Interrupted Time Series Methods for Nonrandom Sampling Study Designs With Known Sampling Weights.

Area of Science:

• Health Services Research
• Biostatistics
• Clinical Trial Design

Background:

• Enhancing participant diversity in research is crucial for generalizability.
• The University of California Irvine Consent-to-Contact (C2C) Registry aims to recruit from disadvantaged neighborhoods.
• Area Deprivation Index (ADI) is used to define neighborhood disadvantage.

Purpose of the Study:

• To assess the effectiveness of recruitment interventions (Facebook vs. postcards) for disadvantaged neighborhoods.
• To estimate the marginal intervention effect on recruitment across all ADI deciles.
• To evaluate effect modification by ADI strata.

Main Methods:

• Utilized a nonrandom sampling design ensuring uniform inclusion across ADI deciles.
• Extended the Robust-Multiple Interrupted Time Series model to incorporate sampling weights.
• Proposed two novel variance estimators accounting for change point uncertainty and model mis-specification.

Main Results:

• Demonstrated the performance of proposed statistical methods through empirical simulation studies.
• Assessed the C2C-RSS design using proposed methods.
• Showed comparable power for the primary endpoint and increased power for the secondary endpoint versus simple random sampling.

Conclusions:

• The developed statistical methods effectively adjust for sampling bias in recruitment studies.
• The novel recruitment strategy shows promise for engaging diverse populations.
• Proposed methods enhance statistical power for assessing intervention effects and effect modification.