Gender-Based Analysis of Risk Factors for Dementia Using Senior Cohort

Jaekue Choi, Lee-Nam Kwon, Heuiseok Lim, Hong-Woo ChunInternational Journal of Environmental Research and Public Health 2020 · DOI

TL;DR

This study leverages machine learning on a large Korean senior cohort to identify gender-specific dementia risk factors. The research reveals that distinct risk factors contribute to dementia in men and women, suggesting that personalized prevention strategies are crucial.

What Problem Does This Paper Address?

Dementia is a growing global health crisis with no fundamental cure, making early prediction and prevention paramount. While numerous studies have explored dementia risk factors, there's a significant gap in understanding how these factors might differ between genders. Given that dementia prevalence varies by sex, it's hypothesized that the underlying risk factors also differ. Existing statistical methods often overlook subtle causal relationships in complex datasets, necessitating more advanced analytical approaches. This paper aims to bridge this gap by employing machine learning to uncover gender-specific dementia risk factors, moving beyond traditional analyses.

Methods

The researchers employed three machine learning models: Support Vector Machine (SVM), Multi-Layer Perceptron (MLP), and Convolutional Neural Networks (CNN). These models were chosen for their proven performance in classification tasks. The study utilized a "backward greedy algorithm" (slightly modified) for feature selection, iteratively removing features to identify the most impactful ones for dementia prediction. The core of their approach involved training these models on a comprehensive dataset to analyze dementia risk factors separately for men and women, as well as for an integrated cohort.

Datasets

The primary dataset used is the Korean National Health Insurance System (NHIS)—Senior Cohort. This is a large-scale, longitudinal database containing health information for a significant portion of the Korean elderly population. Specifically, the study utilized data from 558,147 elderly individuals over 14 years. For the experimental analysis, a subset of 31,443 individuals who underwent medical examinations in 2013 was selected. From this, approximately 3000 dementia and 3000 normal subjects were randomly sampled for the integrated model, and 1500 dementia and 1500 normal subjects for each gender-specific model.

Key Insights and Ideas

  • Gender Matters: The most significant finding is the clear distinction in dementia risk factors between men and women. This underscores the need for sex-specific approaches to dementia prevention and management.
  • Machine Learning's Edge: Machine learning techniques proved adept at uncovering subtle causal relationships within the data that might be missed by traditional statistical methods. This highlights the power of these algorithms for complex health data analysis.
  • Beyond Known Factors: The study not only identified established dementia risk factors but also proposed novel candidate risk factors that warrant further investigation. This opens new avenues for research into the underlying mechanisms of dementia.
  • Integrated vs. Separate Analysis: Combining data from both genders for risk factor analysis actually deteriorated prediction performance. This strongly supports the hypothesis that gender-specific analysis is essential for accurate risk assessment.

Key Results

The MLP model demonstrated the best performance across all evaluation metrics (precision, recall, F-score) for both men and women. For instance, the MLP achieved a precision of 75.3% for men and 81.5% for women, with comparable recall and F-score values. The study identified distinct top 10 risk factors for men and women. For men, "Other mental disorders due to brain damage and dysfunction and to physical disease" ranked highest, while for women, "Cerebral infarction" was the top risk factor. The analysis also highlighted 15 novel candidate risk factors that were not previously well-established in the literature.

Limitations and Future Work

The authors acknowledge that multicollinearity (where features are highly correlated) can lead to false positives, which they attempted to mitigate using Pearson correlation coefficients. A limitation of the study is its focus solely on the elderly population; dementia can also occur in younger individuals. Future work will involve expanding the analysis to include younger cohorts and exploring additional risk factors beyond the top 100 identified. The researchers also plan to investigate other variables like income and residential areas, which were outside the scope of this gender-focused study.

Summary updated

Abstract

Globally, one of the biggest problems with the increase in the elderly population is dementia. However, dementia still has no fundamental cure. Therefore, it is important to predict and prevent dementia early. For early prediction of dementia, it is crucial to find dementia risk factors that increase a person's risk of developing dementia. In this paper, the subject of dementia risk factor analysis and discovery studies were limited to gender, because it is assumed that the difference in the prevalence of dementia in men and women will lead to differences in the risk factors for dementia among men and women. This study analyzed the Korean National Health Information System-Senior Cohort using machine-learning techniques. By using the machine-learning technique, it was possible to reveal a very small causal relationship between data that are ignored using existing statistical techniques. By using the senior cohort, it was possible to analyze 6000 data that matched the experimental conditions out of 558,147 sample subjects over 14 years. In order to analyze the difference in dementia risk factors between men and women, three machine-learning-based dementia risk factor analysis models were constructed and compared. As a result of the experiment, it was found that the risk factors for dementia in men and women are different. In addition, not only did the results include most of the known dementia risk factors, previously unknown candidates for dementia risk factors were also identified. We hope that our research will be helpful in finding new dementia risk factors.
BibTeX
@article{choi2020gender,
  author = {Jaekue Choi and Lee-Nam Kwon and Heuiseok Lim and Hong-Woo Chun},
  title = {{Gender-Based Analysis of Risk Factors for Dementia Using Senior Cohort}},
  journal = {International Journal of Environmental Research and Public Health},
  year = {2020},
  doi = {10.3390/ijerph17197274},
}
Figure