Automated Collection, Filtering, Ranking, and Delivery of Alzheimer’s Disease Clinical Research Intelligence

Gaurab Subedi [1], Jorge Fonseca [1], Amanda Leisgang Osse [2,3], Andrew Hooyman [4], Jefferson Kinney [2,3], Jeffrey Cummings, MD, ScD [2,3]

[1] Department of Computer Science, University of Nevada, Las Vegas, Las Vegas, USA
[2] Kirk Kerkorian School of Medicine, University of Nevada, Las Vegas, Las Vegas, NV, USA
[3] Chambers-Grundy Center for Transformative Neuroscience, Las Vegas, NV, USA
[4] Physical Therapy, Crean College of Health and Behavioral Sciences, Chapman University, Orange, CA, USA

Funded by the Alzheimer’s Drug Discovery Foundation (ADDF).

1. Background and Objective

Why this matters

Alzheimer’s disease research updates are distributed across government agencies, pharmaceutical companies, academic medical centers, clinical trial sources, and biomedical news platforms.

These updates are important for tracking emerging therapies, biomarker advances, regulatory decisions, and clinical trial activity.

Manual monitoring is slow, fragmented, and prone to missed or irrelevant information.

Objective

Develop an automated information surveillance system that continuously aggregates, filters, classifies, and organizes Alzheimer’s disease research updates into a centralized searchable platform.

2. System Architecture

The platform uses a modular scraper architecture with 51 automated source connectors across four major source categories:

Government agencies: NIH, FDA, EMA

Biomedical news outlets: ScienceDaily, Fierce Biotech, Neuroscience News

Pharmaceutical companies: Biogen, Eli Lilly, Eisai, Novartis

Academic medical centers: Columbia, Stanford, Mass General, Brigham and Women’s

Collection Methods

RSS feed parsing: 27 sources

HTML extraction using HTTP requests: 18 sources

Headless browser automation for JavaScript-rendered sites: 6 sources

Processing Workflow

Data Sources → Collection → 58-Keyword AD Relevance Filter → Full-Text Extraction → SQLite Storage → Rule-Based Classification → Searchable Dashboard

Articles are categorized into clinically meaningful domains, including clinical trials, drug approvals, research, regulatory updates, and other Alzheimer’s disease research news.

3. Results

The system includes 51 connectors across 4 Alzheimer’s disease research source types.

The platform scans approximately 80–100 articles daily from 35 active sources.

A total of 562 Alzheimer’s disease-relevant articles were collected, including 189 full-text articles.

The system achieved sub-20 ms API response time with search, analytics, and admin tools.

4. Platform Output

Searchable article dashboard

Ranked Alzheimer’s disease-relevant research feed

Full-text article access for the research team

Source analytics and admin controls

Optional alerts or manual update trigger

5. Conclusion

We developed an automated Alzheimer’s disease research surveillance platform across regulatory, industry, academic, and news sources.

The system reduces manual monitoring by collecting, filtering, classifying, and organizing Alzheimer’s disease-relevant updates.

The platform supports faster awareness of emerging therapies, clinical trials, biomarkers, and regulatory decisions.

6. Future Work

Expand source coverage by adding additional regulatory, industry, academic, and biomedical news sources.

Extend the surveillance framework to Parkinson’s disease and other neurodegenerative diseases.

Add automated email alerts when new Alzheimer’s disease-relevant research updates, clinical trial news, biomarker findings, or regulatory information are detected.

Improve relevance scoring and classification using semantic ranking and researcher feedback.

Contact

Gaurab Subedi