The study was undertaken across the North West of England; a European region with a population of 6,897,905 (2009 [30
]) which includes a mix of affluent, deprived, urban and rural communities [31
]. Survey sample size was calculated to provide estimates of LS and MWB at a sub-regional geography corresponding with local health authority areas (n
24). The target sample was 18,500 individuals aged 16
years and over. Ethical approval was obtained from Liverpool John Moores University.
Households were selected for inclusion using a clustered random sample approach. The national Post Office Address File (a database containing all known UK addresses and postcodes) was employed as the sampling frame and lower super output areas (LSOAs) were the primary sampling units. LSOAs are standard geographical areas in England that contain approximately 1,500 residents [32
]. Each LSOA has an average measure of deprivation routinely calculated across residents based on the Index of Multiple Deprivation (IMD). The IMD is a composite measure that includes 38 indicators relating to health, economic and educational status [33
]. LSOAs within each health area were labelled by national quintile of IMD and a random selection of LSOAs was made for each quintile based on their proportion within the health area. Households were randomly selected within each specific LSOA resulting in a total of 30,884 households. At the point of visit, the household resident aged ≥16
years who was next to have their birthday was selected to participate in the survey. However, for this analysis of adult LS and MWB, only those aged ≥18
years are included. For the purpose of this study IMD quintile for respondents’ LSOA of residence was used as an ecological measure of respondents’ deprivation.
The survey was piloted and data collected between 1 April and 30 June 2009. Sampling times ranged from 9.00
am to 8.00
pm and sampling was undertaken on weekdays and weekends to ensure those in daytime employment or education were included. Trained researchers attempted a maximum of four visits to each selected household; after which another household was selected at random from the same cluster. Of all households selected 60.1
Data were collected using a structured questionnaire containing 44 questions covering relationships, health, lifestyle behaviours and key demographics such as age, gender, ethnicity and postcode of residence (a full list of data items is published elsewhere [34
]). Researchers explained the purpose of the survey, its confidential and anonymous nature and introduced the hand-held computer assisted self-interviewing (CASI) devices as the mechanism for data collection [35
]. Where required (e.g. with some elderly respondents) researchers provided assistance with using the devices. The final sample size was 18,560.
MWB was measured using the Short Warwick-Edinburgh Mental Wellbeing Scale (SWEMWBS). This is a shortened seven-item alternative to the full version of WEMWBS which uses a series of Likert-style items asking people to describe their experience of feelings and thoughts over the last two weeks: 1) I’ve been feeling optimistic about the future; 2) I’ve been feeling useful; 3) I’ve been feeling relaxed; 4) I’ve been dealing with problems well; 5) I’ve been thinking clearly; 6) I’ve been feeling close to other people; 7) I’ve been able to make my own mind up about people [29
]. Responses are scored (none of the time = 1; rarely = 2; some of the time = 3; often = 4; all the time = 5) and summed to provide a total score for each respondent. The SEMWBS has been validated against the full WEMWBS [29
] and been used to measure wellbeing in UK national household surveys [36
LS was measured using the single item question: “All things considered, how satisfied are you with your life as a whole nowadays on a scale of 1 to 10 where one is extremely dissatisfied and 10 is extremely satisfied” [37
]. This is a well-established measure that is widely used in global (e.g. World Values Survey [38
]), European (e.g. European Social Survey [37
]); and UK (e.g. British Social Attitudes Survey [28
]) population surveys. Across surveys that use this question, average LS values for UK populations have been relatively consistent. Some variations in the wording and scale of single item LS questions are used in other social surveys, yet results produced are generally similar [28
For the purposes of analyses respondents were categorised for MWB into moderate (mean +/− one standard deviation (SD) range 23–32), low (below one SD from the mean) and high (above one SD from the mean). LS was categorised into low (scored 1–5), moderate (6–8) and high (9–10).
Employment was categorised as employed (including full-time, part-time and self- employed), unemployed (including seeking work, not seeking work for reasons of disability), retired, and domestic/other (of which those engaged in domestic duties accounted for 77.4
%). Individuals were also categorised according to whether they had a partner (here, in a long term meaningful relationship) and, if so, if that person was employed. Ethnicity was recorded using standard UK ethnic group categories [39
]; however due to small numbers of some ethnic minorities it was only possible to categorise ethnicity into three groups: White, South Asian and Other (of Other, 44.4
% were Black). For health status we used a standard self-assessed five point scale from very good to very bad health (“How is your health in general? Would you say it is…”) [40
]. Smoking status categorised people into five categories; never smokers, ex-occasional smokers, ex-daily smokers, current occasional smokers and current daily smokers. Exercise was measured by asking participants “In the past week, on how many days have you accumulated at least 30 minutes of moderate intensity physical activity such as brisk walking, cycling, sport, exercise, and active recreation? (do not include physical activity that may be part of your job or usual role activities)” [41
For alcohol consumption, participants were asked “In general, how often do you drink alcoholic drinks” (response categories: never, monthly or less, once or twice a week, three or four days a week, daily or almost daily, don’t know/refused) followed by a question asking drinkers how many of a range of drink types they had consumed in the last week. Individuals who reported never drinking were categorised as abstainers and those that reported drinking monthly or less were categorised as not usual drinkers. All other drinkers were categorised based on the amount of alcohol they reported to have consumed in the past week according to UK Department of Health categories. Thus, lower risk drinkers were, for females, those that drank ≤14 units (standards UK drinks; 1 unit
8 grams of pure alcohol) per week and, for males, ≤21 units per week. Increasing risk drinkers were those that consumed 15 to 35 units (females) and 22 to 50 units (males) per week and higher risk drinkers were those that consumed >35 (female) and >50 units (males) per week [42
]. To ensure that drinking reported over the past week was representative of typical drinking behaviour, individuals were asked if last week’s drinking was higher, lower or about typical for their alcohol consumption. Only those who stated that their past weeks’ alcohol consumption was typical were included in the analysis. Those individuals not providing an age (n
286), not answering both LS and MWB questions (n
383) and not providing answers to all other general health, smoking or exercise questions (n
294) were excluded from analyses.
The final sample contained 15,228 individuals who had provided answers across all variables included in analyses (82
% of all those completing the survey). Statistics utilised chi-squared for bivariate exploration of LS and MWB across the entire dataset. Sample size did not permit robust analyses within each deprivation quintile. Therefore, to examine associations with LS and MWB at different levels of deprivation the dataset was split into tertiles (affluent, n
4346; middle, n
5598; deprived, n
5284). Sample sizes vary slightly between tertiles as the catchment of each was matched to national IMD ranges for quintiles 1&2, 3&4 and 5 respectively. For each tertile, multinomial logistic regression (MLR [43
]) was used to identify independent relationships between LS / MWB and demographic, employment and lifestyle factors. Moderate LS / MWB was used as the reference category in order to quantify odds of low or high LS / MWB. All analyses were undertaken in PASW (Predictive Analytics Software) v18.0.