Article Text

A systematic review on tobacco use among civilian populations affected by armed conflict
  1. Janice Lo1,
  2. Preeti Patel2,
  3. Bayard Roberts1
  1. 1Faculty of Public Health and Policy, ECOHOST—The Centre for Health and Social Change, The London School of Hygiene and Tropical Medicine, London, UK
  2. 2Department of War Studies, King's College London, London, UK
  1. Correspondence to Dr Bayard Roberts, Faculty of Public Health and Policy, ECOHOST—The Centre for Health and Social Change, London School of Hygiene and Tropical Medicine. 15–17 Tavistock Place, London WC1H 9SH, UK; bayard.roberts{at}


Objective To systematically examine evidence on tobacco use among conflict-affected civilian populations.

Data sources Primary quantitative and qualitative studies published in English up to April 2014. Bibliographic databases searched were EMBASE, Global Health, MEDLINE, PsycEXTRA, PsycINFO, Web of Science, Cochrane; with the main terms of: (Smoke*, tobacco*, cigarette*, nicotine, beedi, bidi, papirosi, dip, chew, snuff, snus, smokeless tobacco) AND (armed-conflict, conflict-affected, conflict, war, refugee, internally displaced, forcibly displaced, asylum, humanitarian). Grey literature was searched using humanitarian databases, websites and search engines.

Study selection Studies were independently selected by two reviewers, with a study outcome of tobacco use and a population of conflict-affected civilian populations such as internally displaced persons, refugees, residents in conflict-affected areas, residents and returning forcibly displaced populations returning in stabilised and postconflict periods. 2863 studies were initially identified.

Data extraction Data were independently extracted. The Quality Assessment Tool for Quantitative Studies and the Critical Appraisal Skills Programme for qualitative studies were used to assess study quality.

Data synthesis 39 studies met inclusion criteria and descriptive analysis was used. Findings were equivocal on the effect of conflict on tobacco use. Evidence was clearer on associations between post-traumatic stress and other mental disorders with nicotine dependence. However, there were too few studies for definitive conclusions. No study examined the effectiveness of tobacco-related interventions. The quantitative studies were moderate (N=13) or weak (N=22) quality, and qualitative studies were moderate (N=3) or strong (N=2).

Conclusions Some evidence indicates links between conflict and tobacco use but substantially more research is required.

  • Priority/special populations
  • Global health
  • Low/Middle income country

Statistics from

Request Permissions

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.


There are currently over 50 million people who have been forcibly displaced from their home areas by armed conflict. They include over 33 million internally displaced persons (IDPs) who remain within the borders of their countries and over 17 million refugees and stateless persons who are living in other countries.1 ,2 There are many millions more who remain living in conflict-affected areas or in places that were, until recently, beset by conflict. Many conflict-affected persons will have been exposed to violent and traumatic events, placing them at high risk of post-traumatic stress disorder (PTSD).3 Conflict-affected persons can also experience an increased burden of common mental disorders, particularly depression and anxiety, due to a combination of trauma exposure, poor living conditions, loss of livelihoods, impoverishment and other daily stressors.3–5

Tobacco is often used for the perception that it alleviates stress.6 Research in populations living in non-conflict stable settings has shown that greater tobacco use and nicotine dependence are associated with both PTSD and common mental disorders such as depression, anxiety and alcohol use.7–15 Similarly, military personnel may use tobacco for perceived stress relief (as well as due to many other reasons such as social norms and access to cheap cigarettes).16 It might therefore be expected that civilians exposed to conflict may be more vulnerable to heightened tobacco use.

There is also increasing concern over rising non-communicable diseases (NCDs) among conflict-affected populations, including tobacco-related diseases, as the burden of NCDs grows in low and middle income countries.17 In addition, an increasing number of conflicts have taken place in middle-income countries where both NCDs and tobacco use are high. For example, smoking rates in Syria are estimated at around 48% for men and 8.9% for women.18 Other examples include Iraq, Ukraine, Georgia, Serbia and Kosovo which all have high rates of smoking, particularly among men.18 The humanitarian emergency relief model used over the past few decades is based primarily on immediate health actions of communicable disease control and prevention, water and sanitation, food and shelter. However, as NCDs are becoming more prominent among the conflict-affected populations and protracted crises more common—particularly in middle-income countries, there is an increasing need to address NCD care and prevention, particularly given the treatment costs for tobacco-related diseases.17 ,19

Refugees and IDPs are also now increasingly living in urban areas rather than camp and rural settings.17 The marketing and availability of tobacco products is much higher in urban areas than in rural ones. In addition, it is often more difficult for government and humanitarian agencies to identify and quantify these conflict-affected populations in order to provide adequate health services.17

Postconflict environments may also increase the availability of tobacco. For example, postconflict environments provide opportunities for multinational tobacco companies to potentially influence policies that undermine efforts to control tobacco and increase the marketing and availability of tobacco.20 This may be further exacerbated by weak policy enforcement given the weak governance and policing observed in many fragile settings, including the smuggling of tobacco such as occurred in Montenegro and the Democratic Republic of Congo.21 ,22

Despite the increasing recognition of the need to address NCDs among conflict-affected populations and the key NCD risk attributable to tobacco, to the best of our knowledge no systematic review has examined patterns of tobacco use among conflict-affected civilian populations. The overall objective of our study was to systematically examine evidence on tobacco use among conflict-affected civilian populations. The specific objectives were to examine: (1) the patterns of tobacco use; (2) the association of armed conflict and other factors with tobacco use; and (3) to review the quality of evidence found.


A systematic review methodology was used, following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement.23

The outcome of interest was tobacco use. The population of interest was conflict-affected civilian populations such as IDPs, refugees and residents remaining in their home areas in conflict-affected areas, and also resident and forcibly displaced populations living in stabilised and postconflict periods. We use a standard definition of armed conflict as a contested incompatibility which concerns government and/or territory where the use of armed force between two parties, of which at least one is the government of a state, results in at least 25 combatant battle-related deaths per year.24 We define postconflict as up to 10 years after the formal end of the armed conflict. Studies on serving or former military personnel were excluded.

Primary quantitative and qualitative research studies published up to 15 April 2014 from published and grey literature in the English language were searched. Abstract-only studies and conference abstracts were also excluded.

The bibliographic databases used to search for published literature were: EMBASE, Global Health, MEDLINE, PsycEXTRA, PsycINFO, Web of Science, Cochrane. The following main search terms were applied for the bibliographic databases:

(Smoke*, tobacco*, cigarette*, nicotine, beedi, bidi, papirosi, dip, chew, snuff, snus, smokeless tobacco) AND (armed-conflict, conflict-affected, conflict, war, refugee, internally displaced, forcibly displaced, asylum, humanitarian).

The grey literature was searched through the following websites and humanitarian databases: the WHO, United Nations, Médicins Sans Frontières, ReliefWeb, Active Learning Network for Accountability and Performance (ALNAP), and Google, using amended keywords where necessary. Additional studies were also searched using the references of the final selected studies.

The data extraction fields were selected based on the study objectives and included: author details; study setting and population; tobacco outcome measured; patterns of tobacco use and nicotine dependence (eg, prevalence, knowledge on tobacco use, smoking cessation activities); factors associated with tobacco use (eg, conflict-related factors, sociodemographic factors, mental disorders, information on tobacco control); and quality of studies.

Descriptive analysis conducted as a meta-analysis was not possible due to the heterogeneity of study designs and outcomes. Statistical significance was assumed at p<0.05 when describing the results.

The quality of the quantitative studies was assessed using the Quality Assessment Tool for Quantitative Studies, which was developed by the Effective Public Health Practice Project.25 Eight different areas (selection bias, sampling, confounders, blinding, data collection methods, withdrawals and dropouts, intervention integrity, and analyses) were assessed using this tool and each study was given a global quality rating of strong, moderate and weak (‘strong’=no components were rated as ‘weak’; ‘moderate’=one ‘weak’ component rating; ‘weak’=two or more ‘weak’ component rating). The quality of the qualitative studies was assessed using the Critical Appraisal Skills Programme (CASP) tool.26 Six different areas (research question, study design, data collection methods, ethics, analyses and findings) were assessed using this tool, we then categorised scores of 8–10 as strong; 5–7 moderate; and 1–4 weak.


Study selection and characteristics

The bibliographic database search yielded 2863 studies of which 1075 were excluded as they were duplicates. The title/abstract review then resulted in 1647 studies being excluded, with the main reasons for exclusion being non-English language, non-tobacco related, not conflict-specific or not civilian-specific. This resulted in 141 studies eligible for full text review. A final number of 39 studies then met the selection criteria and were included in the analysis (figure 1).27–65

Figure 1

Results of study screening process.

The final selected 39 studies were published between 1995 and 2014, and the median year of publication was 2008. There were 34 quantitative studies, with the majority being cross-sectional studies (n=33) and one case series study (n=1). The rest were qualitative studies (n=5). There was one mixed methods study which was reviewed for both quantitative and qualitative findings and quantitative and qualitative quality assessment.

These studies were from the European region (n=11), of which 8 were from Eastern European countries, the Middle-East (n=11) and North America (n=11), with the remaining studies from Asia, Africa and New Zealand (n=6).

The study populations included refugees (n=18), IDPs (n=2), conflict/postconflict populations, and specific populations, such as former detainees, students and health professionals. The age group of participants ranged from children to the elderly, with 3 studies specifically on children and 10 on adolescents.

Patterns of tobacco use

Tobacco consumption and dependence

Thirty-four studies provided quantitative information on tobacco use, such as prevalence of tobacco use, cigarette consumption and nicotine dependence (the remaining five studies examined risks only or were qualitative). These findings are summarised in table 1.

Table 1

Summary of final selected studies on tobacco use

The smoking prevalence recorded for adults ranged from 4.7% among female Albanian women in Kosovo to 66.4% among Bosnian refugees in the USA.41 For adolescents, the prevalence of smoking ranged from 2.2% for IDP adolescents in northern Uganda56 to around 70% for Israeli youth.43 When stratified by gender, male smoking prevalence ranged from 10.3% among Somaliland youths61 to 70.7% among Serbian adults during war;62 and female smoking prevalence ranged from 0.6% among Palestinian refugees in Gaza54 to 70% among female Israeli youths.43

Six studies measured tobacco consumption as cigarettes used per day. The prevalence of heavy cigarette consumption (>20 cigarettes/day) ranged from 14.1% among Israeli smokers44 to 70.9% among conflict-affected male smokers in Georgia.57

Two studies of conflict-affected Georgians and Bosnian refugees measured nicotine dependence (both using the Fagerström Test for Nicotine Dependence (FTND)).57 ,65 Among conflict-affected male smokers in Georgia, 41.4% showed high nicotine dependence (FTND score ≥6); and 33.6% showed moderate dependence (FTND score 4–5).57 The study of Bosnian refugees examined risk factors for nicotine dependence (see below).65

Most studies did not define specifically the form of tobacco product measured, and so it was assumed to be cigarette smoking. Three studies explicitly measured the prevalence of alternative tobacco products such as waterpipe smoking, cigar, pipe smoking and chewing tobacco.36 ,45 ,64 For example, Giuliani et al36 found that among Somali youths, the prevalence of waterpipe smoking was 2.3%; 1.7% for chewing tobacco; and 2.3% for cigar smoking. For Palestinian refugee students, the prevalence of waterpipe smoking ranged from 11.7% in Gaza to 33.7% in Lebanon; and the prevalence of smokeless tobacco ranged from 5.5% in Syria to 9.1% in West Bank.45

Perception and knowledge of tobacco use

Ten studies explored the perception and knowledge of tobacco use among their study populations. These studies were conducted with Somali and Bosnian refugees in the USA,30 ,36 ,38 ,41 ,42 Bosnian and other refugees in Canada,46 ,63 refugees in Lebanon,50 youths in Iraq,60 and people from postwar Croatia.64

Two studies analysing data from surveys of Somali youths in the USA and youths in Iraq found smokers among Somali youths in the USA perceived smoking as making them ‘look cool’ (p=0.02) and have more friends (p=0.01) compared with non-smokers;36 while smokers among Iraqi youths felt that boys who smoke have more friends (OR=1.32) and are seen to be more attractive (OR=1.15) compared with non-smokers.60

A qualitative study of youths in refugee camps in Lebanon found that boys who smoke were aware of the negative health effects but continued to smoke. Both boys and girls mentioned smoking as a stress relief and escape mechanism. Furthermore, some youths expressed that waterpipe smoking had a higher social acceptance than cigarettes smoking, and was sometimes encouraged by families in social occasions.50

With regard to knowledge on health hazards related to tobacco use, most participants were aware that smoking is harmful for health and some were also aware of the specific risks associated with cardiovascular diseases.30 ,36 ,64 However, adolescent smokers, compared with non-smokers, had significantly less (p=0.01) awareness of the harm of smoking even lower numbers of cigarettes (1–5 cigarettes per day).36 Another study of Bosnian refugees in the USA showed no significant differences between smokers and non-smokers in terms of their risk perception for heart disease and lung cancer.42 A study of Somali refugee youth in the USA showed they believed that waterpipe smoking was less risky than cigarette smoking.36

Among the 10 studies identified on perception and knowledge of tobacco use, 5 were qualitative studies. In these five qualitative studies, participants described how they used smoking as a stress relief and escape mechanism and as a response to the sense of emptiness caused by war. Other common reasons for smoking included peer pressure, perception of being fashionable, family influences, and for fun or socialisation.38 Smoking was also linked to local culture.46 ,50 A study of Bosnian refugees noted how smoking was identified as a tradition,46 whereas a study of Somali refugees noted how tobacco users were perceived to have a less desirable physical appearance, health status and characters.38

Smoking cessation

Only one study (of Serbians before and after the North Atlantic Treaty Organization (NATO) bombing) reported smoking cessation rates.62 The proportion of respondents having tried to quit smoking ranged from 47.7% for males and 62.4% for females, with success rate of 18.4% for males and 20.5% for females.

Other quantitative studies focused on knowledge, interests, reasons and motivations for cessation.31 ,32 ,37 ,63 ,65 A study of Somali adults in the USA found that commonly cited reasons for cessation were related to health reasons, parental disapproval and cigarette price.37 In a study of the elderly in Lebanon, the presence of chronic disease (OR=4.29) and disability (OR=1.79) were shown to be significantly related to smoking cessation, but depression was not significantly related to smoking cessation.31 A study of Bosnian refugees found the interest in smoking reduction or cessation was significantly associated with access to information on smoking cessation provided by healthcare providers, but not the severity of nicotine dependence.65 Another study explored changes in smoking among Bosnians during the war in Bosnia and Herzegovina and found that the common reasons for cessation were related to the supply and price of cigarettes.32

A qualitative study of Somali refugees in the USA found that the most common cessation barriers cited by the participants were the lack of health insurance, not knowing what is effective and where to go, and absence of organised support groups for quitting.38 In a later quantitative study by the same author, it was found that only 11.9% of smokers had ever tried a cessation programme, and there was also a low willingness (6–8%) to participate in cessation programmes.37 A qualitative study with immigrant and refugee youth in Canada observed factors encouraging smoking cessation included health concerns, parental disapproval, high price of cigarettes and disapproval from friends.38 A qualitative study of Palestinian youths found that participants were aware of the negative health effects of smoking but continued to smoke nevertheless.50

The association of armed conflict and other factors with tobacco use

The results on the factors associated with tobacco use are given in table 2 and described below according to factors of armed conflict, mental health, demographic factors and tobacco control.

Table 2

Summary of risk factors associated with tobacco use

Armed conflict and exposure to traumatic events

Fourteen studies examined the relationship of armed conflicts with tobacco use. Eight studies compared tobacco use between conflict-affected populations and non-conflict-affected populations.31 ,35 ,47 ,51 ,53 ,54 ,57 ,59 Two of these studies found that conflict-affected populations had a significantly higher prevalence of tobacco use than non-conflict-affected populations.31 ,35 However, a study of Palestinian refugees observed the overall prevalence of tobacco use was significantly lower in the refugee camps than the neighbouring host areas, but male refugees were found to have a significantly higher prevalence than the host male population.54 The study from Georgia observed a higher prevalence of tobacco use in male IDPs than male former IDPs who had returned to their home areas (returnees), but male returnees were found to have a significantly higher level of nicotine dependence than current IDPs.57

Five studies examined the effects of conflict by comparing tobacco use between preconflict and conflict-affected periods.27 ,30 ,32 ,44 ,62 An increase in tobacco consumption (cigarette per day) was observed by all five studies, but only statistically significantly in two of these studies.27 ,32 The increase in tobacco consumption was found to be significantly affecting females more than males in the study from Israel (OR=2.44, p<0.001).44 A study of Palestinian and Israeli children aged 11–15 years observed that subjective threat from armed conflict was associated with increased tobacco use.39 A qualitative study of Bosnian refugees observed that respondents reported increased tobacco use during exile and after settling in a new country (USA).30

Mental health

A study of adult Bosnian male refugees in the USA found a significant association (χ2=11.3, p<0.005) between smoking and PTSD.33 However, the other study in postwar Kosovo found this relationship to be non-significant (OR=1.2, p=0.36).34

Two studies examined the association between nicotine dependence and mental health.57 ,65 Roberts et al57 observed that nicotine dependence was significantly associated with PTSD (p=0.02) and depression (p=0.03) among conflict-affected persons in Georgia (but no significant association was observed with anxiety disorder). Weaver et al65 observed PTSD severity was significantly and positively associated with nicotine dependence severity (p<0.01) among Bosnian refugees in the USA, and this significant relationship was also found when individual symptoms of PTSD were examined, such as irritability (p<0.05), concentration difficulties (p<0.05), and hypervigilance (p<0.01).

Two studies examined the association between tobacco use and harmful alcohol use. The first reported a significant association between tobacco use and alcohol consumption among Bhutanese refugees in Nepal.49 The second study did not observe a significant association between nicotine dependence and hazardous alcohol use among conflict-affected persons in Georgia.57

Demographic factors

Twenty-two studies examined associations between demographic factors and tobacco use. The most common association studied was between tobacco use and gender (n=18). Thirteen studies found significant association between male gender and higher tobacco use.28 ,34 ,37 ,40 ,44 ,45 ,52 ,54 ,55 ,58 ,60 ,62 ,64 Four studies found no statistical significance between gender and tobacco use,29 ,56 ,61 including one looking at the association between gender and nicotine dependence among Bosnian refugees in the USA.65 A study on Israelis youths found significant association between higher cigarette use and female gender.43 Men were also found to have a longer smoking history than women.28

Eight studies examined the association between age and tobacco use and all observed that higher prevalence of tobacco use, consumption and nicotine dependence were significantly associated with increase with age, particularly between 20 and 50-year olds.34 ,37 ,40 ,44 ,57 ,60 ,64 ,65 However, this association was found to have diminished by 50–60 years of age in two studies.37 ,57

Low education was found to be significantly associated with smoking in the study in Israel,44 but was found to be non-significant in the study of Somali adults in the USA.37 A study on Bosnian refugees in Sweden analysed the association between smoking and urban or rural setting but found no significant association.29

Parental cigarette smoking was found to be significantly associated with smoking in two studies of youths in Iraq and refugees in Canada.60 ,63 Exposure to tobacco use by close friends or families was also significantly related to smoking in the study of refugees in Canada and the study of Somali refugees in the USA.36 ,63

Two qualitative studies explored tobacco use among Somali refugees in the USA and Palestinian youths from a Lebanese refugee camp using focus groups and interviews. The Palestinian youths felt that smoking behaviours among adolescents were under-reported due to fear of social undesirability and parental disapproval.50 Similarly, smoking prevalence among Somali women was perceived to be inaccurate as women tend to hide their smoking status due to social undesirability.38

Tobacco control

One study looked at the association of antismoking messages and smoking status among youths in Somaliland.61 It observed that youths were less likely to smoke if they had heard or seen a few (OR=0.75, 95% CI 0.58 to 0.96) to a lot (OR=0.68, 95% CI 0.55 to 0.84) of antismoking messages in the media. This difference was also observed with religious organisations discouraging young people from smoking (OR=0.70; 95% CI 0.6 to 0.82). Furthermore, youths were also less likely to smoke if they were taught that most people of the respondent's age do not smoke (OR=0.81; 95% CI 0.69 to 0.95) or the negative effects of smoking on appearance (OR=0.62; 95% CI 0.52 to 0.74). However, seeing a lot of antismoking messages at sports events, fairs, concerts or social gathering meant they were actually more likely to smoke (OR=1.3; 95% CI 1.06 to 1.60).

Two studies by Giuliani et al36 ,37 compared the differences between Somali smokers and non-smokers in terms of exposure to protobacco influences. Their study on Somali youth observed that a significantly higher proportion of smokers had seen tobacco advertising targeted at the Somali community compared with non-smokers (p=0.01). The Somali youth smokers also showed greater acceptance (p=0.02) than non-smokers in using or wearing something with a tobacco brand name. However, no significant difference was observed between smokers and non-smokers in terms of seeing athletes or actors using tobacco on TV.36 In their other study on Somali adults, the difference between smokers and non-smokers in seeing tobacco advertisements targeted at the Somali community was not significant (p=0.94). However, smokers were found to show greater acceptance in using or wearing something with a tobacco brand name (p<0.01).37

The study by Khader et al45 examined cigarette promotion among youths in refugee camps and host populations in five Middle Eastern countries. It found the prevalence of having an object with a cigarette brand logo ranged from 12.9% to 22.1%, and the proportion of youths who had a free cigarette offered by a cigarette company representative ranged from 8.6% to 16%.

Quality assessment

Quality assessment was conducted using the Quality Assessment Tool for Quantitative Studies for the 39 quantitative studies, including the quantitative component of mixed methods study (n=35); and using the CASP Qualitative Research Checklist for the qualitative studies, including qualitative component of mixed method study (n=5). The results were classified into ‘weak’, ‘moderate’ and ‘strong’ as described previously.

For the quantitative studies, 13 (37%) of them were rated as moderate quality, and 22 (63%) of them were rated as poor quality. None of these studies were rated as strong quality. The quality scores were low for these studies mainly due to their study designs (most were cross-sectional designs rather than randomised controlled trials or cohort studies) and weak adjustment of confounders. See online supplementary appendix A for further details.

For the qualitative studies, three studies were rated as moderate quality and two were rated as strong quality. In general, there was a lack of evidence in justifying the methods with the research question posed, as well as consideration of the influence of the researchers on study participants. Furthermore, discussion of ethical issues associated with the research was also not described by most studies. See online supplementary appendix B for further details.


To the best of our knowledge, this is the first systematic review to examine patterns of tobacco use among conflict-affected civilian populations. Thirty-nine studies were selected for review. They looked at tobacco use in various conflict-affected settings and among populations such as refugees, IDPs and population living in conflict and postconflict areas.

Large discrepancies on the prevalence of tobacco use were seen among the reviewed studies, ranging from 2.18% to 66.4%.41 ,56 These differences could be explained by the large variations of characteristics between the study populations, such as gender, age, ethnicity and geographic location. This supports the assumption that prevalence of tobacco use among conflict-affected population is likely to be strongly influenced by the pre-existing prevalence of tobacco use prior to conflict.

There were mixed findings on the effects of conflict on tobacco use. Some studies observed a significantly higher prevalence of tobacco use among conflict-affected populations than non-conflict populations, particularly in males. However, other studies found this association to be non-significant. Some studies had also observed an increase in cigarette consumption among the participants during the times of conflict, with one study showing a significantly higher increase in females than males.44 However, due to the limited number of studies, no definite association can be derived. In addition, there was no evidence on how patterns of tobacco use may have changed over time in postconflict settings. This highlights the need for future research to better understand the influence of armed conflict on tobacco use, including in postconflict settings. This includes the need for more in-depth qualitative research to better understand potential explanations for the changes in tobacco use during different conflict periods. Only five qualitative studies were identified in our review, compared with 34 quantitative studies.

Five studies examined the association between mental disorders and tobacco use among conflict-affected populations. In these studies, tobacco use and nicotine dependence were found to be significantly associated with PTSD. However, the number of studies were again too few to provide conclusive evidence. Nonetheless, other systematic reviews from non-conflict-affected populations have highlighted that smoking is associated with mental health disorders such as PTSD, depression and anxiety disorders.11 ,66 It is also known that armed conflicts contribute significantly towards mental health disorders, and conflict-affected populations have high prevalence of PTSD and depression due to the stress from exposure to traumatic events and daily stressors.3 ,4 ,67 Therefore, tobacco use and nicotine dependency could potentially be higher among conflict-affected persons who also suffer from mental disorders, but further research is required on this.

There is potential for tobacco companies to take advantage of conflict-affected or postconflict states where tobacco control legislation and enforcement may be limited due to weak governance. However, the review found few studies researching tobacco policy issues in conflict-affected countries and how they may influence tobacco use. There was some evidence of the tobacco industry targeting youth populations by giving out free cigarettes and tobacco company merchandises in the Middle East, as well as advertising through the internet.36 ,45 Studies which focus specifically on policy issues (and so did not meet our review inclusion criteria of measuring tobacco use) have found that transnational tobacco companies have been heavily targeting the large markets in Middle Eastern countries. For example, in Lebanon, transnational tobacco companies used the issue of cigarette smuggling to negotiate the need for local manufacturing under a weakened government, which in turn provided opportunities for the transnational tobacco companies to gain access into neighbouring countries (which then provided the opportunity for increased cigarette smuggling).This study suggested that cigarette companies were actively supplying cigarettes for the purpose of smuggling.68 A study looking at cigarette smuggling in conflict-affected central and eastern Africa found that cigarette smuggling resulted from weak state capacity, high level of corruption and financial support for armed rebel groups fighting in the conflicts there.21 Similar findings were also observed in Montenegro.22 Smuggling has been an important component of British American Tobacco's market entry strategy in order to gain leverage in negotiating with governments for tax concessions, compete with other transnational tobacco companies, circumvent local import restrictions and unstable political and economic conditions and gain market presence in several conflict-affected countries in Africa.69

There were a number of methodological limitations arising from the studies reviewed. Most of the studies included were cross-sectional studies, which did not allow determination of a causal relationship between armed conflict and tobacco use to be studied. The exposure to conflict, in terms of timing, severity and longevity, was largely impossible to measure. Therefore, it was possible that some studied populations were more affected by conflicts than others, and there was also no clear definition of the ‘prewar’ and ‘postwar’ periods or longevity of refugee or IDP status period. There were different definitions and measurements of tobacco outcomes, with some giving specific types of tobacco products and some just using ‘smoking’ as an outcome. There was also poor reporting in areas such as description of data sources and outcome measures, sample size, statistical methods, limitations and, in particular, biases. Other methodological issues included a lack of experimental intervention or cohort studies which would give better quality evidence and allow a better understanding of causality as well as effectiveness of smoking interventions. While fully recognising the substantial complexity and challenges of conflict and postconflict settings, there is an increasing need for interventions to reduce tobacco use among conflict-affected populations and research to measure the effectiveness of such interventions.

Study limitations

This review only included studies that were in English. Only descriptive analysis was used as the limited number of studies and heterogeneity of exposures and outcomes used precluded a meta-analysis. Owing to the small number of studies identified, the findings on correlates of tobacco use should be treated with caution.


This systematic review highlights the potential influence of armed conflict on tobacco use among conflict-affected civilian populations. Healthcare responses in conflict and postconflict situations will increasingly have to move away from the traditional emergency relief model to address longer term chronic care associated with NCDs. Given the strong association between tobacco use and a number of NCDs, tobacco use is likely to become an increasingly pertinent issue among conflict-affected populations. This review has also highlighted gaps in current research on tobacco use among conflict-affected civilians, and substantially more research is required on this issue to better understand the effects of conflict on tobacco use and effectiveness of interventions to address it.

What this paper adds

  • Experiencing armed conflict could increase tobacco use among conflict-affected civilian populations through exposure to traumatic events, increased stress and elevated mental disorders. The postconflict environment also presents a number of potential risk factors. This is the first systematic review to examine patterns of tobacco use among conflict-affected populations.

  • The study highlights the limited evidence on this subject. This precludes drawing any conclusions on the influence of conflict and postconflict environments on tobacco use, but post-traumatic stress disorder and other mental disorders do show associations with nicotine dependence.

  • Further research is required on this issue, particularly given the rising concern among humanitarian agencies on how to effectively address the increasing challenge of non-communicable disease among conflict-affected populations.


Supplementary materials

  • Supplementary Data

    This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.

    Files in this Data Supplement:


  • Contributors BR, JL and PP were involved in conceptualising and designing the study. JL and PP conducted the data screening, extraction and analysis. JL led the writing, with contributions from BR and PP. BR, JL and PP reviewed and approved the final manuscript. BR is responsible for the work as guarantor.

  • Competing interests None.

  • Provenance and peer review Not commissioned; externally peer reviewed.

  • Data sharing statement The additional unpublished data from the study consist of the raw extraction data held in an Excel database. The authors would be happy to share this and any interested parties can contact the corresponding author to obtain this.