Dynamic self-efficacy and outcome expectancies: prediction of smoking lapse and relapse.

According to social learning models of drug relapse, decreases in abstinence self-efficacy (ASE) and increases in positive smoking outcome expectancies (POEs) should foreshadow lapses and relapse. In this study, the authors examined this hypothesis by using ecological momentary assessment data from 305 smokers who achieved initial abstinence from smoking and monitored their smoking and their ASE and POEs by using palmtop computers. Daily ASE and POEs predicted the occurrence of a 1st lapse on the following day. Following a lapse, variations in daily ASE predicted the onset of relapse, even after controlling for concurrent smoking. ASE and POEs generally neither mediated nor moderated each other's effects. These data emphasize the role of dynamic factors in the relapse process.