How well can electronic health records from primary care identify Alzheimer's disease cases?

Improving genetic risk modeling of dementia from real-world data in underrepresented populations

Genetic risk modeling for dementia offers significant benefits, but studies based on real-world data, particularly for underrepresented populations, are limited. We employ an Elastic Net model for dementia risk prediction using single-nucleotide polymorphisms prioritized by functional genomic data from multiple neurodegenerative disease genome-wide association studies. We compare this model with APOE and polygenic risk score models across genetic ancestry groups (Hispanic Latino American sample: 610 patients with 126 cases; African American sample: 440 patients with 84 cases; East Asian American sample: 673 patients with 75 cases), using electronic health records from UCLA Health for discovery and the All of Us cohort for validation. Our model significantly outperforms other models across multiple ancestries, improving the area-under-precision-recall curve by 31-84% (Wilcoxon signed-rank test p-value <0.05) and the area-under-the-receiver-operating characteristic by 11-17% (DeLong test p-value <0.05) compared to the APOE and the polygenic risk score models. We identify shared and ancestry-specific risk genes and biological pathways, reinforcing and adding to existing knowledge. Our study highlights the benefits of integrating functional mapping, multiple neurodegenerative diseases, and machine learning for genetic risk models in diverse populations. Our findings hold potential for refining precision medicine strategies in dementia diagnosis.

Challenges and solutions of conducting dementia clinical trials: A palliative care at home pilot for persons with dementia

BACKGROUND

Recruitment and retention are common challenges in clinical trials, particularly with older adults and their caregivers who often benefit from palliative care but have significant strain from caregiving. In recent years, there has been an expansion in home-based palliative care programs, especially for patients with dementia. Because these programs often rely on physicians or advanced practice nurses, they are quite costly and may be difficult to staff due to workforce shortages.

METHODS

We created a novel program of home-based palliative care for patients with advanced dementia and their families, which centers around a community health worker, a social worker, and a nurse. We report on challenges our trial encountered and corresponding solutions.

RESULTS

We enrolled 30 patients and their 30 caregivers in our pilot trial of home-based palliative care. We found two significant barriers to enrollment: (1) the electronic health record was insufficient to determine the severity of patients' dementia; and (2) rates of follow-up survey completion were low, with completion rates at 6 months between 14 and 44%. We created an iterative training process to determine dementia severity from electronic health records and applied person-centered approaches to improve survey completion.

CONCLUSIONS

Electronic health records are not set up to include discrete fields for dementia severity, which makes enrollment of older adults with dementia in a clinical trial challenging. The strain of caring for a loved one with advanced dementia may also make participation in health-services research difficult for patients and their families. Novel approaches have the potential to counteract these challenges, improve recruitment and retention, and ultimately improve care for people with dementia and their caregivers.

Sex matters in the association between cardiovascular health and incident dementia: evidence from real world data

BACKGROUND

Cardiovascular health has been associated with dementia onset, but little is known about the variation of such association by sex and age considering dementia subtypes. We assessed the role of sex and age in the association between cardiovascular risk and the onset of all-cause dementia, Alzheimer's disease, and vascular dementia in people aged 50-74 years.

METHODS

This is a retrospective cohort study covering 922.973 Catalans who attended the primary care services of the Catalan Health Institute (Spain). Data were obtained from the System for the Development of Research in Primary Care (SIDIAP database). Exposure was the cardiovascular risk (CVR) at baseline categorized into four levels of Framingham-REGICOR score (FRS): low (FRS < 5%), low-intermediate (5% ≤ FRS < 7.5%), high-intermediate (7.5% ≤ FRS < 10%), high (FRS ≥ 10%), and one group with previous vascular disease. Cases of all-cause dementia and Alzheimer's disease were identified using validated algorithms, and cases of vascular dementia were identified by diagnostic codes. We fitted stratified Cox models using age parametrized as b-Spline.

RESULTS

A total of 51,454 incident cases of all-cause dementia were recorded over a mean follow-up of 12.7 years. The hazard ratios in the low-intermediate and high FRS groups were 1.12 (95% confidence interval: 1.08-1.15) and 1.55 (1.50-1.60) for all-cause dementia; 1.07 (1.03-1.11) and 1.17 (1.11-1.24) for Alzheimer's disease; and 1.34 (1.21-1.50) and 1.90 (1.67-2.16) for vascular dementia. These associations were stronger in women and in midlife compared to later life in all dementia types. Women with a high Framingham-REGICOR score presented a similar risk of developing dementia - of any type - to women who had previous vascular disease, and at age 50-55, they showed three times higher risk of developing dementia risk compared to the lowest Framingham-REGICOR group.

CONCLUSIONS

We found a dose‒response association between the Framingham-REGICOR score and the onset of all dementia types. Poor cardiovascular health in midlife increased the onset of all dementia types later in life, especially in women.

Improving genetic risk modeling of dementia from real-world data in underrepresented populations

BACKGROUND

Genetic risk modeling for dementia offers significant benefits, but studies based on real-world data, particularly for underrepresented populations, are limited.

METHODS

We employed an Elastic Net model for dementia risk prediction using single-nucleotide polymorphisms prioritized by functional genomic data from multiple neurodegenerative disease genome-wide association studies. We compared this model with APOE and polygenic risk score models across genetic ancestry groups, using electronic health records from UCLA Health for discovery and All of Us cohort for validation.

RESULTS

Our model significantly outperforms other models across multiple ancestries, improving the area-under-precision-recall curve by 21-61% and the area-under-the-receiver-operating characteristic by 10-21% compared to the APOEand the polygenic risk score models. We identified shared and ancestry-specific risk genes and biological pathways, reinforcing and adding to existing knowledge.

CONCLUSIONS

Our study highlights benefits of integrating functional mapping, multiple neurodegenerative diseases, and machine learning for genetic risk models in diverse populations. Our findings hold potential for refining precision medicine strategies in dementia diagnosis.

Improving genetic risk modeling of dementia from real-world data in underrepresented populations

BACKGROUND

Genetic risk modeling for dementia offers significant benefits, but studies based on real-world data, particularly for underrepresented populations, are limited.

METHODS

We employed an Elastic Net model for dementia risk prediction using single-nucleotide polymorphisms prioritized by functional genomic data from multiple neurodegenerative disease genome-wide association studies. We compared this model with APOE and polygenic risk score models across genetic ancestry groups, using electronic health records from UCLA Health for discovery and All of Us cohort for validation.

RESULTS

Our model significantly outperforms other models across multiple ancestries, improving the area-under-precision-recall curve by 21-61% and the area-under-the-receiver-operating characteristic by 10-21% compared to the APOE and the polygenic risk score models. We identified shared and ancestry-specific risk genes and biological pathways, reinforcing and adding to existing knowledge.

CONCLUSIONS

Our study highlights benefits of integrating functional mapping, multiple neurodegenerative diseases, and machine learning for genetic risk models in diverse populations. Our findings hold potential for refining precision medicine strategies in dementia diagnosis.

Sildenafil as a Candidate Drug for Alzheimer's Disease: Real-World Patient Data Observation and Mechanistic Observations from Patient-Induced Pluripotent Stem Cell-Derived Neurons

Background

Alzheimer's disease (AD) is a chronic neurodegenerative disease needing effective therapeutics urgently. Sildenafil, one of the approved phosphodiesterase-5 inhibitors, has been implicated as having potential effect in AD.

Objective

To investigate the potential therapeutic benefit of sildenafil on AD.

Methods

We performed real-world patient data analysis using the MarketScan® Medicare Supplemental and the Clinformatics® databases. We conducted propensity score-stratified analyses after adjusting confounding factors (i.e., sex, age, race, and comorbidities). We used both familial and sporadic AD patient induced pluripotent stem cells (iPSC) derived neurons to evaluate the sildenafil's mechanism-of-action.

Results

We showed that sildenafil usage is associated with reduced likelihood of AD across four new drug compactor cohorts, including bumetanide, furosemide, spironolactone, and nifedipine. For instance, sildenafil usage is associated with a 54% reduced incidence of AD in MarketScan® (hazard ratio [HR] = 0.46, 95% CI 0.32- 0.66) and a 30% reduced prevalence of AD in Clinformatics® (HR = 0.70, 95% CI 0.49- 1.00) compared to spironolactone. We found that sildenafil treatment reduced tau hyperphosphorylation (pTau181 and pTau205) in a dose-dependent manner in both familial and sporadic AD patient iPSC-derived neurons. RNA-sequencing data analysis of sildenafil-treated AD patient iPSC-derived neurons reveals that sildenafil specifically target AD related genes and pathobiological pathways, mechanistically supporting the beneficial effect of sildenafil in AD.

Conclusions

These real-world patient data validation and mechanistic observations from patient iPSC-derived neurons further suggested that sildenafil is a potential repurposable drug for AD. Yet, randomized clinical trials are warranted to validate the causal treatment effects of sildenafil in AD.

Early prediction of Alzheimer's disease and related dementias using real-world electronic health records

INTRODUCTION

This study aims to explore machine learning (ML) methods for early prediction of Alzheimer's disease (AD) and related dementias (ADRD) using the real-world electronic health records (EHRs).

METHODS

A total of 23,835 ADRD and 1,038,643 control patients were identified from the OneFlorida+ Research Consortium. Two ML methods were used to develop the prediction models. Both knowledge-driven and data-driven approaches were explored. Four computable phenotyping algorithms were tested.

RESULTS

The gradient boosting tree (GBT) models trained with the data-driven approach achieved the best area under the curve (AUC) scores of 0.939, 0.906, 0.884, and 0.854 for early prediction of ADRD 0, 1, 3, or 5 years before diagnosis, respectively. A number of important clinical and sociodemographic factors were identified.

DISCUSSION

We tested various settings and showed the predictive ability of using ML approaches for early prediction of ADRD with EHRs. The models can help identify high-risk individuals for early informed preventive or prognostic clinical decisions.

Population-based discovery and Mendelian randomization analysis identify telmisartan as a candidate medicine for Alzheimer's disease in African Americans

INTRODUCTION

African Americans (AAs) and European Americans (EAs) differ in Alzheimer's disease (AD) prevalence, risk factors, and symptomatic presentation and AAs are less likely to enroll in AD clinical trials.

METHODS

We conducted race-conscious pharmacoepidemiologic studies of 5.62 million older individuals (age ≥60) to investigate the association of telmisartan exposure and AD outcome using Cox analysis, Kaplan-Meier analysis, and log-rank test. We performed Mendelian randomization (MR) analysis of large ethnically diverse genetic data to test likely causal relationships between telmisartan's target and AD.

RESULTS

We identified that moderate/high telmisartan exposure was significantly associated with a reduced incidence of AD in the AAs compared to low/no telmisartan exposure (hazard ratio [HR] = 0.77, 95% CI: 0.65-0.91, p-value = 0.0022), but not in the non-Hispanic EAs (HR = 0.97, 95% CI: 0.89-1.05, p-value = 0.4110). Sensitivity and sex-/age-stratified patient subgroup analyses identified that telmisartan's medication possession ratio (MPR) and average hypertension daily dosage were significantly associated with a stronger reduction in the incidence of both AD and dementia in AAs. Using MR analysis from large genome-wide association studies (GWAS) (over 2 million individuals) across AD, hypertension, and diabetes, we further identified AA-specific beneficial effects of telmisartan for AD.

DISCUSSION

Randomized controlled trials with ethnically diverse patient cohorts are warranted to establish causality and therapeutic outcomes of telmisartan and AD.

HIGHLIGHTS

Telmisartan is associated with lower risk of Alzheimer's disease (AD) in African Americans (AAs). Telmisartan is the only angiotensin II receptor blockers having PPAR-γ agonistic properties with beneficial anti-diabetic and renal function effects, which mitigate AD risk in AAs. Mendelian randomization (MR) analysis demonstrates the specificity of telmisartan's protective mechanism to AAs.

Assess the documentation of cognitive tests and biomarkers in electronic health records via natural language processing for Alzheimer's disease and related dementias

BACKGROUND

Cognitive tests and biomarkers are the key information to assess the severity and track the progression of Alzheimer's' disease (AD) and AD-related dementias (AD/ADRD), yet, both are often only documented in clinical narratives of patients' electronic health records (EHRs). In this work, we aim to (1) assess the documentation of cognitive tests and biomarkers in EHRs that can be used as real-world endpoints, and (2) identify, extract, and harmonize the different commonly used cognitive tests from clinical narratives using natural language processing (NLP) methods into categorical AD/ADRD severity.

METHODS

We developed a rule-based NLP pipeline to extract the cognitive tests and biomarkers from clinical narratives in AD/ADRD patients' EHRs. We aggregated the extracted results to the patient level and harmonized the cognitive test scores into severity categories using cutoffs determined based on both relevant literature and domain knowledge of AD/ADRD clinicians.

RESULTS

We identified an AD/ADRD cohort of 48,912 patients from the University of Florida (UF) Health system and identified 7 measurements (6 cognitive tests and 1 biomarker) that are frequently documented in our data. Our NLP pipeline achieved an overall F1-score of 0.9059 across the 7 measurements. Among the 6 cognitive tests, we were able to harmonize 4 cognitive test scores into severity categories, and the population characteristics of patients with different severity were described. We also identified several factors related to the availability of their documentation in EHRs.

CONCLUSION

This study demonstrates that our NLP pipelines can extract cognitive tests and biomarkers of AD/ADRD accurately for downstream studies. Although, the documentation of cognitive tests and biomarkers in EHRs appears to be low, RWD is still an important resource for AD/ADRD research. Nevertheless, providing standardized approach to document cognitive tests and biomarkers in EHRS are also warranted.

One-year prediction of cognitive decline following cognitive-stimulation from real-world data

Clinical evidence based on real-world data (RWD) is accumulating exponentially providing larger sample sizes available, which demand novel methods to deal with the enhanced heterogeneity of the data. Here, we used RWD to assess the prediction of cognitive decline in a large heterogeneous sample of participants being enrolled with cognitive stimulation, a phenomenon that is of great interest to clinicians but that is riddled with difficulties and limitations. More precisely, from a multitude of neuropsychological Training Materials (TMs), we asked whether was possible to accurately predict an individual's cognitive decline one year after being tested. In particular, we performed longitudinal modelling of the scores obtained from 215 different tests, grouped into 29 cognitive domains, a total of 124,610 instances from 7902 participants (40% male, 46% female, 14% not indicated), each performing an average of 16 tests. Employing a machine learning approach based on ROC analysis and cross-validation techniques to overcome overfitting, we show that different TMs belonging to several cognitive domains can accurately predict cognitive decline, while other domains perform poorly, suggesting that the ability to predict decline one year later is not specific to any particular domain, but is rather widely distributed across domains. Moreover, when addressing the same problem between individuals with a common diagnosed label, we found that some domains had more accurate classification for conditions such as Parkinson's disease and Down syndrome, whereas they are less accurate for Alzheimer's disease or multiple sclerosis. Future research should combine similar approaches to ours with standard neuropsychological measurements to enhance interpretability and the possibility of generalizing across different cohorts.

Real-world data in primary care: validation of diagnosis of atrial fibrillation in primary care electronic medical records and estimated prevalence among consulting patients'

BACKGROUND

Primary care electronic medical records contain clinical-administrative information on a high percentage of the population. Before this information can be used for epidemiological purposes, its quality must be verified. This study aims to validate diagnoses of atrial fibrillation (AF) recorded in primary care electronic medical records and to estimate the prevalence of AF in the population attending primary care consultations.

METHODS

We performed a cross-sectional validation study of all diagnoses of AF recorded in primary care electronic medical records in Madrid (Spain). We also performed simple random sampling of diagnoses of AF (ICPC-2 code K78) registered by 55 physicians and random age- and sex-matched sampling of the records that included a diagnosis of AF. Electrocardiograms, echocardiograms, and hospital discharge or cardiology clinic reports were matched. Sensitivity, specificity, positive and negative predictive values (PPV and NPV), and overall agreement were calculated using the kappa statistic (κ). The prevalence of AF in the community of Madrid was estimated considering the sensitivity and specificity obtained in the validation. All calculations were performed overall and by sex and age groups.

RESULTS

The degree of agreement was very high (κ = 0.952), with a sensitivity of 97.84%, specificity of 97.39%, PPV of 97.37%, and NPV of 97.85%. The prevalence of AF in the population aged over 18 years was 2.41% (95%CI 2.39-2.42% [2.25% in women and 2.58% in men]). This increased progressively with age, reaching 16.95% in those over 80 years of age (15.5% in women and 19.44% in men).

CONCLUSIONS

The validation results obtained enable diagnosis of AF recorded in primary care to be used as a tool for epidemiological studies. A high prevalence of AF was found, especially in older patients.

Risk of Parkinson's disease in a gout Mediterranean population: A case-control study

INTRODUCTION

High levels of serum urate has been associated to a neuroprotective effect in Parkinson's disease (PD) as an antioxidant agent. However, the relation between gout and PD remains contradictory.

OBJECTIVE

To study if the neuroprotective effect of serum urate is maintained in patients with gout in a large urban Mediterranean population.

METHODS

Primary care based matched case-control study, carried out using an electronic health record database from the public primary care health system of Barcelona. The database contains anonymous data from 1,520,934 patients. All patients, over 40 years old, with a new diagnostic record of PD, or a new prescription of dopaminergic drugs were included (incident cases). We randomly selected four controls for each case, matched by gender and age, with the frequency matching approach. Retrospective data of PD risk factors were also collected for each individual. A multivariate logistic regression model was used to evaluate the association of gout and PD, adjusted by the presence of other risk factors.

RESULTS

A new PD diagnosis was found in 17,629 individuals (incident diagnosis rate of 2.2 per 1000 individuals). Multivariate logistic regression model showed for gout: aOR=0.83 (0.76-0.91). When stratified by age, aOR for those under 75years was 0.99 (0.85-1.16) and 75 or over OR=0.77 (0.70-0.86). Dyslipidemia, hypertension and diabetes mellitus were associated with an increased risk of PD. Tobacco consumption was protective.

CONCLUSION

Our study, the first one made in a Mediterranean population, shows a PD protective effect of gout in both men and women over 75years old.

How good are medical and death records for identifying dementia?

INTRODUCTION

Retrospective studies using administrative data may be an efficient way to assess risk factors for dementia if diagnostic accuracy is known.

METHODS

Within-individual clinical diagnoses of Alzheimer's disease (AD) and all-cause dementia in ambulatory (outpatient) surgery, inpatient, Medicare administrative records and death certificates were compared with research diagnoses among participants of Cache County Study on Memory, Health, and Aging (CCSMHA) (1995-2008, N = 5092).

RESULTS

Combining all sources of clinical health data increased sensitivity for identifying all-cause dementia (71%) and AD (48%), while maintaining relatively high specificity (81% and 93%, respectively). Medicare claims had the highest sensitivity for case identification (57% and 40%, respectively).

DISCUSSION

Administrative health data may provide a less accurate method than a research evaluation for identifying individuals with dementing disease, but accuracy is improved by combining health data sources. Assessing all-cause dementia versus a specific cause of dementia such as AD will result in increased sensitivity, but at a cost to specificity.

Dementia in health claims data: The influence of different case definitions on incidence and prevalence estimates

OBJECTIVES

The epidemiology of dementia subtypes including Alzheimer's disease (AD) and vascular dementia (VD) and their reliance on different case definitions ("algorithms") in health claims data are still understudied.

METHODS

Based on health claims data, prevalence estimates (per 100 persons), incidence rates (IRs, per 100 person-years), and proportions of AD, VD, and other dementias (oD) were calculated. Five algorithms of increasing strictness considered inpatient/outpatient diagnoses (#1, #2), antidementia drugs (#3) or supportive diagnostics (#4, #5).

RESULTS

Algorithm 1 detected 213,409 cases (#2: 197,400; #3: 48,688; #4: 3033; #5: 3105), a prevalence for any dementia of 3.44 and an IR of 1.39 (AD: 0.80/0.21, VD: 0.79/0.31). The prevalence decreased by algorithms for any dementia (#2: 3.19; #3: 0.75; #4: 0.04; #5: 0.05) as did IRs (#2: 1.13; #3: 0.18; #4: 0.05, #5: 0.05). Algorithms 1-2, and 4-5 revealed similar proportions of AD (23.3%-26.6%), VD (19.9%-23.2%), and oD (53.1%-53.8%), algorithm 3 estimated 45% (AD), 12.1% (VD), and 43.0% (oD).

CONCLUSIONS

Health claims data show lower estimates of AD than previously reported, due to markedly lower prevalent/incident proportions of patients with corresponding codes. Using medication in defining dementia potentially improves estimating the proportion of AD while supportive diagnostics were of limited use.

Identification of Major Cognitive Disorders in Self-Reported versus Administrative Health Data: A Cohort Study in Quebec

BACKGROUND

The first imperative in producing the relevant and needed knowledge about major neurocognitive disorder (MNCD) is to identify people presenting with the condition adequately. To document potential disparities between administrative health databases and population-based surveys could help identify specific challenges in this population and methodological shortfalls.

OBJECTIVE

To describe and compare the characteristics of community-dwelling older adults according to four groups: 1) No MNCD; 2) Self-reported MNCD only; 3) MNCD in administrative health data only; 4) MNCD in both self-reported and administrative health data.

METHODS

This retrospective cohort study used the Care Trajectories-Enriched Data (TorSaDE) cohort, a linkage between five waves of the Canadian Community Health Survey (CCHS) and health administrative health data. We included older adults living in the community who participated in at least one cycle of the CCHS. We reported on positive and negative MNCD in self-reported versus administrative health data. We then compared groups' characteristics using chi-square tests and ANOVA.

RESULTS

The study cohort was composed of 25,125 older adults, of which 784 (3.12%) had MNCD. About 70% of people with an MNCD identified in administrative health data did not report it in the CCHS. The four groups present specific challenges related to the importance of perception, timely diagnosis, and the caregivers' roles in reporting health information.

CONCLUSION

To a certain degree, both data sources fail to consider subgroups experiencing issues related to MNCD; studies like ours provide insight to understand their characteristics and needs better.

Prehospital care for ovarian cancer in Catalonia: could we do better in primary care? Retrospective cohort study

OBJECTIVE

To assess the impact of prehospital factors (diagnostic pathways, first presentation to healthcare services, intervals, participation in primary care) on 1-year and 5-year survival in people with epithelial ovarian cancer (EOC).

DESIGN

Retrospective quasi-population-based cohort study.

SETTING

Catalan Integrated Public Healthcare System.

PARTICIPANTS

People with EOC who underwent surgery with a curative intent in public Catalan hospitals between 1 January 2013 and 31 December 2014.

OUTCOME MEASURES

Data from primary and secondary care clinical histories and care processes in the 18 months leading up to confirmation (signs and symptoms at presentation, diagnosis pathways, referrals, diagnosis interval) of the EOC diagnosis (stage, histology type, treatment). Diagnostic process intervals were based on the Aarhus statement. 1-year and 5-year survival analysis was undertaken.

RESULTS

Of the 513 patients included in the cohort, 67.2% initially consulted their family physician, while 36.4% were diagnosed through emergency services. In the Cox models, survival was influenced by advanced stage at 1 year (HR 3.84, 95% CI 1.23 to 12.02) and 5 years (HR 5.36, 95% CI 3.07 to 9.36), as was the type of treatment received, although this association was attenuated over follow-up. Age became significant at 5 years of follow-up. After adjusting for age, adjusted morbidity groups, stage at diagnosis and treatment, 5-year survival was better in patients presenting with gynaecological bleeding (HR 0.35, 95% CI 0.16 to 0.79). Survival was not associated with a starting point involving primary care (HR 1.39, 95% CI 0.93 to 2.09), diagnostic pathways involving referral to elective gynaecological care from non-general practitioners (HR 0.80, 95% CI 0.51 to 1.26), or self-presentation to emergency services (HR 0.82, 95% CI 0.52 to 1.31).

CONCLUSIONS

Survival in EOC is not associated with diagnostic pathways or prehospital healthcare, but it is influenced by stage at diagnosis, administration of primary cytoreduction plus chemotherapy and patient age.

Traumatic Brain Injury and Risk of Alzheimer's Disease and Related Dementias in the Population

BACKGROUND

Epidemiological studies examining associations between traumatic brain injury (TBI) and Alzheimer's disease and related dementias (ADRD) have yielded conflicting results, which may be due to methodological differences.

OBJECTIVE

To examine the relationship between the presence and severity of TBI and risk of ADRD using a population-based cohort with medical record abstraction for confirmation of TBI and ADRD.

METHODS

All TBI events among Olmsted County, Minnesota residents aged > 40 years from 1985-1999 were confirmed by manual review and classified by severity. Each TBI case was randomly matched to two age-, sex-, and non-head injury population-based referents without TBI. For TBI events with non-head trauma, the Trauma Mortality Prediction Model was applied to assign an overall measure of non-head injury severity and corresponding referents were matched on this variable. Medical records were manually abstracted to confirm ADRD diagnosis. Cox proportional hazards models examined the relationship between TBI and severity with risk of ADRD.

RESULTS

A total of 1,418 residents had a confirmed TBI (865 Possible, 450 Probable, and 103 Definite) and were matched to 2,836 referents. When combining all TBI severities, the risk of any ADRD was significantly higher for those with a confirmed TBI compared to referents (HR = 1.32, 95% CI: 1.11, 1.58). Stratifying by TBI severity, Probable (HR = 1.42, 95% CI: 1.05, 1.92) and Possible (HR = 1.29, 95% CI: 1.02-1.62) TBI was associated with an increased risk of ADRD, but not Definite TBI (HR = 1.22, 95% CI: 0.68, 2.18).

CONCLUSION

Our analyses support including TBI as a potential risk factor for developing ADRD.

Prevalence and incidence of young onset dementia and associations with comorbidities: A study of data from the French national health data system

BACKGROUND

Dementia onset in those aged <65 years (young onset dementia, YOD) has dramatic individual and societal consequences. In the context of population aging, data on YOD are of major importance to anticipate needs for planning and allocation of health and social resources. Few studies have provided precise frequency estimates of YOD. The aim of this study is to provide YOD prevalence and incidence estimates in France and to study the contribution of comorbidities to YOD incidence.

METHODS AND FINDINGS

Using data from the French national health data system (Système National des Données de Santé, SNDS) for 76% of the French population aged 40 to 64 years in 2016 (n = 16,665,795), we identified all persons with dementia based on at least 1 of 3 criteria: anti-Alzheimer drugs claims, hospitalization with the International Classification of Diseases-10th Revision (ICD-10) dementia codes (F00 to F03, G30, G31.0, G31.1, or F05.1), or registration for free healthcare for dementia. We estimated prevalence rate (PR) and incidence rate (IR) and estimated the association of comorbidities with incident YOD. Sex differences were investigated. We identified 18,466 (PRstandardized = 109.7/100,000) and 4,074 incident (IRstandardized = 24.4/100,000 person-years) persons with prevalent and incident YOD, respectively. PR and IR sharply increased with age. Age-adjusted PR and IR were 33% (95% confidence interval (CI) = 29 to 37) and 39% (95% CI = 31 to 48) higher in men than women (p < 0.001 both for PR and IR). Cardio- and cerebrovascular, neurological, psychiatric diseases, and traumatic brain injury prevalence were associated with incident YOD (age- and sex-adjusted p-values <0.001 for all comorbidities examined, except p = 0.109 for antihypertensive drug therapy). Adjustment for all comorbidities explained more than 55% of the sex difference in YOD incidence. The lack of information regarding dementia subtypes is the main limitation of this study.

CONCLUSIONS

We estimated that there were approximately 24,000 and approximately 5,300 persons with prevalent and incident YOD, respectively, in France in 2016. The higher YOD frequency in men may be partly explained by higher prevalence of cardiovascular and neurovascular diseases, substance abuse disorders, and traumatic brain injury and warrants further investigation.

Development and Validation of a Model to Identify Alzheimer's Disease and Related Syndromes in Administrative Data

BACKGROUND

Administrative data are used in the field of Alzheimer's Disease and Related Syndromes (ADRS), however their performance to identify ADRS is unknown.

OBJECTIVE

i) To develop and validate a model to identify ADRS prevalent cases in French administrative data (SNDS), ii) to identify factors associated with false negatives.

METHODS

Retrospective cohort of subjects ≥ 65 years, living in South-Western France, who attended a memory clinic between April and December 2013. Gold standard for ADRS diagnosis was the memory clinic specialized diagnosis. Memory clinics' data were matched to administrative data (drug reimbursements, diagnoses during hospitalizations, registration with costly chronic conditions). Prediction models were developed for 1-year and 3-year periods of administrative data using multivariable logistic regression models. Overall model performance, discrimination, and calibration were estimated and corrected for optimism by resampling. Youden index was used to define ADRS positivity and to estimate sensitivity, specificity, positive predictive and negative probabilities. Factors associated with false negatives were identified using multivariable logistic regressions.

RESULTS

3360 subjects were studied, 52% diagnosed with ADRS by memory clinics. Prediction model based on age, all-cause hospitalization, registration with ADRS as a chronic condition, number of anti-dementia drugs, mention of ADRS during hospitalizations had good discriminative performance (c-statistic: 0.814, sensitivity: 76.0%, specificity: 74.2% for 2013 data). 419 false negatives (24.0%) were younger, had more often ADRS types other than Alzheimer's disease, moderate forms of ADRS, recent diagnosis, and suffered from other comorbidities than true positives.

CONCLUSION

Administrative data presented acceptable performance for detecting ADRS. External validation studies should be encouraged.

Electronic Medical Record-Based Case Phenotyping for the Charlson Conditions: Scoping Review

Background

Electronic medical records (EMRs) contain large amounts of rich clinical information. Developing EMR-based case definitions, also known as EMR phenotyping, is an active area of research that has implications for epidemiology, clinical care, and health services research.

Objective

This review aims to describe and assess the present landscape of EMR-based case phenotyping for the Charlson conditions.

Methods

A scoping review of EMR-based algorithms for defining the Charlson comorbidity index conditions was completed. This study covered articles published between January 2000 and April 2020, both inclusive. Embase (Excerpta Medica database) and MEDLINE (Medical Literature Analysis and Retrieval System Online) were searched using keywords developed in the following 3 domains: terms related to EMR, terms related to case finding, and disease-specific terms. The manuscript follows the Preferred Reporting Items for Systematic reviews and Meta-analyses extension for Scoping Reviews (PRISMA) guidelines.

Results

A total of 274 articles representing 299 algorithms were assessed and summarized. Most studies were undertaken in the United States (181/299, 60.5%), followed by the United Kingdom (42/299, 14.0%) and Canada (15/299, 5.0%). These algorithms were mostly developed either in primary care (103/299, 34.4%) or inpatient (168/299, 56.2%) settings. Diabetes, congestive heart failure, myocardial infarction, and rheumatology had the highest number of developed algorithms. Data-driven and clinical rule–based approaches have been identified. EMR-based phenotype and algorithm development reflect the data access allowed by respective health systems, and algorithms vary in their performance.

Conclusions

Recognizing similarities and differences in health systems, data collection strategies, extraction, data release protocols, and existing clinical pathways is critical to algorithm development strategies. Several strategies to assist with phenotype-based case definitions have been proposed.

Filling the gaps in the characterization of the clinical management of COVID-19: 30-day hospital admission and fatality rates in a cohort of 118 150 cases diagnosed in outpatient settings in Spain

BACKGROUND

Currently, there is a missing link in the natural history of COVID-19, from first (usually milder) symptoms to hospitalization and/or death. To fill in this gap, we characterized COVID-19 patients at the time at which they were diagnosed in outpatient settings and estimated 30-day hospital admission and fatality rates.

METHODS

This was a population-based cohort study.

Data were obtained from Information System for Research in Primary Care (SIDIAP)-a primary-care records database covering >6 million people (>80% of the population of Catalonia), linked to COVID-19 reverse transcriptase polymerase chain reaction (RT-PCR) tests and hospital emergency, inpatient and mortality registers. We included all patients in the database who were ≥15 years old and diagnosed with COVID-19 in outpatient settings between 15 March and 24 April 2020 (10 April for outcome studies). Baseline characteristics included socio-demographics, co-morbidity and previous drug use at the time of diagnosis, and polymerase chain reaction (PCR) testing and results.

Study outcomes included 30-day hospitalization for COVID-19 and all-cause fatality.

RESULTS

We identified 118 150 and 95 467 COVID-19 patients for characterization and outcome studies, respectively. Most were women (58.7%) and young-to-middle-aged (e.g. 21.1% were 45-54 years old). Of the 44 575 who were tested with PCR, 32 723 (73.4%) tested positive. In the month after diagnosis, 14.8% (14.6-15.0) were hospitalized, with a greater proportion of men and older people, peaking at age 75-84 years. Thirty-day fatality was 3.5% (95% confidence interval: 3.4% to 3.6%), higher in men, increasing with age and highest in those residing in nursing homes [24.5% (23.4% to 25.6%)].

CONCLUSION

COVID-19 infections were widespread in the community, including all age-sex strata. However, severe forms of the disease clustered in older men and nursing-home residents. Although initially managed in outpatient settings, 15% of cases required hospitalization and 4% died within a month of first symptoms. These data are instrumental for designing deconfinement strategies and will inform healthcare planning and hospital-bed allocation in current and future COVID-19 outbreaks.

Excess cases of influenza and the coronavirus epidemic in Catalonia: a time-series analysis of primary-care electronic medical records covering over 6 million people

OBJECTIVES

There is uncertainty about when the first cases of COVID-19 appeared in Spain. We aimed to determine whether influenza diagnoses masked early COVID-19 cases and estimate numbers of undetected COVID-19 cases.

DESIGN

Time-series study of influenza and COVID-19 cases, 2010-2020.

SETTING

Primary care, Catalonia, Spain.

PARTICIPANTS

People registered in primary-care practices, covering >6 million people and >85% of the population.

MAIN OUTCOME MEASURES

Weekly new cases of influenza and COVID-19 clinically diagnosed in primary care.

ANALYSES

Daily counts of both cases were computed using the total cases recorded over the previous 7 days to avoid weekly effects. Epidemic curves were characterised for the 2010-2011 to 2019-2020 influenza seasons. Influenza seasons with a similar epidemic curve and peak case number as the 2019-2020 season were used to model expected case numbers with Auto Regressive Integrated Moving Average models, overall and stratified by age. Daily excess influenza cases were defined as the number of observed minus expected cases.

RESULTS

Four influenza season curves (2011-2012, 2012-2013, 2013-2014 and 2016-2017) were used to estimate the number of expected cases of influenza in 2019-2020. Between 4 February 2020 and 20 March 2020, 8017 (95% CI: 1841 to 14 718) excess influenza cases were identified. This excess was highest in the 15-64 age group.

CONCLUSIONS

COVID-19 cases may have been present in the Catalan population when the first imported case was reported on 25 February 2020. COVID-19 carriers may have been misclassified as influenza diagnoses in primary care, boosting community transmission before public health measures were taken. The use of clinical codes could misrepresent the true occurrence of the disease. Serological or PCR testing should be used to confirm these findings. In future, this surveillance of excess influenza could help detect new outbreaks of COVID-19 or other influenza-like pathogens, to initiate early public health responses.

Innovative use of data sources: a cross-sectional study of data linkage and artificial intelligence practices across European countries

Background

The availability of data generated from different sources is increasing with the possibility to link these data sources with each other. However, linked administrative data can be complex to use and may require advanced expertise and skills in statistical analysis. The main objectives of this study were to describe the current use of data linkage at the individual level and artificial intelligence (AI) in routine public health activities, to identify the related estimated health indicators (i.e., outcome and intervention indicators) and health determinants of non-communicable diseases and the obstacles to linking different data sources.

Method

We performed a survey across European countries to explore the current practices applied by national institutes of public health, health information and statistics for innovative use of data sources (i.e., the use of data linkage and/or AI).

Results

The use of data linkage and AI at national institutes of public health, health information and statistics in Europe varies. The majority of European countries use data linkage in routine by applying a deterministic method or a combination of two types of linkages (i.e., deterministic & probabilistic) for public health surveillance and research purposes. The use of AI to estimate health indicators is not frequent at national institutes of public health, health information and statistics. Using linked data, 46 health outcome indicators, 34 health determinants and 23 health intervention indicators were estimated in routine. The complex data regulation laws, lack of human resources, skills and problems with data governance, were reported by European countries as obstacles to routine data linkage for public health surveillance and research.

Conclusions

Our results highlight that the majority of European countries have integrated data linkage in their routine public health activities but only a few use AI. A sustainable national health information system and a robust data governance framework allowing to link different data sources are essential to support evidence-informed health policy development. Building analytical capacity and raising awareness of the added value of data linkage in national institutes is necessary for improving the use of linked data in order to improve the quality of public health surveillance and monitoring activities.

Is it time to use real-world data from primary care in Alzheimer's disease?

BACKGROUND

The analysis of real-world data in clinical research is rising, but its use to study dementia subtypes has been hardly addressed. We hypothesized that real-world data might be a powerful tool to update AD epidemiology at a lower cost than face-to-face studies, to estimate the prevalence and incidence rates of AD in Catalonia (Southern Europe), and to assess the adequacy of real-world data routinely collected in primary care settings for epidemiological research on AD.

METHODS

We obtained data from the System for the Development of Research in Primary Care (SIDIAP) database, which contains anonymized information of > 80% of the Catalan population. We estimated crude and standardized incidence rates and prevalences (95% confidence intervals (CI)) of AD in people aged at least 65 years living in Catalonia in 2016.

RESULTS

Age- and sex-standardized prevalence and incidence rate of AD were 3.1% (95%CI 2.7-3.6) and 4.2 per 1000 person-years (95%CI 3.8-4.6), respectively. Prevalence and incidence were higher in women and in the oldest people.

CONCLUSIONS

Our incidence and prevalence estimations were slightly lower than the recent face-to-face studies conducted in Spain and higher than other analyses of electronic health data from other European populations. Real-world data routinely collected in primary care settings could be a powerful tool to study the epidemiology of AD.