Comparing the Fagerström Test and Heaviness of Smoking Index in Predicting Smoking Abstinence in Cancer Patients.

Abstract

Introduction: People with cancer who smoke exhibit greater cigarette dependence than people without cancer who smoke, a crucial factor in smoking cessation. Research is limited on the predictive potential of the Fagerström Test for Cigarette Dependence (FTCD) and the Heaviness of Smoking Index (HSI) on smoking abstinence in cancer patients undergoing smoking cessation treatment.

Aims and methods: We analyzed data from 5934 cancer patients seeking smoking cessation treatment at The University of Texas MD Anderson Cancer Center (female 52.08%; Mean age = 55.52, SD = 11.17). We evaluated the predictive accuracy of FTCD and HSI on abstinence at 3, 6, and 9 months from the first consultation, and assessed the concordance between these tools in measuring cigarette dependence using Cohen's kappa test and different correlation and regression models. We also analyzed variations across sex at birth and race/ethnicity.

Results: Both the FTCD and the HSI demonstrated comparable predictive accuracy for smoking cessation at all follow-ups, with neither showing high accuracy (Areas Under the Curve scores around 0.6). Concordance analysis revealed substantial agreement between FTCD and HSI scores (Cohen's kappa ~ 0.7), particularly at lower levels of dependence. However, this agreement varied by race, with reduced concordance observed in non-Hispanic Blacks.

Conclusions: Our results indicate that both the FTCD and HSI are effective tools for predicting smoking cessation in cancer patients, with the HSI offering a less burdensome assessment option. Nevertheless, the findings suggest the need for tailored approaches in assessing cigarette dependence that could predict smoking cessation more accurately, considering racial differences.

Implications: The burden of assessing cigarette dependence in cancer care settings can be reduced by using the HSI instead of the FTCD. In addition, both instruments could be substantially interchanged and used for meta-analytic studies examining dependence and abstinence, but race/ethnicity should be considered.