The Use of Real-World Data for Studies of Dynamic Disease Processes

Area of Science:

• Biostatistics
• Epidemiology
• Health Services Research

Background:

• Real-world evidence (RWE) from observational data is crucial but challenging to obtain.
• Healthcare visits and treatment changes often correlate with disease activity, introducing bias.
• Disease-related visits and treatment by indication can distort understanding of disease progression and treatment efficacy.

Purpose of the Study:

• To address challenges in obtaining valid RWE from observational health data.
• To demonstrate biases arising from disease-related visits and treatment changes.
• To present joint modeling as a method to overcome these limitations.

Main Methods:

• Utilized joint models for disease, marker, treatment, and observation (visit) processes.
• Employed illustrative multistate models to simulate biases.
• Applied joint models to patient data for psoriatic arthritis.

Main Results:

• Demonstrated how disease-related factors bias observational studies.
• Showcased the utility of joint models in correcting for these biases.
• Provided insights into treatment effects in psoriatic arthritis.

Conclusions:

• Joint models offer a robust approach to analyzing real-world data with complex disease processes.
• This methodology can yield more accurate insights into intervention effects.
• Accurate RWE is essential for informed clinical decision-making and patient care.