|| Checking for direct PDF access through Ovid
In-person retrospective timeline follow-back (TLFB) interviews are a well-established method for collecting self-reports of drug use from patients. However, this method can require significant staff and patient time. In the context of a randomized clinical trial evaluating interim buprenorphine dosing, we examined the feasibility of an interactive voice response (IVR) system for daily monitoring of illicit opioid use during buprenorphine treatment, with a focus on the agreement of illicit opioid use self-report data collected from the concurrent IVR methodology versus retrospective TLFB interviews. Participants (n = 24) received buprenorphine maintenance for 12 weeks and completed nightly IVR calls in which they reported illicit opioid use in the prior 24 hrs. At approximately weekly visits, they provided in-person TLFB reports of illicit opioid use. Levels of data collection were high for both IVR and TLFB methodologies (94.2% vs. 98.5%, respectively) and did not differ. Overall agreement between the 2 methods was high (97%), whereas Cohen’s kappa was moderate (k = 0.60). When self-report data were compared with urinalysis results for illicit opioid use, IVR and TLFB approaches both showed high specificity (∼99%), although sensitivity was greater for the TLFB method (48% and 69% for IVR and TLFB, respectively; p = .003). These pilot data suggest that an automated IVR approach may offer an efficient alternative for monitoring self-reported opioid use, especially in rural or resource-constrained settings. Additional efforts to understand and improve IVR sensitivity are warranted.