Dataset for Integrating Online Stop Smoking Support with Online Talking Therapies

Data collection method:

The trial was a parallel two arm, online randomised controlled feasibility and pilot trial. Participants were recruited within an existing online mental health services platform called Silver Cloud when they accessed their Silver Cloud account for the first time. An invitation to take part in the feasibility study appeared in the online platform, and after consent was taken, they were randomly allocated to the control or intervention condition. An algorithm in Qualtrics contains the ‘Mersenne Twister’, a standard general-purpose pseudorandom number generator to calculate a randomisation sequence. Qualtrics surveys sitting within the Qualtrics platform were linked to the online Silver cloud Platform, to collect baseline information initially. The smoking cessation programme was a set of modules for intervention participants only which they could access along with the mental health support modules provided for their usual care. The control participants received usual care only. All participants had immediate access to the modules after randomisation. On a monthly basis, exports were taken from the Silver Cloud platform containing activity data such as; number of viewings of the modules pages, and time spent within SilverCloud. Exports also included data on the number of clients eligible for the study and mental health assessment data (PHQ_9 and GAD_7). These exports were extracted for all 17 services sites in excel spreadsheets every month and were cumulative and sent to the Bath Research Team. The Trust sites also provided a monthly excel spreadsheet with clinical information and attendance at review meetings. This data was requested by the Bath Team from the service sites on a monthly basis. Follow up data was collected by sending emails to participants at the set time points and 7 Qualtrics questionnaires were designed for this purpose (5 short surveys sent out every 2 weeks for 10 weeks, and a follow up survey at 3 and 6 months. All Qualtrics survey data was exported in csv files and cleaned using R, before being merged into one spreadsheet using email address as key variable, using the join functions in R. The exports from SilverCloud were also in excel files. All site data for each measure (e.g. PHQ_9) was merged into one datafile and then converted from long to wide format. Each merged Silver cloud spreadsheet was then joined to the Qualtrics data using the SilverCloud ID number as the key joining variable. Finally the excel spreadsheets for each of the 17 sites containing additional clinical information were cleaned and merged into one datafile. The merged clinical data file was then joined to the rest of dataset using R joining function. A set of descriptive statistics was calculated using R. Finally, completed case and also imputed data were analysed using logistic regression and linear regression methods in STATA.

Data processing and preparation activities:

Silver Cloud datasets only contained SilverCloud ID numbers and so anonymised already. The data in the excel spreadsheets were in long format and using R, converted to wide format so one row per participant. All Qualtrics surveys were joined using email addresses and then merged with the Silver Cloud data using the Silver Cloud id numbers. All personally identifiable information was taken out. Finally, the data collected from Trust sites were merged into one data set for all 17 sites, PID was removed, and the service site data was joined in R using the SilverCloud ID as the joining variable.

Technical details and requirements:

All analyses were conducted in R (v4.2.0) or STATA (v18.0).

Additional information:

The variables are labelled with an extension to indicate the time points at which the data were collected. e.g. b for baseline, t1, t2, t3,t4,t5 and 3 month or 6 month. The code book presents the names and labels and coding of variables in the order in which they appear in the dataset.