Chapter 2 identifies three classes of interventions1—cognitive training, blood pressure management in people with hypertension, and increased physical activity—that may be included when communicating with the public about interventions for delaying or slowing age-related cognitive decline (ARCD) and preventing, delaying, or slowing mild cognitive impairment (MCI) and clinical Alzheimer’s-type dementia (CATD) (referred to throughout this report by the shorthand “preventing cognitive decline and dementia”). The strength of the evidence for the beneficial effects even of these most promising interventions, however, is low to moderate at best, in part because of methodological limitations noted in Chapter 1 and described in more detail below for many of the studies reviewed in the Agency for Healthcare Research and Quality (AHRQ) systematic review (Kane et al., 2017).
These limitations stem partially from inherent challenges associated with conducting randomized controlled trials (RCTs) on interventions to prevent cognitive decline and dementia. Examples of such challenges include initiation of interventions at later life stages that may be outside the optimal window; follow-up periods of insufficient duration; high attrition; small sample sizes and studies underpowered to detect changes in incidence of MCI and CATD; use of suboptimal and heterogeneous outcome measures and assessment tools; a focus on individual interventions when multiple
1 As discussed in Chapter 1, the scope of the study was limited to individual-level interventions. Societal-level interventions, including public health policies (e.g., those related to access to education and clean air), were not evaluated.
interventions may be the most beneficial and, conversely, difficulty detecting which components in multimodal interventions are most effective in which combination(s); difficulty identifying appropriate control groups; and lack of knowledge regarding interactions between risk factors and interventions.
These methodological challenges limited the ability of the AHRQ systematic review to draw meaningful conclusions from many studies regarding the efficacy of interventions. To help ensure that future studies yield more definitive results, future investments in research on interventions for preventing cognitive decline and dementia would benefit greatly from efforts to rectify these common methodological shortcomings. One approach is to work toward common standards and protocols that might be applied across trials funded by the National Institute on Aging (NIA) and other institutes and interested organizations to make them more comparable. This chapter presents the committee’s perspective on cross-cutting methodological modifications that would improve the utility of future research on preventing cognitive decline and dementia. The committee’s recommendation in this area is presented at the end of the chapter, and draws on the conclusions in each of the sections below. Priority substantive areas for future research for specific intervention domains are identified in Chapter 4.
In addition to the more specific methodological recommendations described below, there are opportunities to make greater use of new adaptive designs for clinical trials. These include, for example, adaptive treatment strategies that permit valid comparisons when subjects are rerandomized to different or enhanced interventions based on their performance in the trial (Brown et al., 2009). Such designs may be highly useful in studying interventions that rely on adherence for success, as well as in multimodal studies, and are successful where reliable measures of outcome can be made during the trial. Other adaptive trial designs alter how participants are allocated to interventions over time, or expand to add new interventions, based on data emerging over the course of the study (Brown et al., 2009). Platform trials allow many therapies to be evaluated simultaneously and are designed with the flexibility to rapidly add, alter, and remove therapies in response to emerging measures of efficacy (Berry et al., 2015; EPAD, 2017).
Despite only modest successes in identifying interventions that can help prevent, delay, or slow MCI and CATD, dementia incidence has been decreasing in the United States (Langa et al., 2017; Rocca et al., 2011). As discussed in Chapter 1, the reasons for this decrease are not fully understood, but they parallel a decrease in cardiovascular morbidity and mortality that is the result of better prevention and treatment for heart disease
and stroke, the latter being a key determinant of dementia risk, including Alzheimer’s and vascular dementias. The secular trend toward a decrease in dementia incidence may make it difficult to demonstrate that prevention strategies are effective because the control group in any study may be experiencing improvements in dementia incidence. Thus, it may be useful in future studies to target interventions to higher-risk populations that may not be affected by these secular trends—specifically, individuals who face the highest burden of disease and those for whom an intervention could have the greatest effect (e.g., APOE-4 positive individuals, people with a strong family history of dementia, those at high risk of vascular disease).
This approach is supported by findings from subgroup analyses in previous studies. In the Prevention of Dementia by Intensive Vascular Care (PreDIVA) trial, for example, which targeted vascular risk factors but failed to detect a reduction in all-cause dementia, the greatest effects were observed in those with uncontrolled blood pressure at baseline who adhered to the intervention (Moll van Charante et al., 2016). Most studies currently do not stratify or enrich according to these kinds of considerations, although it is worth noting that the Dominantly Inherited Alzheimer Network Trials Unit has launched the addition of new preventive treatment arms to its adaptive trial platform, now known as the Next Generation prevention trial, which targets those carrying deterministic gene mutations. It is hoped that by understanding the progression and prevention of the inherited form of the disease, the same treatments deemed effective for this high-risk group can be applied to those at risk of the more common, sporadic form of the disease (Alzheimer’s Association, 2016). Biomarkers may, in the future, help identify other higher-risk populations that can be similarly targeted in intervention studies.
Identification of risk groups that could derive the greatest benefit from specific interventions would inform the tailoring of interventions so they would have the greatest possible impact on population health. For example, beginning interventions earlier in life may be particularly beneficial for those at risk of developing early-onset dementia. Individuals at higher risk for dementia may have the strongest incentive to adopt particular prevention strategies, and this may be an important consideration in public health messaging. This approach of considering individuals’ risk profiles is consistent with the larger movement toward personalized medicine. It should be emphasized, however, that the committee does not mean to suggest that studies in populations not considered at high risk should be discontinued. Rather, the optimum strategy would be to undertake these two approaches in parallel. Lastly, it is worth examining whether the trend toward decreased dementia incidence varies by socioeconomic status, for which disparities across ethnic and racial groups are well documented in the United States (Fiscella et al., 2000; NRC, 2004; Williams, 1999; Williams et
al., 2010). If the secular trends are weaker for those of lower socioeconomic status, perhaps this population might also be considered at high risk for dementia and be a particular target for research on prevention.
CONCLUSION: Identifying and targeting interventions to high-risk populations may increase the likelihood of detecting a beneficial effect of an intervention and provide a more accurate assessment of its efficacy.
The population of older adults in the United States is not only growing but also becoming increasingly diverse (Johnson, 2016). By 2060, nearly one-half of U.S. adults aged 65 and older are expected to be from a nonwhite racial or ethnic background (U.S. Census Bureau, 2012). This demographic shift has significant implications for intervention research targeting cognitive outcomes, since it is well documented that some minority populations (e.g., African Americans, Latinos) have a higher risk of cognitive impairment and dementia (Mayeda et al., 2016; Mehta and Yeo, 2017; NRC, 2004; Steenland et al., 2016), as well as of vascular disease (Kurian and Cardarelli, 2007; Mozaffarian et al., 2016; Winkleby et al., 1998).
Historically, however, minorities have been underrepresented in biomedical research (Oh et al., 2015), and although the National Institutes of Health (NIH) mandates participation of minorities in research studies, clinical trials, including research on interventions for preventing cognitive decline and dementia, have experienced disappointing limits in their minority recruitment and participation. When race was incorporated at all among reported demographic factors in the studies included in the AHRQ systematic review, study populations often were found to be poorly representative of the diversity of the general population. Similar issues with underrepresentation can be observed with respect to such demographic characteristics as socioeconomic status and educational attainment. Education level, like race, is a well-recognized bias in study volunteer samples, with more highly educated individuals being more likely to volunteer (Cobb et al., 2014; Ganguli et al., 1998; Lindsted et al., 1996; Shavers et al., 2002). Importantly, education also is a key modifier in dementia studies (Evans et al., 1997; Fitzpatrick et al., 2004; Kukull et al., 2002; Ott et al., 1995; Stern et al., 1994). Further complicating the picture are the associations among race, education, and socioeconomic status, making it difficult to disentangle the effects of each. For example, an analysis by Yaffe and colleagues (2013) using data from the Health, Aging, and Body Composition study showed that variation in dementia risk between white and black participants was
no longer statistically significant when socioeconomic differences were accounted for during the analysis.
As a result of biases in the enrollment of study participants, the existing body of evidence may not accurately reflect the effectiveness of cognitive interventions in underrepresented populations. To ensure that public health messages promote interventions that are actually effective for the range of populations affected and can be targeted as appropriate, strengthening this evidence base by increasing the participation of underrepresented populations needs to be a priority in future research. Studies such as the Advanced Cognitive Training for Independent and Vital Elderly (ACTIVE) trial that have achieved comparably higher levels of minority participation have shown, as is well documented in the literature, that greater success in recruiting representative populations can be realized by utilizing well-established community-based participatory research approaches that engage and empower minority populations to be more proactive with respect to their health (Ballard et al., 1993; Barnes and Bennett, 2014; Gauthier and Clarke, 1999; Hinton et al., 2010; Levkoff and Sanchez, 2003). Carefully designed observational studies may be more likely than RCTs to have representative distributions of race, ethnicity, education, and socioeconomic status and therefore represent an important opportunity to glean information on the impact of demographic characteristics on the effectiveness of various interventions.
Appropriate funding support for the kinds of studies needed to address these knowledge gaps and careful attention to tailoring outreach efforts and setting exclusion criteria will be key to increasing the participation of underrepresented populations. For example, exclusion of individuals with comorbidities may disproportionately disqualify minorities (Mount et al., 2012). Outreach will need to begin with engagement of the relevant communities to better understand their needs, concerns, and resource limitations. Attention also is necessary to developing interventions and messages that are culturally and linguistically appropriate (Johnson, 2016).
CONCLUSION: Recruiting study populations that better reflect the distributions of race, ethnicity, education, and socioeconomic status in the general population would help ensure the generalizability of clinical trial results to traditionally underrepresented populations.
Although CATD is commonly understood to be a disease of older adults, an increasing body of evidence generated by epidemiologic studies
suggests that neurological changes associated with disease may in fact start in midlife, decades before the onset of symptoms (Ritchie et al., 2015). Thus, MCI and CATD develop over a long period of time, often in conjunction with other comorbid diseases. These findings indicate that intervention studies restricting enrollment to elderly cohorts may be misaligned with the optimum life-stage window for prevention, and interventions targeting risk factors for cardiovascular disease (e.g., blood pressure, diet, activity levels) may be more effective if initiated earlier in life. It should be acknowledged, however, that even if the preventive effects of an intervention are optimized when it is initiated in midlife,2 the intervention could have other effects (e.g., stabilization or even reversal) when started at different life stages. Studies recruiting adults across a range of age groups could help elucidate these life-course effects.
The presumed latency period between onset of neurological changes and clinical symptoms has significant implications for the design of clinical trials, which if conducted in populations of middle-aged adults would require much longer follow-up periods to observe effects on CATD incidence and other cognitive outcomes. Indeed, the ACTIVE trial demonstrated the importance of long follow-up periods for evaluation of sustained intervention effects on cognitive performance, independent of measurements of disease incidence. In that trial, domain-specific improvements in cognitive function were measurable immediately following the intervention. However, there was a notable lag in the observable effects of cognitive training on instrumental activities of daily living—arguably one of the most meaningful types of measures for those concerned about cognitive decline—with differences from the control group apparent only 5 or 10 years later (Rebok et al., 2014; Unverzagt et al., 2012). The 10-year follow-up period employed in the ACTIVE trial was unusually good for this field.
Many of the studies analyzed in the AHRQ systematic review not only targeted older adults but were further limited by short (≤1 year) followup periods, which made it impossible to draw conclusions regarding the impact of the intervention under investigation on ARCD, MCI, and CATD. Recognizing that the optimal time to initiate an intervention aimed at preventing dementia is still unknown (and may vary across interventions) and that there is a nontrivial trade-off in terms of added costs and challenges arising from attrition, the committee believes that evaluating interventions in middle-aged adults and employing follow-up periods with durations similar to those used in the ACTIVE trial would benefit future studies and reduce the likelihood of falsely concluding that an intervention is not effective. Given the practical difficulties and barriers associated with conducting
2 Different age windows are used in studies enrolling participants in midlife, generally ranging from ages 35 to 65.
longer studies, emphasis on integrating cognitive outcome measures into planned studies targeting other conditions that also represent dementia risk factors (i.e., add-on studies) from the beginning will be important. Inclusion of biomarkers as intermediate outcomes also may be of value when evaluating interventions initiated in midlife. Both of these strategies are discussed later in this chapter.
To address the concerns about attrition (mortality and loss to followup) that naturally arise with long-term studies, especially those on populations prone to dementia, such studies would benefit from taking full advantage of current knowledge regarding enhancing adherence to prevent attrition, as well as analytical methods for addressing the potential bias associated with higher attrition rates. Effective strategies for retaining study participants identified in recent reviews include offering incentives (often financial); using systematic methods for contacting subjects and scheduling visits (e.g., follow-up reminders); and offering alternative data collection methods, sites, or times for studies requiring in-person visits (Booker et al., 2011; Brueton et al., 2014; Robinson et al., 2007). Furthermore, such statistical methods as propensity scores, inverse probability weighting, multiple imputation, and modeling of nonresponse may prove to be useful in gauging the sensitivity of findings to lost follow-up (NRC, 2010). Many RCTs now routinely report results from such analyses. Mechanisms for including this information in evaluations of the quality of evidence from studies in systematic reviews would be of value.
Another important consideration in the design of longitudinal studies is the duration of the intervention itself. For some interventions—particularly those targeting healthier lifestyles, such as increased physical activity, improved diet, or participation in cognitively stimulating activities—it may be desirable for study participants to adhere to the intervention indefinitely. In such cases, follow-up at multiple testing points would require an assessment of adherence in addition to the measurement of outcomes. Statistical methods for assessing the mediating effects of adherence and sharpening the causal effects of an intervention may also be useful here (Emsley et al., 2010; Richiardi et al., 2013).
CONCLUSION: Starting interventions at younger ages and lengthening study follow-up periods may increase the likelihood of detecting a beneficial effect on preventing cognitive decline and dementia in future studies, as well as aid in identifying interventions that are not helpful.
Variability in outcome measures employed across different studies has previously been recognized as a significant challenge to evaluating the effects of interventions on cognitive changes over time (IOM, 2015). As noted in the AHRQ systematic review, variability in the criteria used to define cognitive outcomes (e.g., scores on brief cognitive tests, self- or informant-reported complaints of cognitive impairment), as well as the multiplicity of tests used to measure cognitive performance,3 complicates efforts to conduct meta-analyses and precludes quantitative pooling of results. The committee therefore concurs with the suggestion in the AHRQ systematic review that future research in the field would benefit from the use of formal diagnostic guidelines for measuring incident CATD and consistent batteries of validated tests for assessing cognitive function. Harmonizing test batteries, which would help support comparison and pooling across studies, without mandating the use of a single specific test is also a recommendation of the 2016 Alzheimer’s Disease-Related Dementias (ADRD) Summit (NINDS, 2016). The AHRQ systematic review notes as a precedent the development of the Alzheimer’s Disease Assessment Scale-Cognitive Subscale (ADAS-Cog), which is widely used in clinical trials with cognition as an outcome and has improved consistency of methodology across studies (Kane et al., 2017).
The appropriate analysis and interpretation of results of intervention studies depend on the collection of standardized measures at the time of study initiation to assess baseline cognitive status. Not only does this enable investigators to compare changes from baseline between intervention and control arms, but it also may inform the inclusion/exclusion of study participants. Since the effect and efficacy of interventions may differ in cognitively normal and cognitively impaired populations, interventions often are tested in these two populations separately. In a number of studies included in the AHRQ systematic review, however, baseline status was not assessed, or investigators relied on subjective (self- or informant) reports of cognitive status, which is a suboptimal substitute for objective measurements. Rigorous assessment of baseline status requires attention in future trials aimed at assessing the effect of interventions on preventing cognitive decline and dementia, particularly in the design of add-on studies.
3 To facilitate analysis, the AHRQ systematic review grouped neuropsychological tests into broad categories based on what they were being used to measure (e.g., executive function, memory). However, these classifications were acknowledged to be somewhat arbitrary, as some tests could have been assigned to multiple domains.
CONCLUSION: The development and use of consistent cognitive outcome measures would enable comparison across studies and pooling of data in meta-analyses. Routine collection of baseline data using such measures is needed to properly assess the effectiveness of interventions.
Many of the targets for prevention of cognitive decline and dementia are also targets for prevention of cardiovascular and metabolic diseases (which themselves are known risk factors for dementia, as discussed in Chapter 2), creating opportunities to include cognitive outcome measures in clinical trials for interventions targeting these other conditions. The AHRQ systematic review does note a need for clinical trials explicitly designed to study interventions targeting the prevention, delay, and slowing of MCI and CATD. The committee agrees with this assessment but, recognizing the many practical reasons why large primary trials addressing cognition alone are unlikely to be undertaken (e.g., requirements for very long followup periods with associated risk of attrition and sample sizes large enough to enable detection of changes in disease incidence, as discussed above), believes that adding cognitive measures to other trials offers important opportunities to expand the evidence base. This view is consistent with the recommendation made at the 2016 ADRD Summit (NINDS, 2016). Yet while there are economic efficiencies to be gained from add-on studies relative to stand-alone cognition trials (e.g., avoidance of cohort recruitment costs), the addition of cognitive outcomes to a trial entails added cost and effort. Thus, there should be a compelling reason based on observational studies to think that the intervention being studied could also yield benefits for cognition.
In addition, a thoughtful, a priori study design is critical to the successful execution of add-on studies. Otherwise, the results for cognitive benefit may not be definitive, and there is a risk of detracting from the primary aim of the study. Challenges in interpreting results from add-on studies noted in the AHRQ systematic review—unsophisticated outcome measures, a lack of baseline measurements, failure to assign subjects randomly—generally arose from adding cognitive measures post hoc instead of building them carefully into the initial study design. Interinstitute communication within NIH and a standardized approach for a priori planning of cognitive add-on studies, including power considerations, would help overcome issues encountered in previous ancillary studies of cognition.
Cognitive outcome measures have been integrated into a number of antihypertension trials. For example, the Systolic Blood Pressure Interven-
tion Trial: Memory and Cognition in Decreased Hypertension (SPRINT-MIND) substudy of the Systolic Blood Pressure Intervention Trial (SPRINT) (discussed in Chapter 2), which prespecified secondary outcomes related to dementia and MCI (Ambrosius et al., 2014), could be considered a paradigm for future add-on studies. Although a study that included cognitive outcomes as an ancillary measure might not be sufficiently powered to detect dementia prevention effects, use of uniform cognitive measures (discussed earlier in this chapter) would enable pooling of results.
CONCLUSION: Integrating cognitive outcome measures into trials with other primary purposes is a cost-effective means of evaluating the effects of some interventions for preventing cognitive decline and dementia, but it is important to design such studies carefully a priori.
Given the late onset and relative infrequency of dementia diagnoses in the population, long-duration studies with very large sample sizes are required to detect an effect of an intervention on dementia incidence (see Chapter 2). Therefore, the identification and use of biomarkers4 (e.g., brain volume, brain amyloid accumulation) as intermediate outcomes predictive of cognitive decline and dementia is of significant interest, and may allow investigators to conduct smaller and shorter studies while also increasing confidence that an observed result is having a meaningful impact on brain function. To the degree that biomarkers can accurately reflect or predict cognitive function, improvement in biomarker measures could suggest that an intervention is slowing cognitive decline. It should be emphasized, however, that the relationship between biomarkers and clinical outcomes is not always transparent, and changes in biomarkers may not necessarily translate to observable cognitive benefits (Sano, 2016).
A wide range of biomarkers under development may be useful for identifying underlying pathologies across the continuum of dementia and enrich clinical trials (Mattsson et al., 2015). Blood-based biomarkers could be particularly useful since they would allow more frequent and inexpensive assessment in longitudinal studies (O’Bryant et al., 2017). Among the novel
4 The term biomarkers, as used in the context of the AHRQ systematic review, refers to the biological targets of brain imaging and laboratory tests used to assess changes in brain structural characteristics and activity as proxies for functional abnormalities (Kane et al., 2017). Importantly, these are not necessarily biomarkers for Alzheimer’s disease. All biomarkers reported in the studies included in the AHRQ systematic review were based on brain imaging techniques, specifically, magnetic resonance imaging (MRI) or positron emission tomography (PET) scans.
biomarkers under development—all of which have pros and cons for this purpose—are measures of tau deposition (Villemagne et al., 2015), hypometabolism (Mosconi et al., 2008), axonal damage (Mayo et al., 2017), synaptic dysfunction (Hellwig et al., 2015), neuronal damage (Tarawneh et al., 2012), inflammation (Gispert et al., 2016a,b), and changes in functional brain networks (Binnewijzend et al., 2014). Several Cochrane reviews of the literature have sought to systematically assess the use of biomarkers—cerebrospinal fluid beta-amyloid (Ritchie et al., 2014), cerebrospinal fluid tau (Ritchie et al., 2017), and positron emission tomography (PET) imaging (Zhang et al., 2014)—for accurately detecting and diagnosing future dementia in people with MCI. None of the biomarkers reviewed were found to be appropriate for current clinical practice, and all required additional research.
The utility of biomarkers as intermediate outcomes in clinical trials is an active and evolving area of research. Given the potential of valid biomarkers to reduce the length and size of clinical trials (with associated impacts on feasibility and cost), the committee supports continued efforts to further elucidate their relationship to cognitive outcomes. In fact, this is an objective of several studies currently under way, including the Mediterranean-Dietary Approaches to Stop Hypertension (DASH) Intervention for Neurodegenerative Delay (MIND) diet trial (Morris et al., 2014, 2015a,b) and the Exercise in Adults with Mild Memory Problems (EXERT) study on the effects of aerobic exercise in adults with mild memory impairment (NIA and Wake Forest University, 2017). Measuring a limited number of biomarkers in subsets of study participants can reduce the cost associated with data collection and analysis. The collection of numerous indicators is to be avoided without clear theoretical or empirical justification, as it not only would add significantly to the cost of trials but also would have implications for analysis in terms of correction for multiple comparisons. Given the current uncertainty regarding the value of biomarkers, drawing blood for future analysis would enable revisiting them as the field evolves.
CONCLUSION: The inclusion of biomarkers as intermediate outcomes has the potential to reduce significantly the length and cost of future clinical trials for interventions to prevent cognitive decline and dementia. However, this approach requires further development of a set of biomarkers that are useful for tracking response to an intervention or for predicting longer-term outcomes.
The meta-analysis from the AHRQ systematic review provides an indication of intervention domains that offer the greatest promise for preventing cognitive decline and dementia. As discussed in Chapter 2, however, individual interventions within each of these domains were quite variable, and a number of questions remain regarding the effectiveness of specific interventions, including targets and components. Large RCTs that are conducted in clinical care or community settings—sometimes referred to as pragmatic trials5—lend themselves well to head-to-head comparisons of interventions already shown to have efficacy, and also offer an opportunity to assess dose–response relationships and optimal timing of delivery. The PreDIVA multidomain trial6 (Moll van Charante et al., 2016) provides an example of how an intervention can be compared with standard care in the context of a pragmatic trial (in this case, a cluster randomized trial as entire primary care practices, not individuals, were randomized). Leveraging the existing clinical care infrastructure, including electronic medical records, allows for economic efficiencies, particularly in subject recruitment and follow-up. Additionally, because these trials are conducted with more inclusive cohorts, their results can be expected to be more generalizable than those of traditional RCTs (Patsopoulos, 2011), and may better inform clinicians and the public as to how an intervention will perform under their specific life conditions (e.g., in people on similar medication regimes). Pragmatic trials may therefore be an efficient means of addressing knowledge gaps raised in the AHRQ systematic review, such as optimal blood pressure targets in antihypertensive trials and the comparative effectiveness of specific cognitive training applications.
CONCLUSION: Large trials designed to test the effectiveness of an intervention in broad, routine clinical practices or community settings may be more cost effective than traditional RCTs for comparative effectiveness research for interventions that have already been shown to have beneficial effects on cognition.
5 Unlike traditional RCTs, which assess the efficacy of interventions under highly controlled circumstances, pragmatic trials are designed to measure effectiveness, or how an intervention performs when used in usual conditions of care and in varied populations more reflective of the real world (Roland and Torgerson, 1998).
6 In the PreDIVA trial, which is described in more detail in Chapter 4, the intervention group met three times per year with nurses who provided medical and nonmedical assistance to subjects to reduce vascular risk factors. The study’s primary outcome was all-cause dementia (Moll van Charante et al., 2016).
CONCLUSION: The absence of high-strength evidence supporting beneficial effects on cognitive decline and dementia for interventions included in the AHRQ systematic review likely results in part from methodological limitations of past intervention studies. Recognizing the limited pool of resources available for research on ARCD, MCI, and CATD, future research investments will have the greatest impact if directed to a limited number of well-designed trials of sufficient power and duration.
Alzheimer’s Association. 2016. Alzheimer’s Association awards $4.3 million to accelerate the launch of new drug treatment arms in landmark prevention trial. Chicago, IL: Alzheimer’s Association. https://www.alz.org/documents_custom/dia-tu_nextgen_release_042316.pdf (accessed May 8, 2017).
Ambrosius, W. T., K. M. Sink, C. G. Foy, D. R. Berlowitz, A. K. Cheung, W. C. Cushman, L. J. Fine, D. C. Goff, Jr., K. C. Johnson, A. A. Killeen, C. E. Lewis, S. Oparil, D. M. Reboussin, M. V. Rocco, J. K. Snyder, J. D. Williamson, J. T. Wright, Jr., and P. K. Whelton. 2014. The design and rationale of a multicenter clinical trial comparing two strategies for control of systolic blood pressure: The Systolic Blood Pressure Intervention Trial (SPRINT). Clinical Trials 11(5):532-546.
Ballard, E. L., F. Nash, K. Raiford, and L. E. Harrell. 1993. Recruitment of black elderly for clinical research studies of dementia: The CERAD experience. Gerontologist 33(4): 561-565.
Barnes, L. L., and D. A. Bennett. 2014. Alzheimer’s disease in African Americans: Risk factors and challenges for the future. Health Affairs (Project Hope) 33(4):580-586.
Berry, S. M., J. T. Connor, and R. J. Lewis. 2015. The platform trial: An efficient strategy for evaluating multiple treatments. Journal of the American Medical Association 313(16):1619-1620.
Binnewijzend, M. A., S. M. Adriaanse, W. M. Van der Flier, C. E. Teunissen, J. C. de Munck, C. J. Stam, P. Scheltens, B. N. van Berckel, F. Barkhof, and A. M. Wink. 2014. Brain network alterations in Alzheimer’s disease measured by eigenvector centrality in fMRI are related to cognition and CSF biomarkers. Human Brain Mapping 35(5):2383-2393.
Booker, C. L., S. Harding, and M. Benzeval. 2011. A systematic review of the effect of retention methods in population-based cohort studies. BMC Public Health 11:249.
Brown, C. H., T. R. T. Have, B. Jo, G. Dagne, P. A. Wyman, B. Muthén, and R. D. Gibbons. 2009. Adaptive designs for randomized trials in public health. Annual Review of Public Health 30:1-25.
Brueton, V. C., J. F. Tierney, S. Stenning, S. Meredith, S. Harding, I. Nazareth, and G. Rait. 2014. Strategies to improve retention in randomised trials: A Cochrane systematic review and meta-analysis. BMJ Open 4(2).
Cobb, E. M., D. C. Singer, and M. M. Davis. 2014. Public interest in medical research participation: Differences by volunteer status and study type. Clinical and Translational Science 7(2):145-149.
Emsley, R., G. Dunn, and I. R. White. 2010. Mediation and moderation of treatment effects in randomised controlled trials of complex interventions. Statistical Methods in Medical Research 19(3):237-270.
EPAD (European Prevention of Alzheimer’s Dementia Consortium). 2017. Project objectives. http://ep-ad.org/project-objectives (accessed May 8, 2017).
Evans, D. A., L. E. Hebert, L. A. Beckett, P. A. Scherr, M. S. Albert, M. J. Chown, D. M. Pilgrim, and J. O. Taylor. 1997. Education and other measures of socioeconomic status and risk of incident Alzheimer disease in a defined population of older persons. Archives of Neurology 54(11):1399-1405.
Fiscella, K., P. Franks, M. R. Gold, and C. M. Clancy. 2000. Inequality in quality: Addressing socioeconomic, racial, and ethnic disparities in health care. Journal of the American Medical Association 283(19):2579-2584.
Fitzpatrick, A. L., L. H. Kuller, D. G. Ives, O. L. Lopez, W. Jagust, J. C. Breitner, B. Jones, C. Lyketsos, and C. Dulberg. 2004. Incidence and prevalence of dementia in the Cardiovascular Health Study. Journal of the American Geriatrics Society 52(2):195-204.
Ganguli, M., M. E. Lytle, M. D. Reynolds, and H. H. Dodge. 1998. Random versus volunteer selection for a community-based study. Journals of Gerontology, Series A: Biological Sciences and Medical Sciences 53(1):M39-M46.
Gauthier, M. A., and W. P. Clarke. 1999. Gaining and sustaining minority participation in longitudinal research projects. Alzheimer Disease & Associated Disorders 13(Suppl 1):S29-S33.
Gispert, J. D., G. C. Monte, C. Falcon, A. Tucholka, S. Rojas, R. Sanchez-Valle, A. Antonell, A. Llado, L. Rami, and J. L. Molinuevo. 2016a. CSF YKL-40 and pTau181 are related to different cerebral morphometric patterns in early AD. Neurobiology of Aging 38:47-55.
Gispert, J. D., M. Suarez-Calvet, G. C. Monte, A. Tucholka, C. Falcon, S. Rojas, L. Rami, R. Sanchez-Valle, A. Llado, G. Kleinberger, C. Haass, and J. L. Molinuevo. 2016b. Cerebrospinal fluid sTREM2 levels are associated with gray matter volume increases and reduced diffusivity in early Alzheimer’s disease. Alzheimer’s & Dementia: The Journal of the Alzheimer’s Association 12(12):1259-1272.
Hellwig, K., H. Kvartsberg, E. Portelius, U. Andreasson, T. J. Oberstein, P. Lewczuk, K. Blennow, J. Kornhuber, J. M. Maler, H. Zetterberg, and P. Spitzer. 2015. Neurogranin and YKL-40: Independent markers of synaptic degeneration and neuroinflammation in Alzheimer’s disease. Alzheimer’s Research & Therapy 7:74.
Hinton, L., K. Carter, B. R. Reed, L. Beckett, E. Lara, C. DeCarli, and D. Mungas. 2010. Recruitment of a community-based cohort for research on diversity and risk of dementia. Alzheimer Disease and Associated Disorders 24(3):234-241.
IOM (Institute of Medicine). 2015. Cognitive aging: Progress in understanding and opportunities for action. Washington, DC: The National Academies Press.
Johnson, J. 2016. Interventions conducted in minority populations: Individual characteristics. Paper presented at Preventing Dementia and Cognitive Impairment: A Workshop, Washington, DC, October 25.
Kane, R. L., M. Butler, H. A. Fink, M. Brasure, H. Davila, P. Desai, E. Jutkowitz, E. McCreedy, V. Nelson, J. R. McCarten, C. Calvert, E. Ratner, L. Hemmy, and T. Barclay. 2017. Interventions to prevent age-related cognitive decline, mild cognitive impairment, and clinical Alzheimer’s-type dementia. Comparative effectiveness review 188. Rockville, MD: Agency for Healthcare Research and Quality.
Kukull, W. A., R. Higdon, J. D. Bowen, W. C. McCormick, L. Teri, G. D. Schellenberg, G. van Belle, L. Jolley, and E. B. Larson. 2002. Dementia and Alzheimer disease incidence: A prospective cohort study. Archives of Neurology 59(11):1737-1746.
Kurian, A. K., and K. M. Cardarelli. 2007. Racial and ethnic differences in cardiovascular disease risk factors: A systematic review. Ethnicity & Disease 17(1):143-152.
Langa, K. M., E. B. Larson, E. M. Crimmins, J. D. Faul, D. A. Levine, M. U. Kabeto, and D. R. Weir. 2017. A comparison of the prevalence of dementia in the United States in 2000 and 2012. JAMA Internal Medicine 177(1):51-58.
Levkoff, S., and H. Sanchez. 2003. Lessons learned about minority recruitment and retention from the Centers on Minority Aging and Health Promotion. Gerontologist 43(1):18-26.
Lindsted, K. D., G. E. Fraser, M. Steinkohl, and W. L. Beeson. 1996. Healthy volunteer effect in a cohort study: Temporal resolution in the Adventist Health Study. Journal of Clinical Epidemiology 49(7):783-790.
Mattsson, N., M. C. Carrillo, R. A. Dean, M. D. S. Devous, T. Nikolcheva, P. Pesini, H. Salter, W. Z. Potter, R. Sperling, R. J. Bateman, L. J. Bain, and E. Liu. 2015. Revolutionizing Alzheimer’s disease and clinical trials through biomarkers. Alzheimer’s & Dementia: Diagnosis, Assessment & Disease Monitoring 1:412-419.
Mayeda, E. R., M. M. Glymour, C. P. Quesenberry, and R. A. Whitmer. 2016. Inequalities in dementia incidence between six racial and ethnic groups over 14 years. Alzheimer’s & Dementia: The Journal of the Alzheimer’s Association 12(3):216-224.
Mayo, C. D., E. L. Mazerolle, L. Ritchie, J. D. Fisk, and J. R. Gawryluk. 2017. Longitudinal changes in microstructural white matter metrics in Alzheimer’s disease. NeuroImage: Clinical 13:330-338.
Mehta, K. M., and G. W. Yeo. 2017. Systematic review of dementia prevalence and incidence in United States race/ethnic populations. Alzheimer’s & Dementia: The Journal of the Alzheimer’s Association 13(1):72-83.
Moll van Charante, E. P., E. Richard, L. S. Eurelings, J.-W. van Dalen, S. A. Ligthart, E. F. van Bussel, M. P. Hoevenaar-Blom, M. Vermeulen, and W. A. van Gool. 2016. Effectiveness of a 6-year multidomain vascular care intervention to prevent dementia (PreDIVA): A cluster-randomised controlled trial. The Lancet 388(10046):797-805.
Morris, M. C., C. C. Tangney, Y. Wang, L. L. Barnes, D. Bennett, and N. Aggarwal. 2014. MIND diet score more predictive than DASH or Mediterranean diet scores. Alzheimer’s & Dementia: The Journal of the Alzheimer’s Association 10(4):P166.
Morris, M. C., C. C. Tangney, Y. Wang, F. M. Sacks, L. L. Barnes, D. A. Bennett, and N. T. Aggarwal. 2015a. MIND diet slows cognitive decline with aging. Alzheimer’s & Dementia: The Journal of the Alzheimer’s Association 11(9):1015-1022.
Morris, M. C., C. C. Tangney, Y. Wang, F. M. Sacks, D. A. Bennett, and N. T. Aggarwal. 2015b. MIND diet associated with reduced incidence of Alzheimer’s disease. Alzheimer’s & Dementia: The Journal of the Alzheimer’s Association 11(9):1007-1014.
Mosconi, L., A. Pupi, and M. J. De Leon. 2008. Brain glucose hypometabolism and oxidative stress in preclinical Alzheimer’s disease. Annals of the New York Academy of Sciences 1147:180-195.
Mount, D. L., C. Davis, B. Kennedy, S. Raatz, K. Dotson, T. L. Gary-Webb, S. Thomas, K. C. Johnson, and M. A. Espeland. 2012. Factors influencing enrollment of African Americans in the Look AHEAD trial. Clinical Trials 9(1):80-89.
Mozaffarian, D., E. J. Benjamin, A. S. Go, D. K. Arnett, M. J. Blaha, M. Cushman, S. R. Das, S. de Ferranti, J.-P. Després, H. J. Fullerton, V. J. Howard, M. D. Huffman, C. R. Isasi, M. C. Jiménez, S. E. Judd, B. M. Kissela, J. H. Lichtman, L. D. Lisabeth, S. Liu, R. H. Mackey, D. J. Magid, D. K. McGuire, E. R. Mohler, C. S. Moy, P. Muntner, M. E. Mussolino, K. Nasir, R. W. Neumar, G. Nichol, L. Palaniappan, D. K. Pandey, M. J. Reeves, C. J. Rodriguez, W. Rosamond, P. D. Sorlie, J. Stein, A. Towfighi, T. N. Turan, S. S. Virani, D. Woo, R. W. Yeh, and M. B. Turner. 2016. Heart disease and stroke statistics—2016 update. A Report from the American Heart Association 133(4):e38-e360.
NIA (National Institute on Aging) and Wake Forest University. 2017. Exercise in adults with mild memory problems (EXERT). https://www.nia.nih.gov/alzheimers/clinical-trials/exercise-adults-mild-memory-problems-exert (accessed May 29, 2017).
NINDS (National Institute of Neurological Disorders and Stroke). 2016. Report to the National Advisory Neurological Disorders and Stroke Council. Paper read at Alzheimer’s Disease-Related Dementias Summit, Bethesda, MD, March 29-30.
NRC (National Research Council). 2004. Critical perspectives on racial and ethnic differences in health in late life. Washington, DC: The National Academies Press.
NRC. 2010. The prevention and treatment of missing data in clinical trials. Washington, DC: The National Academies Press.
O’Bryant, S. E., M. M. Mielke, R. A. Rissman, S. Lista, H. Vanderstichele, H. Zetterberg, P. Lewczuk, H. Posner, J. Hall, L. Johnson, Y. L. Fong, J. Luthman, A. Jeromin, R. Batrla-Utermann, A. Villarreal, G. Britton, P. J. Snyder, K. Henriksen, P. Grammas, V. Gupta, R. Martins, and H. Hampel. 2017. Blood-based biomarkers in Alzheimer disease: Current state of the science and a novel collaborative paradigm for advancing from discovery to clinic. Alzheimer’s & Dementia: The Journal of the Alzheimer’s Association 13(1):45-58.
Oh, S. S., J. Galanter, N. Thakur, M. Pino-Yanes, N. E. Barcelo, M. J. White, D. M. de Bruin, R. M. Greenblatt, K. Bibbins-Domingo, A. H. B. Wu, L. N. Borrell, C. Gunter, N. R. Powe, and E. G. Burchard. 2015. Diversity in clinical and biomedical research: A promise yet to be fulfilled. PLoS Medicine 12(12):e1001918.
Ott, A., M. M. B. Breteler, F. van Harskamp, J. J. Claus, T. J. M. van der Cammen, D. E. Grobbee, and A. Hofman. 1995. Prevalence of Alzheimer’s disease and vascular dementia: Association with education. The Rotterdam Study. BMJ 310(6985):970-973.
Patsopoulos, N. A. 2011. A pragmatic view on pragmatic trials. Dialogues in Clinical Neuroscience 13(2):217-224.
Rebok, G. W., K. Ball, L. T. Guey, R. N. Jones, H. Y. Kim, J. W. King, M. Marsiske, J. N. Morris, S. L. Tennstedt, F. W. Unverzagt, and S. L. Willis. 2014. Ten-year effects of the Advanced Cognitive Training for Independent and Vital Elderly cognitive training trial on cognition and everyday functioning in older adults. Journal of the American Geriatrics Society 62(1):16-24.
Richiardi, L., R. Bellocco, and D. Zugna. 2013. Mediation analysis in epidemiology: Methods, interpretation and bias. International Journal of Epidemiology 42(5):1511-1519.
Ritchie, C., N. Smailagic, A. H. Noel-Storr, Y. Takwoingi, L. Flicker, S. E. Mason, and R. McShane. 2014. Plasma and cerebrospinal fluid amyloid beta for the diagnosis of Alzheimer’s disease dementia and other dementias in people with mild cognitive impairment (MCI). Cochrane Database of Systematic Reviews 6:CD008782.
Ritchie, C., N. Smailagic, A. H. Noel-Storr, O. Ukoumunne, E. C. Ladds, and S. Martin. 2017. CSF tau and CSF tau/ABeta ratio for the diagnosis of Alzheimer’s disease dementia and other dementias in people with mild cognitive impairment (MCI). Cochrane Database of Systematic Reviews 3:CD010803.
Ritchie, K., C. W. Ritchie, K. Yaffe, I. Skoog, and N. Scarmeas. 2015. Is late-onset Alzheimer’s disease really a disease of midlife? Alzheimer’s & Dementia: Translational Research & Clinical Interventions 1(2):122-130.
Robinson, K. A., C. R. Dennison, D. M. Wayman, P. J. Pronovost, and D. M. Needham. 2007. Systematic review identifies number of strategies important for retaining study participants. Journal of Clinical Epidemiology 60(8):757-765.
Rocca, W. A., R. C. Petersen, D. S. Knopman, L. E. Hebert, D. A. Evans, K. S. Hall, S. Gao, F. W. Unverzagt, K. M. Langa, E. B. Larson, and L. R. White. 2011. Trends in the incidence and prevalence of Alzheimer’s disease, dementia, and cognitive impairment in the United States. Alzheimer’s & Dementia: The Journal of the Alzheimer’s Association 7(1):80-93.
Roland, M., and D. J. Torgerson. 1998. Understanding controlled trials: What are pragmatic trials? BMJ 316(7127):285.
Sano, M. 2016. Methodological considerations pertaining to the prevention of dementia. Paper presented at Preventing Dementia and Cognitive Impairment: A Workshop, Washington, DC, October 25.
Shavers, V. L., C. F. Lynch, and L. F. Burmeister. 2002. Racial differences in factors that influence the willingness to participate in medical research studies. Annals of Epidemiology 12(4):248-256.
Steenland, K., F. C. Goldstein, A. Levey, and W. Wharton. 2016. A meta-analysis of Alzheimer’s disease incidence and prevalence comparing African-Americans and caucasians. Journal of Alzheimer’s Disease 50(1):71-76.
Stern, Y., B. Gurland, T. K. Tatemichi, M. X. Tang, D. Wilder, and R. Mayeux. 1994. Influence of education and occupation on the incidence of Alzheimer’s disease. Journal of the American Medical Association 271(13):1004-1010.
Tarawneh, R., J. M. Lee, J. H. Ladenson, J. C. Morris, and D. M. Holtzman. 2012. CSF VILIP-1 predicts rates of cognitive decline in early Alzheimer disease. Neurology 78(10): 709-719.
Unverzagt, F. W., L. T. Guey, R. N. Jones, M. Marsiske, J. W. King, V. G. Wadley, M. Crowe, G. W. Rebok, and S. L. Tennstedt. 2012. ACTIVE cognitive training and rates of incident dementia. Journal of the International Neuropsychological Society 18(4):669-677.
U.S. Census Bureau. 2012. U.S. Census Bureau projections show a slower growing, older, more diverse nation a half century from now. Washington, DC: U.S. Census Bureau. https://www.census.gov/newsroom/releases/archives/population/cb12-243.html (accessed March 3, 2017).
Villemagne, V. L., M. T. Fodero-Tavoletti, C. L. Masters, and C. C. Rowe. 2015. Tau imaging: Early progress and future directions. The Lancet Neurology 14(1):114-124.
Williams, D. R. 1999. Race, socioeconomic status, and health the added effects of racism and discrimination. Annals of the New York Academy of Sciences 896(1):173-188.
Williams, D. R., S. A. Mohammed, J. Leavell, and C. Collins. 2010. Race, socioeconomic status and health: Complexities, ongoing challenges and research opportunities. Annals of the New York Academy of Sciences 1186:69-101.
Winkleby, M. A., H. C. Kraemer, D. K. Ahn, and A. N. Varady. 1998. Ethnic and socioeconomic differences in cardiovascular disease risk factors: Findings for women from the third National Health and Nutrition Examination Survey, 1988–1994. Journal of the American Medical Association 280(4):356-362.
Yaffe, K., C. Falvey, T. B. Harris, A. Newman, S. Satterfield, A. Koster, H. Ayonayon, and E. Simonsick. 2013. Effect of socioeconomic disparities on incidence of dementia among biracial older adults: Prospective study. BMJ 347:f7051.
Zhang, S., N. Smailagic, C. Hyde, A. H. Noel-Storr, Y. Takwoingi, R. McShane, and J. Feng. 2014. CPIB-PET for the early diagnosis of Alzheimer’s disease dementia and other dementias in people with mild cognitive impairment. Cochrane Database for Systematic Reviews 7:CD010386.