Preliminary examination of the validity of a behavioral mindfulness measure in adults with advanced cancer and their family caregivers.

The breath counting task (BCT) has shown evidence of validity as a behavioral measure of mindfulness in healthy populations but remains largely untested in clinical contexts. The BCT is a computerized measure of present-moment awareness based on breath-counting accuracy. This study provides a preliminary evaluation of its validity in adults with advanced cancer and their family caregivers. Fifty-five patient-caregiver dyads were randomized to a 6-week mindfulness intervention or usual care. Participants completed the BCT and self-report surveys at baseline, post-intervention, and 1-month follow-up. The BCT's construct validity was examined through: (1) sensitivity to mindfulness intervention using linear mixed models, (2) convergent validity via correlations with self-reported mindfulness and theoretically related constructs (e.g., inner peace), and (3) criterion validity via correlations with clinical outcomes (e.g., quality of life). Findings differed for patients and caregivers. Among caregivers, the BCT demonstrated sensitivity to intervention; breath-counting accuracy on the BCT increased over time in the mindfulness condition and remained stable in the usual care condition. Among patients in the mindfulness condition, greater breath-counting accuracy was moderately associated with better quality of life at follow-ups, including a significant correlation at 1 month (r = .57, p < .05), supporting its criterion validity. Evidence of convergent validity was limited. However, for patients and caregivers, greater breath-counting accuracy was moderately associated with higher self-reported mindfulness facets following intervention. Preliminary findings suggest the BCT may capture certain attentional aspects of mindfulness in patients with advanced cancer and caregivers; however, patterns varied across groups, highlighting the need for further evaluation of its validity in clinical contexts.