The short-term stability and reliability of daily estimates of posttraumatic stress disorder symptoms

There is no established guidance on how many days of posttraumatic stress disorder (PTSD) assessments are sufficient to capture reliable and stable estimates of intraindividual mean (iM) and variability (intraindividual standard deviations [iSD]) in intensive longitudinal studies. Thus, this study examined the reliability and short-term stability of daily PTSD symptoms measured using the Primary Care PTSD Screen for DSM-5 (PC-PTSD-5). Participants (N = 70, M_age = 30.44 years, SD = 12.78 72.9% female) completed the PC-PTSD-5 for 21 consecutive days before and after attending four intervention sessions. To examine reliability, generalizability coefficients assessing systematic consistency across days (Rc) and single-day reliability (R1F) were calculated. To examine short-term stability in each phase, we calculated reference iMs and iSDs from 21 days for the preintervention and postintervention phases. We used (a) correlation coefficients (stable: r > .80) and mean absolute differences (stable: < .25) to compare these reference estimates, with estimated values ranging from 2-21 days per participant, and (b) bias and agreement using Bland-Altman analyses. Results indicate that the PC-PTSD-5 yielded varying intraindividual variability estimates in the short term, Rcs_pre = .45-.67, Rcs_post = .40-.55, but good single-day reliability, R1Fs_pre = .72-.78, R1Fs_post = .77-.82. Assessing PTSD using the PC-PTSD-5 for 7-11 days could produce iM and iSD estimates comparable to 21 days. Overall, the PC-PTSD-5 was more reliable for capturing between-person differences than within-person fluctuations. Intensive longitudinal studies could use 7-11 days of daily PC-PTSD-5 assessments to capture stable estimates of average and variable PTSD symptoms.