Estimation of the epidemiology of dementia and associated neuropsychiatric symptoms by applying machine learning to real-world data

[Cognitive, Emotional and Dependency Level on Institutionalized Older Adults]

INTRODUCTION

It is necessary to analyze the aging process in institutionalized older people. For this purpose, a descriptive and analytical epidemiological study was carried out in nursing homes for older adults before the COVID-19 pandemic.

OBJECTIVE

Provide an in-depth insight into cognitive (MEC), emotional (Geriatric Depression Scale), and functional (Barthel Scale) status within the older adult participants.

METHOD

A sample of 973 participants analyzed the relationship between cognitive status, vulnerability to depression, and autonomous performance in Daily Day Activities (DDA) to predict the impact of the comorbidity of these variables. Therefore, in addition to the general distribution of the sample in the previously mentioned dimensions, differences were analyzed according to gender, age, educational level, and geographic area.

RESULTS

The results confirm the hypothesis that a more impaired cognitive state is associated with higher levels of depression and lower functional capacity. The MEC scores have positive and highly significant correlations with Barthel and Yesavage. The relationship between dementia and autonomy is observed for both sexes, while the relationship between dementia and depression is only observed in women. The educational level influences the MEC scores (the more education, the better performance) and the Barthel scores (the less education, the greater dependency). Statistically significant differences were also found depending on the area of residence location.

CONCLUSIONS

The more deteriorated cognitive state will be associated with a higher level of depression and lower functional capacity in daily life activities.

Comorbidity and household income as mediators of gender inequalities in dementia risk: a real-world data population study

BACKGROUND

Low household income (HI), comorbidities and female sex are associated with an increased risk of dementia. The aim of this study was to measure the mediating effect of comorbidity and HI on the excess risk due to gender in relation to the incidence and prevalence of dementia in the general population.

METHODS

A retrospective and observational study using real-world data analysed all people over 60 who were registered with the Basque Health Service in Gipuzkoa. The study measured HI level, the Charlson comorbidity index (CCI), age and sex. The prevalence and incidence of dementia were analysed using logistic regression and Poisson regression models, respectively, adjusted by HI, sex, comorbidity and age. We estimated the combined mediation effect of HI and comorbidity on the prevalence of dementia associated with gender.

RESULTS

Of the 221,777 individuals, 3.85% (8,549) had a diagnosis of dementia as of 31 December 2021. Classification by the CCI showed a gradient with 2.90% in CCI 0-1, 10.60% in CCI 2-3 and 18.01% in CCI > 3. Both low HI and gender were associated with a higher crude prevalence of dementia. However, in the CCI-adjusted model, women had an increased risk of dementia, while HI was no longer statistically significant. The incidence analysis produced similar results, although HI was not significant in any model. The CCI was significantly higher for men and for people with low HI. The mediation was statistically significant, and the CCI and HI explained 79% of the gender effect.

CONCLUSIONS

Comorbidity and low HI act as mediators in the increased risk of dementia associated with female sex. Given the difference in the prevalence of comorbidities by HI, individual interventions to control comorbidities could not only prevent dementia but also reduce inequalities, as the risk is greater in the most disadvantaged population.

Excess healthcare costs of mental disorders in children, adolescents and young adults in the Basque population registry adjusted for socioeconomic status and sex

BACKGROUND

Mental illnesses account for a considerable proportion of the global burden of disease. Economic evaluation of public policies and interventions aimed at mental health is crucial to inform decisions and improve the provision of healthcare services, but experts highlight that nowadays the cost implications of mental illness are not properly quantified. The objective was to measure the costs of excess use of all healthcare services by 1- to 30-year-olds in the Basque population as a function of whether or not they had a mental disorder diagnosis.

METHODS

A real-world data study was used to identify diagnoses of mental disorders and to measure resource use in the Basque Health Service Registry in 2018. Diagnoses were aggregated into eight diagnostic clusters: anxiety, attention deficit hyperactivity disorder, conduct disorders, mood disorders, substance use, psychosis and personality disorders, eating disorders, and self-harm. We calculated the costs incurred by each individual by multiplying the resource use by the unit costs. Annual costs for each cluster were compared with those for individuals with no diagnosed mental disorders through entropy balancing and two-part models which adjusted for socioeconomic status (SES).

RESULTS

Of the 609,381 individuals included, 96,671 (15.9%) had ≥ 1 mental disorder diagnosis. The annual cost per person was two-fold higher in the group diagnosed with mental disorders (€699.7) than that with no diagnoses (€274.6). For all clusters, annual excess costs associated with mental disorders were significant. The adjustment also evidenced a social gradient in healthcare costs, individuals with lower SES consuming more resources than those with medium and higher SES across all clusters. Nonetheless, the effect of being diagnosed with a mental disorder had a greater impact on the mean and excess costs than SES.

CONCLUSIONS

Results were consistent in showing that young people with mental disorders place a greater burden on healthcare services. Excess costs were higher for severe mental disorders like self-harm and psychoses, and lower SES individuals incurred, overall, more than twice the costs per person with no diagnoses. A socioeconomic gradient was notable, excess costs being higher in low SES individuals than those with a high-to-medium SES. Differences by sex were also statistically significant but their sizes were smaller than those related to SES.

Incidence of mental disorders in the general population aged 1-30 years disaggregated by gender and socioeconomic status

PURPOSE

The objective of this study was to estimate the incidence and age of onset of mental disorders diagnosed by gender and socioeconomic status (SES) in children, adolescents, and young adults up to 30 years of age in the whole population of the Basque Country (Spain).

METHODS

All mental health diagnoses documented in Basque Health Service records from 1 January 2003 to 31 December 2018, were classified into eight clusters: anxiety, attention deficit hyperactivity disorder (ADHD), conduct disorders, depression, psychosis/personality disorders, substance use, eating disorders, and self-harm. We calculated incidence and cumulative incidence for each cluster, disaggregated by gender, and socioeconomic status (SES). Poisson regression analyses were performed.

RESULTS

Overall, 9,486,853 person-years of observation were available for the 609,281 individuals included. ADHD and conduct disorders were diagnosed in the first decade, anxiety and depression disorders in the second and third decades, and psychosis/personality and substance use in the third. The cumulative incidence at 18 years of age for any type of disorder was 15.5%. The group with low SES had a statistically significantly higher incidence of all eight clusters. The incidence of ADHD, conduct disorders, depression, psychosis/personality disorders, and substance use was higher in males and that of anxiety, eating disorders and self-harm was higher in females.

CONCLUSIONS

The incidence of mental disorders is high among children, adolescents, and young adults in the Basque Country underlining the need for preventive interventions. Marked differences by gender and SES highlight mental health inequalities, especially for depression and psychosis in low SES males.

Budget impact analysis of the use of Souvenaid in patients with prodromal Alzheimer’s Disease in Spain

Introduction

The effectiveness, safety, and cost-effectiveness of the use of Souvenaid for Alzheimer’s disease (AD) have been previously evidenced. To complete the economic analysis, there is a need to assess whether society can afford it. The objective of this study was to carry out a budget impact analysis of the use of Souvenaid in Spain under the conditions of the LipiDidiet clinical trial from a societal perspective.

Methods

We built a population model that took into account all the cohorts of individuals with AD, their individual progression, and the potential impact of Souvenaid treatment on their trajectories. Patient progression data were obtained from mixed models. The target population was estimated based on the population forecast for 2020–2035 and the incidence of dementia. Individual progression to dementia measured by the Clinical Dementia Rating-Sum of Boxes was reproduced using mixed models. Besides the costs of treatment and diagnosis, direct costs of medical and non-medical care and indirect costs were included.

Results

The epidemiological indicators and the distribution of life expectancy by stages validated the model. From the third year (2022), the differences in the cost of dementia offset the incremental cost of diagnosis and treatment. The costs of dependency reached €500 million/year while those of the intervention were limited to €40 million.

Conclusions

Souvenaid, with modest effectiveness in delaying dementia associated with AD, achieved a positive economic balance between costs and savings. Its use in the treatment of prodromal AD would imply an initial cost that would be ongoing, but this would be offset by savings in the care system for dependency associated with dementia from the third year. These results were based on adopting a societal perspective taking into account the effect of treatment on the use of health, social, and family resources.

Supplementary Information

The online version contains supplementary material available at 10.1186/s13195-022-01111-7.

Applied machine learning in Alzheimer's disease research: omics, imaging, and clinical data

Alzheimer's disease (AD) remains a devastating neurodegenerative disease with few preventive or curative treatments available. Modern technology developments of high-throughput omics platforms and imaging equipment provide unprecedented opportunities to study the etiology and progression of this disease. Meanwhile, the vast amount of data from various modalities, such as genetics, proteomics, transcriptomics, and imaging, as well as clinical features impose great challenges in data integration and analysis. Machine learning (ML) methods offer novel techniques to address high dimensional data, integrate data from different sources, model the etiological and clinical heterogeneity, and discover new biomarkers. These directions have the potential to help us better manage the disease progression and develop novel treatment strategies. This mini-review paper summarizes different ML methods that have been applied to study AD using single-platform or multi-modal data. We review the current state of ML applications for five key directions of AD research: disease classification, drug repurposing, subtyping, progression prediction, and biomarker discovery. This summary provides insights about the current research status of ML-based AD research and highlights potential directions for future research.

Use of Machine Learning and Artificial Intelligence Methods in Geriatric Mental Health Research Involving Electronic Health Record or Administrative Claims Data: A Systematic Review

Introduction: Electronic health records (EHR) and administrative healthcare data (AHD) are frequently used in geriatric mental health research to answer various health research questions. However, there is an increasing amount and complexity of data available that may lend itself to alternative analytic approaches using machine learning (ML) or artificial intelligence (AI) methods. We performed a systematic review of the current application of ML or AI approaches to the analysis of EHR and AHD in geriatric mental health. Methods: We searched MEDLINE, Embase, and PsycINFO to identify potential studies. We included all articles that used ML or AI methods on topics related to geriatric mental health utilizing EHR or AHD data. We assessed study quality either by Prediction model Risk OF Bias ASsessment Tool (PROBAST) or Quality Assessment of Diagnostic Accuracy Studies (QUADAS-2) checklist. Results: We initially identified 391 articles through an electronic database and reference search, and 21 articles met inclusion criteria. Among the selected studies, EHR was the most used data type, and the datasets were mainly structured. A variety of ML and AI methods were used, with prediction or classification being the main application of ML or AI with the random forest as the most common ML technique. Dementia was the most common mental health condition observed. The relative advantages of ML or AI techniques compared to biostatistical methods were generally not assessed. Only in three studies, low risk of bias (ROB) was observed according to all the PROBAST domains but in none according to QUADAS-2 domains. The quality of study reporting could be further improved. Conclusion: There are currently relatively few studies using ML and AI in geriatric mental health research using EHR and AHD methods, although this field is expanding. Aside from dementia, there are few studies of other geriatric mental health conditions. The lack of consistent information in the selected studies precludes precise comparisons between them. Improving the quality of reporting of ML and AI work in the future would help improve research in the field. Other courses of improvement include using common data models to collect/organize data, and common datasets for ML model validation.