J Am Geriatr Soc
· 2026 Jun · PMID 41987487
·
Full text
BACKGROUND: Food insecurity affects many older adults and is associated with a range of adverse health outcomes. However, most prior research has relied on cross-sectional data, and evidence on predictors of subsequent f...BACKGROUND: Food insecurity affects many older adults and is associated with a range of adverse health outcomes. However, most prior research has relied on cross-sectional data, and evidence on predictors of subsequent food insecurity among Medicare beneficiaries remains limited. METHODS: We conducted a retrospective longitudinal cohort study of 12,029 individuals using 2017-2022 Medicare Current Beneficiary Survey data. We applied linear lagged dependent variable models to estimate predicted probabilities of food insecurity in the subsequent year, based on the prior year's demographics, socioeconomic characteristics, health conditions, and food insecurity status. Food insecurity was assessed using the USDA's 6-item short-form survey. Demographic and socioeconomic characteristics were measured from self-reported survey responses, and health conditions were identified using medical claims data. RESULTS: In adjusted linear probability models, the strongest predictor of subsequent-year food insecurity was prior-year food insecurity, which is associated with a 43 percentage point (p.p.) higher probability of food insecurity in the subsequent year (95% CI: 38.7-47.2). Other important predictors included self-reported "poor" general health (6.5 p.p. [0.5-12.4]), inability to afford dental care (6.4 p.p [3.1-9.7]), difficulty paying medical bills (5.2 p.p. [1.5-8.9]), not filling prescriptions due to costs (4.4 p.p. [2.2-6.5]), and Hispanic ethnicity (4.1 p.p. [1.7-6.6]). CONCLUSIONS: Among older Medicare beneficiaries, prior food insecurity and markers of acute financial strain, including difficulty affording dental care or prescriptions, strongly predict future food insecurity, even after adjusting for their income. Incorporating these indicators into screening efforts may improve identification of beneficiaries at risk of food insecurity. Furthermore, linking screening efforts with navigation and benefits enrollment support, alongside broader policy reforms to reduce financial strain, may help mitigate food insecurity and its downstream health consequences in this population.
Ma JE, Uejo E, Quenstedt S
… +12 more, Preiss D, Sperling J, Locklear TD, Gambino J, Bowling CB, Chadwick AL, Cox CE, Lorenz KA, Matallana L, Winger JG, Bosworth HB, Steinhauser K
J Am Geriatr Soc
· 2026 Apr · PMID 41987429
·
Full text
BACKGROUND: Chronic pain is highly prevalent among older adults with advanced chronic kidney disease (CKD stage 4, stage 5, and end-stage kidney disease [ESKD]). Yet, pain management involves a delicate balance between a...BACKGROUND: Chronic pain is highly prevalent among older adults with advanced chronic kidney disease (CKD stage 4, stage 5, and end-stage kidney disease [ESKD]). Yet, pain management involves a delicate balance between alleviating symptoms and avoiding harm related to impaired renal drug clearance and the high risk of medication side effects. Because little is known about how patients and clinicians navigate these complex pain management conversations, we examined patient and provider perspectives on communication and decision making in chronic pain and advanced kidney disease. METHODS: We conducted a qualitative study using semi-structured interviews based on the Ottawa Decision Support Framework. Participants included older adults (age ≥ 65) with both advanced CKD and chronic pain lasting ≥ 3 months and physicians and advanced practice providers from primary care, geriatrics, nephrology, and palliative care. We used thematic analysis to summarize major themes on communication and decision making. RESULTS: We interviewed 48 participants, including 24 older adults with advanced kidney disease and chronic pain and 24 clinicians, with 6 clinicians from each specialty. Three major themes about barriers to effective communication emerged: (1) treatment complexity and uncertainty; (2) fragmentation of care across specialties and the care team; and (3) divergent treatment preferences between patients and clinicians. Communication strategies to overcome these barriers included: open communication, multidisciplinary care team collaboration, patient advocacy, and relationship- and values-centered decision making. CONCLUSIONS: This study highlighted key barriers and potential communication strategies among older adults with chronic pain and advanced kidney disease. These findings can inform the development of targeted interventions that support patients and clinicians in navigating these complex conversations and decisions.
More than 6 million persons living with dementia (PLWD) in the United States rely on the emergency department (ED) for unscheduled care, with up to half discharged home after treatment. The ED-to-home transition poses si...More than 6 million persons living with dementia (PLWD) in the United States rely on the emergency department (ED) for unscheduled care, with up to half discharged home after treatment. The ED-to-home transition poses significant challenges for PLWD and their care partners (referred to as "dyads"), contributing to high rates of ED revisits and adverse outcomes. The Community Paramedic-led Transitions Intervention (CPTI) was developed to address these challenges by adapting the validated Care Transitions Intervention for the ED setting. Delivered by trained community paramedics, CPTI is a short-term 30-day program that includes one home visit and up to three follow-up phone calls. Using a coaching model, paramedic coaches work with members of the dyad to strengthen their knowledge, skills, and confidence to manage their health and successfully navigate the health care system. CPTI is being implemented as part of Emergency Departments LEading the Transformation of Alzheimer's and Dementia Care (ED-LEAD), a cluster-randomized pragmatic trial testing 3 interventions designed to improve outcomes for PLWD discharged home from the ED across 14 health systems and 79 EDs nationwide. This paper describes the CPTI model as implemented within ED-LEAD, detailing its theoretical foundation, structure, training curriculum, workflow integration, and implementation monitoring. This framework can provide a model for health systems, provider groups, and emergency medical service agencies interested in adopting this innovative approach and implementing the CPTI. Insights from its implementation within ED-LEAD will guide future efforts to improve post-ED outcomes and continuity of care for PLWD and their care partners.
Rosenberg MW, Li E, Xue WL
… +4 more, Eckstrom E, Howe M, Smith AK, Kotwal AA
J Am Geriatr Soc
· 2026 Jun · PMID 41981819
·
Full text
BACKGROUND: Emotional and financial mistreatment among older adults is often under-recognized among clinicians. Exploring whether sensory impairment is linked to an elevated risk of elder financial or emotional mistreatm...BACKGROUND: Emotional and financial mistreatment among older adults is often under-recognized among clinicians. Exploring whether sensory impairment is linked to an elevated risk of elder financial or emotional mistreatment could help inform targeted screening strategies. METHODS: We used nationally-representative, cross-sectional data from the National Social Life, Health, and Aging Project, including interview rounds with available assessments of recent elder mistreatment (Round 1 (R1): 2005-06 and Round 4 (R4): 2021-23). Elder mistreatment included reported emotional mistreatment (being insulted or feeling controlled) and financial mistreatment (money or belongings taken without permission) within the last 12 months. Assessments of sensory impairment differed slightly by round of data collection and included self-reported overall vision, olfaction, and hearing, self-reported hearing loss that interferes with communication, and objective vision (visual acuity) and olfaction (sniff test). We used multivariable logistic regressions to determine the association of self-reported and objective individual and multi-sensory impairments with mistreatment risk. Analyses for multisensory impairment were conducted using the full sample, while analyses for individual impairments were stratified by round of data collection. RESULTS: The total sample included 5231 participants; R1 included 2461 participants who were on average 70 years old (SD 7) and 53% female, and R4 included 2770 participants who were on average 71 years old (SD 9) and 55% female. Multisensory impairment was linked to a higher adjusted probability of experiencing either emotional or financial mistreatment (0 impairments: 20% vs. 1 impairment: 23% vs. 2+ impairments: 28%, p < 0.01). Respondents with self-reported vision impairment had a higher adjusted probability of financial mistreatment (8% vs. 3%, p < 0.001) whereas objective vision impairment had no association. Self-reported hearing impairment that interfered with communication was associated with financial mistreatment (12% vs. 7%, p = 0.03). CONCLUSION: Integrating assessments for multisensory impairment, vision, and hearing loss may add to elder mistreatment prevention and detection efforts.
BACKGROUND: Polypharmacy affects nearly 50% of older adults, increasing risks of adverse drug events, falls, and hospitalizations. While patient-centered educational interventions show promise, optimal delivery formats r...BACKGROUND: Polypharmacy affects nearly 50% of older adults, increasing risks of adverse drug events, falls, and hospitalizations. While patient-centered educational interventions show promise, optimal delivery formats remain unclear. METHODS: This pragmatic quasi-experimental study was conducted May to September 2025 in Michigan using venue-based allocation. Participants received identical educational content through either in-person presentation (n = 60) or self-paced packet with pamphlet, video, and multimedia formats (n = 80). Educational components included the World Health Organization Five Moments for Medication Safety framework and a study-developed drug list collection tool. The primary outcome was medication management confidence (5-point Likert scale). Secondary outcomes included polypharmacy knowledge, self-efficacy, and tool adoption intentions. Paired t-tests assessed pre-post changes, and analysis of covariance compared formats adjusting for age, sex, race, thyroid disorder, arthritis, and baseline confidence. RESULTS: The sample (N = 140, mean age 70.0 ± 12.0 years, 57.1% female, 82.1% White) showed significant improvement in medication management confidence (mean change = 0.24 points, 95% CI: 0.12 to 0.36, p < 0.001, d = 0.33). Both formats demonstrated significant within-group improvements (in-person: 0.32 points, p = 0.005; self-paced: 0.19 points, p = 0.010). After covariate adjustment, no significant difference existed between formats (adjusted difference: 0.01 points, p = 0.911). Knowledge of polypharmacy definition improved from 82.9% to 97.1%. Post hoc power analysis indicated the study was powered to detect moderate between-format differences (d ≥ 0.48); the observed between-format effect (d = 0.18) was smaller, suggesting any true difference is likely small rather than moderate or large. CONCLUSIONS: Both formats produced significant improvements in medication management confidence and knowledge, with no significant difference between delivery formats after adjustment. These findings support flexible implementation strategies, allowing delivery approaches based on resource availability, patient needs, and setting with potential for scalable dissemination to populations facing transportation, scheduling, or mobility barriers.
Ten Hoope SMK, Marongiu S, Siegert CEH
… +3 more, Heerdink ER, Janssen MJA, Karapinar-Çarkit F
J Am Geriatr Soc
· 2026 May · PMID 41968599
·
Publisher ↗
BACKGROUND: The increasing prevalence of patients with hyperpolypharmacy (> 10 medications) has made medication reviews increasingly complex. ChatGPT-4-Turbo (ChatGPT), a large language model, has demonstrated potential...BACKGROUND: The increasing prevalence of patients with hyperpolypharmacy (> 10 medications) has made medication reviews increasingly complex. ChatGPT-4-Turbo (ChatGPT), a large language model, has demonstrated potential in healthcare applications and could potentially support medication reviews. OBJECTIVES: This study aimed to evaluate the agreement between medication reviews conducted by ChatGPT compared to healthcare professionals (HCPs) in older people. Secondary objectives included: the validity of additional interventions detected by ChatGPT, its ability to structure diagnoses to medication use, and laboratory target values based on patient characteristics. METHODS: In this retrospective proof-of-concept study, 51 medication reviews previously conducted by a geriatric internist and hospital pharmacist were re-evaluated using ChatGPT. ChatGPT was trained on polypharmacy guidelines, was then provided with the same primary data as HCPs, and was asked to perform medication reviews. Two pharmacists scored the agreement between ChatGPT and HCPs. The structuring of information and additional interventions suggested by ChatGPT were reviewed within an expert team. Descriptive statistics were used. OUTCOMES: The primary outcome was the percentage agreement between the interventions suggested by ChatGPT compared to HCPs. Secondary outcomes included the proportion of valid and incorrect interventions suggested by ChatGPT and its ability to structure patient information. RESULTS: HCPs suggested 183 interventions and ChatGPT 202 interventions. ChatGPT achieved a 27.7% agreement with interventions of HCPs. It identified 19 additional valid interventions which HCPs missed (7.6%), but also proposed 84 incorrect interventions (33.7%). While ChatGPT demonstrated strong capability in structuring patient data (86.4% correct diagnoses linked to medication), it struggled with contextualizing appropriate laboratory target values based on patient characteristics (46.7%). CONCLUSION: ChatGPT had low agreement with HCPs, but found additional interventions that HCPs missed. ChatGPT lacks clinical decision-making capabilities based on individual patient contexts in older people. ChatGPT may, however, serve as a support tool to structure diagnoses and medication lists.
de Moraes GS, Avelino-Silva TJ, Covinsky KE
… +5 more, Carpenter CR, Umoh ME, Morinaga CV, Curiati PK, Aliberti MJR
J Am Geriatr Soc
· 2026 May · PMID 41958224
·
Full text
BACKGROUND: Delirium predicts adverse outcomes in older emergency department (ED) patients, but many acutely ill patients without delirium have underlying cognitive impairment that goes unrecognized. Whether cognitive im...BACKGROUND: Delirium predicts adverse outcomes in older emergency department (ED) patients, but many acutely ill patients without delirium have underlying cognitive impairment that goes unrecognized. Whether cognitive impairment screening improves risk prediction beyond delirium remains uncertain. We compared practical bedside screening strategies for delirium, cognitive impairment, or both for predicting 90-day functional decline and mortality in older adults admitted from the ED. METHODS: A prospective cohort comprising patients aged ≥ 65 years admitted from the ED of a large hospital in São Paulo, Brazil. Trained professionals screened for delirium using the brief Confusion Assessment Method (bCAM) and for cognitive impairment using the 10-point Cognitive Screener (10-CS). Patients were classified as having normal cognition (bCAM negative, 10-CS > 5), delirium (bCAM positive), or cognitive impairment without delirium (bCAM negative, 10-CS ≤ 5). Blinded investigators assessed decline in basic activities of daily living (ADL) and mortality within 90 days of admission. Fine-Gray models (death as a competing risk) and Cox models estimated associations with outcomes, adjusting for sociodemographic and clinical factors. RESULTS: Among 830 patients (mean age = 80 ± 9 years; women = 47%), 427 (51.5%) had normal cognition, 171 (20.6%) had delirium, and 232 (27.9%) had cognitive impairment without delirium. Among delirium-negative patients with cognitive impairment, 52% had no documented dementia diagnosis or reported memory problems. Compared with normal cognition, cognitive impairment without delirium was associated with 90-day functional ADL decline (sub-HR = 1.60; 95% CI = 1.03-2.49) and mortality (HR = 2.31; 95% CI = 1.18-4.51), with risks similar to those observed in delirium. A staged strategy (bCAM first, then 10-CS if bCAM negative) showed higher discrimination than delirium-only or 10-CS-only screening. CONCLUSIONS: Cognitive impairment without delirium is common, often unrecognized, and predicts 90-day adverse outcomes in older patients admitted from the ED. A brief staged screening strategy integrating delirium assessment with cognitive impairment testing among delirium-negative patients may enhance early detection of cognitive vulnerability and support care planning.
J Am Geriatr Soc
· 2026 Jun · PMID 41958179
·
Full text
INTRODUCTION: Monitoring adverse events (AEs) assures patient safety in clinical trials. Studies involving seriously ill and older adults can have a high volume of AEs related to underlying courses of illness rather than...INTRODUCTION: Monitoring adverse events (AEs) assures patient safety in clinical trials. Studies involving seriously ill and older adults can have a high volume of AEs related to underlying courses of illness rather than study procedures. Trials not regulated by the Food and Drug Administration (FDA) can incur excessive reporting burden beyond that required for compliance or patient safety assurance. In 2021, a working group proposed an alternative approach to AE evaluation, emphasizing routine, aggregate reporting, and reserving expedited reporting of individual AEs for serious events at least possibly related to study procedures and not aligned with goals of care. Another outcome was a call to operationalize the approach for eventual implementation. METHODS: A panel comprising 31 leaders from government agencies, professional organizations, aging and serious illness research communities, research administration, and medical ethics met five times in 2024-2025. The panel advised on deliverables, assessed concept acceptability, and reviewed tools; a multidisciplinary project team generated content and incorporated feedback. Research compliance experts reviewed the final decision algorithm. RESULTS: The resulting Decision-Making for Adverse Event Reporting Toolkit Tailored to Social, Behavioral and Health Services Studies involving Seriously Ill and Older Adult Populations (DART) is a novel research tool to support researchers in AE evaluation. The decision algorithm assesses: (1) relatedness, (2) goals of care alignment, (3) seriousness, and (4) greater risk of harm. The toolkit includes a case series, frequently asked questions, and reporting form. Feedback emphasized DART applies to any study population and intervention type not regulated by the FDA. CONCLUSIONS: DART provides a novel solution to AE reporting challenges faced by aging and serious illness researchers. While its design centers on seriously ill older adults, DART can likely be applied to research in any population and intervention type not regulated by the FDA. Future pilot testing will inform implementation.
Webber C, Konikoff L, Li W
… +7 more, Grubic N, Maxwell CJ, Bush SH, Casey G, Isenberg SR, Tanuseputro P, Qureshi D
J Am Geriatr Soc
· 2026 May · PMID 41958168
·
Publisher ↗
BACKGROUND: Transfers from long-term care (LTC) to hospital near the end of life can be burdensome for LTC residents. Dementia and frailty are common in this population, yet their relationship to burdensome transitions r...BACKGROUND: Transfers from long-term care (LTC) to hospital near the end of life can be burdensome for LTC residents. Dementia and frailty are common in this population, yet their relationship to burdensome transitions remains unclear. This study examined the joint associations between dementia, frailty, and end-of-life burdensome transitions among LTC residents. METHODS: We conducted a population-based retrospective cohort study using health administrative data, capturing LTC residents aged ≥ 65 who died in Ontario, Canada, January 1, 2015-March 31, 2020. Dementia was identified using a validated algorithm, supplemented with Resident Assessment Instrument-Minimum Dataset 2.0 (RAI-MDS) data. Frailty (non-frail, pre-frail, frail) was assessed using a 72-item frailty index developed for use with the RAI-MDS. Burdensome transitions were defined as 3+ transfers to hospital (≥ 2 if due to pneumonia, urinary tract infection, sepsis, dehydration) in the last 90 days of life (last 180 days as secondary outcome). Poisson regression was used to estimate adjusted relative risks (RRs) and 95% confidence intervals (CIs) for burdensome transitions by dementia and frailty status. RESULTS: Among 88,507 LTC decedents, most were frail (67.6%) and had dementia (87.5%). Approximately 4.7% of residents experienced a burdensome transition in the last 90 days of life (8.8% in the last 180 days). Dementia was associated with a reduced risk of burdensome transitions in the last 90 days of life across frailty strata (non-frail: RR = 0.77, 95% CI, 0.64-0.92; pre-frail: RR = 0.86, 95% CI, 0.77-0.97; frail: RR = 0.79, 95% CI, 0.71-0.89). Similar findings were observed when examining burdensome transitions in the last 180 days of life. CONCLUSION: Dementia and frailty are independently associated with a lower likelihood of burdensome transitions near the end of life in LTC residents. Further research is needed to explore the appropriateness of hospital transfers at the end of life, considering the perspectives of LTC residents, family, and healthcare providers.
Wang BX, Lindenberg JH, Herzig SJ
… +2 more, Schonberg MA, Anderson TS
J Am Geriatr Soc
· 2026 Jun · PMID 41952328
·
Full text
BACKGROUND: Risks related to long-term opioid therapy for chronic pain are high and may increase over time with aging. Deprescribing may be a beneficial intervention for older adults prescribed chronic opioids. METHODS:...BACKGROUND: Risks related to long-term opioid therapy for chronic pain are high and may increase over time with aging. Deprescribing may be a beneficial intervention for older adults prescribed chronic opioids. METHODS: Semi-structured interviews with hypothetical clinical cases of older adults prescribed opioids for chronic pain: (1) low-risk case: a patient prescribed low-dose opioids without concerns; (2) moderate-risk case: a patient with multimorbidity and concurrent benzodiazepine use prescribed moderate opioid doses; (3) high-risk case: a patient prescribed high-dose opioids with signs of an opioid use disorder (OUD). PCPs were asked, in an open-ended fashion, to discuss whether they would initiate a deprescribing conversation, how they would approach deprescribing, and how they would approach a patient who declined recommendations to deprescribe. PARTICIPANTS AND SETTING: PCPs from a Massachusetts health system. RESULTS: 18 PCPs participated (56% female, 78% academic). More than half of PCPs would initiate a deprescribing conversation across the three cases. PCPs' approach to deprescribing and mitigating risks differed based on clinical risk. In low and moderate-risk cases, PCPs emphasized a patient-directed taper plan and education on opioid risks. In the high-risk case, some PCPs were uncertain about initiating a deprescribing conversation due to concerns about the patient's mental health and the risk of illicit opioid use. Naloxone was infrequently recommended across the three cases, but in the high-risk case, approximately half of PCPs suggested medications for OUD. CONCLUSIONS: PCPs reported that they would often initiate opioid deprescribing conversations with older adults, but were less confident in managing older adults with signs of OUD. PCPs require additional support to implement successful conversations on opioid deprescribing with older adults.
Markland AD, Kennedy R, Crowe M
… +5 more, Williams B, Munoz C, Elgayar S, Burgio KL, Howard VJ
J Am Geriatr Soc
· 2026 Apr · PMID 41952255
·
Full text
BACKGROUND: To understand cross-sectional associations between lower urinary tract symptoms (LUTS) and specific cognitive domains, we examined cognitive performance in older adults with and without LUTS. METHODS: The stu...BACKGROUND: To understand cross-sectional associations between lower urinary tract symptoms (LUTS) and specific cognitive domains, we examined cognitive performance in older adults with and without LUTS. METHODS: The study comprised a subset of Black and White participants, aged 45 and older, enrolled in the REasons for Geographic and Racial Differences in Stroke (REGARDS) national, longitudinal cohort study 2003-2007. Computer-assisted telephone interviews were conducted for follow-up every 6 months including an established cognitive battery: Six-Item Screener (SIS) for cognitive impairment, animal naming, letter F naming, word list learning, and delayed recall for verbal fluency/executive function and verbal memory. Validated LUTS questionnaires were administered beginning in 2020 at 6-month intervals. Separate linear and logistic regression models for men and women estimated associations between LUTS and cognitive battery scores, adjusting for covariates. RESULTS: From the 10,464 participants (6026 women, 4438 men), 70% of women reported LUTS (mean age 69 ± 8 years) versus 62% of men (mean age 63 ± 7 years). More Black men reported LUTS compared to White men, 69% versus 59% (p < 0.001), respectively, without differences among Black and White women. Men and women with LUTS had decreased odds of being cognitively intact based on the SIS, OR 0.65, 95% CI 0.50, 0.85 and OR 0.64, 95% CI 0.46, 0.78, respectively. Men with LUTS had lower scores for Animal Naming (β -0.42, 95% CI -0.75, -0.08) and Delayed Recall (β -0.18, 95% CI -0.30, -0.05). Women with LUTS had lower test scores for Letter F Naming (β -0.29, 95% CI -0.55, -0.04), Word List Learning (β -0.40, 95% CI -0.68, -0.11) and Delayed Recall (β -0.13, 95% CI -0.25, -0.02). CONCLUSIONS: Older adults with LUTS have subtle differences in cognitive function, especially in new learning, verbal memory, and verbal fluency/executive function that should be considered in treatment decisions.
Stenlund S, Koga HK, James P
… +4 more, Farmer J, McGrath CB, Grodstein F, Kubzansky LD
J Am Geriatr Soc
· 2026 May · PMID 41947615
·
Publisher ↗
BACKGROUND: Previous studies suggest that higher optimism is associated with better cognitive function and slower cognitive decline in aging. Using data from the Health and Retirement Study, a nationally representative s...BACKGROUND: Previous studies suggest that higher optimism is associated with better cognitive function and slower cognitive decline in aging. Using data from the Health and Retirement Study, a nationally representative sample of older U.S. adults, we examined whether optimism was associated with lower risk of developing dementia in different population groups and if associations were maintained after accounting for initial health status and other potential confounders and across multiple sensitivity analyses. METHODS: Optimism was measured using the validated Life Orientation Test-Revised in 9071 cognitively healthy individuals within 2 years of obtaining each person's first measure of cognitive function. Dementia was identified by an algorithm developed to perform well across major racial and ethnic groups, obtained at each of eight waves of data collection from 2006 to 2020. Cox proportional hazard models were used, and sensitivity analyses addressed major concerns such as reverse causation. RESULTS: We observed that a 1-standard deviation increase in optimism was associated with a lower hazard of developing dementia (hazard ratio = 0.85, 95% confidence interval 0.82-0.88), after adjusting for age, sex, race and ethnicity, education, depression, and major health conditions over follow-up ranging up to 14 years. When stratifying by race and ethnicity, we observed similar associations in the Non-Hispanic White and Black sub-populations. Associations did not substantially change when health behaviors were included in the models, when we removed the first 2 years of follow-up to mitigate concerns about potential reverse causation, or when we excluded individuals with the poorest mental health. CONCLUSION: Higher optimism was associated with a lower incidence of dementia. These findings suggest a potential value of optimism in supporting healthy aging, which could be considered in future research on dementia prevention initiatives.