Abstract

Background: To analyse trends in socio-economic inequalities in the prevalence of diabetes among men and women aged ≥35 years in Spain during the period 1987–2006. Methods: We analysed trends in the age-standardized prevalence of self-reported diabetes and obesity in relation to level of education using data from the Spanish National Health Survey for the years 1987, 1993, 1995, 1997, 2001, 2003 and 2006 (86 345 individuals aged ≥35 years). To assess the relationship between education level and diabetes and obesity, we computed the Slope Index of Inequality and the Relative Index of Inequality (RII) for each year. Additional models were fit to take into account mediator variables in socio-economic position (SEP) diabetes inequalities. Results: The prevalence of self-reported diabetes was higher among persons of low educational level, increasing more rapidly over time among people with lower education level (5.0–12.6% in men, and 8.4–13.1% in women between 1987 and 2006) than among those with higher education level (6.3–8.7% in men and 3.8–4.0% in women). Relative inequalities showed a weak tendency to increase. In women, the RII of self-reported diabetes increased from 3.04 (1.95–4.74) in 1987 to 4.28 (2.98–6.13) in 2006, while in men were constant since 1993. Trends in SEP inequalities in diabetes prevalence were attenuated when mediator variables were taken into account in women but not in men. Conclusion: SEP inequalities in diabetes existed >20 years ago and have increased, especially among women. These patterns may be explained by trends in health behaviours and obesity, but only to a limited extent.

Introduction

Diabetes has become an important worldwide health problem due to its high prevalence and associated mortality rate. In Europe in 2000, 6.5% and 5.1% of all deaths among men and women, respectively, were due to diabetes.1 Moreover, the global burden of diabetes is expected to increase from 171.2 to 366.2 million cases between 2000 and 2030 (2.8–4.4% of total population).2

Several European studies have observed health inequalities related to socio-economic position (SEP),3 and individuals with low SEP or who live in deprived areas have higher Type 2 Diabetes Mellitus (T2DM) incidence, prevalence, and mortality.4–6 Further, it has been reported the variation in the magnitude of these inequalities between different regions of Europe.4–6

Icks et al.7 and Imkampe et al.8 examined trends in SEP inequalities in T2DM and observed an increase in inequality among Germans during the 1990s, and a similar increase among English women but not men in the period 1994–2006. The authors argue that these trends are driven by SEP-related differences in obesity, health behaviours or improvements in diagnosis. While, these studies have described trends in SEP inequalities in T2DM in central and northern Europe, no similar work has been reported for southern European countries, which generally have different social characteristics. For example, south European countries have higher prevalence of obesity and higher prevalence of diabetes as well as different diet patterns and other lifestyles, such as less physical activity.9,10

At the beginning of the 21st century, Spain was one of the European countries with the largest inequalities in obesity and in T2DM in women.11 Since the prevalence of T2DM is increasing,12 it is important to understand how trends in inequality in T2DM prevalence can be expected to develop. As a complement studies performed in northern Europe, we report on trends in inequalities in Spain. Especially because of differences in lifestyles in southern European countries,9,13,14 such a trend study may help to better understand an European overview of the trends in SEP-related inequalities in T2DM. Therefore, the aim of this study was to analyse trends in SEP-related inequalities in the prevalence of T2DM in Spanish men and women aged ≥35 years during the period 1987–2006.

Methods

Design, study population and information sources

We used a trend design analysing seven versions of the Spanish National Health Survey (NHS). NHS is a nationally representative survey of the Spanish population, which aims to provide data about nation's health; the prevalence of specific health conditions and the prevalence of risk factors for disease. The study population consists of non-institutionalized men and women living in Spain in the years of the surveys. Subjects were selected by means of a stratified multi-stage sampling strategy and the information was collected through personal interviews in the subjects’ homes. More information about the methodology used in these surveys is described in detail elsewhere.15 In this study, we used data from seven different NHS over 20 years (1987, 1993, 1995, 1997, 2001, 2003 and 2006). A subsample of people aged ≥35 years was taken (individuals <35 years old were excluded because diabetes is less frequent and type 1 diabetes represents a higher proportion of cases at young ages). Due to their relatively small sample sizes, data from the 1995 and 1997 surveys were analysed jointly; this approach is valid since these surveys used the same methodology and sampling scheme.15 Survey sample sizes were 17 855 individuals in 1987, 13 025 in 1993, 7851 in 1995–97, 13 593 in 2001, 14 208 in 2003, 19 813 in 2006.

Variables

Dependent

Diabetes and obesity were examined as dependent variables. T2DM status was self-reported, based on the question ‘Has your doctor told you that you have diabetes?’. The same question was used in each NHS, except for 2006, when the diabetes-related question was ‘Have you had diabetes ever in your life?’ Individuals with body mass index (BMI, computed from self-reported height and weight) ≥30 kg/m2 were considered to be obese.

Independent

The main independent variable, education level (as an indicator of SEP), was determined on the basis of the reclassification of national educational schemes into three categories according to the International Standard Classification of Education (ISCED): no formal education (ISCED 1), primary education (ISCED 2) and secondary or higher education (ISCED 3-4-5-6). Data for educational level was missing in 0.6% of subjects.

Age was analysed as a potential confounding variable, and a series of further variables were included in the analysis as possible mediators of the relationship between diabetes and SEP4: for the 1993–2006 surveys, these variables were smoking (daily, non-daily, former or never smoker), workplace physical activity (measured using a Likert scale ranging from 1 to 4 corresponding to less and more effort) and leisure time physical activity (regular or not); for the 2001–06 surveys the frequency of vegetable and sweet consumption (daily, ≥3 times, 1–2 times or less than once per week, or never). All these variables were self-reported using the same question in each year.

Data analysis

All analyses were performed separately for men and women. Sampling weights derived from the sample design were used in all calculations.15

The age-standardized prevalence of T2DM and obesity was calculated for each survey year and each educational level using the direct method16 with the 2006 survey population as the reference population. The Slope Index of Inequality (SII) and the Relative Index of Inequality (RII) were used to quantify socio-economic inequalities in T2DM and obesity.17 These can be interpreted as the ratio and absolute difference, respectively, in prevalence at the extremes of the spectrum of SEP (highest compared with lowest). The association between T2DM and obesity and educational level for each survey year was calculated using a robust Poisson regression model. Educational level was introduced as a continuous variable with values for each educational level corresponding to the cumulative proportion of its population, so the variable ranges between 0 and 1. The RII corresponds to the exponent of the coefficient of the term for educational term derived from the regression model. The SII was derived form the RII and the overall prevalence rate (PR) using the following formula:18 

formula
where PR is the overall prevalence rate and RII is the Relative Index of Inequality.

Finally, to examine how the RII of T2DM changed according to the possible mediator variables, we fit multivariate robust Poisson regression models, adjusted for possible mediator variables. BMI was entered in the model as a quadratic factor. We checked for effect modification between education and each possible mediator variables including the interaction term in the robust Poisson regression models for each year.

Results

Trends in characteristics of the Spanish sample between 1987 and 2006 are shown in table 1. We observed an increase in the proportion of men and women aged >65 years from 19.9% and 23.4% in 1987 to 24.4% and 30.1%, respectively, in 2006. The proportion of men and women aged ≥35 years without formal education decreased from 39.5% and 53.7% in 1987 to 12.3% and 18.6%, respectively, by 2006. We also observed changes in the lifestyle and dietary habits during the study period: the proportion of daily smokers decreased among men but increased slightly among women; other unhealthy behaviours, such as sweet food consumption have increased while healthy behaviours, such as leisure time physical activity and vegetable consumption increased or remained stable; finally, the proportions of men and women whose work involves little physical activity increased from 34.1% and 26.3% in 1993 to 37.9% and 28.7% in 2006, respectively.

Table 1

Study sample stratified by independent variables (total number of cases and column percentages) by year of survey

 Men
 
Women
 
1987 1993 1995–97 2001 2003 2006 1987 1993 1995–97 2001 2003 2006 
Number of cases 8312 6099 3672 6390 6751 9477 9543 6926 4179 7203 7457 10 336 
Diabetes             
    Yes 4.8 5.2 6.4 8.0 8.4 9.4 7.0 7.7 8.5 8.7 8.9 8.5 
Obesity             
    Yes 9.4 12.0 14.7 15.2 16.8 19.1 12.5 14.6 17.6 19.6 17.7 18.9 
Age group (years)             
    35–64 80.1 78.6 76.8 73.7 75.0 75.6 76.6 74.4 71.6 69.2 69.1 69.9 
    >65 19.9 21.3 23.2 26.3 25.0 24.4 23.4 25.6 28.4 30.8 30.9 30.1 
Educational level             
    No formal education 39.5 19.1 19.5 15.8 15.0 12.3 53.7 29.0 25.2 22.2 21.0 18.6 
    Primary 36.4 53.2 51.7 52.5 57.1 51.0 33.8 55.1 56.5 55.9 57.4 53.8 
    Secondary or higher 23.1 27.2 28.1 31.4 28.0 36.7 11.6 15.0 17.7 21.7 21.6 27.6 
Smoking             
    Daily smoker – 41.6 40.6 36.9 32.5 30.1 – 11.3 13.8 17.4 17.3 17.8 
    Non-daily smoker – 3.5 2.7 2.6 2.9 2.6 – 2.0 1.2 1.4 1.6 1.8 
    Former smoker – 28.6 32.1 34.2 34.4 37.9 – 4.5 6.2 8.8 11.2 13.5 
    Never smoker – 25.6 24.5 26.0 30.2 29.4 – 81.3 78.7 72.2 69.9 66.9 
Physical activity in work place             
    Little effort – 34.1 35.7 35.5 35.6 37.9 – 26.3 22.0 25.3 28.6 28.7 
Physical activity in the leisure time             
    Yes – 42.9 52.7 53.8 59.8 60.2 – 31.0 43.5 46.2 63.5 57.1 
Vegetables consumption             
    Never or almost never – – – 1.5 2.6 1.7 – – – 0.5 0.8 0.7 
Sweets consumption             
    Every day – – – 28.7 30.0 31.4 – – – 30.7 30.0 33.7 
 Men
 
Women
 
1987 1993 1995–97 2001 2003 2006 1987 1993 1995–97 2001 2003 2006 
Number of cases 8312 6099 3672 6390 6751 9477 9543 6926 4179 7203 7457 10 336 
Diabetes             
    Yes 4.8 5.2 6.4 8.0 8.4 9.4 7.0 7.7 8.5 8.7 8.9 8.5 
Obesity             
    Yes 9.4 12.0 14.7 15.2 16.8 19.1 12.5 14.6 17.6 19.6 17.7 18.9 
Age group (years)             
    35–64 80.1 78.6 76.8 73.7 75.0 75.6 76.6 74.4 71.6 69.2 69.1 69.9 
    >65 19.9 21.3 23.2 26.3 25.0 24.4 23.4 25.6 28.4 30.8 30.9 30.1 
Educational level             
    No formal education 39.5 19.1 19.5 15.8 15.0 12.3 53.7 29.0 25.2 22.2 21.0 18.6 
    Primary 36.4 53.2 51.7 52.5 57.1 51.0 33.8 55.1 56.5 55.9 57.4 53.8 
    Secondary or higher 23.1 27.2 28.1 31.4 28.0 36.7 11.6 15.0 17.7 21.7 21.6 27.6 
Smoking             
    Daily smoker – 41.6 40.6 36.9 32.5 30.1 – 11.3 13.8 17.4 17.3 17.8 
    Non-daily smoker – 3.5 2.7 2.6 2.9 2.6 – 2.0 1.2 1.4 1.6 1.8 
    Former smoker – 28.6 32.1 34.2 34.4 37.9 – 4.5 6.2 8.8 11.2 13.5 
    Never smoker – 25.6 24.5 26.0 30.2 29.4 – 81.3 78.7 72.2 69.9 66.9 
Physical activity in work place             
    Little effort – 34.1 35.7 35.5 35.6 37.9 – 26.3 22.0 25.3 28.6 28.7 
Physical activity in the leisure time             
    Yes – 42.9 52.7 53.8 59.8 60.2 – 31.0 43.5 46.2 63.5 57.1 
Vegetables consumption             
    Never or almost never – – – 1.5 2.6 1.7 – – – 0.5 0.8 0.7 
Sweets consumption             
    Every day – – – 28.7 30.0 31.4 – – – 30.7 30.0 33.7 

The categories do not sum 100% because of their missing values. Men and women ≥35 years.

Men with lower SEP and women with higher SEP smoke more. Smoking increased over time in women with primary or no formal education [percentage of daily smokers increased from 9.6% and 5.1% (1993) to 18.6% and 19.4% in women with primary and no formal education, respectively] but decreased in women with at least secondary level education (from 23.5% in 1993 to 21.3% in 2006). People with higher SEP undertook more leisure time physical activity, while the percentage of people doing less physical activity decreased over time irrespective of education levels, although a greater decrease was observed in people without a formal education. While consumption of sweet foods in 1987 was greatest among people with higher SEP, we observe a significant increase over time in sweet food consumption in people with no formal education, while levels of consumption of this type of food remained stable over time (table 2).

Table 2

Age-adjusted prevalence (%) of each health behaviour in each educational level by year of Survey

 Men
 
Women
 
1993 1995–97 2001 2003 2006 1993 1995–97 2001 2003 2006 
Daily smokers (%)           
    No formal education 41.8 41.5 43.3 37.9 39.8* 5.1 9.7 9.5 12.8 19.4** 
    Primary 40.3 41.4 36.9 33.2 31.0** 9.6 12.4 17.0 18.1 18.6** 
    Secondary or higher 38.9 36.1 33.2 27.1 25.9** 23.5 25.9 24.0 21.1 21.3** 
Who undertake little work time physical activity (%)           
    No formal education 30.7 30.6 29.3 31.7 27.4* 31.1 22.6 22.9 25.3 25.9** 
    Primary 29.8 30.8 29.7 29.5 30.8 23.6 18.2 21.1 21.6 22.4 
    Secondary or higher 45.3 46.9 45.0 48.7 51.0** 29.6 26.0 29.6 38.3 34.7** 
Who undertake little leisure time physical activity (%)           
    No formal education 69.8 59.3 59.6 24.1 52.0** 81.2 71.8 67.0 23.3 51.2** 
    Primary 57.7 51.2 50.1 37.9 41.2** 68.3 56.5 54.0 37.4 41.4** 
    Secondary or higher 53.0 34.0 31.2 54.4 27.9** 48.4 38.4 40.0 45.5 33.7** 
Who eat vegetables less than once per week (%)           
    No formal education – – 12.1 12.0 11.4 – – 9.5 9.3 3.7 
    Primary – – 6.2 7.1 5.4** – – 2.5 2.5 2.4* 
    Secondary or higher – – 3.9 5.0 3.0 – – 1.7 1.7 1.5 
Who eat sweet foods every day (%)           
    No formal education – – 25.1 31.5 39.3 – – 27.2 28.3 35.0 
    Primary – – 28.8 29.7 33.0** – – 32.1 29.8 35.3** 
    Secondary or higher – – 31.2 30.3 30.3** – – 29.6 30.9 33.5** 
 Men
 
Women
 
1993 1995–97 2001 2003 2006 1993 1995–97 2001 2003 2006 
Daily smokers (%)           
    No formal education 41.8 41.5 43.3 37.9 39.8* 5.1 9.7 9.5 12.8 19.4** 
    Primary 40.3 41.4 36.9 33.2 31.0** 9.6 12.4 17.0 18.1 18.6** 
    Secondary or higher 38.9 36.1 33.2 27.1 25.9** 23.5 25.9 24.0 21.1 21.3** 
Who undertake little work time physical activity (%)           
    No formal education 30.7 30.6 29.3 31.7 27.4* 31.1 22.6 22.9 25.3 25.9** 
    Primary 29.8 30.8 29.7 29.5 30.8 23.6 18.2 21.1 21.6 22.4 
    Secondary or higher 45.3 46.9 45.0 48.7 51.0** 29.6 26.0 29.6 38.3 34.7** 
Who undertake little leisure time physical activity (%)           
    No formal education 69.8 59.3 59.6 24.1 52.0** 81.2 71.8 67.0 23.3 51.2** 
    Primary 57.7 51.2 50.1 37.9 41.2** 68.3 56.5 54.0 37.4 41.4** 
    Secondary or higher 53.0 34.0 31.2 54.4 27.9** 48.4 38.4 40.0 45.5 33.7** 
Who eat vegetables less than once per week (%)           
    No formal education – – 12.1 12.0 11.4 – – 9.5 9.3 3.7 
    Primary – – 6.2 7.1 5.4** – – 2.5 2.5 2.4* 
    Secondary or higher – – 3.9 5.0 3.0 – – 1.7 1.7 1.5 
Who eat sweet foods every day (%)           
    No formal education – – 25.1 31.5 39.3 – – 27.2 28.3 35.0 
    Primary – – 28.8 29.7 33.0** – – 32.1 29.8 35.3** 
    Secondary or higher – – 31.2 30.3 30.3** – – 29.6 30.9 33.5** 

The categories not sum 100% because of their missing values. Men and women ≥35 years

*Trend P > 0.05, **trend P < 0.01

Trends in RII, SII and age-standardized prevalence of T2DM for each educational level are shown in table 3. In 2006, T2DM prevalence was higher among people without a formal education (men: 12.6, 95% CI 10.4–15.4; women: 13.1 95% CI 11.3–15.3) than among those with at least a secondary education (men: 8.7, 95% CI 7.4–10.2; women: 4.0, 95% CI 3.0–5.4). Moreover, this inequality has increased over time, with RII for T2DM increasing from 0.91 (95% CI 0.56–1.50) and 3.04 (95% CI 1.95–4.74) in men and women, respectively, in 1987 to 1.45 (95% CI 1.01–2.09) and 4.28 (95% CI 2.98–6.13), respectively, in 2006. During this period, the greatest inequalities in T2DM were observed in 2003, with a slight decrease in 2006. A similar pattern of results was observed for obesity (table 3). Among men, the magnitude of inequality in obesity was quite similar to that observed for T2DM, while the social gradient among women in recent years appears to be larger for T2DM than obesity. Between 1987 and 2006, absolute differences changed concordantly with relative differences. Among men, the SII increased from −0.45 (95% CI −2.71 to 1.92) to 3.45 (95% CI 0.09 to 12.91). Among women, it increased from 7.07 (95% CI 4.51–9.12) to 10.56 (95% CI 8.46–12.23) (table 3).

Table 3

Age-standardized prevalence of T2DM/obesity (%) and association of education level with T2DM/obesity prevalence (Relative Index of Inequalities (RII) and Slope Index of Inequalities (SII) and their 95% CI) by year of Survey

Educational level / year 1987 1993 1995–97 2001 2003 2006 
Diabetes—Men       
    No formal education % (95% CI) 5.0 (4.2–5.8) 6.6 (5.1–8.6) 6.9 (5.1–10.3) 10.9 (8.7–14.4) 14.0 (11.4–17.1) 12.6 (10.4–15.4) 
    Primary % (95% CI) 5.5 (4.6–6.6) 6.0 (5.1–7.0) 7.5 (6.2–9.0) 8.4 (7.4–9.5) 8.8 (7.9–9.9) 10.1 (9.3–11.1) 
    Secondary higher % (95% CI) 6.3 (4.7–8.3) 3.5 (2.4–5.1) 4.4 (2.8–6.7) 7.0 (5.3–9.0) 6.1 (4.8–7.7) 8.7 (7.4–10.2) 
    RII (95% CI)** 0.91 (0.56–1.50) 1.93 (1.19–3.14) 1.91 (1.18–3.08) 2.05 (1.38–3.05) 2.19 (1.43–3.36) 1.45 (1.01–2.09) 
    SII (95% CI) −0.45 (−2.71 to 1.92) 3.30 (0.91–5.38) 4.00 (1.06–6.53) 5.51 (2.55–8.10) 6.27 (2.97–9.09) 3.45 (0.09–12.91) 
Diabetes—Women       
    No formal education % (95% CI) 8.4 (7.7–9.3) 11.8 (10.3–13.7) 12.5 (10.2–16.0) 12.7 (10.9–15.5) 14.9 (12.9–17.4) 13.1 (11.3–15.3) 
    Primary % (95% CI) 6.0 (5.0–7.1) 6.3 (5.5–7.3) 7.0 (6.0–8.1) 7.0 (6.3–7.9) 7.5 (6.7–8.3) 7.5 (6.8–8.2) 
    Secondary higher % (95% CI) 3.8 (2.2–6.2) 2.9 (1.5–5.3) 2.3 (1.0–4.5) 3.5 (2.1–5.5) 2.2 (1.1–4.0) 4.0 (3.0–5.4) 
    RII (95% CI)** 3.04 (1.95–4.74) 4.89 (3.22–7.43) 5.17 (3.37–7.92) 6.09 (4.18–8.85) 7.54 (5.09–11.16) 4.28 (2.98–6.13) 
    SII (95% CI) 7.07 (4.51–9.12) 10.17 (8.10–11.75) 11.49 (9.22–13.19) 11.76 (9.44–13.50) 13.63 (11.95–14.87) 10.56 (8.46–12.23) 
Obesity—Men       
    No formal education % (95% CI) 11.1 (9.7–12.5) 13.6 (11.2–16.5) 18.7 (14.3–24.8) 25.5 (19.8–32.8) 20.6 (17.4–24.6) 20.4 (17.0–24.6) 
    Primary % (95% CI) 10.4 (9.0–11.9) 13.6 (12.3–15.0) 15.1 (13.3–17.2) 15.8 (14.5–17.3) 18.0 (16.6–19.4) 21.0 (19.7–22.4) 
    Secondary higher % (95% CI) 6.4 (5.0–8.2) 8.0 (6.3–10.1) 11.4 (8.9–14.6) 11.8 (9.9–14.1) 11.4 (9.6–13.5) 16.8 (15.1–18.6) 
    RII (95% CI)* 1.99 (1.41–2.81) 2.15 (1.56–2.95) 2.14 (1.53–3.00) 2.35 (1.75–3.15) 2.58 (1.94–3.43) 1.66 (1.28–2.14) 
    SII (95% CI) 6.22 (3.20–8.93) 8.76 (5.25–11.85) 10.67 (6.16–14.70) 12.25 (9.72–16.68) 14.83 (10.74–18.43) 9.62 (4.76–14.09) 
Obesity—Women       
    No formal education % (95% CI) 17.2 (15.6–18.9) 24.5 (21.1–28.4) 24.7 (20.3–30.5) 36.0 (29.9–43.5) 30.6 (26.8–35.1) 32.6 (28.4–37.4) 
    Primary % (95% CI) 11.2 (9.6–13.1) 14.4 (13.0–16.0) 18.0 (16.1–20.1) 20.2 (18.7–21.8) 17.4 (16.2–18.7) 19.8 (18.6–21.1) 
    Secondary higher % (95% CI) 4.8 (3.3–7.2) 4.9 (3.1–7.7) 8.2 (5.5–12.0) 10.1 (7.8–13.0) 8.6 (6.5–11.3) 10.4 (8.8–12.2) 
    RII (95% CI)* 3.98 (2.78–5.68) 4.77 (3.47–6.57) 4.05 (3.00–5.49) 4.35 (3.37–5.62) 4.99 (3.82–6.51) 4.25 (3.35–5.39) 
    SII (95% CI) 14.96 (11.77–17.51) 19.08 (16.14–24.49) 21.26 (17.60–24.35) 24.55 (21.26–27.36) 23.71 (20.83–26.12) 23.40 (20.42–25.97) 
Educational level / year 1987 1993 1995–97 2001 2003 2006 
Diabetes—Men       
    No formal education % (95% CI) 5.0 (4.2–5.8) 6.6 (5.1–8.6) 6.9 (5.1–10.3) 10.9 (8.7–14.4) 14.0 (11.4–17.1) 12.6 (10.4–15.4) 
    Primary % (95% CI) 5.5 (4.6–6.6) 6.0 (5.1–7.0) 7.5 (6.2–9.0) 8.4 (7.4–9.5) 8.8 (7.9–9.9) 10.1 (9.3–11.1) 
    Secondary higher % (95% CI) 6.3 (4.7–8.3) 3.5 (2.4–5.1) 4.4 (2.8–6.7) 7.0 (5.3–9.0) 6.1 (4.8–7.7) 8.7 (7.4–10.2) 
    RII (95% CI)** 0.91 (0.56–1.50) 1.93 (1.19–3.14) 1.91 (1.18–3.08) 2.05 (1.38–3.05) 2.19 (1.43–3.36) 1.45 (1.01–2.09) 
    SII (95% CI) −0.45 (−2.71 to 1.92) 3.30 (0.91–5.38) 4.00 (1.06–6.53) 5.51 (2.55–8.10) 6.27 (2.97–9.09) 3.45 (0.09–12.91) 
Diabetes—Women       
    No formal education % (95% CI) 8.4 (7.7–9.3) 11.8 (10.3–13.7) 12.5 (10.2–16.0) 12.7 (10.9–15.5) 14.9 (12.9–17.4) 13.1 (11.3–15.3) 
    Primary % (95% CI) 6.0 (5.0–7.1) 6.3 (5.5–7.3) 7.0 (6.0–8.1) 7.0 (6.3–7.9) 7.5 (6.7–8.3) 7.5 (6.8–8.2) 
    Secondary higher % (95% CI) 3.8 (2.2–6.2) 2.9 (1.5–5.3) 2.3 (1.0–4.5) 3.5 (2.1–5.5) 2.2 (1.1–4.0) 4.0 (3.0–5.4) 
    RII (95% CI)** 3.04 (1.95–4.74) 4.89 (3.22–7.43) 5.17 (3.37–7.92) 6.09 (4.18–8.85) 7.54 (5.09–11.16) 4.28 (2.98–6.13) 
    SII (95% CI) 7.07 (4.51–9.12) 10.17 (8.10–11.75) 11.49 (9.22–13.19) 11.76 (9.44–13.50) 13.63 (11.95–14.87) 10.56 (8.46–12.23) 
Obesity—Men       
    No formal education % (95% CI) 11.1 (9.7–12.5) 13.6 (11.2–16.5) 18.7 (14.3–24.8) 25.5 (19.8–32.8) 20.6 (17.4–24.6) 20.4 (17.0–24.6) 
    Primary % (95% CI) 10.4 (9.0–11.9) 13.6 (12.3–15.0) 15.1 (13.3–17.2) 15.8 (14.5–17.3) 18.0 (16.6–19.4) 21.0 (19.7–22.4) 
    Secondary higher % (95% CI) 6.4 (5.0–8.2) 8.0 (6.3–10.1) 11.4 (8.9–14.6) 11.8 (9.9–14.1) 11.4 (9.6–13.5) 16.8 (15.1–18.6) 
    RII (95% CI)* 1.99 (1.41–2.81) 2.15 (1.56–2.95) 2.14 (1.53–3.00) 2.35 (1.75–3.15) 2.58 (1.94–3.43) 1.66 (1.28–2.14) 
    SII (95% CI) 6.22 (3.20–8.93) 8.76 (5.25–11.85) 10.67 (6.16–14.70) 12.25 (9.72–16.68) 14.83 (10.74–18.43) 9.62 (4.76–14.09) 
Obesity—Women       
    No formal education % (95% CI) 17.2 (15.6–18.9) 24.5 (21.1–28.4) 24.7 (20.3–30.5) 36.0 (29.9–43.5) 30.6 (26.8–35.1) 32.6 (28.4–37.4) 
    Primary % (95% CI) 11.2 (9.6–13.1) 14.4 (13.0–16.0) 18.0 (16.1–20.1) 20.2 (18.7–21.8) 17.4 (16.2–18.7) 19.8 (18.6–21.1) 
    Secondary higher % (95% CI) 4.8 (3.3–7.2) 4.9 (3.1–7.7) 8.2 (5.5–12.0) 10.1 (7.8–13.0) 8.6 (6.5–11.3) 10.4 (8.8–12.2) 
    RII (95% CI)* 3.98 (2.78–5.68) 4.77 (3.47–6.57) 4.05 (3.00–5.49) 4.35 (3.37–5.62) 4.99 (3.82–6.51) 4.25 (3.35–5.39) 
    SII (95% CI) 14.96 (11.77–17.51) 19.08 (16.14–24.49) 21.26 (17.60–24.35) 24.55 (21.26–27.36) 23.71 (20.83–26.12) 23.40 (20.42–25.97) 

Men and women ≥35 years

*P-value of interaction between SEP and year <0.05; **P-value of interaction between SEP position and year <0.01

In 2006, the RII for T2DM among women decreased from 4.28 (95% CI 2.98–6.13) in the age-adjusted model to 2.74 (95% CI 1.78–4.21) after additional adjustment for BMI, and to 2.44 (95% CI 1.60–3.73) after adjustment for age, BMI, lifestyle and diet (figure 1). Using the formula proposed by Brotman et al.,19 BMI, lifestyle and diet, explain 39% of the SEP inequalities in T2DM in women in the year 2006. These risk factors, particularly BMI, explain part of the social differences in the trends observed in inequalities in prevalence of T2DM among women, but not among men. No interactions were found since all the interaction terms between BMI, lifestyle and diet and SEP were not statistically significant (P > 0.05).

Figure 1

Relative Index of Inequalities (RII) of the association between educational level and T2DM (multivariate log-binomial regression model), models adjusted by possible mediators for year of survey Men and women ≥35 years. This analyses includes only individuals not missing in BMI categories aPlus smoking status, physical activity at work, physical activity in leisure time: bPlus intake vegetables and intake sweet food

Figure 1

Relative Index of Inequalities (RII) of the association between educational level and T2DM (multivariate log-binomial regression model), models adjusted by possible mediators for year of survey Men and women ≥35 years. This analyses includes only individuals not missing in BMI categories aPlus smoking status, physical activity at work, physical activity in leisure time: bPlus intake vegetables and intake sweet food

Discussion

One of the main results of this study is the observation that T2DM prevalence has increased in all educational levels. In general, SEP inequalities in T2DM existed >20 years ago and have increased, especially among women. Since 1993, new inequalities seem to have emerged among men. Moreover, SEP-related inequalities in T2DM prevalence were greater among women than among men throughout the period examined. Among women these relative inequalities and increases were attenuated when BMI and other risk factors were taken into account.

Limitations

A possible limitation of this study is the fact that self-reported data were used as a measure of the prevalence of T2DM. It is well known that health survey respondents report only diagnosed T2DM, which represents between 30% and 50% of total T2DM prevalence.20,21 However, educational level may not be associated with undiagnosed T2DM22 and self-reported T2DM in health interview surveys thus seem to be a good instrument to evaluate SEP inequalities in T2DM prevalence.23 Moreover, since the majority of cases are diagnosed by general practitioners and there were no major inequalities in access in Spain between 1993 and 2006,24 we do not expect major SEP-related inequalities in diagnosis.

Another limitation could be the fact that the question used to collect data on T2DM status in the 2006 survey was different to that used in previous years. However, estimates of inequalities in T2DM and obesity, and the question used to collect data on the latter, were consistent across all survey years, suggesting that this question did not markedly affect the estimate of inequalities (table 3). Finally, another limitation could be that the confounding effect of obesity may have been underestimated by the fact that weight loss is recommended in people with T2DM. As a result, in the analyses of our data, we may have underestimated the association between obesity and T2DM, and therefore the contribution that obesity makes to inequalities in T2DM. It is worth to mention that although we have considered the risk factors as possible mediators (obesity, BMI, lifestyles and nutrition), some of them may not be part of the causal chain between SEP and T2DM and therefore should be better considered as confounders.25

Differences in SEP inequalities in T2DM between men and women

In agreement with previous studies, SEP-related inequalities in T2DM were found in both men and women. As in other European countries, inequalities in T2DM prevalence in Spain are more marked among women than men,6 as observed for other chronic diseases. It has been suggested that this may be due to corresponding gender differences in health behaviours.6,8 However, we observed that differences in inequality remain even after adjusting for some of these health behaviours. This may be due to other psychosocial factors, such as power inequalities expressed as stress or work-related factors (overtime, shift-work, tense working conditions, low salaries, etc.).26,27 In this sense, women with lower SEP have higher prevalence of obesity, lower levels of physical activity and poorer health habits than similar men. This is not the case in higher SEP groups, where domestic responsibilities may be more evenly shared between men and women or may be done by someone else. That implies that women of these groups have more free time (e.g. to do physical or other healthy activities).

Trends in SEP inequalities in T2DM

In this study, SEP inequalities in T2DM in Spain have emerged in men and grown in women since 1987, in agreement with studies carried out in London and Germany.7,8 In the German study,7 the increase in T2DM inequalities was due to an increase in the prevalence of T2DM among disadvantaged SEP groups and a decrease among advantaged SEP groups, which could perhaps be attributable to a better adherence to diet and physical activity recommendations in the latter, improvement in treatment options for deprived SEP groups, or better detection of disease among deprived SEP groups.8 In our study, the increase in T2DM inequalities among women is partly explained by increased obesity-related inequalities, as shown by the multivariate models. Therefore, understanding trends in the inequalities of obesity may help to understand the trends in inequalities in T2DM. During the 20 years of these surveys, changes in the patterns of physical activity, diet and other social behaviours have occurred in the population, and these may influence trends in obesity. In the following sections, we hypothesize how these trends could effect to SEP inequalities in Spain.

Changes in eating behaviours

Changes in the patterns of food consumption may have had different effects according to SEP, with lower SEP groups having more limited access to healthy food. Spain has experienced important socio-economic transitions and consequently dietary changes in the post-dictatorship period. The Spanish diet is now characterized by a very high level of energy intake from fat, increased meat consumption, high fruit consumption and low vegetable consumption.9 Increasing numbers of people started to eating outside home, and availability of cheap fast-food increased. The frequency of fast-food consumption has been associated with obesity and T2DM.28 Lower SEP groups tend to have greater fast-food consumption than higher SEP groups.29 Consumption of healthy foods depends on availability30 as well as their proximity and variety on offer.29 Studies carried out in the USA have shown that deprived neighbourhoods tend to have more food stores,31 which tends to attract more fast-food outlets and convenience stores compared with more advantaged areas, which attract the best restaurants and freshest products.32 In addition, food purchase choices are partly made on the basis of cost time availability and low-cost foods are generally energy dense and nutrient poor.31,32

Changes in physical activity

It should be noted that physical activity is an economic and time cost activity, which creates social inequalities in the use of sports facilities.32,33 Interventions to promote physical activity should focus on modifiable determinants such as social support or facilities, providing information, behavioural management skills and other resources.34 Changes in patterns of work and leisure time physical activity may also have different effects according to SEP. Spain has the lowest rate of physical activity in the European Union10 and physical activity prevalence has been decreasing between 1995 and 2008 because of reduced occupational physical activity,35 and a lack of compensation through increased leisure time physical activity.14 Women with deprived SEP have the worst indicators in leisure time physical activity and that may be due to greater domestic responsibilities, poorer health and more financial problems.36 In our study, we found that inequalities in physical activity among women are constant throughout the study period. So the increase in SEP inequalities in T2DM can not be explained by changes in physical activity. For this reason, more studies focused on this issue are needed.

Changes in social structure

In Spain, changes in social structure and rapid changes in the distribution of education levels may perhaps contribute to the greater increase in T2DM inequalities observed among women than among men. Between 1987 and 1993, inequalities among Spanish women increased more than among men for some common chronic illness, including T2DM.37 During this period, women have entered the educational system.13 Therefore, since low educational level probably reflects a lower SEP in 2006 than in 1987, women with a low level of education are becoming an increasingly marginal population,38 and this might contribute to our finding of worse health behaviours from 1993 to 2006 among women with a lower level of education.

Changes in mortality

Since, increases in T2DM prevalence in developed countries have been found to be mainly due to decreases in mortality,12 changes in mortality among people with diabetes could play an additional role. Borrell et al.39 found that inequalities in T2DM mortality between 1992 and 2003 tended to increase among men and decrease among women, although these trends were not statistically significant. A decrease in inequalities in women mortality over time might result in an increase in inequalities in the prevalence of T2DM. The reduction in SEP-related inequalities in T2DM mortality in women since 1987 is partly due to improved access and use healthcare services among women of lower SEP.24

Conclusions

SEP inequalities in T2DM existed >20 years ago and have increased, especially among women. These trends could be partly explained by inequalities in obesity conditioned by a combination of changes in eating behaviours, physical activity or patterns of mortality. Underlying these changes are far-reaching transitions that Spain has experienced since 1975.

The American Diabetes Association's recommendations for prevention of T2DM are healthy eating habits and physical activity (http://www.diabetes.org/). However, programmes to reduce social inequalities need to go further and focus on structural changes such as personal contexts, and social environments, and have to take personal situation into account.40,41 The results of this study suggest that the socio-economic inequalities in T2DM and obesity among Spanish women, which are now greater than in other European countries, have existed for >20 years.

Key points

What is already known on this subject What this study adds

  • SEP inequalities in the prevalence of T2DM are present throughout Europe.

  • In Germany as well in the UK, SEP inequalities in the prevalence of T2DM have been rising in recent years.

  • In Spain, a country in Southern Europe, T2DM prevalence has increased during the past 20 years in all educational levels.

  • SEP inequalities in T2DM exist and have increased during this period, especially among women.

  • In women, BMI and other risk factors explain part of the SEP inequalities in the prevalence of T2DM and account for part of the increase in these inequalities.

Acknowledgements

The authors are grateful to the Ministerio de Sanidad y Política Social for providing data from the Spanish National Health Survey, on which this study is based. This article forms part of the doctoral dissertation of Albert Espelt at the Universitat Pompeu Fabra of Barcelona.

Conflicts of interest: None declared.

References

1
Roglic
G
Unwin
N
Bennett
PH
, et al.  . 
The burden of mortality attributable to diabetes: realistic estimates for the year 2000
Diabetes Care
 , 
2005
, vol. 
28
 (pg. 
2130
-
5
)
2
Wild
S
Roglic
G
Green
A
, et al.  . 
Global prevalence of diabetes: estimates for the year 2000 and projections for 2030
Diabetes Care
 , 
2004
, vol. 
27
 (pg. 
1047
-
53
)
3
Mackenbach
JP
Stirbu
I
Roskam
AJ
, et al.  . 
Socioeconomic inequalities in health in 22 European countries
N Engl J Med
 , 
2008
, vol. 
358
 (pg. 
2468
-
81
)
4
Espelt
A
Arriola
L
Borrell
C
, et al.  . 
Socioeconomic Position and Type 2 Diabetes Mellitus in Europe 1999- 2009: a Panorama of Inequalities
Curr Diabetes Rev
 , 
2011
, vol. 
7
 (pg. 
1
-
11
)
5
Agardh
E
Allebeck
P
Hallqvist
J
, et al.  . 
Type 2 diabetes incidence and socio-economic position: a systematic review and meta-analysis
Int J Epidemiol
 , 
2011
, vol. 
40
 (pg. 
804
-
18
)
6
Espelt
A
Borrell
C
Roskam
AJ
, et al.  . 
Socioeconomic inequalities in diabetes mellitus across Europe at the beginning of the 21st century
Diabetologia
 , 
2008
, vol. 
51
 (pg. 
1971
-
9
)
7
Icks
A
Moebus
S
Feuersenger
A
, et al.  . 
Diabetes prevalence and association with social status–Widening of a social gradient?: German national health surveys 1990-1992 and 1998
Diabetes Res Clin Pract
 , 
2007
, vol. 
78
 (pg. 
293
-
7
)
8
Imkampe
AK
Gulliford
MC
Increasing socio-economic inequality in type 2 diabetes prevalence–Repeated cross-sectional surveys in England 1994-2006
Eur J Public Health
 , 
2010
, vol. 
21
 (pg. 
484
-
90
)
9
Moreno
LA
Sarria
A
Popkin
BM
The nutrition transition in Spain: a European Mediterranean country
Eur J Clin Nutr
 , 
2002
, vol. 
56
 (pg. 
992
-
1003
)
10
Vaz de
A
Graca
P
Afonso
C
, et al.  . 
Physical activity levels and body weight in a nationally representative sample in the European Union
Public Health Nutr
 , 
1999
(pg. 
105
-
13
)
11
Roskam
A
Kunst
A
Espelt
E
, et al.  . 
Social Inequalities in the prevalence of diabetes. En: Cross-national Comparisions of Socioeconomic differences in Overweight and Obesity
 , 
2009
Rotterdam
Erasmus Universiteit Rotterdam
12
Lipscombe
LL
Hux
JE
Trends in diabetes prevalence, incidence, and mortality in Ontario, Canada 1995-2005: a population-based study
Lancet
 , 
2007
, vol. 
369
 (pg. 
750
-
6
)
13
Carrasco-Portino
M
Ruiz-Cantero
MT
Gil-Gonzalez
D
, et al.  . 
[Gender development inequalities epidemiology in Spain (1990-2000)]
Rev Esp Salud Publica
 , 
2008
, vol. 
82
 (pg. 
283
-
99
)
14
Meseguer
CM
Galan
I
Herruzo
R
Rodriguez-Artalejo
F
Trends in leisure time and occupational physical activity in the Madrid region, 1995-2008
Rev Esp Cardiol
 , 
2011
, vol. 
64
 (pg. 
21
-
7
)
15
Ministerio de Sanidad y Consumo
Encuesta Nacional de Salud de España
  
Available at: http://www.msc.es/estadEstudios/estadisticas/encuestaNacional/home.htm. 2010 (September 2010, date last accessed)
16
Rué
M
Borrell
C
Los métodos de estandarización de tasas. Revisiones de Salud Pública
Revisiones de Salud Pública
 , 
1993
, vol. 
3
 (pg. 
263
-
95
)
17
Shaw
M
Galobardes
B
Lawlor
D
, et al.  . 
The Handbook of Inequality and Socioeconomic Position: Concepts and Measures
 , 
2007
Bristol, UK
The Policy Press
18
Ezendam
N
Educational inequalities in cancer mortality differ greatly between countries around the Baltic Sea
Eur J Cancer
 , 
2008
, vol. 
44
 (pg. 
454
-
64
)
19
Brotman
DJ
Mediators of the association between mortality risk and socioeconomic status
JAMA
 , 
2006
, vol. 
296
 (pg. 
763
-
4
)
20
Franse
LV
Di
BM
Shorr
RI
, et al.  . 
Type 2 diabetes in older well-functioning people: who is undiagnosed? Data from the Health, Aging, and Body Composition study
Diabetes Care
 , 
2001
, vol. 
24
 (pg. 
2065
-
70
)
21
Rathmann
W
Haastert
B
Icks
A
, et al.  . 
High prevalence of undiagnosed diabetes mellitus in Southern Germany: target populations for efficient screening. The KORA survey 2000
Diabetologia
 , 
2003
, vol. 
46
 (pg. 
182
-
9
)
22
Wilder
RP
Majumdar
SR
Klarenbach
SW
Jacobs
P
Socio-economic status and undiagnosed diabetes
Diabetes Res Clin Pract
 , 
2005
, vol. 
70
 (pg. 
26
-
30
)
23
Espelt
A
Goday
A
Franch
J
Borrell
C
Validity of self-reported diabetes in Health Interview Surveys for measuring social inequalities in the prevalence of diabetes
J Epidemiol Community Health
 , 
2011
 
doi:10.1136/jech.2010.112698 [Epub ahead of print, 17 April 2011]
24
Palència
L
Espelt
A
Rodríguez-Sanz
M
, et al.  . 
Trends in Social Class Inequalities in use of health care services within the Spanish National Health System, 1993-2006
Eur J Health Economics
 , 
2011
 
[Epub ahead of print, 10 October 2011]
25
Babyak
MA
Understanding confounding and mediation
Evid Based Ment Health
 , 
2009
, vol. 
12
 (pg. 
68
-
71
)
26
Kroenke
CH
Spiegelman
D
Manson
J
, et al.  . 
Work characteristics and incidence of type 2 diabetes in women
Am J Epidemiol
 , 
2007
, vol. 
165
 (pg. 
175
-
83
)
27
Norberg
M
Stenlund
H
Lindahl
B
, et al.  . 
Work stress and low emotional support is associated with increased risk of future type 2 diabetes in women
Diabetes Res Clin Pract
 , 
2007
, vol. 
76
 (pg. 
368
-
77
)
28
Pereira
MA
Kartashov
AI
Ebbeling
CB
, et al.  . 
Fast-food habits, weight gain, and insulin resistance (the CARDIA study): 15-year prospective analysis
Lancet
 , 
2005
, vol. 
365
 (pg. 
36
-
42
)
29
Thornton
L
Bentley
R
Kavanagh
A
Fast food purchasing and access to fast food restaurants: a multilevel analysis of VicLanes
Int J Behav Nutr Phys Act
 , 
2009
(pg. 
6
-
28
)
30
Jago
R
Baranowski
T
Baranowski
J
Fuit and vegetable availability: a micro enivronmental mediating variable?
Public Health Nutr
 , 
2010
, vol. 
10
 (pg. 
681
-
9
)
31
Sharkey
JR
Horel
S
Neighborhood socioeconomic deprivation and minority composition are associated with better potential spatial access to the ground-truthed food environment in a large rural area
J Nutr
 , 
2008
, vol. 
138
 (pg. 
620
-
7
)
32
Drewnowski
A
Obesity, diets, and social inequalities
Nutr Rev
 , 
2009
, vol. 
67
 
Suppl 1
(pg. 
S36
-
9
)
33
Lovasi
GS
Hutson
MA
Guerra
M
Neckerman
KM
Built environments and obesity in disadvantaged populations
Epidemiol Rev
 , 
2009
, vol. 
31
 (pg. 
7
-
20
)
34
Kahn
EB
Ramsey
LT
Brownson
RC
, et al.  . 
The effectiveness of interventions to increase physical activity. A systematic review
Am J Prev Med
 , 
2002
, vol. 
22
 (pg. 
73
-
107
)
35
Brownson
RC
Boehmer
TK
Luke
DA
Declining rates of physical activity in the United States: what are the contributors?
Annu Rev Public Health
 , 
2005
, vol. 
26
 (pg. 
421
-
43
)
36
Droomers
M
Schrijvers
CT
Mackenbach
JP
Educational level and decreases in leisure time physical activity: predictors from the longitudinal GLOBE study
J Epidemiol Community Health
 , 
2001
, vol. 
55
 (pg. 
562
-
8
)
37
Regidor
E
Gutierrez-Fisac
JL
Dominguez
V
Calle
ME
Navarro
P
Comparing social inequalities in health in Spain: 1987 and 1995/97
Soc Sci Med
 , 
2002
, vol. 
54
 (pg. 
1323
-
32
)
38
Regidor
E
Martinez
D
Astasio
P
, et al.  . 
Trends of socioeconomic inequalities and socioeconomic inequalities in self-perceived health in Spain. Gac.Sanit
Maggio
 , 
2006
, vol. 
20
 
3
(pg. 
178
-
82
)
39
Borrell
C
Azlor
E
Rodriguez-Sanz
M
, et al.  . 
Trends in socioeconomic mortality inequalities in a southern European urban setting at the turn of the 21st century
J Epidemiol Community Health
 , 
2008
, vol. 
62
 (pg. 
258
-
66
)
40
Giachello
AL
Arrom
JO
Davis
M
, et al.  . 
Reducing diabetes health disparities through community-based participatory action research: the Chicago Southeast Diabetes Community Action Coalition
Public Health Rep
 , 
2003
, vol. 
118
 (pg. 
309
-
23
)
41
Comisión para reducir las desigualdades en salud en España
Propuesta de políticas e intervenciones para reducir las desigualdades en salud en España
Gac Sanit
 , 
2012
 
doi:10.1016/j.gaceta.2011.07.024

Supplementary data

Comments

0 Comments