What can the 2001, 2011 and 2021 UK censuses tell us about health inequalities among young people? A cross-sectional study using censuses from England, Wales, and Scotland

doi:10.21203/rs.3.rs-4724409/v1

Download PDF

Research Article

What can the 2001, 2011 and 2021 UK censuses tell us about health inequalities among young people? A cross-sectional study using censuses from England, Wales, and Scotland

https://doi.org/10.21203/rs.3.rs-4724409/v1

This work is licensed under a CC BY 4.0 License

You are reading this latest preprint version

Background

Health inequalities are avoidable and unjust differences in people’s health across population groups. The WHO has estimated that up to 70% of premature deaths among adults are related to behaviours initiated in adolescence. However, evidence on health inequalities among adolescents is scarce compared to that among adults and younger children. This study examined the association between self-rated health and a family’s socioeconomic position in young people (10-24-year-olds).

Methods

The outcome variable is the single question of self-reported general health. The study population included young people aged 10 to 24 years. We analysed 2001, 2011 and 2021 census microdata from England and Wales and 2001 and 2011 data from Scotland. We used descriptive methods to determine differences in general health status by sociodemographic variables and logistic regression analysis to calculate the odds of reporting poor health according to the National Statistics Socio-economic Classification of the family reference person. Logistic regression models controlled for the effects of age, gender, household deprivation, and UK region.

Results

Compared to young people from the households where the reference person belonged to the managerial or professional occupational group, all other groups except those self-employed in 2001 had greater odds of reporting poor health. These differences were statistically significant (p < 0.001). The odds were highest among young people from households where the reference person had never worked or was long-term unemployed: 2.7 times in 2001, 3.1 times in 2011 and 3.6 times in 2021. The odds of reporting poor health increased with age and were greater among girls than among boys. Young people from households that were deprived in the dimensions of education or housing had greater odds of reporting poor health than did those from nondeprived groups.

Conclusions

Health inequalities in the self-reported health of 10–24-year-olds can be observed using UK census data. We observed a similar social gradient of self-reported poor health across the 2001, 2011 and 2021 datasets; as disadvantage increased, the share of young people with fair or poor health increased.

Young people

Adolescents

UK censuses

Self-reported health

Socioeconomic status

National Statistics Socio-economic Classification (NS-SEC)

This study investigates health inequalities among young people in England, Wales and Scotland by utilising data from the UK Censuses. Public Health Scotland defines health inequalities as “differences in people’s health across the population and between specific population groups, that are systematic, avoidable, and unfair” [1]. Until recently, overall population health has improved; however, people from lower socioeconomic (SE) groups still suffer more disease and die earlier than those from higher SE backgrounds [2–5]. The WHO has estimated that globally, approximately 70% of premature deaths among adults are related to behaviours initiated during adolescence [6]. However, different factors have led to the collection of evidence on health inequalities mostly for younger children and adults [2, 4, 7]. Compared to older age groups, clinical diagnoses are not common in adolescents [8–10], and it is thought that emphasis on early childhood has helped to reduce the mortality of children under five years of age [2]. However, many chronic disease risk factors are already present during adolescence [11], and self-reported health (SRH) starts to decline during adolescence alongside the increasing number of nonspecific symptoms [8, 12–14]. The SRH measure has been shown to capture functional and psychosocial aspects of individuals’ lives, predict future health outcomes, and has proven to be a useful measure for researching young people’s health [8, 10, 12, 13, 15].

Another reason for the lack of evidence is the difficulty of measuring young people’s socioeconomic position (SEP). Individuals’ SEP is usually measured by objective characteristics such as educational level, income, or occupation-based indicators; however, these indicators may not be appropriate at this life stage [16–19]. A widely used strategy is to use measures that reflect a household’s SEP, such as parental occupation [3, 16, 19, 20], household income, parental education or employment status [17, 19, 21–23]. The Family Affluence Scale (FAS) was developed to measure adolescents’ SEP and has since been widely used [24–26]. Evidence also indicates that adolescents’ SEP when assessed by subjective measures such as perceived family wealth [19, 20, 22, 27, 28], academic performance or pocket money [18, 19, 29], or subjective social status [19–22, 28, 30, 31] is a valid indicator in health inequality research [2–4, 6]. Some studies have shown that youth are in a period of relative health equality [32, 33]. However, most evidence indicates that health inequalities are present in childhood and adolescence [3, 19, 21, 27, 28, 34–36] and persist into adulthood [36–40].

This study investigated health inequalities based on the UK censuses; we utilised the UK National Statistics Socio-economic Classification (NS-SEC) of the household reference person (HRP) as a measure of young people’s SEP. The SEP includes the components of occupation, income, gender, and social class and is acknowledged as one of the structural drivers of health inequalities in the Commission on Social Determinants of Health (CSDH) Conceptual Framework [41, 42]. Additionally, labour market position captures factors such as the level and stability of income, prospects of advancement and stability, and an overall reflection of a person’s place in society [17, 43–45]. The NS-SEC consists of the groupings of occupations that share similar employment relations, such as type of contract or independence at work, and employment-related benefits [17, 45, 46]. The NS-SEC measure was developed to provide SEP for all population groups regardless of their employment status [45]. Family has been a traditional unit of analysis in sociology; family members have shared resources and decision-making processes and are thought to share the same class fate [45, 47]. The family context is a recognised determinant of health [4, 48], and evidence indicates that the occupational status of the HRP or parental SEP is related to different health outcomes and self-assessed health [3, 7, 20, 21, 31, 49–51].

Few studies have utilised data from the UK Censuses to study health inequalities among young people [51]. Studies that included the adult population have revealed health inequalities across UK regions, NS-SEC categories and gender, ethnic groups or population groups with specific health conditions [52–56]. Among children and adolescents, Spencer et al. [51] found statistically significant differences in the prevalence of long-term limiting illness by parental NS-SEC. The aim of this study was to examine the association between the NS-SEC of the HRP and the SRH of 10–24-year-olds while controlling for the effects of age, gender, household deprivation measures, and the UK region.

Design and sample

This cross-sectional study utilised data from the 2001, 2011, and 2021 UK censuses. For the 2001 Census, UK Data Archive (UKDA) Study 7207 Census 2001: Small Area Microdata (SAM) was used [57]. This dataset is an approximately 3% random sample of the 2001 Census of the population of England, Wales, Scotland, and Northern Ireland. For 2011, separate files for England and Wales (UKDA Study 7682, 2011 Census Microdata Local Authority data) [58] and Scotland (UKDA Study 7835, 2011 Census Microdata Individual Safeguarded Sample, Scotland Grouped Local Authority) [59] were obtained and combined into the merged dataset. The 2011 studies are a 5% random sample of the 2011 censuses. For the 2021 Census, the UKDA Study 9154 (Safeguarded Individual Microdata Sample at Region Level, England and Wales) was used [60]. This dataset is a 5% sample of individuals from Wales, regions within England, and Inner and Outer London. Northern Ireland was excluded because comparable household-level deprivation variables were not available. The 2021 census was delayed in Scotland because of the COVID-19 pandemic, and the data were not yet available for this study.

Health outcome

The single, self-reported general health question measured health. General health was captured for the first time in the 2001 Census: "Over the last twelve months would you say your health has on the whole been good, fairly good or not good? The 2011 and 2021 censuses utilised the European Union Statistics on Income and Living Conditions (EU-SILC) general health question: “How is your health in general? Is it very good, good, fair, bad, or very bad?” [61]. To compare the results of the 2001, 2011, and 2021 data analysis, we applied the ONS method [61] that requires the general health variables to be coded into binary variables of good (0) and bad (1) health (Box 1). We used all categories of general health variables for exploratory and bivariate analysis. In this manuscript, the term ‘bad health’ (used interchangeably with ‘poor health’) includes categories of ‘not good’ health in the 2001 data and ‘fair’, ‘bad’ and ‘very bad’ health in the 2011 and 2021 data. The ‘good health’ category included ‘good’ and ‘fairly good’ health in the 2001 data and ‘very good’ and ‘good’ health in the 2011 and 2021 data.

Socioeconomic position

The selection of variables was theoretically informed. The NS-SEC of the HRP variable was utilised to measure a young person’s SEP; it is a nominal or categorical measure that is composed of the groupings of occupations that share similar employment relations, such as type of contract or independence at work, and employment-related benefits [17, 45, 46]. The NS-SEC variable includes 14 functional and three residual operational categories that can be collapsed into eight, five and three class versions [45, 62]. We utilised a five-class version with residual categories of 1) never worked and long-term unemployed, 2) full-time students, and 3) ‘not classifiable occupations’ in the 2001 dataset (Box 2). This decision was based on the exploratory analysis of health status across 14 functional categories and on having an optimal number of respondents across all categories. Drever et al. [54] excluded respondents from residual categories because they are not a ‘class’ in the sociological sense, and people may belong to these groups for a variety of reasons. However, we decided to include all respondents who had information on the NS-SEC of the HRPs to observe the maximum number of population groups.

Covariates

Demographics: age and sex

Young people are defined as 10-24-year-olds, and adolescents are defined as 10-19-year-olds [6]. In the 2001 and 2011 datasets, age was included as a categorical variable. The age groups in the 2001 dataset were 10–15, 16–19 and 20–24, and those in the 2011 dataset were 10–15, 16–18 and 19–24. In the 2021 dataset, age was included as a continuous variable; we recoded this into categorical variable of 10–15, 16–18, and 19–24 years to reflect standard school-leaving age. We kept age groups separate to account for possible reporting differences. We assumed that the majority of the 10–16-year-olds were living with their parents and were in full-time education. Therefore, their health information was likely provided by an adult family member. The same reasoning would likely apply to many 16- to 18-year-olds. We used ‘gender’ and ‘sex’ interchangeably. Both sex and gender shape health, with ‘sex’ referring to biological factors and ‘gender’ referring to the social constructs of roles and behaviours [42, 63, 64]. Exploratory analysis indicated that the share of poor health increases with age and is greater among girls. To facilitate the interpretation of the results, we computed an interaction variable of age and gender, which takes the following values: boys aged 10–15, girls aged 10–15, boys aged 16–18, girls aged 16–18, boys aged 20–24, and girls aged 20–24. The 16–19-year-old and 20–24-year-old age groups in the 2001 dataset were included in the 16–18-year-old and 19–24-year-old categories, respectively.

Household deprivation: housing and educational dimensions

Household deprivation variables are derived variables included in the UKDA datasets. A household is deprived in the educational dimension if no person aged 16 to pensionable age has at least level two education (level 1 in Scotland), and no person aged 16 to 18 is a full-time student¹. A household is deprived in the housing dimension if it is either overcrowded, in a shared dwelling, does not have sole use of a bath/shower or toilet, or does not have central heating. These variables are binary: deprived (1) or not deprived (0). Housing and educational attainment are often used as indicators of SEP and are related to an individual’s occupation [17, 43]. We included these variables as control variables, where education represents the household’s knowledge-related assets and housing is related to material resources [17, 43].

Region

The census microdata includes information on the UK constituent countries, regions, and local authorities. It is known from existing evidence that there are significant differences in health and mortality across the different parts of the UK [53, 55]. After exploratory analysis, Northern Ireland was excluded because household deprivation measures were not included. The following categorisation was calculated based on existing knowledge and exploratory analysis: 1) North and North‒East England, 2) South and East England, 3) London, 4) Wales, 5) Glasgow city, and 6) Other Scotland.

Statistical analysis

All analyses were conducted in Stata version 16, separately for the 2001, 2011, and 2021 datasets. The first step consisted of univariate and bivariate analyses to explore variable distributions. Second, we used bivariate analysis with chi-square tests to estimate associations between SRH and explanatory variables. Third, we applied bivariate logistic regression to estimate the differences in odds of reporting poor health by the NS-SEC of the HRP. Fourth, we applied stepwise logistic regression to calculate odds ratios of reporting bad health by the NS-SEC of the HRP [65]. We utilised forward stepwise modelling to observe the change in odds of reporting bad health when covariates were added to the model one at a time [66, 67]. The selection and sequence of included variables were theoretically informed; covariates were included stepwise in the following sequence. Model 1: binary logistic regression model with NS-SEC of the HRP variable (7 categories in 2011 and 2021, 8 categories in 2001). Model 2: Model 1 plus age and sex interaction variable (6 categories). Model 3: Model 2 plus household deprivation in the educational dimension. Model 4: Model 3 plus household deprivation in the housing dimension. Model 5: Model 4 plus region (4 categories in 2021, 6 categories in 2001 and 2011). A likelihood ratio test was performed to determine the best fit of each nested model; these results are not reported here but are available on request. For both bivariate and logistic regression models, we calculated the binomial response variable and n-1 binary variables for multinomial covariates (n is the number of categories of the variable). All tests were performed at the 0.05 level of statistical significance (p < 0.05).

After exploratory data analysis, the datasets were limited to 10–24-year-olds, and the following observations were removed:

1) respondents without a valid response to the general health question (not usually residents),

2) NS-SEC HRP not applicable,

3) household deprivation indicators of housing and education not applicable (not in a household),

4) respondents from Northern Ireland.

In all datasets, respondents who were deprived in the level of education or housing were overlapping. All respondents had information on gender and region. The population base includes both usual and short-term residents [68–71]. The number of excluded observations and the population base of the samples are shown in Supplementary Table 1 (Additional File 1). The final numbers of observations for the data analysis were 1) 459,923 for the 2001 dataset, 2) 558,230 for the 2011 dataset, and 3) 505,133 for the 2021 dataset.

Figure 1 presents the share of respondents who reported fair, bad, or very bad health by the NS-SEC of the HRP. In 2001, ‘bad’ health was captured by the single category of ‘not good’ health, and we use the term ‘bad’ or ‘poor’ for this when discussing the results. All differences are statistically significant at the 99% confidence level (p < 001).

Three main findings were observed and are visualised in Fig. 1. First, the share of respondents with ‘bad health’ was lower in 2001 than in 2011 and 2021; this observation can be attributed to the percentage of respondents who reported having ‘fair’ health. Second, we observed a similar SE gradient of health status across NS-SEC groups three decades apart. The highest share of young people who reported very bad, bad, or fair health was highest among those whose HRP was classified as ‘never worked or long-term unemployed’ (NWLTU). The respective shares were 3.9% in 2001, 10.6 in 2011, and 14.5% in 2021. The only comparable share was observed in 2001 among young people whose HRP belonged to ‘not classifiable occupation’ category (4.1%). As expected, the lowest share of young people with very bad, bad, or fair health was observed among those whose HRP was in the ‘managerial and professional’ occupational category. Their rates were 1.3% in 2001, 2.9% in 2011, and 4% in 2021. Third, the share of young people who reported poor health increased across all NS-SEC groups between 2011 and 2021. The distribution by all SRH categories and covariates are shown in Supplementary Tables 2 and 3 (Additional File 1).

Second, we applied bivariate logistic regression to investigate the statistical significance of differences in reporting bad health between the NS-SEC groups. The odds and 95% confidence intervals of reporting bad health relative to young people from managerial and professional households are presented in Fig. 2 as unadjusted models. Compared with young people from households where HRP had a managerial or professional occupation, young people from all other NS-SEC groups had greater odds of reporting bad health. This was observed across the 2001, 2011 and 2021 datasets, except for young people from the households of self-employed reference persons in 2001. They had slightly lower odds of reporting bad health; however, this difference was not statistically significant.

Third, we applied stepwise logistic regression to control for the effect of age, gender, household deprivation indicators of education and housing, and region. The odds and 95% confidence intervals of reporting bad health relative to young people from the households where the reference person belonged to the managerial or professional occupational group are presented in Fig. 2 as adjusted models. Supplementary Table 4 (Additional File 1) shows the odds ratios and 95% confidence intervals for the unadjusted and adjusted models.

Multivariate logistic regression revealed a similar pattern in the odds of reporting bad health across the 2001, 2011, and 2021 datasets. Controlling for the effect of covariates resulted in a slight decrease in the odds of reporting bad health by the NS-SEC of the HRP; however, the gradient by employment conditions did not change. According to the 2001 full model, the odds of reporting bad health were 11% greater for young people with an HRP from lower supervisory or technical occupations, 15% greater in the intermediate occupational group, 29% greater for semi-routine or routine households, and 57% greater for young people whose HRP was a full-time student. The highest odds relative to young people from managerial or professional households were observed for respondents from the NWLTU households (2.7 times higher odds) and for young people whose HRP’s occupation was not classifiable (3.1 times higher odds). The lower odds of reporting poor health for respondents with an HRP from the self-employed occupational group were not statistically significant. In 2011, the odds of reporting bad health were 31% higher for young people from households with reference person from the intermediate occupational group. The odds were 10% higher for young people with self-employed HRP, 36% higher for those in lower supervisory and technical occupational groups, 74% for those in semi-routine or routine group, and 13% for full-time students. The highest odds were observed for young people from NWLTU households; they were 3.1 times more likely to report poor health. In 2021, the odds were 8% greater for young people from households of small employers and own account workers, 18% greater for full-time student households, 24% greater for households where HRP has lower supervisory or technical occupation 34% greater for those from intermediate occupations, and 72% greater for those from semi-routine or routine occupation households. Again, the observed odds were highest for young people from NWLTU households (3.6 times higher).

This study investigated health inequalities among young people by using data from the 2001, 2011, and 2021 UK censuses. Two significant findings stand out. First, we observed similar patterns of health status across three datasets collected ten years apart. The findings confirm the graded relationship between the NS-SEC of the HRP and the health status of young people; the prevalence of poor health increased as the employment conditions of the HRP worsened. This relationship did not change significantly after the effects of age, gender, household deprivation, and the UK region were controlled for. Importantly, the patterning of health by SEP was similar despite the lack of direct comparability of the data across the three censuses due to changes in general health questions and the basis of the NS-SEC measure. Second, the share of young people with bad health was strikingly greater in households where the reference person had never worked, was long-term unemployed, or whose job was not classifiable in 2001. Moreover, their levels were significantly greater than those of all other NS-SEC groups, including young people from households where the HRP was classified as having semi-routine or routine occupation.

These findings are in line with existing evidence on higher levels of poor SRH among young people with less advantaged SEP [3, 19–22, 24, 34, 35, 50, 72]. Socioeconomic inequalities in health can arise through different pathways [5, 73–75]. This study utilised the NS-SEC of the HRP as an indicator of SEP. This measure represents the combination of occupational groups and statuses that share similar employment relations [17, 45]. Additionally, employment provides both financial and non-financial resources that are linked to health outcomes [17, 43]. Therefore, our findings should be interpreted as mostly in line with the materialist model that focuses on aspects such as employment status or housing conditions in explaining health inequalities. Within this framework, inequalities are thought to arise due to differences in access to health-supporting resources and better control over the physical home environment [5, 48, 73, 74]. This hypothesis also supports our findings that young people from households who are deprived in the educational or housing dimension experience higher levels of poor health than their nondeprived peers.

Indeed, occupational position comprises factors such as human capital, income, access to health-supporting resources, and employment relations; therefore, it can be seen as a reflection of a person’s place in society that conditions people’s life chances and patterns of action [17, 43–45]. Reuter et al. [76] showed that health inequalities can be observed among 15–24-year-olds due to stratified working hazards comprising job demands and company characteristics. Additionally, existing evidence shows that young people who live in households that are deprived in the level of education [18, 19, 22, 36] or housing [36] report higher levels of poor health. Educational attainment and housing are frequently used as proxies for social position, with education being associated with knowledge-related assets and factors such as cultural capital, health behaviours, or material resources [18, 19, 34, 36, 43, 77]. A higher educational level of mothers has been found to have a protective effect on health in unemployed families [72]. Housing deprivation measures mostly material aspects of SE position, with factors such as housing tenure or dwelling characteristics being associated with health status [36, 43, 48, 78, 79].

Another point of discussion is that we are utilising the SEP of the HRP and not that of the young people themselves. However, the family is a traditional unit of analysis in sociology because of the shared resources and decision-making processes, and therefore, it is believed that family members share the same class ‘fate’ [45, 47]. For example, research by Sleskova et al. [80] showed that parental unemployment affects adolescents’ health, and research on young people’s perceptions of health has demonstrated that factors such as competing priorities, poor housing, or lack of time impact families' ability to make healthy choices [81–83]. Given the complexity of factors that are associated with a person’s employment conditions and relations, we would also like to draw upon behavioural and cultural explanations of health inequalities [3, 48, 74, 75, 84] and on a psychosocial model that focuses on factors such as relationships and emotions [48, 74, 75]. The former emphasises behaviours and the influence of community, while the latter discusses how perceived inequalities can lead to chronic stress, which can cause illnesses through neuroendocrine pathways [48, 74, 75, 85]. Therefore, we can argue that our findings support the role of material, psychosocial and behavioural pathways in explaining health inequalities among young people, as they are situated in different socioeconomic environments that shape their opportunities and behaviours.

Regarding covariates, the results of this study confirm the existing knowledge on the relationships between health status and age, gender, and deprivation. Our results also support the evidence of higher rates of self-reported poor health among girls and as young people grow older [48, 86–89]. The share of respondents who reported fair or bad health was greater among older age groups [13, 35, 72, 88] and among girls than among boys [13, 35, 72, 87–89].

Methodological considerations

The comparability of the 2001, 2011, and 2021 results is impacted by changes in the general health question and the basis of the NS-SEC classification. The general health question was changed to improve the comparability with other EU member states and provide a stronger indicator of functional health status after the 2001 census [61]. Analysis of the effects of the question change showed that the proportion of respondents who will be classified as having poor health is likely to increase [61]. Compared to 2001 data, we observed a higher share of respondents who were categorised as having poor health. This increase can be attributed to the number of respondents who reported ‘fair’ health. The NS-SEC measure was developed from the Goldthorpe Schema and has been used in official statistics since 2001 [62]. The 2001 census used occupation codes as defined by the SOC2000 while the 2011 census used codes defined by SOC2010 [90]. The 2021 Census was rebased on the 2020 Standard Occupational Classification [62]; therefore, results are not directly comparable across the years. Despite these methodological challenges, the observed social gradients in SRH were similar across the three datasets. Importantly, random samples from the censuses enabled us to complete this study; this also implies that any differences from the whole census are due to chance.

Strengths and weaknesses

The major strength of this study is the inclusion of different population groups regardless of their labour market participation. We utilised a single self-reported general health question to measure health. Occupation-based classifications often exclude unemployed people, which is thought to result in underestimation of SE differences in health [17, 20]. The limitation of the study is the cross-sectional design, which does not allow for causal inferences. Additionally, we excluded Northern Ireland because of the incompatibility of deprivation measures. For the 2021 Census, only data from England and Wales are available.

Our findings suggest that health inequalities among 10–24-year-olds can be observed by using data from the UK Censuses and by using the NS-SEC of the HRP. We observed a similar gradient of poor health across the 2001, 2011 and 2021 data. As disadvantage increases, the share of those reporting poor health also increases; these findings apply after the effects of age, gender, household deprivation and UK region are controlled for. These findings are significant in the context of the census questionnaire, which is often completed by the HRP and not by the young people themselves, which implies that at least among 10-15-year-olds, young people’s perceptions of SEP should not impact health assessments. This finding lends support to the view of family as a unit of shared resources and perceived life chances that places young people from different SE backgrounds on more or less favourable pathways of health. Although the cross-sectional design does not allow us to draw conclusions about causality, the similar patterns of health across the three UK census datasets confirm the role of social determinants in young people’s health.

HRP

household reference person

NS-SEC

The National Statistics Socio-economic Classification (NS-SEC)

NWLTU

never worked or long-term unemployed

socioeconomic

SEP

socioeconomic position

SRH

self-reported health

Ethics approval and consent to participate

This study is part of a KMs PhD project that used safeguarded datasets available through the UK Data Service. Safeguarded microdata samples are designed to protect the confidentiality of individuals and are available from the UK Data Service under the UK Data Service End User Licence. KM accepted the End User Licence Agreement. The census form informs respondents that the data will be used for statistical purposes. The safeguarded census microdata files do not enable the identification of individuals, as the amount of detail available is restricted to a nondisclosive level, and individual respondents only appear in one file.

Consent for publication

Not applicable

Availability of data and materials

The safeguarded datasets analysed during the current study are available through the UK Data Service by accepting the End User Licence Agreement. Please see references to the datasets used.

Competing interests

None.

Funding

This study is part of the PhD project KM is undertaking at the University of St Andrews as a member of staff. Therefore, there was no specific funding associated with completing this study. JI is supported by the Medical Research Council (MC_UU_00022/1) and the Scottish Government Health and Social Care Directorate (SHPSU16).

Authors' contributions

KM conceptualised the study and refined it based on feedback from FS and JI. KM completed the data analysis and discussed the results with FS, JI and AJW. KM interpreted the results and wrote the paper. FS, JI and AJW critically revised the study for intellectual content and data analysis results. All authors have made contributions to the content of the paper and approved the final manuscript.

Acknowledgements

This research was conducted using datasets from the UK Data Service, the principal repository for economic, population, and social research data in the UK and supported by the ESRC, UKRI councils, and European based projects (https://ukdataservice.ac.uk). We thank the UK Data Service and its funders for making these data available and the people who completed the census questionnaires.

Health inequalities: What are they? How do we reduce them? Edinburgh: NHS Health Scotland; 2015. [https://www.healthscotland.scot/media/1086/health-inequalities-what-are-they-how-do-we-reduce-them-mar16.pdf]
Bundy DAP, de Silva N, Horton S, Patton GC, Schultz L, Jamison DT, Abubakara A, Ahuja A, Alderman H, Allen N et al: Investment in child and adolescent health and development: key messages from Disease Control Priorities, 3rd Edition. The Lancet 2018, 391(10121):687-699.
Hargreaves DS, Marbini AD, Viner RM: Inequality trends in health and future health risk among English children and young people, 1999–2009. Arch Dis Child 2013, 98(11):850-855.
Viner RM, Ozer EM, Denny S, Marmot M, Resnick M, Fatusi A, Currie C: Adolescence and the social determinants of health. The Lancet 2012, 379(9826):1641-1652.
Wilkinson RG, Marmot M: Social determinants of health: the solid facts. 2nd edition: World Health Organization; 2003.
WHO: The Second Decade: Improving Adolescence Health and Development. In., vol. 50; 2001.
Elgar FJ, Pförtner T-K, Moor I, De Clercq B, Stevens GWJM, Currie C: Socioeconomic inequalities in adolescent health 2002–2010: A time-series analysis of 34 countries participating in the Health Behaviour in School-Aged Children study. The Lancet 2015, 385(9982):2088-2095.
Bauldry S, Shanahan MJ, Boardman JD, Miech RA, Macmillan R: A life course model of self-rated health through adolescence and young adulthood. Soc Sci Med 2012, 75(7):1311-1320.
Kestilä L, Martelin T, Rahkonen O, Härkänen T, Koskinen S: The contribution of childhood circumstances, current circumstances and health behaviour to educational health differences in early adulthood. BMC Public Health 2009, 9(1):164.
Power C, Matthews S, Manor O: Inequalities in self-rated health: explanations from different stages of life. Lancet 1998, 351(9108):1009-1014.
Lynch J, Smith GD: A Life Course Approach to Chronic Disease Epidemiology. Annual Review of Public Health 2005, 26(1):1-35.
Breidablik H-J, Meland E, Lydersen S: Self-rated health in adolescence: A multifactorial composite. Scand J Public Health 2008, 36(1):12-20.
Cavallo F, Dalmasso P, Ottová-Jordan V, Brooks F, Mazur J, Välimaa R, Gobina I, Gaspar de Matos M, Raven-Sieberer U, Group PHF: Trends in self-rated health in European and North-American adolescents from 2002 to 2010 in 32 countries. The European Journal of Public Health 2015, 25(suppl_2):13-15.
Haugland S, Wold B, Stevenson J, Aaroe LE, Woynarowska B: Subjective health complaints in adolescence: a cross-national comparison of prevalence and dimensionality. The European Journal of Public Health 2001, 11(1):4-10.
Jylhä M: What is self-rated health and why does it predict mortality? Towards a unified conceptual model. Social Science and Medicine 2009, 69(3):307-316.
Currie C, Molcho M, Boyce W, Holstein B, Torsheim T, Richter M: Researching health inequalities in adolescents: the development of the Health Behaviour in School-Aged Children (HBSC) family affluence scale. Soc Sci Med 2008, 66(6):1429-1436.
Galobardes B, Lynch J, Smith GD: Measuring socioeconomic position in health research. 2007.
Koivusilta LK, Rimpelä AH, Kautiainen SM: Health inequality in adolescence. Does stratification occur by familial social background, family affluence, or personal social position? BMC Public Health 2006, 6:1-14.
Moor I, Kuipers MAG, Lorant V, Pförtner TK, Kinnunen JM, Rathmann K, Perelman J, Alves J, Robert PO, Rimpelä A et al: Inequalities in adolescent self-rated health and smoking in Europe: Comparing different indicators of socioeconomic status. Journal of Epidemiology and Community Health 2019, 73(10):963-970.
Pförtner T-K, Günther S, Levin KA, Torsheim T, Richter M: The use of parental occupation in adolescent health surveys. An application of ISCO-based measures of occupational status. J Epidemiol Community Health 2014.
Elgar FJ, McKinnon B, Torsheim T, Schnohr CW, Mazur J, Cavallo F, Currie C: Patterns of socioeconomic inequality in adolescent health differ according to the measure of socioeconomic position. Social Indicators Research 2016, 127(3):1169-1180.
Moreno-Maldonado C, Rivera F, Ramos P, Moreno C: Measuring the Socioeconomic Position of Adolescents: A Proposal for a Composite Index. Social Indicators Research 2018, 136(2):517-538.
Smith NR, Lewis DJ, Fahy A, Eldridge S, Taylor SJC, Moore DG, Clark C, Stansfeld SA, Cummins S: Individual socio-demographic factors and perceptions of the environment as determinants of inequalities in adolescent physical and psychological health: The Olympic Regeneration in East London (ORiEL) study. BMC Public Health 2015, 15(1):1-18.
Currie C, Alemán Díaz AY, Bosáková L, de Looze M: The international Family Affluence Scale (FAS): Charting 25 years of indicator development, evidence produced, and policy impact on adolescent health inequalities. SSM - Population Health 2024, 25:101599.
Inchley JC, Willis M, Mabelis J, Brown J, Currie DB: Inequalities in health complaints: 20-year trends among adolescents in Scotland, 1998–2018. Front Psychol 2023, 14:1095117.
Nielsen L, Damsgaard MT, Meilstrup C, Due P, Madsen KR, Koushede V, Holstein BE: Socioeconomic differences in emotional symptoms among adolescents in the Nordic countries: Recommendations on how to present inequality. Scand J Public Health 2015, 43(1):83-90.
Torchyan AA, Bosma H: Socioeconomic inequalities in health among Armenian adolescents. International Journal of Environmental Research and Public Health 2020, 17(11):1-9.
Torsheim T, Nygren JM, Rasmussen M, Arnarsson AM, Bendtsen P, Schnohr CW, Nielsen L, Nyholm M: Social inequalities in self-rated health: A comparative cross-national study among 32,560 Nordic adolescents. Scand J Public Health 2018, 46(1):150-156.
Hagquist CEI: Health inequalities among adolescents—the impact of academic orientation and parents' education. Eur J Public Health 2007, 17(1):21-26.
Joffer J, Flacking R, Bergstrom E, Randell E, Jerden L: Self-rated health, subjective social status in school and socioeconomic status in adolescents: a cross-sectional study. BMC Public Health 2019, 19.
Piko B, Fitzpatrick KM: Does class matter? SES and psychosocial health among Hungarian adolescents. Soc Sci Med 2001, 53(6):817-830.
West P: Health inequalities in the early years: Is there equalisation in youth? Social Science and Medicine 1997, 44(6):833-858.
West P, Sweeting H: Evidence on equalisation in health in youth from the West of Scotland. Soc Sci Med 2004, 59(1):13-27.
Chen E, Martin AD, Matthews KA: Socioeconomic status and health: do gradients differ within childhood and adolescence? Soc Sci Med 2006, 62(9):2161-2170.
Inchley J, Currie D, Young T, Samdal O, Torsheim T, Augustson L, Mathison F, Aleman-Diaz A, Molcho M, Weber M et al: Growing up unequal : gender and socioeconomic differences in young people ’ s health and well-being. Health Behaviour in School-Aged Children (HBSC) study: International Reportfrom the 2013/2014 survey. In.; 2016.
Rahkonen O, Arber S, Lahelma E: Health inequalities in early adulthood: a comparison of young men and women in Britain and Finland. Soc Sci Med 1995, 41(2):163-171.
Burdette AM, Needham BL, Taylor MG, Hill TD: Health Lifestyles in Adolescence and Self-rated Health into Adulthood. Journal of Health and Social Behavior 2017, 58(4):520-536.
Due P, Krølner R, Rasmussen M, Andersen A, Trab Damsgaard M, Graham H, Holstein BE: Pathways and mechanisms in adolescence contribute to adult health inequalities. Scand J Public Health 2011, 39(6_suppl):62-78.
Power C, Matthews S, Manor O: Inequalities in self rated health in the 1958 birth cohort: lifetime social circumstances or social mobility? Bmj 1996, 313(7055):449-453.
Power C, Matthews S, Manor O: Inequalities in self-rated health: explanations from different stages of life. The Lancet 1998, 351(9108):1009-1014.
Moreau C, Blum R, Mmari K, Hunersen K, Mafuta E, Lulebi A, Wilopo S, Stones W, Shervington D, Michielsen K et al: Gender and Health in Very Young Adolescents. J Adolesc Health 2021, 69(1S):S3-S4.
Solar O, Irwin A: A conceptual framework for action on the social determinants of health. In.: WHO Document Production Services; 2010.
Galobardes B, Shaw M, Lawlor DA, Lynch JW: Indicators of socioeconomic position (part 1). Journal of epidemiology and community health 2006, 60(1):7.
Goldthorpe JH: Rent, Class Conflict, and Class Structure: A Commentary on Sørensen. American Journal of Sociology 2000, 105(6):1572-1582.
Rose DP, David J. : The National Statistics Socio-economic Classification: unifying official and sociological approaches to the conceptualisation and measurement of social class. ISER Working Paper Series, No. 2001-04, University of Essex, Institute for Social and Economic Research (ISER), Colchester. In.; 2001.
Erikson R, Goldthorpe JH, Portocarero L: Intergenerational Class Mobility in Three Western European Societies: England, France and Sweden. The British Journal of Sociology 1979, 30(4):415-441.
Rose DP, David J. O'Reilly, Karen: The National Statistics Socio-economic Classification : Origins , Development and Use; 2005.
Chen E, Matthews KA, Boyce WT: Socioeconomic differences in children's health: how and why do these relationships change with age? Psychological bulletin 2002, 128(2):295.
Chung A, Backholer K, Wong E, Palermo C, Keating C, Peeters A: Trends in child and adolescent obesity prevalence in economically advanced countries according to socioeconomic position: a systematic review. Obesity Reviews 2016, 17(3):276-295.
Holstein BE, Jørgensen SE, Due P, Damsgaard MT, Rasmussen M: Short report: persistent social inequality in poor self-rated health among adolescents in Denmark 1991–2014. Eur J Public Health 2018, 28(6):1114-1116.
Spencer NJ, Blackburn CM, Read JM: Prevalence and social patterning of limiting long-term illness/disability in children and young people under the age of 20 years in 2001: UK census-based cross-sectional study. Child: Care, Health and Development 2010, 36(4):566-573.
Allik M, Brown D, Dundas R, Leyland AH: Differences in ill health and in socioeconomic inequalities in health by ethnic groups: a cross-sectional study using 2011 Scottish census. Ethn Health 2022, 27(1):190-208.
Doran T, Drever F, Whitehead M: Is there a north-south divide in social class inequalities in health in Great Britain? Cross sectional study using data from the 2001 census. British Medical Journal 2004, 328(7447):1043-1045.
Drever F, Doran T, Whitehead M: Exploring the relation between class, gender, and self rated general health using the new socioeconomic classification. A study using data from the 2001 census. Journal of Epidemiology and Community Health 2004, 58(7):590-596.
Popham F: Is there a "Scottish effect" for self reports of health? Individual level analysis of the 2001 UK census. BMC Public Health 2006, 6:191-191.
Rydzewska E H-MLHA, Cooper SA: Health of people with autism: findings from Scotland's 2011 census - Enlighten: Publications. In: 2016; 2016.
University of Manchester, Cathie Marsh Centre for Census and Survey Research, Office for National Statistics, Census Division. (2023). Census 2001: Small Area Microdata (SAM) . [data collection]. UK Data Service. SN: 7207, DOI: http://doi.org/10.5255/UKDA-SN-7207-1. In.; 2023.
Office for National Statistics. (2023). 2011 Census Microdata Individual Safeguarded Sample (Local Authority): England and Wales. [data collection]. UK Data Service. SN: 7682, DOI: http://doi.org/10.5255/UKDA-SN-7682-1. In.; 2023.
National Records of Scotland. (2023). 2011 Census Microdata Individual Safeguarded Sample (Local Authority): Scotland. [data collection]. UK Data Service. SN: 7835, DOI: http://doi.org/10.5255/UKDA-SN-7835-1. In.; 2023.
2021 Census: Safeguarded Individual Microdata Sample at Region Level (England and Wales). [data collection]. UK Data Service. SN: 9154, DOI: http://doi.org/10.5255/UKDA-SN-9154-1. In.; 2023.
Smith M, White C: An investigation into the impact of question change on estimates of general health status and healthy life expectancy. Health Stat Q 2009(41):28-41.
SOC 2020 Volume 3: the National Statistics Socio-economic Classification (NS-SEC rebased on the SOC 2020) [https://www.ons.gov.uk/methodology/classificationsandstandards/standardoccupationalclass
ificationsoc/soc2020/soc2020volume3thenationalstatisticssocioeconomicclassificationnssecrebasedonthesoc2020]
Krieger N: Genders, sexes, and health: what are the connections—and why does it matter? Int J Epidemiol 2003, 32(4):652-657.
Patwardhan V, Gil GF, Arrieta A, Cagney J, DeGraw E, Herbert ME, Khalil M, Mullany EC, O'Connell EM, Spencer CN et al: Differences across the lifespan between females and males in the top 20 causes of disease burden globally: a systematic analysis of the Global Burden of Disease Study 2021. Lancet Public Health 2024, 9(5):e282-e294.
Sperandei S: Understanding logistic regression analysis. Biochem Med (Zagreb) 2014, 24(1):12-18.
Chowdhury MZI, Turin TC: Variable selection strategies and its importance in clinical prediction modelling. Fam Med Community Health 2020, 8(1):e000262.
Singh-Manoux A, Martikainen P, Ferrie J, Zins M, Marmot M, Goldberg M: What does self rated health measure? Results from the British Whitehall II and French Gazel cohort studies. J Epidemiol Community Health 2006, 60(4):364-372.
2001 Samples of Anonymised Records. Guide to the 2001 Samples. Version 2.1 September 2014 Restructured version by Cathie Marsh Institute for Social Research, University of Manchester Under the auspices of Census Support, UK Data Service. In.
2011 Census Microdata Individual Safeguarded Samples – User Guide. UK Data Archive Study Number 7682 - 2011 Census Microdata Individual Safeguarded Sample (Local Authority). In.
2011 Census Microdata Individual Safeguarded Samples – User Guide. UK Data Archive Study Number 7835 - 2011 Census Microdata Individual Safeguarded Sample (Local Authority): Scotland. In.
Office for National Statistics (ONS), released 7 September 2023, ONS website, methodology article. User guide to microdata samples for Census 2021, England and Wales. In.
Moreno-Maldonado C, Jiménez-Iglesias A, Rivera F, Moreno C: Characterization of Resilient Adolescents in the Context of Parental Unemployment. Child Indicators Research 2020, 13(2):681-702.
Moor I, Spallek J, Richter M: Explaining socioeconomic inequalities in self-rated health: A systematic review of the relative contribution of material, psychosocial and behavioural factors. Journal of Epidemiology and Community Health 2016, 71(6):565-575.
Pearce A, Dundas R, Whitehead M, Taylor-Robinson D: Pathways to inequalities in child health. Arch Dis Child 2019, 104(10):998-1003.
Bartley M: Health Inequality: an Introduction to Concepts, Theories and Methods. Second Edition; 2017.
Reuter M, Pischke CR, Rigo M, Diehl K, Spallek J, Richter M, Hövener C, Dragano N: Health inequalities among young workers: the mediating role of working conditions and company characteristics. International Archives of Occupational and Environmental Health 2023, 96(10):1313-1324.
Manor O, Matthews S, Power C: Comparing measures of health inequality. Social Science and Medicine 1997, 45(5):761-771.
Dunn JR: Housing and inequalities in health: A study of socioeconomic dimensions of housing and self reported health from a survey of Vancouver residents. Journal of Epidemiology and Community Health 2002, 56(9):671-681.
Shaw M: Housing and Public Health. Annual Review of Public Health 2004, 25(Volume 25, 2004):397-418.
Sleskova M, Salonna F, Geckova AM, Nagyova I, Stewart RE, van Dijk JP, Groothoff JW: Does parental unemployment affect adolescents’ health? J Adolesc Health 2006, 38(5):527-535.
Backett-Milburn K, Cunningham-Burley S, Davis J: Contrasting lives, contrasting views? understandings of health inequalities from children in differing social circumstances. Soc Sci Med 2003, 57(4):613-623.
Dow M, Murrin C, O'Malley G, Brinkley A, Bel‐Serrat S: A qualitative study exploring the perceptions of health among pre‐teen girls from disadvantaged communities in Dublin. Children & Society 2022.
Fairbrother H, Woodrow N, Crowder M, Holding E, Griffin N, Er V, Dodd-Reynolds C, Egan M, Lock K, Scott S et al: 'It All Kind of Links Really': Young People's Perspectives on the Relationship between Socioeconomic Circumstances and Health. International Journal of Environmental Research and Public Health 2022, 19(6).
Richter M, Erhart M, Vereecken CA, Zambon A, Boyce W, Gabhainn SN: The role of behavioural factors in explaining socio-economic differences in adolescent health: A multilevel study in 33 countries. Soc Sci Med 2009, 69(3):396-403.
Marmot M, and Richard G. Wilkinson (eds): 'Social organization, stress, and health', in Michael Marmot, and Richard Wilkinson (eds), Social Determinants of Health, 2nd edn (Oxford, 2005; online edn, Oxford Academic, 1 Sept. 2009), accessed 4 June 2024. In: Social Determinants of Health. edn. Edited by Marmot M, Wilkinson RG: Oxford University Press; 2005.
Cavallo F, Zambon A, Borraccino A, Raven-Sieberer U, Torsheim T, Lemma P, Grp HPH: Girls growing through adolescence have a higher risk of poor health. Qual Life Res 2006, 15(10):1577-1585.
Eriksson J, Dellve L, Eklöf M, Hagberg M: Early inequalities in excellent health and performance among young adult women and men in Sweden. Gender medicine 2007, 4(2):170-182.
Jerdén L, Burell G, Stenlund H, Weinehall L, Bergström E: Gender Differences and Predictors of Self-Rated Health Development Among Swedish Adolescents. J Adolesc Health 2011, 48(2):143-150.
Vingilis E, Wade TJ, Adlaf E: What factors predict student self-rated physical health? J Adolesc 1998, 21(1):83-97.
National Statistics Socio-economic Classification (NS-SeC) [https://www.scotlandscensus.gov.uk/metadata/national-statistics-socio-economic-classification-ns-sec/#panel-1]

In England and Wales, Level 2 refers to at least 5 or more GCSEs. In Scotland, Level 1 refers to the Scottish Standard Grade or equivalent qualifications.

Boxes 1 and 2 are available in the Supplementary File section.

No competing interests reported.

Download PDF

Editor invited by journal
12 Jul, 2024
Editor assigned by journal
12 Jul, 2024
Submission checks completed at journal
12 Jul, 2024
First submitted to journal
11 Jul, 2024

You are reading this latest preprint version

What can the 2001, 2011 and 2021 UK censuses tell us about health inequalities among young people? A cross-sectional study using censuses from England, Wales, and Scotland

Status:

Version 1

Abstract

Background

Methods

Results

Conclusions

Figures

Background

Methods

Design and sample

Health outcome

Socioeconomic position

Covariates

Statistical analysis

Results

Discussion

Methodological considerations

Strengths and weaknesses

Conclusions

Abbreviations

Declarations

References

Footnotes

Boxes

Additional Declarations

Supplementary Files

Status:

Version 1