Is patient dropout from a longitudinal study of lung function predictable and reversible?

Area of Science:

• Epidemiology
• Pulmonary Medicine
• Biostatistics

Background:

• Longitudinal studies are crucial for understanding lung function development.
• Participant attrition, or dropout, poses a significant challenge to the validity and generalizability of longitudinal research.
• Predictive factors for dropout and effective re-recruitment strategies are not well-established for lung function studies.

Purpose of the Study:

• To identify factors at enrollment that predict dropout from a longitudinal lung function study.
• To develop and describe methods for re-recruiting participants who have previously dropped out.

Main Methods:

• A birth cohort study comparing enrollment details of participants who dropped out versus those retained at 6 and 11 years.
• Logistic regression analysis to identify independent risk factors for dropout.
• Description of strategies used for re-contacting dropouts, including study records and public databases.

Main Results:

• Dropout rates were substantial: 49% at 6 years and 23% at 11 years.
• Predictors of dropout at 6 years included no other sibling enrolled, being first-born, mother's lower education, and mother born overseas.
• Predictors of dropout at 11 years included mother born overseas, paternal smoking, no other sibling enrolled, and being first-born.
• Re-recruitment was successful for 47% using existing study data and 43% using public databases for those lost at 6 years but found at 11 years.

Conclusions:

• Enrollment characteristics can predict dropout in longitudinal lung function studies.
• Proactive strategies utilizing existing and public data can effectively re-recruit previously dropped-out participants.
• These findings can improve the design and retention rates of future longitudinal cohort studies.