Validation of women's self-reported smoking status in a Sydney colposcopy clinic

This study validated a self-reported smoking prevalence questionnaire against urinary cotinine levels among women referred to a Sydney based public hospital colposcopy clinic over the period November 1997 to June 1998. Of 213 eligible women, 160 (75%) agreed to participate. Of these, 151 (94%) completed the smoking prevalence item and 130 (86%) also provided a urine specimen. 40% (95% CI: 33%-49%) of respondents self-reported as current smokers. Observed agreement between self-reported and biochemically validated smoking status was 94% compared with 52% chance agreement (kappa = 0.87, p<0.001). This high level of agreement was also robust to sensitivity analyses using a more conservative cutoff for the urinary cotinine levels. We conclude that self-reported smoking status provides valid estimates of actual rates of smoking among women referred with abnormal cervical smears.