Background & Scientific Rationale
Clinical trials often have limited sample sizes, narrow inclusion criteria, and short follow-up periods. Consequently, they may fail to capture the full spectrum of treatment outcomes and safety signals once a drug is widely used. Real-world big data, such as UK Biobank, allows for large-scale, long-term observation across diverse patient populations. Leveraging these rich data resources, our team will use advanced epidemiological methods and machine learning to examine post-marketing outcomes, providing robust and timely evidence on treatment effectiveness, safety, and cost-effectiveness for both physical and mental health conditions.
Research Questions
1. What are the incidence, prevalence, and risk factors for a range of physical and mental conditions?
2. How do these diseases progress over time, and what are the long-term treatment outcomes?
3. Which factors influence variability in treatment response (efficacy and safety)?
4. Can we identify high-risk patient subgroups and develop prediction tools to optimize disease management?
5. What is the cost-effectiveness of different treatments at the population level?
Objectives
1. Evaluate incidence, prevalence, risk factors, progression, and treatment effects for conditions including cancers, cardiovascular, endocrine, neurological, mental health, and more.
2. Apply machine learning to identify distinct phenotype clusters and to develop clinical prediction tools.
3. Validate findings using traditional statistical models and advanced epidemiological approaches.
4. Assess population-level cost-effectiveness of treatments to guide evidence-based policy and shared decision-making.