Smoking in movies and initiation of smoking in adolescents

Cigarette smoking has a long celebrated history in cinema. Everyone from the glamorous Audrey Hepburn (Breakfast at Tiffany’s), to the legendary Marlon Brando (A Streetcar Named Desire) has been seen “lighting-up” on the silver screen. Smoking is the ultimate cinematic device to portray a character as “sexy”, “rebellious”, “sophisticated” and dare I say it “cool”.

In countries such as the UK where sponsorship and advertising by tobacco companies has been banned, portrayal of smoking in films is the “last frontier” in smoking promotion. In their 2015 report “Smoke-free Movies” the World Health Organisation (WHO) highlight that in 2014 alone 44% of all films, and 36% of films rated for young people contained images of smoking. In response to these figures the WHO recommends that films containing smoking should be rated as adult films. This suggestion is in light of recent evidence indicating a 37% risk for adolescent smoking initiation due to exposure to smoking imagery in films (Sargent et al., 2012).

Despite evidence suggesting an association between exposure to smoking imagery in films and increased uptake of smoking in adolescence, no evidence to date has examined the magnitude of this effect. The present study (Leonardi-Bee et al., 2016) aims to provide a systematic review and meta-analysis of this association.

In 2014, the WHO recommended that films containing smoking should be adult rated.

Methods

Inclusion of studies

Studies were epidemiological (cross sectional and cohort) that reported the association between exposure to smoking imagery in films and smoking initiation in adolescents (aged 10-19 years). Longitudinal associations were only included if studies recruited adolescents who were never-smokers at baseline.

Search strategy

MEDLINE, EMBASE, PsycINFO and International Bibliography of the Social Sciences databases were searched using MeSH and text words for ‘smoking initiation’ and ‘movie’ as search terms.

Screening and data extraction

Papers were independently reviewed and data were extracted by two authors.

Data were extracted on:

Study design
Data collection
Definition of exposure
Smoking uptake outcome
Inclusion and exclusion criteria
Country, settings
Demographics of participants
Limitations of studies.

The quality of included studies was assessed using the Newcastle-Ottawa Quality Assessment Scale.

Data synthesis

Random effects meta-analyses were used to estimate pooled relative risk of the effect of exposure to smoking imagery and smoking initiation. Effect estimates were adjusted for socioeconomic status and demographics. Continuous and categorical measures of smoking exposure were pooled in meta-analyses.

Results

A total of 17 studies were selected for inclusion in the review and meta-analyses:

9 studies were cross sectional design and 8 were longitudinal
7 studies were conducted in the US, 2 in Mexico, 3 in the UK, 2 in Germany, 1 in India and 6 in a group of 6 European countries
The age of participants ranged from 7 to 19 years
The median sample size for studies was 4,919 for cross sectional and 2,298 for longitudinal studies

Most studies examined smoking exposure in top grossing films, using a measure based on the total of smoking occurrences in a viewing of the film. The majority of studies classified exposure into quartiles and 5 studies analysed exposure as a continuous variable.

Systematic review

All cross sectional studies reporting “ever tried smoking” as their outcome variable
Longitudinal studies used smoking initiation in adolescents who had never smoked at baseline
All studies reported adjustment for socioeconomic status, with other common confounders including age, sex, school performance, parental/sibling smoking status, parenting style and sensation seeking
7 out of the 9 cross sectional studies were deemed to be of a high quality
There was no evidence of publication bias in cross sectional studies, but some evidence of bias in longitudinal studies

Meta analyses

Cross sectional studies found that higher exposure to smoking in films significantly increased the risk of having ever tried smoking by 1.93 (95% CI 1.66 to 2.25; I²=60%)
Longitudinal studies found that higher exposure to smoking in films significantly increased the risk of smoking initiation by 1.46 (95% CI 1.23 to 1.73; I²=90%)
A sub-group analysis comparing risk estimates between study country found a significant difference between pooled estimates from longitudinal studies, where relative risks were lower in the US than elsewhere
A post-hoc subgroup analysis of longitudinal studies found a larger magnitude of effect in studies that quantified exposure using quartiles vs. continuous measures. No difference was observed for cross sectional studies

Main findings

The study is the first meta-analysis of longitudinal studies examining the association between exposure to smoking imagery in films and risk of becoming a smoker in adolescents. The study also extends previous meta-analyses of cross sectional studies of this association.

The main finding from longitudinal data suggests that adolescents who are most exposed to smoking imagery in films are 40% more likely to become smokers than those with the lowest exposure.

The analysis of cross sectional data indicates a slightly lower risk for smoking initiation (just under a two-fold increase) than reported in previous meta-analyses.

The cross-sectional association between young people reporting having seen smoking imagery in films and smoking status was greater than the prospective association.

Interpretation and implications

Previous studies have established an association between smoking imagery in films and smoking uptake. However, this body of work has been limited to the use of cross sectional and longitudinal studies carried out in the same cohort. Such studies are prone to bias and do not represent independent studies. The present meta-analysis strengthens the evidence base for the association between exposure to smoking imagery and adolescent smoking uptake, by reporting the magnitude of effect for this association.

The present study also adjusts for socioeconomic status among other potential confounders. This is particularly relevant as previous work has indicated that adolescent sensation seeking (Heatherton et al., 2009), the nature of the smoking exposure (Sargent et al., 2007) and parental restriction on viewing adult films and peer networks (Wills et al., 2007) may all mediate the association between exposure to smoking imagery and smoking initiation.

Interestingly, the participant population age ranged from 7 to 19 years. The WHO defines adolescence as the period following childhood that occurs from ages 10-19 years. Therefore three of the included studies appear to include both children and adolescents. At present it is not known if exposure to smoking imagery in films is associated with smoking initiation in both children and adolescents equally.

The majority of studies included in the present review and meta-analysis estimated exposure to smoking imagery in top grossing or popular contemporary films. Whilst there is a current trend for a decrease in the levels of smoking content in films (Lyons et al., 2010) this does not account for young people viewing a wider range of films. Included studies rarely examined exposure to smoking in films watched by adolescents using TV, DVD and streaming services.

This study does not consider the breadth of adolescent movie viewing across different TV and streaming platforms, which is now commonplace.

The included studies did not focus on the breadth of adolescent movie viewing across different TV and streaming platforms, which is now commonplace.

Conclusion

Findings from the present study add to previous research, supporting the notion that smoking imagery in films increases risk of smoking for young people. These findings also lend support to the recommendations of the WHO, who call for restrictions on adolescent exposure to smoking imagery in films. The authors of the present paper suggest several strategies for preventing exposure to smoking imagery including:

Classifying films with smoking content as suitable for over 18 year olds only,
Broadcasting TV films with smoking content after the watershed, and
Showing anti-tobacco messages before and after the broadcast of films with smoking exposure.

As the authors of the present paper point out, films are not the only platform via which exposure to smoking imagery may occur. Young people are frequently exposed to cigarette smoking via TV and music videos. Therefore methods of preventing and/or controlling exposure may need to be widened beyond the scope of the silver screen.