Erin Laura Ferguson
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Erin Laura Ferguson has a total of 88 co-authors. Here are the top co-authors by number of shared publications:
- Medellena Maria Glymour
Department of Epidemiology, Boston University School of Public Health, MA
9 shared publications Jingxuan Wang
Department of Epidemiology and Biostatistics University of California San Francisco California USA
4 shared publications- Jingxuan Wang
Department of Epidemiology and Biostatistics, University of California , San Francisco, CA
4 shared publications Peter Buto
Department of Epidemiology, Brown University , Providence, RI
4 shared publicationsRuijia Chen
Department of Epidemiology, Boston University , Boston, MA
4 shared publications- Catherine Schaefer
Kaiser Permanente Division of Research, Pleasanton, CA
3 shared publications
Latest Publications
- Open Access
Associations of hearing loss with social isolation, loneliness, and depressive symptoms among older adults in the Health, Aging, and Body Composition Study
Background: Hearing loss (HL) may be a risk factor for poor psychosocial outcomes among older adults, but evidence remains mixed. We assessed associations of self-reported and objective HL with and without hearing aid use with social contact, loneliness, and depression pooled across 6 years of follow-up. Methods: We studied 2049 Black and White adults from the Health, Aging, and Body Composition study aged 70-79 at recruitment. Self-reported HL and audiometric HL with and without hearing aid use were assessed at analytic baseline (Year 5, 2001-2002). Outcomes were frequency of contact with family and friends (<weekly vs. at least weekly), depressive symptoms (CESD-10), and loneliness (CESD-10 item “I felt lonely”) measured across 6 annual visits. Adjusted for demographic and clinical variables, we used generalized linear regression with generalized estimating equations to assess associations with outcomes pooled across six follow-up waves. Results: Self-reported HL (16%) was associated with more depressive symptoms (ß=0.13 SD; 95%CI:0.03,0.24), but no other outcome. Objective HL without hearing aid use (11%) was associated with infrequent contact with friends (OR=1.38; 95%CI:1.07,1.78) and more depressive symptoms (ß=0.19 SD; 95%CI:0.07,0.31); objective HL with hearing aid use (9%) was not associated with these outcomes. Objective HL, regardless of hearing aid use, was borderline associated with more frequent feelings of loneliness. Discussion: Objective HL without hearing aid use may be an important risk factor for isolation from friendship networks and depressive symptoms among older adults. Self-reported HL and objective HL with hearing aid use may also be linked to some adverse psychosocial outcomes.
Open AccessAssociations of hearing loss with social isolation, loneliness, and depressive symptoms among older adults in the Health, Aging, and Body Composition Study
Background: Hearing loss (HL) may be a risk factor for poor psychosocial outcomes among older adults, but evidence remains mixed. We assessed associations of self-reported and objective HL with and without hearing aid use with social contact, loneliness, and depression pooled across 6 years of follow-up. Methods: We studied 2049 Black and White adults from the Health, Aging, and Body Composition study aged 70-79 at recruitment. Self-reported HL and audiometric HL with and without hearing aid use were assessed at analytic baseline (Year 5, 2001-2002). Outcomes were frequency of contact with family and friends (<weekly vs. at least weekly), depressive symptoms (CESD-10), and loneliness (CESD-10 item “I felt lonely”) measured across 6 annual visits. Adjusted for demographic and clinical variables, we used generalized linear regression with generalized estimating equations to assess associations with outcomes pooled across six follow-up waves. Results: Self-reported HL (16%) was associated with more depressive symptoms (ß=0.13 SD; 95%CI:0.03,0.24), but no other outcome. Objective HL without hearing aid use (11%) was associated with infrequent contact with friends (OR=1.38; 95%CI:1.07,1.78) and more depressive symptoms (ß=0.19 SD; 95%CI:0.07,0.31); objective HL with hearing aid use (9%) was not associated with these outcomes. Objective HL, regardless of hearing aid use, was borderline associated with more frequent feelings of loneliness. Discussion: Objective HL without hearing aid use may be an important risk factor for isolation from friendship networks and depressive symptoms among older adults. Self-reported HL and objective HL with hearing aid use may also be linked to some adverse psychosocial outcomes.
- Open Access
- 1 July 2026
- Published by Wiley in Alzheimer's & Dementia: Translational Research & Clinical Interventions
- Vol. 12 (3) , e70306
Memantine, acetylcholinesterase inhibitors, and quality‐of‐care outcomes among patients with cognitive impairment: an emulated trial design
INTRODUCTION: We evaluated the effect of acetylcholinesterase inhibitors (AChEIs) and memantine on non‐cognitive outcomes by implementing emulated trials in electronic health records. METHODS: Individuals diagnosed with cognitive impairment who initiated AChEIs or memantine were matched to up to five non‐initiators. Associations of medication initiation with 11 non‐cognitive outcomes were estimated using Cox proportional hazards models. We calculated e‐values for each hazard ratio (HR) and compared their magnitude to socioeconomic predictors of memantine and AChEI use in a US‐based cohort. RESULTS: AChEI initiation was associated with nine of 11 adverse outcomes. For example, the HR for antipsychotic use = 1.67 (95% confidence interval [CI]: 1.51 to 1.84) and depression HR = 1.86 (1.67 to 2.06). Results for memantine were similar. E‐values varied from 1.45 to 2.11 for memantine and 1.40 to 2.44 for AChEIs. Many potential confounders had odds ratios (ORs) > 1.4 for initiation, including marital status (ORs = 1.36 to 6.67). DISCUSSION: Memantine and AChEIs were associated with adverse outcomes, and residual confounding was unlikely to fully account for all the observed associations.
- 1 July 2026
- Published by Wiley in Alzheimer's & Dementia: Translational Research & Clinical Interventions
- Vol. 12 (3) , e70306
Open AccessMemantine, acetylcholinesterase inhibitors, and quality‐of‐care outcomes among patients with cognitive impairment: an emulated trial design
INTRODUCTION: We evaluated the effect of acetylcholinesterase inhibitors (AChEIs) and memantine on non‐cognitive outcomes by implementing emulated trials in electronic health records. METHODS: Individuals diagnosed with cognitive impairment who initiated AChEIs or memantine were matched to up to five non‐initiators. Associations of medication initiation with 11 non‐cognitive outcomes were estimated using Cox proportional hazards models. We calculated e‐values for each hazard ratio (HR) and compared their magnitude to socioeconomic predictors of memantine and AChEI use in a US‐based cohort. RESULTS: AChEI initiation was associated with nine of 11 adverse outcomes. For example, the HR for antipsychotic use = 1.67 (95% confidence interval [CI]: 1.51 to 1.84) and depression HR = 1.86 (1.67 to 2.06). Results for memantine were similar. E‐values varied from 1.45 to 2.11 for memantine and 1.40 to 2.44 for AChEIs. Many potential confounders had odds ratios (ORs) > 1.4 for initiation, including marital status (ORs = 1.36 to 6.67). DISCUSSION: Memantine and AChEIs were associated with adverse outcomes, and residual confounding was unlikely to fully account for all the observed associations.
- Open Access
Social and Cardiovascular Risk Factors as Predictors of the Progression from Mild Cognitive Impairment to Dementia in a Large EHR Database
Background and Objectives: Progression from mild cognitive impairment (MCI) to Alzheimer’s Disease and Related Dementias (AD/ADRD) varies widely across individuals, yet the mechanisms underlying this heterogeneity remain unclear. Identifying clinical and social determinants influencing this transition could enable earlier intervention. While cardiovascular and social risk factors are established contributors to dementia incidence, their role in progression from MCI to dementia may differ. Few studies using real world clinical data have evaluated these potential determinants of MCI progression. Methods: Using electronic health records (EHR) from patients with incident MCI at UCSF Health (2010-2024), we evaluated cardiovascular (blood pressure [BP], body mass index [BMI], and type II diabetes) and social (marital status, language preference, race/ethnicity, and neighborhood disadvantage) risk factors for rate of progression from MCI to AD/ADRD. Covariate-adjusted Cox proportional hazards models estimated hazard ratios for incident AD/ADRD, with evaluation of interactions by sex. Results: Among 6,529 patients, higher systolic BP was associated with AD/ADRD incidence (HR per 10 mmHg: 1.09, 95% CI: 1.05–1.14). BMI was inversely associated with incidence in both males (HR: 0.94; 95% CI: 0.92–0.97) and females (HR:0.98; 95% CI: 0.96–0.99). Compared to married individuals, widowed patients had a higher hazard of progression (HR: 1.15; 95% CI: 1.00–1.32). Spanish-speaking (HR: 1.38; 95% CI: 1.04–1.81), Chinese-speaking (HR: 1.19; 95% CI: 1.00–1.42), and “Other non-English” speaking patients (HR:1.24; 95% CI: 1.03–1.51) had a higher hazard of progression compared to English speakers. Latinx (HR:1.22; 95% CI: 1.01–1.48) and Asian patients (HR:1.14, 95% CI: 1.00-1.30; p=0.04) also had higher hazards of progression compared to White patients. Neighborhood disadvantage was not significantly associated with disease progression. Discussion: Cardiovascular and social factors independently influence dementia progression, with some sex-specific patterns. Integrating clinical and social indicators highlights the potential of EHR data to identify high-risk patients earlier in the care continuum and support equitable dementia prevention.
Open AccessSocial and Cardiovascular Risk Factors as Predictors of the Progression from Mild Cognitive Impairment to Dementia in a Large EHR Database
Background and Objectives: Progression from mild cognitive impairment (MCI) to Alzheimer’s Disease and Related Dementias (AD/ADRD) varies widely across individuals, yet the mechanisms underlying this heterogeneity remain unclear. Identifying clinical and social determinants influencing this transition could enable earlier intervention. While cardiovascular and social risk factors are established contributors to dementia incidence, their role in progression from MCI to dementia may differ. Few studies using real world clinical data have evaluated these potential determinants of MCI progression. Methods: Using electronic health records (EHR) from patients with incident MCI at UCSF Health (2010-2024), we evaluated cardiovascular (blood pressure [BP], body mass index [BMI], and type II diabetes) and social (marital status, language preference, race/ethnicity, and neighborhood disadvantage) risk factors for rate of progression from MCI to AD/ADRD. Covariate-adjusted Cox proportional hazards models estimated hazard ratios for incident AD/ADRD, with evaluation of interactions by sex. Results: Among 6,529 patients, higher systolic BP was associated with AD/ADRD incidence (HR per 10 mmHg: 1.09, 95% CI: 1.05–1.14). BMI was inversely associated with incidence in both males (HR: 0.94; 95% CI: 0.92–0.97) and females (HR:0.98; 95% CI: 0.96–0.99). Compared to married individuals, widowed patients had a higher hazard of progression (HR: 1.15; 95% CI: 1.00–1.32). Spanish-speaking (HR: 1.38; 95% CI: 1.04–1.81), Chinese-speaking (HR: 1.19; 95% CI: 1.00–1.42), and “Other non-English” speaking patients (HR:1.24; 95% CI: 1.03–1.51) had a higher hazard of progression compared to English speakers. Latinx (HR:1.22; 95% CI: 1.01–1.48) and Asian patients (HR:1.14, 95% CI: 1.00-1.30; p=0.04) also had higher hazards of progression compared to White patients. Neighborhood disadvantage was not significantly associated with disease progression. Discussion: Cardiovascular and social factors independently influence dementia progression, with some sex-specific patterns. Integrating clinical and social indicators highlights the potential of EHR data to identify high-risk patients earlier in the care continuum and support equitable dementia prevention.
- 2 March 2026
- Published by Oxford University Press (OUP) in American Journal of Epidemiology
- Vol. 195 (6) , 1739-1748
Referral of patients with cognitive impairment to specialty memory care: associations with patient-centered outcomes and specificity of diagnoses
Primary care providers may consider referring patients with cognitive impairment to clinicians specializing in memory care. We evaluated whether referrals are associated with quality-of-care and diagnostic outcomes, comparing estimates based on traditional and instrumental variable (IV) analyses. Analyses included individuals diagnosed with memory loss, mild cognitive impairment, or Alzheimer’s disease and related dementias in a single healthcare system after 2005. Electronic health records were used to define referral to specialists and primary care provider preference (instrument) for referring. We modeled traditional and IV associations between referral to care and 14 patient-centered outcomes over 5 years of follow-up using adjusted Aalen additive hazards models. Overall, 1019 (15%) older adults were referred at diagnosis. Preference strongly predicted actual referral (F-statistic = 637). Referral was observationally associated with increased cumulative hazard of receiving a more specific cognitive diagnosis (hazard difference at year 5: 0.09 [95% CI, 0.04-0.15]) and depression (0.09 [95% CI, 0.01-0.18]). Using IV, referral was significantly associated with decreased hazard of International Classification of Diseases-defined weight loss (–0.35 [95% CI, –0.60 to –0.09]); other estimates were imprecise and consistent with possible benefits or harms. Given barriers in accessing specialty care, it is critical to further investigate how specialty care affects outcomes of individuals with living with cognitive impairment.
- 2 March 2026
- Published by Oxford University Press (OUP) in American Journal of Epidemiology
- Vol. 195 (6) , 1739-1748
Referral of patients with cognitive impairment to specialty memory care: associations with patient-centered outcomes and specificity of diagnoses
Primary care providers may consider referring patients with cognitive impairment to clinicians specializing in memory care. We evaluated whether referrals are associated with quality-of-care and diagnostic outcomes, comparing estimates based on traditional and instrumental variable (IV) analyses. Analyses included individuals diagnosed with memory loss, mild cognitive impairment, or Alzheimer’s disease and related dementias in a single healthcare system after 2005. Electronic health records were used to define referral to specialists and primary care provider preference (instrument) for referring. We modeled traditional and IV associations between referral to care and 14 patient-centered outcomes over 5 years of follow-up using adjusted Aalen additive hazards models. Overall, 1019 (15%) older adults were referred at diagnosis. Preference strongly predicted actual referral (F-statistic = 637). Referral was observationally associated with increased cumulative hazard of receiving a more specific cognitive diagnosis (hazard difference at year 5: 0.09 [95% CI, 0.04-0.15]) and depression (0.09 [95% CI, 0.01-0.18]). Using IV, referral was significantly associated with decreased hazard of International Classification of Diseases-defined weight loss (–0.35 [95% CI, –0.60 to –0.09]); other estimates were imprecise and consistent with possible benefits or harms. Given barriers in accessing specialty care, it is critical to further investigate how specialty care affects outcomes of individuals with living with cognitive impairment.
- 21 August 2025
- Published by Oxford University Press (OUP) in American Journal of Epidemiology
- Vol. 194 (12) , 3630-3639
Trends in the distribution of P values in epidemiology journals: a statistical, P -curve, and simulation study
Epidemiologists have advocated for reporting confidence intervals and deemphasizing P values to address long-standing concerns about null-hypothesis statistical-significance testing, P hacking, and reproducibility. It is unknown if efforts to reduce reliance on P values have altered the distribution of P values. For 21 332 abstracts published 2000 to 2024 in 4 major epidemiology journals, two-sided P values (N = 25 288) were calculated from estimates and confidence intervals scraped using ChatGPT’s 4o-mini model. We evaluated trends over time to determine whether the empirical distribution of P values changed. We fitted to expected P-value distributions and simulated these distributions with and without assuming changes in statistical power over time. Average P values decreased from 2000 to 2024; the fraction of P values just below the .05 threshold also decreased. Fits to models indicate that statistical power increased. Increasing power would reduce average P value while also decreasing P values near the .05 threshold—precisely the trends observed in epidemiology journals. Although the frequency of P values near the .05 threshold has declined modestly, this likely reflects increases in statistical power rather than decreases in P hacking.
- 21 August 2025
- Published by Oxford University Press (OUP) in American Journal of Epidemiology
- Vol. 194 (12) , 3630-3639
Trends in the distribution of P values in epidemiology journals: a statistical, P -curve, and simulation study
Epidemiologists have advocated for reporting confidence intervals and deemphasizing P values to address long-standing concerns about null-hypothesis statistical-significance testing, P hacking, and reproducibility. It is unknown if efforts to reduce reliance on P values have altered the distribution of P values. For 21 332 abstracts published 2000 to 2024 in 4 major epidemiology journals, two-sided P values (N = 25 288) were calculated from estimates and confidence intervals scraped using ChatGPT’s 4o-mini model. We evaluated trends over time to determine whether the empirical distribution of P values changed. We fitted to expected P-value distributions and simulated these distributions with and without assuming changes in statistical power over time. Average P values decreased from 2000 to 2024; the fraction of P values just below the .05 threshold also decreased. Fits to models indicate that statistical power increased. Increasing power would reduce average P value while also decreasing P values near the .05 threshold—precisely the trends observed in epidemiology journals. Although the frequency of P values near the .05 threshold has declined modestly, this likely reflects increases in statistical power rather than decreases in P hacking.
- Open Access
- 22 July 2025
- Published by Wolters Kluwer Health in Neurology
- Vol. 105 (2) , e213855
Statin Initiation and Dementia Incidence in a Large Health Care System From 1997 to 2020
Previous research of associations between statins and Alzheimer disease and Alzheimer disease-related dementias (AD/ADRDs) has been limited by short follow-up, small samples, and confounding. We aimed to estimate the association between the 1st statin prescription and incident AD/ADRD among members of a large population-based cohort of older adults. We used a cohort study design emulating a target trial using data from Kaiser Permanente Northern California (KPNC), an integrated health care delivery system. Participants were born before 1951 and KPNC members for 4+ years during 1997-2010. Embedded subsamples included sociodemographic and genetic data. Statin initiators were matched at first prescription ("baseline") with up to 5 "noninitiators" based on age and low-density lipoprotein cholesterol (LDL-C). Participants with extreme propensity scores were excluded. The outcome was time to incident AD/ADRD diagnosis, censoring, or the administrative end of study (December 31, 2020). Cox proportional hazard models were used to estimate hazard ratios for statin initiation on AD/ADRD incidence. Follow-up time was divided at the first year of follow-up to account for increased AD/ADRD detection in the first year due to increased interaction with the health care system after a statin prescription. Among eligible participants (n = 705,061), 264,294 individuals (37.5% of eligible participants) initiated any statin during 2001-2010 ("initiators"), of whom 249,613 (94.4%) were matched with 255,937 unique noninitiators to create the analytic sample (322,358 unique participants; mean age at baseline = 67.4 years; 55.1% female). The average follow-up was 11.8 years. In the first year after initiating statins, AD/ADRD diagnoses were elevated by 46% (hazard ratio [HR] = 1.46, 95% CI 1.42-1.53) compared with noninitiators. After 1 year, statin initiators experienced no difference in AD/ADRD incidence (full sample: HR = 1.00, 95% CI 0.99-1.01; subsample with survey covariates: HR = 1.01, 95% CI 0.98-1.06; subsample with survey and genetic covariates: HR = 0.97, 95% CI 0.91-1.07). Adjustment for sociodemographic covariates and apolipoprotein E e4 allele count did not materially change the findings. In this large emulated target trial, statin initiation was inconsistent with more than a 3% increase or decrease in the hazard of AD/ADRD after the first year of follow-up. This intent-to-treat analysis does not directly quantify effects of long-term exposure to statins. Associations in the first year likely reflect increased medical observation immediately after statin initiation. This emulated trial provides Class II evidence that statin initiation is not associated with AD/ADRD or AD incidence after the first year of follow-up.
- 22 July 2025
- Published by Wolters Kluwer Health in Neurology
- Vol. 105 (2) , e213855
Open AccessStatin Initiation and Dementia Incidence in a Large Health Care System From 1997 to 2020
Previous research of associations between statins and Alzheimer disease and Alzheimer disease-related dementias (AD/ADRDs) has been limited by short follow-up, small samples, and confounding. We aimed to estimate the association between the 1st statin prescription and incident AD/ADRD among members of a large population-based cohort of older adults. We used a cohort study design emulating a target trial using data from Kaiser Permanente Northern California (KPNC), an integrated health care delivery system. Participants were born before 1951 and KPNC members for 4+ years during 1997-2010. Embedded subsamples included sociodemographic and genetic data. Statin initiators were matched at first prescription ("baseline") with up to 5 "noninitiators" based on age and low-density lipoprotein cholesterol (LDL-C). Participants with extreme propensity scores were excluded. The outcome was time to incident AD/ADRD diagnosis, censoring, or the administrative end of study (December 31, 2020). Cox proportional hazard models were used to estimate hazard ratios for statin initiation on AD/ADRD incidence. Follow-up time was divided at the first year of follow-up to account for increased AD/ADRD detection in the first year due to increased interaction with the health care system after a statin prescription. Among eligible participants (n = 705,061), 264,294 individuals (37.5% of eligible participants) initiated any statin during 2001-2010 ("initiators"), of whom 249,613 (94.4%) were matched with 255,937 unique noninitiators to create the analytic sample (322,358 unique participants; mean age at baseline = 67.4 years; 55.1% female). The average follow-up was 11.8 years. In the first year after initiating statins, AD/ADRD diagnoses were elevated by 46% (hazard ratio [HR] = 1.46, 95% CI 1.42-1.53) compared with noninitiators. After 1 year, statin initiators experienced no difference in AD/ADRD incidence (full sample: HR = 1.00, 95% CI 0.99-1.01; subsample with survey covariates: HR = 1.01, 95% CI 0.98-1.06; subsample with survey and genetic covariates: HR = 0.97, 95% CI 0.91-1.07). Adjustment for sociodemographic covariates and apolipoprotein E e4 allele count did not materially change the findings. In this large emulated target trial, statin initiation was inconsistent with more than a 3% increase or decrease in the hazard of AD/ADRD after the first year of follow-up. This intent-to-treat analysis does not directly quantify effects of long-term exposure to statins. Associations in the first year likely reflect increased medical observation immediately after statin initiation. This emulated trial provides Class II evidence that statin initiation is not associated with AD/ADRD or AD incidence after the first year of follow-up.
- Open Access
Diagnostic Reversion in Dementia Care: A Real-World Analysis of Mild Cognitive Impairment Diagnoses Following Dementia in a Large Electronic Medical Record System
INTRODUCTION If mild cognitive impairment (MCI) is diagnosed after dementia, it suggests either the dementia diagnosis was premature, or the MCI diagnosis is incorrect. We investigated the prevalence and predictors of such “diagnostic reversion”—MCI diagnosis following dementia diagnosis—in a large academic health system. METHODS Among 5,965 patients aged 50+ with incident dementia in UCSF Health electronic health records, we identified “reverters” with a subsequent MCI diagnosis. We used Group LASSO–regularized logistic regression and random forest models to identify predictors. RESULTS Reversion occurred in 13.7% of patients. Lower odds were observed among older adults (OR=0.95/year; 95% CI: 0.92–0.98), while higher odds were found among Spanish speakers (OR=2.26; 95% CI: 1.28–4.00), those with greater cardiovascular risk (OR=1.16; 95% CI: 1.01–1.33), and higher Charlson comorbidity burden (OR=1.09; 95% CI: 1.05–1.14). DISCUSSION Diagnostic reversion is common and socially patterned, suggesting contributions from misdiagnosis, clinical uncertainty, or variability in clinical presentation and care setting.
Open AccessDiagnostic Reversion in Dementia Care: A Real-World Analysis of Mild Cognitive Impairment Diagnoses Following Dementia in a Large Electronic Medical Record System
INTRODUCTION If mild cognitive impairment (MCI) is diagnosed after dementia, it suggests either the dementia diagnosis was premature, or the MCI diagnosis is incorrect. We investigated the prevalence and predictors of such “diagnostic reversion”—MCI diagnosis following dementia diagnosis—in a large academic health system. METHODS Among 5,965 patients aged 50+ with incident dementia in UCSF Health electronic health records, we identified “reverters” with a subsequent MCI diagnosis. We used Group LASSO–regularized logistic regression and random forest models to identify predictors. RESULTS Reversion occurred in 13.7% of patients. Lower odds were observed among older adults (OR=0.95/year; 95% CI: 0.92–0.98), while higher odds were found among Spanish speakers (OR=2.26; 95% CI: 1.28–4.00), those with greater cardiovascular risk (OR=1.16; 95% CI: 1.01–1.33), and higher Charlson comorbidity burden (OR=1.09; 95% CI: 1.05–1.14). DISCUSSION Diagnostic reversion is common and socially patterned, suggesting contributions from misdiagnosis, clinical uncertainty, or variability in clinical presentation and care setting.
- 23 June 2025
- Published by Oxford University Press (OUP) in American Journal of Epidemiology
- Vol. 194 (9) , 2477-2488
Sociodemographic, health-related, and clinical characteristics and their associations with mortality among All of Us participants compared with the United States general population
The large and diverse All of Us Research Program offers tremendous opportunities for health research. However, results may not be generalizable to the US population due to the program’s targeted recruitment efforts. We compared All of Us participants to those from the nationally representative continuous 1999-2018 National Health and Nutrition Examination Survey (NHANES) with respect to overall mortality rates; the distribution of sociodemographic, health-related, and clinical characteristics; the association of each characteristic with mortality estimated using Cox proportional hazards models; and population attributable fractions (PAFs) for each characteristic and mortality. All of Us participants were older, less likely to be non-Hispanic White, had more years of education, and had a higher prevalence of major chronic conditions than NHANES. Mortality rates were generally lower for All of Us participants, especially at older ages. The direction of associations in All of Us and NHANES matched for almost all comparisons but differed in magnitude for some conditions, primarily clinical diagnoses. PAFs were generally higher in All of Us. Predictors of mortality in All of Us do not consistently generalize to the US population. Analytical approaches are needed to address nonrepresentativeness and mitigate potential biases associated with the selection into the All of Us cohort.
- 23 June 2025
- Published by Oxford University Press (OUP) in American Journal of Epidemiology
- Vol. 194 (9) , 2477-2488
Sociodemographic, health-related, and clinical characteristics and their associations with mortality among All of Us participants compared with the United States general population
The large and diverse All of Us Research Program offers tremendous opportunities for health research. However, results may not be generalizable to the US population due to the program’s targeted recruitment efforts. We compared All of Us participants to those from the nationally representative continuous 1999-2018 National Health and Nutrition Examination Survey (NHANES) with respect to overall mortality rates; the distribution of sociodemographic, health-related, and clinical characteristics; the association of each characteristic with mortality estimated using Cox proportional hazards models; and population attributable fractions (PAFs) for each characteristic and mortality. All of Us participants were older, less likely to be non-Hispanic White, had more years of education, and had a higher prevalence of major chronic conditions than NHANES. Mortality rates were generally lower for All of Us participants, especially at older ages. The direction of associations in All of Us and NHANES matched for almost all comparisons but differed in magnitude for some conditions, primarily clinical diagnoses. PAFs were generally higher in All of Us. Predictors of mortality in All of Us do not consistently generalize to the US population. Analytical approaches are needed to address nonrepresentativeness and mitigate potential biases associated with the selection into the All of Us cohort.
- Open Access
- 26 February 2025
- Published by Wiley in Alzheimer's & Dementia
- Vol. 21 (2) , e14575
Independent associations of high‐density lipoprotein cholesterol and triglyceride levels with Alzheimer's disease and related dementias
INTRODUCTION: We evaluated the independent associations between high‐density lipoprotein cholesterol (HDL‐C) and triglyceride (TG) levels with Alzheimer's disease and related dementias (ADRD). METHODS: Among 177,680 members of Kaiser Permanente Northern California who completed a survey on health risks, we residualized TGs and HDL‐C conditional on age, sex, and body mass index. We included these residuals individually and concurrently in Cox models predicting ADRD incidence. RESULTS: Low (hazard ratio [HR] 1.06, 95% confidence interval [CI] 1.02–1.10) and high quintiles (HR 1.07, 95% CI 1.03–1.12) of HDL‐C residuals were associated with an increased risk of ADRD compared to the middle quintile. Additional adjustment for TGs attenuated the association with high HDL‐C (HR 1.03, 95% CI 0.99–1.08). Low TG residuals were associated with an increased ADRD risk (HR 1.10, 95% CI 1.06–1.15); high TG residuals were protective (HR 0.92, 95% CI 0.88–0.96). These estimates were unaffected by HDL‐C adjustment. DISCUSSION: Low HDL‐C and TG levels are independently associated with increased ADRD risk. The correlation with low TG level explains the association of high HDL‐C with ADRD. Highlights: Strong correlations between lipid levels are important considerations when investigating lipids as late‐life risk factors for Alzheimer's disease and related dementias (ADRD). Low levels of high‐density lipoprotein cholesterol (HDL‐C) and triglycerides (TGs) were independently associated with an increased risk of ADRD. We found no evidence for an association between high HDL‐C and increased ADRD risk after adjustment for TGs. High levels of TGs were consistently associated with a decreased risk of ADRD. There may be interaction between TG and HDL‐C levels, where both low HDL‐C and TG levels increase the risk of ADRD compared to average levels of both.
- 26 February 2025
- Published by Wiley in Alzheimer's & Dementia
- Vol. 21 (2) , e14575
Open AccessIndependent associations of high‐density lipoprotein cholesterol and triglyceride levels with Alzheimer's disease and related dementias
INTRODUCTION: We evaluated the independent associations between high‐density lipoprotein cholesterol (HDL‐C) and triglyceride (TG) levels with Alzheimer's disease and related dementias (ADRD). METHODS: Among 177,680 members of Kaiser Permanente Northern California who completed a survey on health risks, we residualized TGs and HDL‐C conditional on age, sex, and body mass index. We included these residuals individually and concurrently in Cox models predicting ADRD incidence. RESULTS: Low (hazard ratio [HR] 1.06, 95% confidence interval [CI] 1.02–1.10) and high quintiles (HR 1.07, 95% CI 1.03–1.12) of HDL‐C residuals were associated with an increased risk of ADRD compared to the middle quintile. Additional adjustment for TGs attenuated the association with high HDL‐C (HR 1.03, 95% CI 0.99–1.08). Low TG residuals were associated with an increased ADRD risk (HR 1.10, 95% CI 1.06–1.15); high TG residuals were protective (HR 0.92, 95% CI 0.88–0.96). These estimates were unaffected by HDL‐C adjustment. DISCUSSION: Low HDL‐C and TG levels are independently associated with increased ADRD risk. The correlation with low TG level explains the association of high HDL‐C with ADRD. Highlights: Strong correlations between lipid levels are important considerations when investigating lipids as late‐life risk factors for Alzheimer's disease and related dementias (ADRD). Low levels of high‐density lipoprotein cholesterol (HDL‐C) and triglycerides (TGs) were independently associated with an increased risk of ADRD. We found no evidence for an association between high HDL‐C and increased ADRD risk after adjustment for TGs. High levels of TGs were consistently associated with a decreased risk of ADRD. There may be interaction between TG and HDL‐C levels, where both low HDL‐C and TG levels increase the risk of ADRD compared to average levels of both.
- 25 February 2025
- Published by Wolters Kluwer Health in Neurology
- Vol. 104 (4) , e213355
Cholesterol Variability and Dementia Risk
- 25 February 2025
- Published by Wolters Kluwer Health in Neurology
- Vol. 104 (4) , e213355
Cholesterol Variability and Dementia Risk