The present study uses self‐reported data from monthly cross-sectional household surveys of representative samples of the population of adults in England collected as part of the ongoing Smoking and Alcohol Toolkit Study. The methodology for the Toolkit Study has been described in detail by Fidler et al (16) and to date, it has been used as the basis for approximately 85 peer-reviewed publications. Briefly, each month (wave), a new sample of approximately 1700 adults aged 16 years or older in England completes a face-to-face computer-assisted survey. After stratification by geo-demographic classification of population, small geographical areas containing approximately 300 households are allocated randomly to interviewers who visit households within the locality and conduct computer‐assisted face‐to‐face interviews with one member of a household in those areas until a pre‐specified quota tailored to the area is fulfilled. This form of sampling has benefits over conventional quota sampling, because the allocation of small areas to interviewers reduces the impact of selection bias resulting from the selection of properties (17). Response rates cannot be calculated because of the lack of a definitive gross sample: all units fulfilling the criteria of a given quota are interchangeable within the areas. For the present study, measures of mental health were added to the survey for 24 monthly waves from January 2016 to December 2017.
All measures were self-reported. Generally, questions were completed in an interview with the interviewer recording the responses; the mental health questions were self-completed by the interviewee on a laptop following familiarisation using similar example questions.
Sociodemographic measures included age (16-24; 25-34; 35-44; 45-54; 55-64; 65 and over), gender and occupational grade (coded AB for higher or intermediate managerial, administrative or professional occupation; C1 for Supervisory or clerical and junior managerial, administrative or professional; C2 for skilled manual workers; D for semi-skilled and unskilled manual workers; E for state pensioners, casual and lowest grade workers, unemployed with state benefits only (18)).
Smoking-related measures included smoking status, assessed using “Which of the following best applies to you? Please note we are referring to cigarettes and other kinds of tobacco that you set light to and NOT electronic or 'heat-not-burn' cigarettes. a) I smoke cigarettes (including hand-rolled) every day; b) I smoke cigarettes (including hand-rolled), but not every day; c) I do not smoke cigarettes at all, but I do smoke tobacco of some kind (e.g. pipe, cigar or shisha); d) I have stopped smoking completely in the last year; e) I stopped smoking completely more than a year ago; f) I have never been a smoker (i.e. smoked for a year or more).” Those who had stopped smoking more than a year ago (response e) or had never been smokers (f) were excluded as were the 186 smokers who reported currently smoking other tobacco products (c). Those selecting responses a or b were categorised as current smokers; current smokers and those who had stopped in the last year (d) were categorised as past-year smokers. Those who had stopped smoking may have included a small number of respondents who had smoked products other than cigarettes. Past-year cigarette smokers reported type of cigarette smoked (hand-rolled, manufactured or both), strengths of urges to smoke (19) and heaviness of smoking index (cigarettes per day and time to first cigarette (20)); current cigarette smokers also reported motivation to stop smoking (21) and average spend on tobacco per week.
Harm reduction measures for all past-year smokers included number of attempts to quit smoking in the past year and whether abstinence from smoking of one month or longer was achieved during the past year which was derived from questions about how long each of up to three quit attempts undertaken in the last year lasted before going back to smoking. Current smokers were additionally asked whether they were currently attempting to cut down on the number of cigarettes and whether they were currently using nicotine replacement therapy (NRT) or e-cigarettes/vaping products (EC) for cutting down, temporary abstinence or other reasons.
Mental health measures included three indicators of mental health problems:
Ever diagnosis: Participants were asked: “Since the age of 16, which of the following, if any, has a doctor or health professional ever told you that you had? Depression; Anxiety; Obsessive Compulsive Disorder; Panic Disorder or a phobia; Post-traumatic Stress Disorder; Psychosis; Personality Disorder; Attention Deficit Hyperactivity Disorder; An Eating Disorder; Alcohol Misuse or Dependence; Drug Use or Dependence; Problem Gambling; None of these; Don’t know; Prefer not to say.” Question wording was adapted from previous surveys (22, 23), the list of diagnoses based on a report on adult psychiatric morbidity in England (24). Response options excluding the final three were presented in a randomised order. Individual diagnoses were dummy-coded. For analysis, a composite measure of ‘Any diagnosis’ was derived.
Past-month distress: Measured using the K6 screener for mental distress in the past 30 days (25, 26): “During the past 30 days, about how often, if at all, did you feel… nervous; hopeless; restless or fidgety; so depressed that nothing could cheer you up; that everything was an effort; worthless?” The options were presented in a randomised order and for each the respondent indicated one of the following: “All of the time (scored 4); Most of the time (3); Some of the time (2); A little of the time (1); None of the time (0)”; these options were presented in this or the reverse order for the entire K6. Additional response options “Don’t know; Prefer not to say” were provided. The K6 has been validated and used in a number of population surveys (25-35). A sum score with a possible range from 0 to 24 was calculated and following common practice, scores of 13 and higher categorised as serious distress (25, 29). Some publications have additionally used a category of moderate distress (scores 5-12) (31) which was also used.
Past-year treatment: Respondents who selected any of the responses for ever diagnosis were asked for those they selected: “In the last 12 months, which of the following conditions, if any, have you had any treatment or taken any prescribed medication for?”.
In 2016 and 2017, 40,831 adults were surveyed of whom 7,651 were past-year (current and recent-ex) smokers who were asked the mental health questions. Smokers who exclusively smoked tobacco products (pipes, cigars) other than cigarettes were excluded (n=186). Those who did not complete the mental health questions or selected ‘don’t know’ or ‘prefer not to say’ in response to any of them were also excluded, leaving 6,280 past-year smokers with information on mental health. Finally, those with missing data on any of the other variables included in the present analysis were excluded, leaving 6,071 past-year smokers for the main analyses, of whom 5,637 were current cigarette smokers and 434 recent ex-smokers. Type of cigarette smoked was missing for 164, reducing the sample to 5,907 past-year smokers and 5,506 current cigarette smokers if type of cigarette was included in analysis.
Past-year smokers who completed the mental health information were compared with those who did not complete the mental health information or responded don’t know or prefer not to say.
To address research question 1 (mental health), weighted proportions were used to describe the prevalence of ever having had any mental health diagnosis since the age of 16, the prevalence of moderate and serious distress in the past month and the prevalence of treatment for mental health problems in the past year overall and by diagnosis.
To address research question 2 (smoking behaviour by mental health), weighted proportions were calculated and chi-square statistics were used to compare smoking status, type of cigarette smoked, dependence (as measured using urges to smoke and heaviness of smoking) of past-year cigarette smokers with and without any diagnosis, with minimal, moderate and serious past-month distress and with and without past-year treatment. For chi-square tests, Cramer’s V was used as effect size. Cramer’s V can range from 0 (no association) to 1 (perfect association). Additionally, for current cigarette smokers, spend on tobacco was compared using analyses of variance (ANOVAs) and independent samples median tests.
To address research question 3 (harm reduction behaviour), weighted proportions for cutting down, using e-cigarettes, using NRT, using either e-cigarettes or NRT, motivation to stop smoking and having achieved at least one month of abstinence in the past year were calculated for current cigarette smokers and compared using chi-square statistics with Cramer’s V as effect size.
To address research question 4 (quit attempts), the mean number of quit attempts made in the past year was compared among past-year smokers with and without mental health problems. A dichotomised version of this measure (at least one quit attempt versus none) was used to calculate prevalence and 95% confidence intervals of having made at least one quit attempt and as outcome in logistic regressions assessing its association with ever diagnosis, past-month levels of distress and past-year treatment among past-year smokers. Unadjusted models were followed by models adjusting for socio-demographics. Additional models adjusted for motivation to stop smoking and type of cigarette; as this information can only be collected from current cigarette smokers, any quit attempt had been unsuccessful.
The analyses plan was pre-registered on the Open Science Framework (https://osf.io/9pbv8/). The following changes were made from the registration: Research question 1 was added to provide a fuller description of the population. The present research questions 2 and 3 were initially presented as a single research question. To increase consistency, smokers of other tobacco products such as pipes or cigars were excluded from all analysis.