Smoking prevalence and the changing risk profiles in the UK ethnic and migrant minority populations: implications for stop smoking services

Abstract

Objectives

Smoking is the leading risk factor for disability-adjusted life-years, yet evidence with which to establish the smoking rates of people with different ethnic backgrounds and how they are changing in relation to recent migration is lacking. The objective is to provide current information on the changing risk profiles of the UK population.

Study design

Observational study using cross-sectional surveys.

Methods

Data from the Integrated Household Survey (pooled for the years 2009/10–2011/12), obtained under Special Licence, and the GP Patient Survey (2012) have been used to establish smoking prevalence in a wider range of ethnic groups in England and Wales, including the ‘mixed’ groups and amongst East European migrants, and how such prevalence differs across socio-economic classes.

Results

Smoking prevalence is substantially higher amongst migrants from East European countries (that for males exceeding 50% from three such countries and for females over 33% from four countries) and from Turkey and Greece, compared with most other non-UK born groups, and amongst ethnic groups is elevated in the ‘mixed’ groups. Rates are highest in the Gypsy or Irish Traveller group, 49% (of 162) and 46% (of 155) for males and females respectively. Across ethnic groups, rates are almost always higher in the UK born than non-UK born population with the notable exception of the ‘White Other’ group, with Prevalence Ratios (PRs) indicating a larger migrant-non-migrant differential amongst females (e.g. Indians 2.95 (2.33–3.73); Black Caribbeans 3.28 (2.73–3.94). Age-adjusted rates show the persistence of these differentials in females across age groups, though young males (18–29) in seven minority ethnic groups show lower rates in the UK-born groups. The ‘White’ and ‘Chinese’ groups show a strong socio-economic gradient in smoking which is absent in the South Asian groups and diminished in the ‘mixed’ and black groups.

Conclusions

Given the evidence that smoking behaviour is significantly different in some of the new groups, notably East European migrants, stop smoking services are failing to optimize the acceptability and, consequently, favourable outcomes for these programmes. These services need to be adapted to the particular patterns of smoking behaviour and language skills within different communities of descent.

Introduction

Using the findings of the Global Burden of Disease Study, 2010,¹ investigators report that the leading risk factor for disability-adjusted life-years (DALYS) was the harmful effects of tobacco including second-hand smoke (as it had been in 1990), accounting for 11.8% (10.5–13.3) of all DALYS in the UK in 2010. This is despite a 41% (95% UI [Uncertainty Interval] 35–46) decrease in attributable DALYS. These statistics, described as ‘a more advanced epidemic than most high-income countries’, underline the need for the active pursuit of tobacco control policies in the UK. While smoking cessation services were introduced through the National Health Service in 1999,² the NHS National Institute of Clinical Excellence has emphasized that reducing the prevalence of smoking among people in ‘routine and manual groups’ (the NS-SEC condensed category, combining lower supervisory and technical occupations and semi-routine and routine occupations), some minority ethnic groups, and disadvantaged communities will help reduce health inequalities more than any other measure to improve the public's health.³ Yet accurate prevalence data is needed to target campaigns for minority ethnic communities in the local population and to provide tailored advice, counselling and support in the clients' chosen languages.

In 2011 Karlsen and colleagues noted that evidence with which to establish the smoking rates of people with different ethnic backgrounds and how they are changing is lacking.⁴ Hitherto, the most frequently cited data on the prevalence of smoking amongst minority ethnic groups was that collected in the 2004 Health Survey for England (HSE)⁵ and previously the 1999 HSE,⁶ enhanced by ethnic boosts. Both surveys reported data for the ‘Black Caribbean’, ‘Black African’, ‘Indian’, ‘Pakistani’, ‘Bangladeshi’, ‘Chinese’, and ‘Irish‘ groups, the 2004 survey adding ‘Black Africans’. These investigators provided more up-to-date data, pooled from the HSE for 2006–08, but conceded that methodological issues limited their ability to distinguish between groups which may be heterogeneous: ‘Changing migration patterns are likely to affect the smoking profile of the migrant population in the UK’. There is now empirical evidence to address this issue, enabling smoking prevalence for the many diverse heterogeneous – so-called polyethnic – minorities concealed in some of the residual ‘other’ ethnic categories, the ‘mixed/multiple’ census categories, and new groups added to the 2011 Census (‘Gypsy or Irish Traveller’, ‘Arab’, and, in Scotland, ‘Polish’) to be established.⁷

This selectivity in reported groups is now a significant drawback because of the impact of increasing ethnic and cultural diversity of the population resulting from increasing migration to the UK from an ever larger range of countries, many with no connection to Britain's colonial past; increased population mixing and inter-ethnic union formation; and the increasingly diverse ways in which the population chooses to describe its ethnic identity. For example, the number of usual residents in England and Wales born outside the UK increased from 4.6 million to 7.5 million during 2001–11: in 2011 there were 1,114,368 migrants from the EU countries that acceded between April 2001–March 2011 living in England and Wales,⁸ including 579,000 born in Poland. In England and Wales around 1.22 million persons were enumerated as ‘mixed/multiple’ in the 2011 Census,⁷ an 85% increase on the 660,000 enumerated in the 2001 Census, demographers projecting the figure to rise to 1.6 million by 2031.⁹ The seven minority ethnic groups identified in the 2004 HSE accounted for only 5.49 m persons, or 50.2% of the population of 10.94 m who were not white British in 2011.

There is evidence that the ethnicity risk profiles for the currently smoking population are changing commensurately, resulting over the last decade from globalized migration from a number of countries with high smoking rates but also the continually increasing ‘mixed’ population which has traditionally had high smoking prevalence rates. Small-scale research studies have shown high smoking rates amongst East European migrants, and groups such as Turks and Greeks.10, 11 Pooled data from the General Household Survey, 2001–2005, revealed that mixed ‘White and Black African’ men had the second highest smoking prevalence rate of 38% (after Bangladeshi men, 45%) while mixed ‘White and Asian’ women (33%) and mixed ‘White and Black Caribbean’ women (29%) had the highest female smoking rates.¹²

The time-limitedness of the conventional categories used to report smoking prevalence now needs to be addressed to help the new Clinical Commissioning Groups¹³ commission services for a population of rapidly changing diversity. While some of these new population groups comprise only a small proportion of local authority populations (for example, in the top five ranking local authorities, the Polish-born group in 2011 comprised only 4.1–6.4% of the population) they may have a disproportionate effect on stop smoking services because of their substantially higher smoking prevalence rates. Moreover, in some London boroughs which have been the recipients of high volumes of international migration over the last decade or two,⁸ the cumulative impact of these new groups may be significant though possibly concealed in total population prevalence data because of low rates amongst females in the Asian/Asian British, Black African and other groups. Given accumulating evidence that smoking behaviour is significantly different in some of these new groups, notably East European migrants, stop smoking services may be failing to optimize the acceptability and, consequently, favourable outcomes for these programmes.