Ambulatory assessment of affective instability in borderline personality disorder - The effect of the sampling frequency

Affective instability is an essential criterion for borderline personality disorder (BPD) but studies using diaries have rendered conflicting results. Discrepant findings may be caused by differences in the time-based design, especially in the diverse length of the time intervals between self-reports. Even though there is consensus that the time-based design should fit the temporal dynamics of the processes of interest, no general conventions exist. We used 24 h ambulatory monitoring to repeatedly assess subjective ratings of distress in 50 BPD-patients and in 50 healthy controls. We investigated if the chosen time-based design with a time interval of 15 min between self-reports (1) reveals within subject variability in BPD-patients and (2) taps the process of interest. Using graphical and statistical evaluation, we demonstrate that the time-based design (15 min/24 h) does catch within-person variability in the BPD-patients. Comparison of the original data with randomly distributed data (simulation) and autocorrelation analyses prove that we tapped a specific process and did not randomly pick states of distress. Using increasing lags between self-reports reveals that short intervals, especially, (15 min, 30 min) tap a specific process. We recommend using short time intervals to study temporal dynamics of affective instability.