Abstract
BACKGROUND: The Bluetooth iCOquit enables remote biochemical verification of smoking status, but its validity among adults attempting to quit smoking is unclear. This study 1) compared the iCOquit, piCO, and Vitalograph sensors to identify device-specific bias, 2) assessed the diagnostic accuracy of the iCOquit for the overall sample and within specific subgroups (sex, race, smoking rate, menthol use), and 3) assessed the validity of iCOquit readings against standardized CO canisters. METHODS: iCOquit devices were tested with human breath samples from individuals seeking treatment for combustible tobacco use (N = 93) attending an in-person clinic visit. Participants provided breath samples via the iCOquit, piCO, and Vitalograph (order randomized). iCOquit devices were also tested using 5 and 10 parts per million (ppm) canisters. RESULTS: The iCOquit underestimated CO and categorized more participants as abstinent relative to the other CO sensors with human breath samples. The results suggested the iCOquit could not be used interchangeably with the other CO devices. Using a cut-off of < 6 ppm, the diagnostic accuracy of the iCOquit (specificity = 94%; sensitivity = 85%) did not vary across demographic/smoking subgroups. Canister tests with the iCOquit suggested good precision (< 1 ppm). CONCLUSIONS: The iCOquit is an affordable option for the remote measurement of CO that provides a reasonably accurate assessment of smoking status of those attempting to quit smoking using abstinence cut-off criteria of < 6 ppm. However, compared to other CO monitors, the iCOquit may underestimate CO, thereby increasing error in assessing abstinence.