Measuring craving: an attempt to connect subjective craving with cue reactivity

Better insight into craving may contribute to the development of more efficient relapse prevention strategies. Inconsistent findings on the relation between craving and relapse may be due to difficulties in the measurement of craving. These difficulties are accounted for by 3 interrelated problems: lack of consensus regarding the definition of craving, the use of different time frames (craving now vs craving in the past), and lack of concordance between self-reported craving and psychophysiological measures of cue reactivity. The aim of this study is to develop and validate a new self-report questionnaire for the assessment of the core aspects of craving, taking into account different time frames and emphasizing the psychophysiological aspects of craving. It is hypothesized that this questionnaire will show higher concordance with measures of cue reactivity than existing self-report craving questionnaires.

Based on a semantic mapping sentence, a 24-item self-report questionnaire was developed: the Jellinek Alcohol Craving Questionnaire (JACQ). The questionnaire was tested in 2 samples of treatment-seeking alcohol-dependent patients (sample A, n = 251; sample B, n = 48). Psychometric properties were examined in sample A and cross-validated in sample B. The associations with psychophysiological and neuroendocrine measures of cue reactivity were studied in sample B.

The JACQ consists of 1 dimension with the following 4 aspects: (1) emotional urge, (2) physical sensations, (3) temptation to drink, and (4) uncontrolled thoughts. All (sub)scales had a good internal consistency (alpha = 0.77-0.95) and were highly intercorrelated (r = 0.57-0.86). Craving-past correlated low with craving-now (r = 0.32). Craving-now (sub)scales showed a moderate association with heart rate (0.46-0.49), but not with respiration rate, skin conductance, or cortisol production in saliva following cue exposure. In contrast, craving-past (sub)scales showed a moderate association with cortisol production in saliva (0.15-0.42) following cue exposure. Remarkably, the physical symptoms scale did not show stronger association with psychophysiological and neuroendocrine measures than the other subscales.

The JACQ reliably measures 1 dimension including 4 aspects of craving for distinct time frames. Despite the presence of a special subscale for physical sensations, only moderate relationships were found between self-reported craving and biological indicators of cue reactivity. These findings suggest individual differences between alcoholics in the detection and/or reporting of cue-related psychophysiological signs as indicators of cue reactivity. Further research should focus on the nature of these individual differences. In addition, more research is needed on the relative predictive validity of subjective phenotypic indicators of craving (such as self-reported craving) versus objective endophenotypic indicators of craving (such as physiological measures of cue reactivity) for example in the prediction of relapse.