Search | Search by Center | Search by Source | Keywords in Title
Arthur PB, Mann WC, Gitlin L. Staying Blind in a Colorful World: Minimizing Collection Bias in Trials with Intensive Interventions. Poster session presented at: Gerontological Society of America Annual Scientific Meeting; 2013 Nov 21; New Orleans, LA.
Blinding in clinical trials is critical, though challenging due to complexities of behavioral interventions with older adults and the contexts in which they are implemented. This poster reports on blinding methods used in a home-based intervention and their effectiveness. Methods: In this single-blind parallel randomized controlled trial, participants are randomized into attention control or home-based intervention groups. Attention control participants receive caregiver education telephone calls for 4 months while intervention participants receive in-home interventions for 4 months. A blinded data collector takes assessments at baseline, 4 month, and 8 month (long term) periods. The data collector is blinded to address treatment bias and 6 study controls are implemented to ensure continued blinding. Results: To date, 38 standardized and manualized baseline interviews (T1) and 6 followup interviews (T2) have been completed in a trial using a behavioral intervention to address challenging behaviors in Veterans with dementia. The data collector has reported continued blinding to treatment assignment for all 38 participants. Discussion: To ensure blinding, caregivers are notified of data collector blinding and asked not to reveal their assignment. The data collector does not have access to study data that could compromise blinding. Study phone calls are performed in locked offices. The data collector is not present at interventionist meetings. An unblinded monitor distributes study-related messages to the correct party. The data collector records their "guess" of group allocation after each contact with scripted evidence. To date, none of the 6 procedures have been violated and study blinding has been successful.