Evaluating Community-Based Translational Interventions Using Historical Controls: Propensity Score vs. Disease Risk...

Area of Science:

• Public Health
• Epidemiology
• Health Services Research

Background:

• Community-based interventions often use one-arm study designs due to ethical and practical constraints.
• Evaluating the effectiveness of these translational health initiatives presents significant challenges.

Purpose of the Study:

• To assess the effectiveness of the Special Diabetes Program for Indians Diabetes Prevention (SDPI-DP) lifestyle intervention.
• To compare propensity score (PS) and disease risk score (DRS) matching for adjusting confounder imbalance in one-arm studies using historical controls.

Main Methods:

• Utilized data from the Diabetes Prevention Program (DPP) placebo group as a historical control.
• Employed propensity score (PS) and disease risk score (DRS) matching techniques to adjust for baseline differences between the intervention and control groups.
• Calculated hazard ratios (HR) to estimate the intervention's effect on diabetes risk.

Main Results:

• The unadjusted hazard ratio (HR) for diabetes risk was 0.35 for the SDPI-DP intervention group compared to the control.
• Adjusted HRs, accounting for diabetes risk factors, were attenuated towards 1, ranging from 0.56 to 0.69.
• Disease risk score (DRS) matching resulted in a larger number of matched participants and smaller standard errors compared to propensity score (PS) matching.

Conclusions:

• Publicly available randomized clinical trial data can serve as a valuable historical control for evaluating one-arm community interventions.
• Appropriate statistical methods, such as DRS matching, are critical for balancing confounders and accurately estimating intervention effectiveness.
• This approach supports the evaluation of translational public health initiatives in real-world community settings.