Adult Lifelines participants were asked whether they (had) smoked any tobacco. Using the self-reported data from 1A, 1B, 1C, 2A, 2B, and 3A, researchers from the UMCG Department of Epidemiology calculated various derivative variables for smoking behavior (sections: lifestyle & environment (Smoking & tobacco use) and secondary & linked variables).
These derivative variables can be requested from Lifelines (data@lifelines.nl) or requested in the Lifelines catalogue.
Please note: these are updated variables from the previously released Smoking derivatives from the baseline assessment.v The data on smoking habits are collected in Lifelines questionnaires from multiple time-points. As a result, these data may contain inconsistencies and/or missing data. In October 2019, the baseline questionnaire has been validated resulting in the first version of the smoking derivatives. The updated smoking derivates incorporate the follow-up questionnaires, resulting in changes in the validated baseline data.

As a preparation step all smoking variables from the different timepoints are combined into one file. Additionally, duplicates are removed and variables are checked on correctness.

After the overall data file has been generated, the derivative variables are calculated. This is done in an extensive syntax which can be found on the Lifelines workspace and cluster environment to read all details. In short:

• Smoking duration is calculated by combining the total years a participant smoked, taking start and stop ages into account
• Pack years is calculated by combining the smoking duration and the amount smoked per day (of the different smoking options). Where one pack year equals 20 cigarettes per day

Most of the additional validations was done manually by inspecting the full longitudinal dataset of each subject separately. Because of this it is far too complicated to write all of them down. However, the general principles were as follows:

1. If the full smoking history at timepoint 1A was completed, the start and stop age at later time-points will be adjusted to those given at timepoint 1A.
2. If the full smoking history at timepoint 1A was not completed, the lowest start and highest stop age given at all timepoints will be taken.
3. If a participant did not fill in the amount of smoking and (s)he was a non-smoker at all other timepoints, this participant will be recoded to a non-smoker for this timepoint as well.
4. If a participant did not fill in the amount of smoking and smoked before and/or after this timepoint, the amount of smoking will be calculated as the mean of before and after. (if one is missing then the other is taken)
5. If a participant reported that (s)he is a never smoker at some timepoint but at another timepoint the data shows that this is not correct, the participant will be recoded to an exsmoker with the appropriate start and stop ages.
6. If a participant reported smoking cigars at one timepoint but at all other timepoints (s)he never smoked cigars, then the number of cigars are changed to number of cigarettes.
7. If the stop age of a participant is missing, this can sometimes be calculated by taking the (rounded) age halfway between the questionnaire at which (s)he smoked and the uestionnaire at which (s)he no longer smoked.
8. Missing values for the question “do you smoke now?” will be estimated using possible start and stop ages at later timepoints or by taking the answer from the questionnaire before the missing one.
9. If a participant completed the questionnaires in the wrong order (e.g. 2A before 1C) the derivative variables are calculated manually

* Please note that in contrast to the previous calculation, cigars are now regarded to as 1 cigarette (and not as 3)