treatment with small-molecule inhibitors. Sci Adv 5, eaau3333.

47.

Crider

, Bailey

, Berry

(2011) Folic acid food fortification-its history, effect, concerns, and future directions. Nutrients 3, 370–384.

48.

Lonn

, Yusuf

, Arnold

, Sheridan

, Pogue

, Micks

, McQueen

, Probstfield

, Fodor

, Held

, Genest

Jr , Heart Outcomes Prevention Evaluation (HOPE) 2 Investigators (2006) Homocysteine lowering with folic acid and B vitamins in vascular disease. N Engl J Med 354, 1567–1577.

49.

Tucker

, Qiao

, Scott

, Rosenberg

, Spiro.

(2005) High homocysteine and low B vitamins predict cognitive decline in aging men: The veterans affairs normative aging study. Am J Clin Nutr 82, 627–635.

50.

Hainsworth

, Yeo

, Weekman

, Wilcock

(2016) Homocysteine, hyperhomocysteinemia and vascular contributions to cognitive impairment and dementia (VCID). Biochim Biophys Acta 1862, 1008–1017.

51.

Tun

, Lachman

(2010) The association between computer use and cognition across adulthood: Use it so you won’t lose it? Psychol Aging 25, 560–568.

52.

Hunsaker

, Hargittai

(2018) A review of internet use among older adults. New Media Soc 20, 3937–3954.

53.

Holwerda

, Deeg

DJH

, Beekman

ATF

, van Tilburg

, Stek

, Jonker

, Schoevers

(2014) Feelings of loneliness, but not social isolation, predict dementia onset: Results from the Amsterdam study of the elderly (AMSTEL). J Neurol Neurosurg Psychiatry 85, 135–142.

54.

Rafnsson

, Orrell

, d’Orsi

, Hogervorst

, Steptoe

(2017) Loneliness, social integration, and incident dementia over 6 years: Prospective findings from the English longitudinal study of ageing. J Gerontol B Psychol Sci Soc Sci 75, 114–124.

55.

Chopik

(2016) The benefits of social technology use among older adults are mediated by reduced loneliness. Cyberpsychol Behav Soc Netw 19, 551–556.

56.

Wilson

, Moulton

(2010) Loneliness among older adults: A national survey of adults 4545+. Knowledge Networks and Insight Policy Research: AARP.

57.

Anderson

, Thayer

(2018) Loneliness and social connections: A national survey of adults 45 and older. GfK Custom Research: AARP.

58.

United States Environmental Protection Agency (2020) Air Quality-National Summary.

59.

Killin

, Starr

, Shiue

, Russ

(2016) Environmental risk factors for dementia: A systematic review. BMC Geriatr 16, 175.

60.

Block

, Elder

, Auten

, Bilbo

, Chen

J-C

, Cory-Sletcha

, Costa

, Diaz-Sanchez

, Dorman

, Gold

, Gray

, Jeng

, Kaufman

, Kleinman

, Kirshner

, Lawler

, Miller

, Nadadur

, Ritz

, Semmens

, Tonelli

, Veronesi

, Wright

(2012) The outdoor air pollution and brain health workshop. Neurotoxicology 33, 972–984.

61.

Cacciottolo

, Wang

, Driscoll

, Woodward

, Saffari

, Reyes

, Serre

, Vizuete

, Sioutas

, Morgan

, Gatz

, Chui

, Shumaker

, Resnick

, Espeland

, Finch

, Chen

(2017) Particulate air pollutants, APOE alleles and their contributions to cognitive impairment in older women and to amyloidogenesis in experimental models. Transl Psychiatry 7, e1022.

62.

Carey

, Anderson

, Atkinson

, Beevers

, Cook

, Strachan

, Dajnak

, Gulliver

, Kelly

(2018) Are noise and air pollution related to the incidence of dementia? A cohort study in London, England. BMJ Open 8, e022404.

63.

Chen

, Kwong

, Copes

, Hystad

, van Donkelaar

, Tu

, Brook

, Goldberg

, Martin

, Murray

, Wilton

, Kopp

, Burnett

(2017) Exposure to ambient air pollution and the incidence of dementia: A population-based cohort study. Environ Int 108, 271–277.

64.

Shih

, Hu

, Weisskopf

, Schwartz

(2007) Cumulative lead dose and cognitive function in adults: A review of studies that measured both blood lead and bone lead. Environ Health Perspect 115, 483–492.

65.

Weisskopf

, Proctor

, Wright

, Schwartz

, Spiro III

, Sparrow

, Nie

, Hu

(2007) Cumulative lead exposure and cognitive performance among elderly men. Epidemiology 18, 59–66.

66.

Weisskopf

, Wright

, Schwartz

, Spiro III

, Sparrow

, Aro

, Hu

(2004) Cumulative lead exposure and prospective change in cognition among elderly men: The VA normative aging study. Am J Epidemiol 160, 1184–1193.

67.

Pirkle

, Brody

, Gunter

, Kramer

, Paschal

, Flegal

, Matte

(1994) The decline in blood lead levels in the United States. The National Health and Nutrition Examination Surveys (NHANES). JAMA 272, 284–291.

68.

Brault

, Stern

, Raglin

(2007) Evaluation report covering disability. American Community Survey Content Test Report P.4. US Census Bureau. https://www.census.gov/content/dam/Census/library/working-papers/2007/acs/2007_Brault_01.pdf

69.

Livingston

, Huntley

, Sommerlad

, Ames

, Ballard

, Banerjee

, Brayne

, Burns

, Cohen-Mansfield

, Cooper

, Costafreda

, Dias

, Fox

, Gitlin

, Howard

, Kales

, Kivimäki

, Larson

, Ogunniyi

, Orgeta

, Ritchie

, Rockwood

, Sampson

, Samus

, Schneider

, Selbæk

, Teri

, Mukadam

(2020) Dementia prevention, intervention, and care: 2020 report of the commission. Lancet 396, 413–446.