Agreement of offspring-reported parental smoking status: the RHINESSA generation study

Abstract

Background: With increasing interest in exposure effects across generations, it is crucial to assess the validity of information given on behalf of others. Aims: To compare adult’s report of their parent’s smoking status against parent’s own report and examine predictors for discrepant answers. Methods: We studied 7185 offspring (18-51 years) and one of their parents, n = 5307 (27-67 years) participating in the Respiratory Health in Northern Europe, Spain and Australia (RHINESSA) generation study. Information about parent’s smoking status during offspring’s childhood and mother’s smoking status during pregnancy was obtained by questionnaires from parents and their offspring. We calculated sensitivity, specificity and Cohen’s Kappa [κ] for agreement using parent’s own report as the gold standard. We performed logistic regression to examine if offspring’s sex, age, educational level, asthma status, own smoking status or parental status, as well as the parent’s sex and amount of smoking during childhood predicted disagreement. Results: The sensitivity for offspring’s correct report of parent’s smoking status during childhood (0-10 years) was 0.82 (95% CI 0.81-0.84), specificity was 0.95 (95% CI 0.95-0.96) and a good agreement was observed, κ = 0.79 (95% CI 0.78-0.80). Offspring’s report of mothers’ smoking status during pregnancy showed a lower sensitivity, 0.66 (95% CI 0.60-0.71), a slightly lower specificity, 0.92 (95% CI 0.90-0.95) and a good agreement, κ = 0.61 (95% CI 0.55-0.67). In multivariate logistic regression analysis, offspring not having children was a predictor for discrepant answers (odds ratio [OR] 2.11 [95% CI 1.21-3.69]). Low amount of parents’ tobacco consumption, < 10 cigarettes/day (OR 2.72 [95% CI 1.71-4.31]) also predicted disagreement compared to ≥10 cigarettes per day, and so did offspring’s reports of fathers’ smoking status (OR 1.73 [95% CI 1.09-2.74]) compared to mothers’ smoking status. Offspring’s sex, asthma status, educational level, smoking status or age was not related to discrepant answers. Conclusions: Adults report their parent’s smoking status during their childhood, as well as their mother’ smoking status when pregnant with them, quite accurately. In the absence of parents’ direct report, offspring’s reports could be valuable.