Objectives The purpose of this study was to investigate smoking prevalence of Tri-Service recruits, and changes in smoking behaviour at 3-year follow-up, by trade group and gender. Associations with educational attainment and deprivation were also assessed.
Methods Analysis of a survey into the health behaviours of 10 531 recruits in 1998/1999. A follow-up 3 years later measured changes in behaviour. Correlation and multiple regression was used to investigate the relationship between smoking prevalence in each trade group and both educational attainment and deprivation, using Index of Multiple Deprivation 2004 (IMD 2004) scores.
Results Army recruits exhibited a significantly higher smoking prevalence (45%) than Royal Navy recruits (34%) and Royal Air Force (RAF) recruits (31%). There were marked differences between smoking levels amongst officer cadets (12%, 20% and 10% in the Navy, Army and RAF, respectively) and other rank trade groups (24–56%), with the exception of the Marines (13%). At follow up, smoking had generally increased, and in some parts of the infantry had risen to 66%. There was a clear correlation between smoking at enlistment and both educational attainment (correlation coefficient=0.7, p<0.005) and deprivation score (correlation coefficient=0.8, p<0.005).
Conclusions There were clear differences between Services, rank and trade groups in smoking prevalence at recruitment. Smoking levels increased in the 3 years after recruitment to the Armed Forces. Deprivation was more important than educational attainment in determining the smoking status of recruits.
Statistics from Altmetric.com
The prevalence of smoking amongst recruits to the Armed Forces increases after joining.
The majority of those who start smoking after joining had previously smoked.
Smoking prevalence amongst recruits is positively correlated with both educational attainment and deprivation score.
Deprivation is more important than educational attainment in explaining smoking behaviour amongst recruits.
In 1998 the Government published a white paper on tobacco—‘Smoking Kills’.1 It noted that at a time when smoking was the principal cause of premature deaths in the UK, the numbers who smoked had stopped falling. It also highlighted that smoking hits the worst off people hardest. The 2004 white paper Choosing Health2 had as it's first overarching priority to reduce the numbers of people who smoke. The then government set a new Public Service Agreement with the aim of reducing the overall prevalence of smoking to 21% or less by 2010. This agreement recognised the increased risk of cancer in manual socio-economic groups, with an additional target of reducing smoking in routine and manual groups to 26% or less by 2010. By 2007, data from the General Household Survey in England3 showed that 25% of manual groups were smokers, suggesting that this target was met. However, a reduction to 16% in the proportion that smoked in the non-manual group led to a significant difference between socio-economic groups.
Since the early 1990s, the prevalence of cigarette smoking has been higher among those aged 20–24 than among those in other age groups.3 The tobacco control strategy published in 20104 has the primary objective of stopping the inflow of young people recruited to smoking. While gender differences in smoking behaviour are decreasing, differences between socio-economic groups are not.3 Smoking remains the main cause of preventable disease and premature death in the UK. It is estimated to cause one third of all cancers and, in England alone, over 80 000 deaths per year are due to smoking.4
Since the military population is young and often drawn from deprived areas, smoking in the Armed Forces is an important issue about which there is a body of literature. Estimates of the prevalence of smoking in the Armed Forces range from 28%5 to 35%.6 ,7 A tri-service survey conducted in 19897 reported prevalences of 36% in the Royal Navy (RN), 41% in the Army, and 26% in the Royal Air Force (RAF). One follow-up study found that smoking fell from 28% to 23% over the course of 3 years.8 Another reported that the prevalence of smoking decreased in lower ranks between 1998 and 2004 by 5.1% in 20–24-year olds and 6.3% in 35–49 year olds.9 A study of smoking on deployment found that pre-deployment smoking rates of 29% rose to 38% by the sixth week of deployment.10 Sixty-four percent of the new smokers were personnel restarting an old habit. Smoking amongst young soldiers (aged 15–18 years) has been monitored by a series of questionnaires administered in 1959, 1966 and 1971.11 Although the rates of smoking reported were very high, there was a clear reduction in recruits smoking over this time period. When the survey was repeated in 1988, the prevalence of regular smoking amongst young soldiers was found to be 45%.12 More recent collection of lifestyle data on a small sample of Army recruits found that 40% were current smokers and 11% were ex-smokers.13 A study of Royal Marines (RM) recruits found much lower levels of smoking, at 13%.14
The primary aim of this paper is to describe smoking behaviour in tri-Service recruits to the UK Armed Forces at enrolment and after 3 years of follow-up, and to examine smoking commencement and cessation patterns during this time. The data allows prevalence to be estimated for each trade group, and this can be examined in relation to educational attainment and average deprivation score. Whilst the prevalence data are now somewhat out of date, these results still contribute to our understanding of smoking uptake and cessation behaviour in individual trade groups, and by gender, and how smoking prevalence may be influenced by predisposing factors such as education and deprivation. This is the first step in the effective targeting of heath promotion and smoking cessation services. The importance of tackling smoking in the Armed Forces is twofold—in the short term we require our Armed Forces to be fit for deployment, but we have a duty of care towards our current personnel in terms of their long-term health outcomes, of which smoking is the most important modifiable factor.
In 1998/1999 the Defence Dental Service conducted a survey into the health behaviour of Armed Forces recruits. The target sample was 100% of the Royal Navy intake that year, 80% of the RAF intake and 25% of the Army intake. The sampling in the Army was normally every fourth training course, except where that would leave some trades under represented in which case data on those trades was collected as and when they were enlisted. The recruits were asked to complete a questionnaire (authorised by the tri-Service ethics committee) whilst waiting for their initial dental examination. The questionnaire included oral health behaviour and belief questions from a recognised profiling tool,15 and questions on educational attainment and postcode at the time of recruitment. The questionnaire was piloted on an intake of recruits at RAF Halton. Three years after enlistment these recruits were followed up at a regular dental inspection, but due to the vagaries of response to recall reminders this data collection spanned 2001 to 2003, although 2002 as the ‘3-year point’ is referred to throughout this paper.
Where there were anomalies between the Service and trade group recorded, these were resolved wherever possible by referring to the Service number. The percentage who reported smoking was tabulated by Service, officer/rank status, trade group and gender. This was done for the original sample at enlistment in 1998/1999, the subset of those who were successfully followed up at enlistment, and for the follow-up sample in 2001–2003 itself, to enable changes over time to be compared for equivalent groups; 95% CIs are also shown. Commencement and cessation behaviour was examined in more detail by officer/rank status and trade group by looking at the previous smoking history of those who had started smoking between enlistment and follow-up. Prevalence of smoking at enlistment was also tabulated by age and educational attainment.
To enable the role of broader socio-economic factors to be examined, a measure of deprivation was generated for each recruit using the postcode recorded as their address immediately prior to enlistment. Postcode data was matched to Index of Multiple Deprivation (IMD) 2004 scores.16 IMD 2004 is a measure of multiple deprivation made up of seven indices—income deprivation, employment deprivation, health deprivation and disability, education, skills and training deprivation, barriers to housing and services, living environment deprivation and crime. The Local Authority with the lowest average IMD score (ie, least deprived) is Hart in Hampshire scoring 4.2 compared to the Liverpool Unitary Authority which is the most deprived with an IMD score of 49.8.16 As IMD scores are not comparable across UK countries, this analysis was restricted to recruits with English postcodes. IMD scores were then compared by Service using a Kruskal-Wallis test for equality of populations. The percentage smoking in each trade group was plotted against the percentage with no further education, and the median IMD score in each trade group; correlation coefficients were calculated to describe the strength of the linear associations. Finally, a multiple linear regression model was fitted to assess the relative importance of educational attainment, IMD score, gender and age in explaining the percentage smoking in each trade group.
The total number of questionnaires returned was 10 532 representing a response rate of 99.5%; one individual was excluded from further analyses because their Service could not be determined. At recall in 2001–3, it was possible to collect follow up data on 1809 personnel of the original Army sample of 3596 (50%), 2116/4248 (50%) of the original RN and 1804/2687 (67%) of the original RAF sample. Of the 10 531 recruits included in the analysis, trade group was missing for 402, and gender was missing for 88.
Smoking amongst Army recruits (45%) was more common than amongst recruits to the Navy (34%) or the RAF (31%) in 1998/1999 (Table 1). When the data was re-analysed to show the smoking behaviour at enlistment of just those who were still contactable in 2002, the percentage who reported smoking in each Service was slightly less than for the whole sample but by 2002 the percentage smoking had increased by 7% in the Navy, 6% in the Army and 2% in the RAF.
When the data was stratified by the trade the recruits were intending to follow (Table 2), the most obvious difference at enlistment was between officer cadets and other ranks. Although considerably more Army officers (20%) smoked than officers in the RAF (10%) and Navy (12%), officers were less likely to smoke than any of the other trade groups and this was true within each Service. In the Army smoking was most prevalent in the Royal Logistics Corp (RLC) (56%), Royal Armoured Corp (RAC) (55%) and Infantry (44–51%). In the Navy, Short Engagement Seamen (44%), Engineers (40%), and Mechanics (40%) smoked more than other trades. In the RAF, Communications personnel reported the highest prevalence at 39%. Those military regiments where there is a greater emphasis on physical fitness reported lower levels of smoking than other non-officer trade groups in the same Service; for example, 30% of recruits to the Parachute Regiment and 13% of Royal Marines reported smoking.
Smoking increased during the first 3 years of service in all trade groups with the exceptions of the Intelligence Corp, RAF medical staff and RN Communications (Table 2). The greatest absolute increase in smoking during the first 3 years of service occurred in the Light Division of the Infantry and Junior Artillery personnel. The greatest percentage increase was amongst Royal Marine and RN officers.
When recruits smoking commencement and cessation behaviour was investigated over the period of follow-up (Table 3), it appeared that only 7% of recruits who had previously never smoked had started to do so after enlistment; they were slightly outnumbered by those who had been smoking when they enlisted but had since stopped (12% of those smoking at enlistment in 1998). The group that mostly explained the increase in smoking prevalence was those who reported having previously smoked, had given up by the time they enlisted but who later restarted their habit. Forty three percent of ex-smokers were smoking at follow-up.
Table 4 shows the prevalence of smoking at enlistment and follow-up for each Service by gender. Based on the initial sample at recruitment, the prevalence amongst females in the RAF was higher than amongst males; in the subsample who were followed up this difference halved after 3 years. The prevalence amongst males and females in the Navy was similar at recruitment and follow-up. Smoking prevalence at enlistment was slightly higher amongst males than females in the Army, but the difference increased at follow-up.
Table 5 shows the prevalence of smoking at enlistment and follow-up by trade group and gender, for selected trades with a high percentage of females. In the Navy, female medics were nearly twice as likely to smoke as their male counterparts. In all the Army trades shown, females smoked more than the males. In the RAF the gender differences at enlistment were most extreme. In each Service, the notably higher prevalence of smoking amongst female medics compared with male medics diminished over follow-up (generally due to increases in male smoking). Further analysis of smoking commencement and cessation behaviour by gender revealed that, although certain trade groups saw an increase in female smoking prevalence, the figures were on the whole stable and, particularly in the RAF, some trade groups saw a decrease in females smoking levels, although this was more than offset by the increase in smoking amongst males.
An assessment of the prevalence of smoking by age group (Table 6) suggested a trend towards decreasing prevalence of smoking with increasing age group from 40% amongst 16–17 year olds to 34% amongst 30+ years. In particular, there was a low prevalence amongst 22–23 year olds (statistically significantly lower than amongst the three younger age groups 16–17, 18–19 and 20–21 year olds). The hypothesis that this was due to an influx of non-smoking graduates was tested; the 22–23 year old age group still had a lower smoking prevalence even with graduates excluded, so although they explain part of the reduction in this age group they do not account for all of it.
Table 6 also examines the relationship between smoking prevalence and educational attainment. In all age groups there was a trend towards decreasing smoking prevalence with higher educational attainment and this was clearest in the youngest age groups. There were some notable exceptions, for example, having National Vocational Qualifications (NVQs) over no qualifications was in some age groups associated with an increased prevalence of smoking, and similarly with a degree compared to A’ levels as the highest educational qualification.
It is possible that broader socio-economic factors underlie the apparent relationships between smoking and some of the variables examined in this paper. To further examine this we used postcode data collected as part of the dental survey to generate a measure of deprivation corresponding to each recruit. As explained in the Methods section, this analysis is restricted to recruits from England. Table 7 presents IMD 2004 scores for English recruits by Service, along with smoking prevalence. The percentages smoking by Service presented in Table 7 are somewhat lower than those in Table 1 for the UK. The median and inter-quartile ranges of the IMD scores suggest that recruits to the Army come from more deprived backgrounds than recruits to the Navy, and that recruits to the RAF come from the least deprived backgrounds. There was strong statistical evidence of a difference in IMD scores between Services (Kruskal-Wallis test: χ22=144, p<0.001).
Smoking prevalence in each trade group (across all Services) was plotted against both education and socio-economic status. In Figure 1 the percentage smoking is plotted against the percentage with no further education (correlation coefficient 0.7, p<0.005). In Figure 2, the percentage smoking is plotted against the median IMD score in each trade group (correlation coefficient 0.8, p<0.005).
These explanatory variables (percentage without further education and median IMD score) were entered into a stepwise regression model, along with the median age and percentage female in each trade group, to explain the percentage smoking in each trade group. The results suggested that median IMD is positively associated with the percentage smoking in each trade group (p<0.005) and that an increase in 1 year in the median age of a trade group is associated with a decrease of 3.4% in the percentage smoking (p=0.01). Between them these variables explain 71% of the variation in smoking prevalence between the trade groups. There was no evidence that the percentage without further education and the percentage who were female contributed further to explaining smoking prevalence.
This paper addresses smoking prevalence, and changes in smoking prevalence, based on a survey of Armed Forces recruits (1998/1999) with 3-year follow-up to approximately 2002. Whilst not specifically checked, the proformas were felt to be consistent with the recruits responses at their initial dental examination and oral inspection. Data on cigarette consumption is generally less robust but no attempt was made to collect this. The sample size at enlistment was very large, much larger than even the General Household Survey for the age groups of interest to the Services (although estimates based on subgroups may be less reliable). Several factors contributed to the decreased sample size at follow-up—the drop-out rate of recruits during training, early discharge from Service, and the level of administrative efficiency of dental centre staff in recalling patients and presenting the questionnaire. The data is now considerably out of date and is therefore not intended to estimate current prevalence amongst recruits. However, time lags in reporting longitudinal data are inevitable and the General Household Surveys shows that prevalence levels are only changing slowly. The dataset remains valuable for several reasons. Firstly it enables analysis by age, gender and not only Service but trade group; secondly, it includes a follow-up period of 3 years so that factors associated with smoking cessation and uptake after joining can be examined and finally it contains information on educational level and postcode data which can be linked to a deprivation score.
As expected, smoking at recruitment was higher amongst the Army (45%) than the Navy (34%) and the RAF (31%). An analysis of those in the follow-up sample showed that smoking increased during the first 3 years of Service in the majority of trade groups. This was mostly due to ex-smokers recommencing smoking. The greatest absolute difference in smoking during follow-up was in the Light Division of the Infantry and junior Artillery personnel. Although the Tri-service survey conducted in 19897 was not restricted to recruits (so includes older personnel), the prevalences we report at 3-years follow-up are similar to those found in this survey. The prevalence we found amongst Army recruits is also similar to that reported by Lewthwaite and Graham for Army recruits in 198812 (45%), and our estimate at enlistment of 13% for Royal Marines corresponds directly to the published estimate.14 Apart from the RM, there were significant differences between the smoking behaviour of officers and other ranks but, even amongst officers, smoking prevalence had increased by 2002. The relationship between smoking prevalence and gender is confounded by trade group. In the RAF, females reported a higher level of smoking prevalence than males. Although the reverse was true in the Army, it's likely that this was because the infantry represents a large proportion of Army recruits; our analysis of IMD scores showed that they typically come from the more deprived parts of the country but, at the time of data collection, did not include females. This finding may reflect the type of female attracted to a career in the Services. The national data indicate lower levels of smoking in the older population and to some extent this is born out by this study, but the age spectrum was too small to be able to draw valid conclusions.
The relationship between socio-economic background and smoking prevalence is well-documented and persistent.3 ,17 While the prevalence of smoking amongst recruits at enlistment is likely to be associated with their socio-economic background, it has generally been difficult in the past to classify the socio-economic status of military personnel for the purposes of research, since such frameworks are derived from an individual's occupation. A mapping from military rank to Registrar General's Social Class was proposed in 199918 but has not been widely used. In contrast, measures of deprivation are based on the geographical area in which an individual lives. This is a more useful measure than socio-economic status for recruits, who enter a very limited number of ranks but come from a wide range of areas with differing levels of deprivation.
The differences in smoking prevalence between the Services and between the trade groups probably reflect the technical ability and qualifications necessary to meet the respective recruiting standards. The Army in particular recruits strongly from the socially deprived populations of Scotland, the North East and North West, especially to the infantry. Generally, the inference is that the more educated an individual the more likely they are to recognise the risks of smoking. However, it is possible that level of educational attainment is acting as a proxy for the socio-economic background of that individual and it may well be that the latter is the more important factor. Indeed, we found that the relationship between smoking and IMD score was stronger than that between smoking and education.
The Services have long had a culture of smoking. In the 1950s cigarettes were still issued free in some parts of the world to Service personnel and even now Royal Naval ships outside UK waters and personnel posted overseas have access to duty free cigarettes. The culture allows frequent smoking breaks and leads to non-smokers feeling ostracised from their smoking peer group. It is therefore not surprising that those who have previously smoked decide to resume the habit in order to integrate socially. Boos and Croft10 demonstrated the effect operational detachments could have on raising smoking prevalence and this may well explain some of the increase in smoking after enlistment.
The vast majority of service personnel leave the Services long before the health detriment of smoking becomes apparent but, given the widely known health effects of smoking, the Armed Forces should prioritise tobacco control, whether by a ‘whole population’ approach (using economic tools such as withdrawing duty free cigarette allowance, and tackling the underlying culture that smoking is normal and acceptable) or a more targeted approach. Our findings suggest that certain groups within the Army should be targeted for smoking cessation, and that those most at risk of starting smoking after recruitment are ex-smokers.
Service recruits exhibit smoking behaviour that reflects the socio-economic background from which they are drawn. There are significant differences in smoking prevalence between the Services, between officers and other ranks, and between individual trade groups. It appears that over the first few years of Service life smoking prevalence increases, mostly as a result of those who had previously given up resuming smoking.
Editors Note: Although this paper analyses what is now a historical dataset (from 1998–1999), the messages contained within it remain pertinent to both the Journal’s readership and those they treat and hence publication is justified.
Competing interests None.
Provenance and peer review Not commissioned; externally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.