Current use of estrogen-containing oral contraceptives or hormone therapy and risk of COVID-19 infection and hospitalization: a population-based cohort study.

This population-based cohort study evaluated the association between current use of oral contraceptives (OC) among women under 50 years (n=306,541), and hormone therapy (HT) among women aged 50 or older (n=323,203), and COVID-19 infection and hospitalization. Current OC/HT use was recorded monthly using prescription dispensing data. COVID-19 infections were identified March 2020-February 2021. COVID-19 infection and hospitalization were identified through diagnosis codes and laboratory tests. Weighted generalized estimating equations models estimated multivariable-adjusted odds ratios (aORs) for COVID-19 infection associated with time-varying OC/HT use. Among women with COVID-19, logistic regression models evaluated OC/HT use and COVID-19 hospitalization. Over 12 months, 11,727 (3.8%) women <50 years and 8,661 (2.7%) women ≥50 years experienced COVID-19 infections. There was no evidence of an association between OC use and infection (aOR=1.05; 95%CI: 0.97, 1.12). There was a modest association between HT use and infection (aOR=1.19; 95%CI: 1.03, 1.38). Women using OC had a 39% lower risk of hospitalization (aOR=0.61; 95%CI: 0.38, 1.00), but there was no association of HT use with hospitalization (aOR=0.89; 95%CI: 0.51, 1.53). These findings do not suggest a meaningfully greater risk of COVID-19 infection associated with OC or HT use. OC use may be associated with lower COVID-19 hospitalization risk.

Association Between Diabetes Severity and Risks of COVID-19 Infection and Outcomes.

BACKGROUND: Little is known about whether diabetes increases the risk of COVID-19 infection and whether measures of diabetes severity are related to COVID-19 outcomes. OBJECTIVE: Investigate diabetes severity measures as potential risk factors for COVID-19 infection and COVID-19 outcomes. DESIGN, PARTICIPANTS, MEASURES: In integrated healthcare systems in Colorado, Oregon, and Washington, we identified a cohort of adults on February 29, 2020 (n = 1,086,918) and conducted follow-up through February 28, 2021. Electronic health data and death certificates were used to identify markers of diabetes severity, covariates, and outcomes. Outcomes were COVID-19 infection (positive nucleic acid antigen test, COVID-19 hospitalization, or COVID-19 death) and severe COVID-19 (invasive mechanical ventilation or COVID-19 death). Individuals with diabetes (n = 142,340) and categories of diabetes severity measures were compared with a referent group with no diabetes (n = 944,578), adjusting for demographic variables, neighborhood deprivation index, body mass index, and comorbidities. RESULTS: Of 30,935 patients with COVID-19 infection, 996 met the criteria for severe COVID-19. Type 1 (odds ratio [OR] 1.41, 95% CI 1.27-1.57) and type 2 diabetes (OR 1.27, 95% CI 1.23-1.31) were associated with increased risk of COVID-19 infection. Insulin treatment was associated with greater COVID-19 infection risk (OR 1.43, 95% CI 1.34-1.52) than treatment with non-insulin drugs (OR 1.26, 95% 1.20-1.33) or no treatment (OR 1.24; 1.18-1.29). The relationship between glycemic control and COVID-19 infection risk was dose-dependent: from an OR of 1.21 (95% CI 1.15-1.26) for hemoglobin A1c (HbA1c) < 7% to an OR of 1.62 (95% CI 1.51-1.75) for HbA1c ≥ 9%. Risk factors for severe COVID-19 were type 1 diabetes (OR 2.87; 95% CI 1.99-4.15), type 2 diabetes (OR 1.80; 95% CI 1.55-2.09), insulin treatment (OR 2.65; 95% CI 2.13-3.28), and HbA1c ≥ 9% (OR 2.61; 95% CI 1.94-3.52). CONCLUSIONS: Diabetes and greater diabetes severity were associated with increased risks of COVID-19 infection and worse COVID-19 outcomes.

Investigation of the association of military employment and Parkinson's disease with a validated Parkinson's disease case-finding strategy.

INTRODUCTION: Persons with military involvement may be more likely to have Parkinson's disease (PD) risk factors. As PD is rare, case finding remains a challenge, contributing to our limited understanding of PD risk factors. Here, we explore the validity of case-finding strategies and whether military employment is associated with PD. MATERIALS AND METHODS: We identified Adult Changes in Thought (ACT) study participants reporting military employment as their longest or second longest occupation. We used self-report and prescription fills to identify PD cases and validated this case-finding approach against medical record review. RESULTS: At enrollment, 6% of 5,125 eligible participants had military employment and 1.8% had prevalent PD; an additional 3.5% developed PD over follow-up (mean: 8.3 years). Sensitivity of our case-finding approach was higher for incident (80%) than prevalent cases (54%). Specificity was high (>97%) for both. Military employment was not associated with prevalent PD. Among nonsmokers, point estimates suggested an increased risk of incident PD with military employment, but the result was non-significant and based on a small number of cases. CONCLUSIONS: Self-report and prescription medications can accurately identify incident PD cases relative to the reference method of medical record review. We found no association between military employment and PD.

Association of Lifetime TBI and Military Employment with Late Life ADL Functioning: A Population-Based Prospective Cohort Study.

OBJECTIVE: To determine associations of traumatic brain injury (TBI) and military employment with activities of daily living (ADL) in late life. DESIGN: Population-based prospective cohort study with biennial follow-up and censoring at the time of dementia diagnosis. SETTING: Community-based integrated health care delivery system. PARTICIPANTS: Participants (N=4953) were men (n=2066) and women (n=2887) aged ≥65 years who were dementia free. INTERVENTIONS: Not applicable. MAIN OUTCOME MEASURES: ADL difficulties at baseline and accumulation during follow-up. RESULTS: TBI with loss of consciousness (LOC) before the age of 40 years was associated with slightly higher ADL difficulty at baseline for women (rate ratio [RR], 1.44; 95% confidence interval [CI], 1.08-1.93; P=.01). For men, TBI with LOC at any age was associated with greater ADL difficulty at baseline (age <40y: RR, 1.58; 95% CI, 1.20-2.08; P=.001; age ≥40y: RR, 2.14; 95% CI, 1.24-3.68; P=.006). TBI with LOC was not associated with the rate of accumulation of ADL difficulties over time in men or women. There was no evidence of an association between military employment and either outcome, nor of an interaction between military employment and TBI with LOC. Findings were consistent across a variety of sensitivity analyses. CONCLUSIONS: Further investigation into factors underlying greater late life functional impairment among survivors of TBI is warranted.

Associations between physical function and device-based measures of physical activity and sedentary behavior patterns in older adults: moving beyond moderate-to-vigorous intensity physical activity.

BACKGROUND: Research supports that moderate-to-vigorous intensity physical activity (MVPA) is key to prolonged health and function. Among older adults, substantial changes to MVPA may be infeasible, thus a growing literature suggests a shift in focus to whole-day activity patterns. METHODS: With data from 795 older adults aged 65-100 in the Adult Changes in Thought Activity Monitoring study, we used linear regression to estimate associations between ActiGraph and activPAL measured activity patterns - including light intensity physical activity, steps, standing, and sedentary behaviors - and physical function as measured by a short Performance-based Physical Function (sPPF) score (range 0-12), a composite score based on three standardized physical performance tasks: gait speed, timed chair stands, and grip strength. We examined whether relationships persisted when controlling for MVPA or differed across age, gender, or quartiles of MVPA. RESULTS: In models unadjusted for MVPA, a 1-standard deviation (SD) increment of daily sitting (1.9 h more), mean sitting bout duration (8 min longer average), or time spent in sedentary activity (1.6 h more) was associated with ~ 0.3-0.4 points lower mean sPPF score (all p < 0.05). A 1-SD increment in daily steps (~ 3500 more steps) was associated with ~ 0.5 points higher mean sPPF score (95% CI: 0.22 to 0.73). MVPA adjustment attenuated all relationships. The association between physical function and steps was strongest among adults aged 75+; associations of worse function with greater sedentary behavior were more pronounced in participants with the lowest levels of MVPA. CONCLUSIONS: We found associations between function and activity metrics other than MVPA in key subgroups, findings that support research on broader activity patterns and may offer ideas regarding practical intervention opportunities for improving function in older adults.

Clinical risk prediction models and informative cluster size: Assessing the performance of a suicide risk prediction algorithm.

Clinical visit data are clustered within people, which complicates prediction modeling. Cluster size is often informative because people receiving more care are less healthy and at higher risk of poor outcomes. We used data from seven health systems on 1,518,968 outpatient mental health visits from January 1, 2012 to June 30, 2015 to predict suicide attempt within 90 days. We evaluated true performance of prediction models using a prospective validation set of 4,286,495 visits from October 1, 2015 to September 30, 2017. We examined dividing clustered data on the person or visit level for model training and cross-validation and considered a within cluster resampling approach for model estimation. We evaluated optimism by comparing estimated performance from a left-out testing dataset to performance in the prospective dataset. We used two prediction methods, logistic regression with least absolute shrinkage and selection operator (LASSO) and random forest. The random forest model using a visit-level split for model training and testing was optimistic; it overestimated discrimination (area under the curve, AUC = 0.95 in testing versus 0.84 in prospective validation) and classification accuracy (sensitivity = 0.48 in testing versus 0.19 in prospective validation, 95th percentile cut-off). Logistic regression and random forest models using a person-level split performed well, accurately estimating prospective discrimination and classification: estimated AUCs ranged from 0.85 to 0.87 in testing versus 0.85 in prospective validation, and sensitivity ranged from 0.15 to 0.20 in testing versus 0.17 to 0.19 in prospective validation. Within cluster resampling did not improve performance. We recommend dividing clustered data on the person level, rather than visit level, to ensure strong performance in prospective use and accurate estimation of future performance at the time of model development.

Impact of initiatives to reduce prescription opioid risks on medically attended injuries in people using chronic opioid therapy.

PURPOSE: The purpose of the study is to determine whether initiatives to improve the safety of opioid prescribing decreased injuries in people using chronic opioid therapy (COT). METHODS: We conducted an interrupted time series analysis using data from Group Health (GH), an integrated health care delivery system in the United States. In 2007, GH implemented initiatives which substantially reduced daily opioid dose and increased patient monitoring. Among GH members age 18 or older receiving COT between 2006 and 2014, we compared injury rates for patients in GH's integrated group practice (IGP; exposed to the initiatives) vs patients cared for by contracted providers (not exposed). Injuries were identified using a validated algorithm. We calculated injury incidence during the baseline (preintervention) period from 2006 to 2007; the dose reduction period, 2008 to 2010; and the risk stratification and monitoring period, 2010 to 2014. Using modified Poisson regression, we estimated adjusted relative risks (RRs) representing the relative change per year in injury rates. RESULTS: Among 21 853 people receiving COT in the IGP and 8260 in contracted care, there were 2679 injuries during follow-up. The baseline injury rate was 1.0% per calendar quarter in the IGP and 0.9% in contracted care. Risk reduction initiatives did not decrease injury rates: Within the IGP, the RR in the dose reduction period was 1.01 (95% CI, 0.95-1.07) and in the risk stratification and monitoring period, 0.99 (95% CI, 0.95-1.04). Injury trends did not differ between the two care settings. CONCLUSIONS: Risk reduction initiatives did not decrease injuries in people using COT.

Patterns of antihypertensive and statin adherence prior to dementia: findings from the adult changes in thought study.

BACKGROUND: Detecting patients with undiagnosed dementia is an important clinical challenge. Changes in medication adherence might represent an early sign of cognitive impairment. We sought to examine antihypertensive and statin adherence trajectories in community-dwelling older adults, comparing people who went on to develop dementia to those who did not. METHODS: We analyzed data from Adult Changes in Thought (ACT), a population-based cohort study embedded within an integrated healthcare delivery system. Analyses included 4368 participants aged ≥65 years who had at least one follow-up visit. Research-quality dementia diagnoses were used to identify cases. We selected non-dementia control visits matched on age, sex, and study cohort that occurred at similar ACT follow-up time as the case's dementia onset; we treated this as the index date. Participants were included if they were prevalent users of either a statin or antihypertensive medication on the first day of follow up - 3 years prior to the index date. Using prescription fill dates and days supply, we calculated daily binary medication availability measures for each participant ('days covered') over 3 years leading up to the index date. We used group-based trajectory models to identify patterns of antihypertensive and statin adherence, and used conditional logistic regression to examine associations between adherence trajectories and dementia. RESULTS: Four trajectories were identified for antihypertensive users (292 cases, 3890 control visits), including near perfect (n = 1877, 36.6% cases, 45.5% controls), high (n = 1840, 43.2% cases, 44.1% controls), moderate (n = 365, 18.5% cases, 8.0% controls) and early poor adherence (n = 100, 1.7% cases, 2.4% controls). Odds of dementia was 3 times greater for those with moderate antihypertensive adherence compared to those with near perfect adherence (adjusted OR 3.0, 95% CI 2.0, 4.3). Four trajectories were identified for statin users (148 cases, 1131 control visits), including high (n = 1004, 75.0% cases, 79.0% controls), moderate (n = 192, 19.6% cases, 14.4% controls), early poor (n = 43, 2.0% cases, 3.5% controls), and delayed poor adherence (n = 40, 3.4% cases, 3.1% controls). No association was detected between statin adherence trajectories and dementia. CONCLUSIONS: Patterns of medication adherence may be useful to identify a subset of people at higher likelihood of developing dementia.

Proton pump inhibitor use and the risk of fractures among an older adult cohort.

PURPOSE: The purpose of the study is to determine if the use of a proton pump inhibitor (PPI) is associated with an increased fracture risk, as some prior studies have suggested. METHODS: This retrospective cohort study included data on 4438 participants aged 65 and older who had no fracture in the year prior to baseline and had ≥5 years of enrollment history in Kaiser Permanente Washington, an integrated healthcare delivery system in Seattle, WA, during 1994 to 2014. Time-varying cumulative exposure to PPIs was determined from automated pharmacy data by summing standard daily doses (SDDs) across fills, and patients were categorized as no use (reference group, ≤30 SDD), light use (31-540 SDD), moderate use (541-1080 SDD), and heavy use (≥1081 SDD). Incident fractures were assessed using International Classification of Diseases, Ninth Revision codes from electronic medical records. Potential confounders, chosen a priori, were assessed at baseline and at each 2-year follow-up visit. Fracture risk was analyzed using a Cox proportional hazards model. RESULTS: Over a mean follow-up of 6.1 years, 802 (18.1%) participants experienced a fracture. No overall association was found between PPI use and fracture risk. Adjusted hazard ratios comparing users to the referent category were 1.08 (95% CI 0.83-1.40) for light users, 1.31 (95% CI 0.86-1.95) for moderate users, and 0.95 (95% CI 0.68-1.34) for heavy users. Among patients with SSD > 30, no appreciable increase in fracture risk was present in persons with recent versus distant use (adjusted hazard ratio of 1.14 [95% CI 0.91-1.42]). CONCLUSIONS: No association was observed between PPI use and fracture risk among older adults.

Impact of Opioid Dose Reduction and Risk Mitigation Initiatives on Chronic Opioid Therapy Patients at Higher Risk for Opioid-Related Adverse Outcomes.

Objective: We aimed to determine if opioid risk reduction initiatives including dose reduction and risk mitigation strategies for chronic noncancer pain patients receiving chronic opioid therapy (COT) had a differential impact on average daily opioid doses of COT patients at higher risk for opioid-related adverse outcomes compared with lower-risk patients. Design: Interrupted time series. Setting: Group Health Cooperative (GH), a health care delivery system and insurance within Washington State, between 2006 and 2014. Population: GH enrollees on COT defined as receiving a supply of 70 or more days of opioids within 90 days using electronic pharmacy data for filled prescriptions. Methods: We compared the average daily morphine equivalent doses (MED) of COT patients with and without each of the following higher-risk characteristics: mental disorders, substance use disorders, sedative use, and male gender. Results: In all four pairwise comparisons, the higher-risk subgroup had a higher average daily MED than the lower-risk subgroup across the study period. Adjusted for covariates, modest differences in the annual rate of reduction in average daily MED were noted between higher- and lower-risk subgroups in three pairwise comparisons: those with mental disorders vs without (-8.2 mg/y vs -5.2 mg/y, P = 0.005), with sedative use vs without (-9.2 mg/y vs -5.8 mg/y, P = 0.004); mg), in men vs women (-8.8 mg/y vs -5.9 mg/y, P = 0.01). Conclusion: Using clinical policy initiatives in a health care system, dose reductions were achieved among COT patients at higher risk for opioid-related adverse outcomes that were at least as large as those among lower-risk patients.

Impact of an opioid risk reduction initiative on motor vehicle crash risk among chronic opioid therapy patients.

PURPOSE: Although prescription opioids have been associated with higher motor vehicle crash (MVC) risk, it is unknown whether health system initiatives to better manage chronic opioid therapy (COT) can reduce MVC risk at the population level. METHODS: We conducted an interrupted time series population-level cohort study at Group Health (GH), between January 2006 and September 2014, comparing MVC risk among COT patients who were GH members receiving care in either group practice or contracted care settings. Group practice COT risk reduction initiatives were implemented in two phases: (1) altered prescribing expectations and (2) multifaceted initiatives. These initiatives did not exist in the contracted care network. We compared the adjusted quarterly rate of MVC between group practice and contracted care patients over time using a modified Poisson regression model for a binary outcome. RESULTS: A total of 32 691 COT patients (27.4% from contracted care) met eligibility criteria and experienced a total of 1956 MVCs during study follow-up (mean, 8.1 quarters per person), of which 810 were serious injury crashes. Crash rates were not significantly different between the patient groups within any of the time periods. Analyses stratified by concurrent prescription of a sedative hypnotic or benzodiazepine found no significant difference between the group practice and contracted care patients. There was a modest elevation of MVC risk for high-dose patients relative to former COT patients who stopped receiving opioids. CONCLUSIONS: The risk of MVC was not mitigated in a large cohort of COT patients exposed to a health plan policy initiative that substantially lowered mean opioid dose. Copyright © 2016 John Wiley & Sons, Ltd.

Natural Language Processing to identify pneumonia from radiology reports.

PURPOSE: This study aimed to develop Natural Language Processing (NLP) approaches to supplement manual outcome validation, specifically to validate pneumonia cases from chest radiograph reports. METHODS: We trained one NLP system, ONYX, using radiograph reports from children and adults that were previously manually reviewed. We then assessed its validity on a test set of 5000 reports. We aimed to substantially decrease manual review, not replace it entirely, and so, we classified reports as follows: (1) consistent with pneumonia; (2) inconsistent with pneumonia; or (3) requiring manual review because of complex features. We developed processes tailored either to optimize accuracy or to minimize manual review. Using logistic regression, we jointly modeled sensitivity and specificity of ONYX in relation to patient age, comorbidity, and care setting. We estimated positive and negative predictive value (PPV and NPV) assuming pneumonia prevalence in the source data. RESULTS: Tailored for accuracy, ONYX identified 25% of reports as requiring manual review (34% of true pneumonias and 18% of non-pneumonias). For the remainder, ONYX's sensitivity was 92% (95% CI 90-93%), specificity 87% (86-88%), PPV 74% (72-76%), and NPV 96% (96-97%). Tailored to minimize manual review, ONYX classified 12% as needing manual review. For the remainder, ONYX had sensitivity 75% (72-77%), specificity 95% (94-96%), PPV 86% (83-88%), and NPV 91% (90-91%). CONCLUSIONS: For pneumonia validation, ONYX can replace almost 90% of manual review while maintaining low to moderate misclassification rates. It can be tailored for different outcomes and study needs and thus warrants exploration in other settings.

Association of Military Employment With Late-Life Cognitive Decline and Dementia: A Population-Based Prospective Cohort Study.

INTRODUCTION: As the number of U.S. veterans over age 65 has increased, interest in whether military service affects late-life health outcomes has grown. Whether military employment is associated with increased risk of cognitive decline and dementia remains unclear. MATERIALS AND METHODS: We used data from 4,370 participants of the longitudinal Adult Changes in Thought (ACT) cohort study, enrolled at age 65 or older, to examine whether military employment was associated with greater cognitive decline or higher risk of incident dementia in late life. We classified persons as having military employment if their first or second-longest occupation was with the military. Cognitive status was assessed at each biennial Adult Changes in Thought study visit using the Cognitive Abilities Screening Instrument, scored using item response theory (CASI-IRT). Participants meeting screening criteria were referred for dementia ascertainment involving clinical examination and additional cognitive testing. Primary analyses were adjusted for sociodemographic characteristics and APOE genotype. Secondary analyses additionally adjusted for indicators of early-life socioeconomic status and considered effect modification by age, gender, and prior traumatic brain injury with loss of consciousness TBI with LOC. RESULTS: Overall, 6% of participants had military employment; of these, 76% were males. Military employment was not significantly associated with cognitive change (difference in modeled 10-year cognitive change in CASI-IRT scores in SD units (95% confidence interval [CI]): -0.042 (-0.19, 0.11), risk of dementia (hazard ratio [HR] [95% CI]: 0.92 [0.71, 1.18]), or risk of Alzheimer's disease dementia (HR [95% CI]: 0.93 [0.70, 1.23]). These results were robust to additional adjustment and sensitivity analyses. There was no evidence of effect modification by age, gender, or traumatic brain injury with loss of consciousness. CONCLUSIONS: Among members of the Adult Changes in Thought cohort, military employment was not associated with increased risk of cognitive decline or dementia. Nevertheless, military veterans face the same high risks for cognitive decline and dementia as other aging adults.

Increased COVID-19 Infection Risk Drives Racial and Ethnic Disparities in Severe COVID-19 Outcomes.

COVID-19 inequities have been well-documented. We evaluated whether higher rates of severe COVID-19 in racial and ethnic minority groups were driven by higher infection rates by evaluating if disparities remained when analyses were restricted to people with infection. We conducted a retrospective cohort study of adults insured through Kaiser Permanente (Colorado, Northwest, Washington), follow-up in March-September 2020. Laboratory results and hospitalization diagnosis codes identified individuals with COVID-19. Severe COVID-19 was defined as invasive mechanical ventilation or mortality. Self-reported race and ethnicity, demographics, and medical comorbidities were extracted from health records. Modified Poisson regression estimated adjusted relative risks (aRRs) of severe COVID-19 in full cohort and among individuals with infection. Our cohort included 1,052,774 individuals, representing diverse racial and ethnic minority groups (e.g., 68,887 Asian, 41,243 Black/African American, 93,580 Hispanic or Latino/a individuals). Among 7,399 infections, 442 individuals experienced severe COVID-19. In the full cohort, severe COVID-19 aRRs for Asian, Black/African American, and Hispanic individuals were 2.09 (95% CI: 1.36, 3.21), 2.02 (1.39, 2.93), and 2.09 (1.57, 2.78), respectively, compared to non-Hispanic Whites. In analyses restricted to individuals with COVID-19, all aRRs were near 1, except among Asian Americans (aRR 1.82 [1.23, 2.68]). These results indicate increased incidence of severe COVID-19 among Black/African American and Hispanic individuals is due to higher infection rates, not increased susceptibility to progression. COVID-19 disparities most likely result from social, not biological, factors. Future work should explore reasons for increased severe COVID-19 risk among Asian Americans. Our findings highlight the importance of equity in vaccine distribution.

Diabetic Retinopathy and Dementia Association, Beyond Diabetes Severity.

PURPOSE: To investigate whether associations between diabetic retinopathy (DR) and dementia and Alzheimer's disease (AD) remain significant after controlling for several measures of diabetes severity. DESIGN: Retrospective cohort study. METHODS: Adult Changes in Thought (ACT) is a prospective cohort study of adults aged ≥65 years, randomly selected and recruited from the membership rolls of Kaiser Permanente Washington, who are dementia free at enrollment and followed biennially until incident dementia. The ACT participants were included in this study if they had type 2 diabetes mellitus at enrollment or developed it during follow-up, and data were collected through September, 2018 (3516 person-years of follow-up). Diabetes was defined by ≥ 2 diabetes medication fills in 1 year. Diagnosis of DR was based on International Classification of Diseases Ninth and Tenth Revision codes. Estimates of microalbuminuria, long-term glycemia, and renal function from longitudinal laboratory records were used as indicators of diabetes severity. Alzheimer's disease and dementia were diagnosed using research criteria at expert consensus meetings. RESULTS: A total of 536 participants (median baseline age 75 [interquartile range 71-80], 54% women) met inclusion criteria. Significant associations between DR >5 years duration with dementia (hazard ratio 1.81 [95% CI 1.23, 2.65]) and AD (1.80 [1.15, 2.82]) were not altered by adjustment for estimates of microalbuminuria, long-term glycemia, and renal function (dementia: 1.69 [1.14, 2.50]; AD: 1.73 [1.10, 2.74]). CONCLUSIONS: Among people with type 2 diabetes, DR itself appears to be an important biomarker of dementia risk in addition to glycemia and renal complications.

Association of Traumatic Brain Injury with Late Life Neuropathological Outcomes in a Community-Based Cohort.

BACKGROUND: Prior studies into the association of head trauma with neuropathology have been limited by incomplete lifetime neurotrauma exposure characterization. OBJECTIVE: To investigate the neuropathological sequelae of traumatic brain injury (TBI) in an autopsy sample using three sources of TBI ascertainment, weighting findings to reflect associations in the larger, community-based cohort. METHODS: Self-reported head trauma with loss of consciousness (LOC) exposure was collected in biennial clinic visits from 780 older adults from the Adult Changes in Thought study who later died and donated their brain for research. Self-report data were supplemented with medical record abstraction, and, for 244 people, structured interviews on lifetime head trauma. Neuropathology outcomes included Braak stage, CERAD neuritic plaque density, Lewy body distribution, vascular pathology, hippocampal sclerosis, and cerebral/cortical atrophy. Exposures were TBI with or without LOC. Modified Poisson regressions adjusting for age, sex, education, and APOE ɛ4 genotype were weighted back to the full cohort of 5,546 participants. RESULTS: TBI with LOC was associated with the presence of cerebral cortical atrophy (Relative Risk 1.22, 95% CI 1.02, 1.42). None of the other outcomes was associated with TBI with or without LOC. CONCLUSION: TBI with LOC was associated with increased risk of cerebral cortical atrophy. Despite our enhanced TBI ascertainment, we found no association with the Alzheimer's disease-related neuropathologic outcomes among people who survived to at least age 65 without dementia. This suggests the pathophysiological processes underlying post-traumatic neurodegeneration are distinct from the hallmark pathologies of Alzheimer's disease.

Complex modeling with detailed temporal predictors does not improve health records-based suicide risk prediction.

Suicide risk prediction models can identify individuals for targeted intervention. Discussions of transparency, explainability, and transportability in machine learning presume complex prediction models with many variables outperform simpler models. We compared random forest, artificial neural network, and ensemble models with 1500 temporally defined predictors to logistic regression models. Data from 25,800,888 mental health visits made by 3,081,420 individuals in 7 health systems were used to train and evaluate suicidal behavior prediction models. Model performance was compared across several measures. All models performed well (area under the receiver operating curve [AUC]: 0.794-0.858). Ensemble models performed best, but improvements over a regression model with 100 predictors were minimal (AUC improvements: 0.006-0.020). Results are consistent across performance metrics and subgroups defined by race, ethnicity, and sex. Our results suggest simpler parametric models, which are easier to implement as part of routine clinical practice, perform comparably to more complex machine learning methods.

Angiotensin-converting enzyme inhibitor use and pneumonia risk in community-dwelling older adults: results from a population-based case-control study.

PURPOSE: To test whether angiotensin-converting enzyme (ACE) inhibitor use is associated with decreased risk of community-acquired pneumonia in older adults. METHODS: We analyzed data from a nested case-control study of community-dwelling, immunocompetent adults aged 65-94 within an integrated healthcare delivery system. Cases of ambulatory and hospitalized pneumonia from 2000 to 2003 were identified from International Classification of Disease, version 9, codes and validated using medical record review. Controls were matched to cases by age, sex, and calendar year. Using health plan pharmacy data, we defined current use as filling ≥2 prescriptions during the 180 days prior to the case's diagnosis date. We calculated standardized doses per day using World Health Organization defined daily doses. Multivariable conditional logistic regression estimated adjusted odds ratios (ORs) for pneumonia in relation to ACE inhibitor use, adjusting for comorbidity, functional and cognitive status, and other covariates from medical record review and pharmacy data. RESULTS: Current use of ACE inhibitors was seen in 23% (242/1039) of cases and 21% (433/2022) of controls. Lisinopril accounted for 95% of prescriptions. The OR for pneumonia comparing current use to no current use was 0.99 (95% confidence interval [CI] 0.83-1.19). The OR for use of more than two standardized daily doses per day was 1.39 (95% CI 0.93-2.06) compared to no current use. CONCLUSIONS: ACE inhibitor use is not associated with reduced pneumonia risk in community-dwelling older adults.

Agreement of Step-Based Metrics From ActiGraph and ActivPAL Accelerometers Worn Concurrently Among Older Adults.

Purpose: Our study evaluated the agreement of mean daily step counts, peak 1-min cadence, and peak 30-min cadence between the hip-worn ActiGraph GT3X+ accelerometer, using the normal filter (AGN) and the low frequency extension (AGLFE), and the thigh-worn activPAL3 micro (AP) accelerometer among older adults. Methods: Nine-hundred and fifty-three older adults (≥65 years) were recruited to wear the ActiGraph device concurrently with the AP for 4-7 days beginning in 2016. Using the AP as the reference measure, device agreement for each step-based metric was assessed using mean differences (AGN - AP and AGLFE - AP), mean absolute percentage error (MAPE), and Pearson and concordance correlation coefficients. Results: For AGN - AP, the mean differences and MAPE were: daily steps -1,851 steps/day and 27.2%, peak 1-min cadence -16.2 steps/min and 16.3%, and peak 30-min cadence -17.7 steps/min and 24.0%. Pearson coefficients were .94, .85, and .91 and concordance coefficients were .81, .65, and .73, respectively. For AGLFE - AP, the mean differences and MAPE were: daily steps 4,968 steps/day and 72.7%, peak 1-min cadence -1.4 steps/min and 4.7%, and peak 30-min cadence 1.4 steps/min and 7.0%. Pearson coefficients were .91, .91, and .95 and concordance coefficients were .49, .91, and .94, respectively. Conclusions: Compared with estimates from the AP, the AGN underestimated daily step counts by approximately 1,800 steps/day, while the AGLFE overestimated by approximately 5,000 steps/day. However, peak step cadence estimates generated from the AGLFE and AP had high agreement (MAPE ≤ 7.0%). Additional convergent validation studies of step-based metrics from concurrently worn accelerometers are needed for improved understanding of between-device agreement.

Machine Learning Prediction of Suicide Risk Does Not Identify Patients Without Traditional Risk Factors.

Objective: To determine whether predictions of suicide risk from machine learning models identify unexpected patients or patients without medical record documentation of traditional risk factors.Methods: The study sample included 27,091,382 outpatient mental health (MH) specialty or general medical visits with a MH diagnosis for patients aged 11 years or older from January 1, 2009, to September 30, 2017. We used predicted risk scores of suicide attempt and suicide death, separately, within 90 days of visits to classify visits into risk score percentile strata. For each stratum, we calculated counts and percentages of visits with traditional risk factors, including prior self-harm diagnoses and emergency department visits or hospitalizations with MH diagnoses, in the last 3, 12, and 60 months.Results: Risk-factor percentages increased with predicted risk scores. Among MH specialty visits, 66%, 88%, and 99% of visits with suicide attempt risk scores in the top 3 strata (respectively, 90th-95th, 95th-98th, and ≥ 98th percentiles) and 60%, 77%, and 93% of visits with suicide risk scores in the top 3 strata represented patients who had at least one traditional risk factor documented in the prior 12 months. Among general medical visits, 52%, 66%, and 90% of visits with suicide attempt risk scores in the top 3 strata and 45%, 66%, and 79% of visits with suicide risk scores in the top 3 strata represented patients who had a history of traditional risk factors in the last 12 months.Conclusions: Suicide risk alerts based on these machine learning models coincide with patients traditionally thought of as high-risk at their high-risk visits.