Validity of self-reported mammography uptake in the Belgian health interview survey: selection and reporting bias.

BACKGROUND: The validity of self-reported mammography uptake is often questioned. We assessed the related selection and reporting biases among women aged 50-69 years in the Belgian Health Interview Survey (BHIS) using reimbursement data for mammography stemming from the Belgian Compulsory Health Insurance organizations (BCHI). METHODS: Individual BHIS 2013 data (n = 1040) were linked to BCHI data 2010-13 (BHIS-BCHI sample). Being reimbursed for mammography within the last 2-years was used as the gold standard. Selection bias was assessed by comparing BHIS estimates reimbursement rates in BHIS-BCHI with similar estimates from the Echantillon Permanent/Permanente Steekproef (EPS), a random sample of BCHI data, while reporting bias was investigated by comparing self-reported versus reimbursement information in the BHIS-BCHI. Reporting bias was further explored through measures of agreement and logistic regression. RESULTS: Mammography uptake rates based on self-reported information and reimbursement from the BHIS-BCHI were 75.5% and 69.8%, respectively. In the EPS, it was 64.1%. The validity is significantly affected by both selection bias {relative size = 8.93% [95% confidence interval (CI): 3.21-14.64]} and reporting bias [relative size = 8.22% (95% CI: 0.76-15.68)]. Sensitivity was excellent (93.7%), while the specificity was fair (66.4%). The agreement was moderate (kappa = 0.63). Women born in non-EU countries (OR = 2.81, 95% CI: 1.54-5.13), with high household income (OR = 1.27, 95% CI: 1.02-1.60) and those reporting poor perceived health (OR = 1.41, 95% CI: 1.14-1.73) were more likely to inaccurately report their mammography uptake. CONCLUSIONS: The validity of self-reported mammography uptake in women aged 50-69 years is affected by both selection and reporting bias. Both administrative and survey data are complementary when assessing mammography uptake.

Is the European Health Interview Survey online yet? Response and net sample composition of a web-based data collection.

BACKGROUND: The European Health Interview Survey (EHIS) provides cross-national data on health status, health care and health determinants. So far, 10 of the 30 member states (MS) opted for web-based questionnaires within mixed-mode designs but none used it as the sole mode. In the context of future EHIS, the response rate and net sample composition of a web-only approach was tested. METHODS: A Belgian study with a target sample size of 1000 (age: 16-85) was organized using the EHIS wave 3 model questionnaire. The sample was selected according to a multistage, clustered sampling procedure with geographical stratification. Field substitution was applied; non-participating households were replaced by similar households regarding statistical sector, sex and age. There was one reminder letter and a €10 conditional incentive. RESULTS: Considering all substitutions, a 16% response rate was obtained after sending one reminder. Elderly, Brussels Capital inhabitants, people living without a partner and those with a non-Belgian nationality were less responding. By design, there were no differences between the initial and final net sample regarding substitution characteristics. Nevertheless, people living without a partner, non-Belgians and lower educated people remain underrepresented. CONCLUSION: There was a low response rate, particularly for some population groups. The response rate was lower than those of MS using mixed-mode designs including web, especially these comprising interviewer-based approaches. Despite the long and complex questionnaire, there was a low break off rate. So far, web-only data collection is not an acceptable strategy for population-based health surveys but efforts to increase the response should be further explored.

A Novel Approach to Optimize Vitamin D Intake in Belgium through Fortification Based on Representative Food Consumption Data.

BACKGROUND: Food fortification is a promising means to improve vitamin D intake of a population. Careful selection of food vehicles is needed to ensure that nearly all individuals within the population benefit from the fortification program. OBJECTIVES: The aim of the study was to develop and apply a model that simultaneously selects the optimal combination of food vehicles and defines the optimal fortification level that adequately increases vitamin D intake in the population without compromising safety. METHODS: Food consumption data from the Belgian Food Consumption Survey 2014 (n = 3200; age 3-64 y) were used. The optimization model included 63 combinations of 6 potential vehicles for food fortification, namely "bread," "breakfast cereals," "fats and oils," "fruit juices," "milk and milk beverages," and "yogurt and cream cheese." The optimization procedure was designed to minimize inadequate or excessive vitamin D intake in each of the food combinations. This allowed the relative ranking of the different combinations according to their fortification utility. The estimated average requirement and upper intake level were used as thresholds. An age-specific and population-based approach enabled the sensitivity of the population subgroups to adverse health effects to be taken into account. Feasibility, technical aspects, and healthiness of the food vehicles were used to select the optimal combination. RESULTS: Multiple combinations of food vehicles significantly reduced the prevalence of inadequate vitamin D intake within the Belgian population (from 92-96% to <2%). Taking other aforementioned criteria into account, the fortification of "milk and milk beverages" and "bread" with 6.9 µg vitamin D/100 kcal was proposed as an optimal fortification scenario. CONCLUSIONS: The optimization model allows identification of an effective fortification scenario to improve vitamin D intake within the Belgian population based on acceptable risks of inadequate and excessive intake. The model can be extended to other micronutrients and other populations.

The genetic structure of the Belgian population.

BACKGROUND: National and international efforts like the 1000 Genomes Project are leading to increasing insights in the genetic structure of populations worldwide. Variation between different populations necessitates access to population-based genetic reference datasets. These data, which are important not only in clinical settings but also to potentiate future transitions towards a more personalized public health approach, are currently not available for the Belgian population. RESULTS: To obtain a representative genetic dataset of the Belgian population, participants in the 2013 National Health Interview Survey (NHIS) were invited to donate saliva samples for DNA analysis. DNA was isolated and single nucleotide polymorphisms (SNPs) were determined using a genome-wide SNP array of around 300,000 sites, resulting in a high-quality dataset of 189 samples that was used for further analysis. A principal component analysis demonstrated the typical European genetic constitution of the Belgian population, as compared to other continents. Within Europe, the Belgian population could be clearly distinguished from other European populations. Furthermore, obvious signs from recent migration were found, mainly from Southern Europe and Africa, corresponding with migration trends from the past decades. Within Belgium, a small north-west to south-east gradient in genetic variability was noted, with differences between Flanders and Wallonia. CONCLUSIONS: This is the first study on the genetic structure of the Belgian population and its regional variation. The Belgian genetic structure mirrors its geographic location in Europe with regional differences and clear signs of recent migration.

Mixing mixed-mode designs in a national health interview survey: a pilot study to assess the impact on the self-administered questionnaire non-response.

BACKGROUND: Many population health surveys consist of a mixed-mode design that includes a face-to-face (F2F) interview followed by a paper-and-pencil (P&P) self-administered questionnaire (SAQ) for the sensitive topics. In order to alleviate the burden of a supplementary P&P questioning after the interview, a mixed-mode SAQ design including a web and P&P option was tested for the Belgian health interview survey. METHODS: A pilot study (n = 266, age 15+) was organized using a mixed-mode SAQ design following the F2F interview. Respondents were invited to complete a web SAQ either immediately after the interview or at a later time. The P&P option was offered in case respondents refused or had previously declared having no computer access, no internet connection or no recent usage of computers. The unit response rate for the web SAQ and the overall unit response rate for the SAQ independent of the mode were evaluated. A logistic regression analysis was conducted to explore the association of socio-demographic characteristics and interviewer effects with the completed SAQ mode. Furthermore, a logistic regression analysis assessed the differential user-friendliness of the SAQ modes. Finally, a logistic multilevel model was used to evaluate the item non-response in the two SAQ modes while controlling for respondents' characteristics. RESULTS: Of the eligible F2F respondents in this study, 76% (107/140) agreed to complete the web SAQ. Yet among those, only 78.5% (84/107) actually did. At the end, the overall (web and P&P) SAQ unit response rate reached 73.5%. In this study older people were less likely to complete the web SAQ. Indications for an interviewer effect were observed as regard the number of web respondents, P&P respondents and respondents who refused to complete the SAQ. The web SAQ scored better in terms of user-friendliness and presented higher item response than the P&P SAQ. CONCLUSIONS: The web SAQ performed better regarding user-friendliness and item response than the P&P SAQ but the overall SAQ unit response rate was low. Therefore, future research is recommended to further assess which type of SAQ design implemented after a F2F interview is the most beneficial to obtain high unit and item response rates.

Consumption of ultra-processed food products and diet quality among children, adolescents and adults in Belgium.

PURPOSE: To assess the dietary share of ultra-processed foods (UPF) among Belgian children, adolescents and adults and associations with diet quality. METHODS: Data from the national Food Consumption Surveys 2004 (N = 3083; ≥ 15 years) and 2014-2015 (N = 3146; 3-64 years) were used. Two 24-h recalls (dietary records for children) were used for data collection. Foods consumed were classified by the level of processing using the NOVA classification. The usual proportion of daily energy intake from UPF was determined using SPADE (Statistical Program to assess dietary exposure). RESULTS: In 2014/2015, 36.4% of foods consumed were ultra-processed, while 42.4% were unprocessed/minimally processed. The usual proportion of daily energy intake from UPF was 33.3% (95% CI 32.1-35.0%) for children, 29.2% (95% CI 27.7-30.3%) for adolescents and 29.6% (95% CI 28.5-30.7%) for adults. There were no differences in UPF consumption between 2004 and 2014/2015. The products contributing most to UPF consumption were processed meat (14.3%), cakes, pies, pastries (8.9%), sweet biscuits (7.7%) and soft drinks (6.7%). The UPF dietary share was significantly lower during consumption days when participants met the WHO salt intake recommendation (≤ 5 g/day) and when saturated fat was ≤ 10% of their total energy intake. The dietary share of unprocessed/minimally processed foods was significantly higher during consumption days when participants met the WHO salt and fruit/vegetable intake (≥ 400 g/day) recommendations and when saturated fat was ≤ 10% of their total energy intake. CONCLUSIONS: The UPF dietary share is substantial and associated with lower diet quality. Internationally recommended policies to limit UPF accessibility and marketing need to be implemented to reduce UPF consumption.

Do sociodemographic characteristics associated with the use of CAM differ by chronic disease?

BACKGROUND: Complementary and alternative medicine (CAM) is oftenused to alleviate the discomfort, disability and pain involved in many chronic diseases. Besides this, females, middle-aged and higher educated people are also known to use CAM the most. This study explores whether the sociodemographic characteristics associated with CAM use differ by type of disease. METHODS: The following data were taken from the Belgian Health Interview Survey 2013 for the individuals aged 15+ years (n = 8942): sociodemographic characteristics, past 12-month diseases (using a list) and contact with a homeopath, chiropractor, acupuncturist and/or osteopath (CAM-therapists) in the past year. The association between CAM use and disease, controlled for gender, age, education and conventional medicine use, was assessed through logistic regressions. When interactions with the sociodemographic characteristics were found, stratified regressions were conducted. RESULTS: People with musculoskeletal diseases [odds ratio (OR) = 2.6], allergy (OR = 1.4) and severe headache (OR = 1.5) had higher odds of using CAM in the past year with statistical significance. For musculoskeletal diseases, the odds of using CAM was higher, with statistical significance, for every sociodemographic subclass. For allergy, CAM use was higher among men, people aged 45+ years and lower educated people, while for severe headache CAM use was higher among women, people aged 45+ years and higher educated people, all with statistical significance. CONCLUSIONS: Sociodemographic characteristics associated with CAM use differ by diseases. The role of CAM in disease management cannot be ignored. Making physicians aware for which disease CAM is used and by whom, may facilitate disease management.

Contribution of chronic conditions to gender disparities in health expectancies in Belgium, 2001, 2004 and 2008.

Background: We aimed to investigate the contribution of chronic conditions to gender differences in disability-free life expectancy (DFLE) and life expectancy with disability (LED) in Belgium in 2001, 2004 and 2008. Methods: Data on disability and chronic conditions from participants of the 2001, 2004 and 2008 Health Interview Surveys in Belgium were used to estimate disability prevalence by cause using the attribution method. Disability prevalence was applied to life tables to estimate DFLE and LED using the Sullivan method. Decomposition techniques were used to assess the contribution of mortality and disability and further of causes of death and disability to gender disparities in DFLE and LED. Results: Higher LE, DFLE and LED were observed for women compared with men in all years studied. A decrease in the gender gap in LE (2001: 5.9; 2004: 5.6; 2008: 5.3) was observed in our cross-sectional approach followed by a decrease in gender differences in DFLE (2001: 1.9; 2004: 1.3; 2008: 0.5) and increase in LED (2001: 4.0; 2004: 4.4; 2008: 4.8). The higher LED in women was attributed to their lower mortality due to lung/larynx/trachea cancer, ischaemic heart diseases, and external causes (2001 and 2004) and higher disability prevalence due to musculoskeletal conditions (2008). Higher DFLE was observed in women owing to their lower mortality from lung/larynx/trachea cancer, ischaemic heart diseases, digestive cancer and chronic respiratory diseases. Conclusion: To promote healthy ageing of populations, priority should be given to reduce the LED disadvantage in women by targeting non-fatal diseases, such as musculoskeletal conditions.

Joint Data Analysis in Nutritional Epidemiology: Identification of Observational Studies and Minimal Requirements.

Background: Joint data analysis from multiple nutrition studies may improve the ability to answer complex questions regarding the role of nutritional status and diet in health and disease. Objective: The objective was to identify nutritional observational studies from partners participating in the European Nutritional Phenotype Assessment and Data Sharing Initiative (ENPADASI) Consortium, as well as minimal requirements for joint data analysis. Methods: A predefined template containing information on study design, exposure measurements (dietary intake, alcohol and tobacco consumption, physical activity, sedentary behavior, anthropometric measures, and sociodemographic and health status), main health-related outcomes, and laboratory measurements (traditional and omics biomarkers) was developed and circulated to those European research groups participating in the ENPADASI under the strategic research area of "diet-related chronic diseases." Information about raw data disposition and metadata sharing was requested. A set of minimal requirements was abstracted from the gathered information. Results: Studies (12 cohort, 12 cross-sectional, and 2 case-control) were identified. Two studies recruited children only and the rest recruited adults. All studies included dietary intake data. Twenty studies collected blood samples. Data on traditional biomarkers were available for 20 studies, of which 17 measured lipoproteins, glucose, and insulin and 13 measured inflammatory biomarkers. Metabolomics, proteomics, and genomics or transcriptomics data were available in 5, 3, and 12 studies, respectively. Although the study authors were willing to share metadata, most refused, were hesitant, or had legal or ethical issues related to sharing raw data. Forty-one descriptors of minimal requirements for the study data were identified to facilitate data integration. Conclusions: Combining study data sets will enable sufficiently powered, refined investigations to increase the knowledge and understanding of the relation between food, nutrition, and human health. Furthermore, the minimal requirements for study data may encourage more efficient secondary usage of existing data and provide sufficient information for researchers to draft future multicenter research proposals in nutrition.

Contribution of chronic conditions to smoking differences in life expectancy with and without disability in Belgium.

Background: Smoking is the leading cause of premature mortality and morbidity. This study aimed at assessing the impact of smoking on life expectancy (LE) and LE with (LED) and without disability (DFLE). We further estimated the contribution of disability and mortality and their causes to differences in LED and DFLE by smoking. Methods: Data on disability, chronic conditions, and smoking from 17 148 participants of the 1997, 2001, 2004 Belgian Health Interview Surveys were used to estimate causes of disability using the attribution method. A 10-year mortality follow-up of survey participants was used. The Sullivan method was applied to estimate LED and DFLE. The contribution of disability and mortality and of causes of disability and death to smoking differences in LED and DFLE was assessed using decomposition methods. Results: Never smokers live longer than daily smokers. DFLE advantage at age 15 of +8.5/+4.3 years (y) in men/women never compared with daily smokers was the result of lower mortality (+6.2y/+3y) and lower disability (2.3y/1.3y). The extra 0.3y/1.6y LED in never smokers was due to lower mortality (+2.6y/+2.9y) and lower disability (-2.3y/-1.3y). Lower mortality from lung/larynx/trachea cancer, chronic respiratory, and ischaemic heart diseases was the main contributor to higher LED and DFLE in never smokers. Lower disability from musculoskeletal conditions in men and chronic respiratory diseases in women increased LED and DFLE in never smokers. Conclusions: Mortality and disability advantage among never smokers contributed to longer DFLE, while mortality advantage contributed to their longer LED.

Is a Health Interview Survey an appropriate tool to assess domestic violence?

Background: The aim of this study is to assess if a Health Interview Survey (HIS) targeting the general population is an appropriate tool to collect valid data on domestic violence. Studying item non-response on the question on domestic violence and its association with socio-demographic and health characteristics compared with victims of domestic violence can contribute to this. Methods: Cross-sectional data from the Belgian HIS 2013 were analysed. A question whether the perpetrator of a violent event was a member of the respondents' household was embedded in a general topic on violence in the self-administered questionnaire. This study is limited to people aged 15+ that at least completed the first question of this topic. Socio-demographic characteristics of item non-respondents and of victims of domestic violence were explored and the association with health status was assessed through ORs calculated via logistic regression. Results: The year prevalence of domestic violence is 1.1%. Although the question on domestic violence yields a high level of non-response (62%), this does not hinder the further completion of the questionnaire. When compared with victims of domestic violence, those not responding on the question on the perpetrator have better (mental) health. When compared with those not being victim of domestic violence, victims report poorer physical and mental health. Conclusion: An HIS can be an appropriate tool to assess domestic violence in the general population and its association with health. However, a solution should be found for the high item non-response on the question on the perpetrator of the violent event.

Additional weighting for education affects estimates from a National Health Interview Survey.

Background: National Health Interview Surveys are used to produce country-wide results for a substantial number of health indicators. However, if some educational groups are underrepresented in the sample, estimates may be biased. This study investigated the impact of the use of post-stratification weights that adjust for the population distribution by education on estimates from the Belgian Health Interview Survey 2013. Methods: For 25 health-related indicators that match the European Core Health Indicator shortlist, estimates were computed using two different sets of post-stratification weights: one based on age group, gender and province only and the other one including also education. The Census 2011 was used as auxiliary data source. Statistical differences between the two estimates were assessed with the Delta method. Results: If education is not included as post-stratification weighting factor, low educational groups (ISCED 0-2) represent 31.1% of the total study population aged 25 years and older. If education is taken into account this proportion rises to 40.3%. The use of post-stratification weights adjusting for the population distribution by education has an impact on several survey estimates. The most pronounced effect is an increase in the estimated proportion of people with diabetes (+0.73%; 95% CI 0.19-1.27; relative increase +11.6%), asthma (+0.52%; 95% CI, 0.06-0.98; relative increase +12.4%) and difficulties to cover their health expenses (+2.31%; 95% CI, 1.52-3.10; relative increase +9.4%). Conclusions: Including education in the calculation of post-stratification weights reduces bias due to educational differences in survey participation. Auxiliary information used to calculate post-stratification weights for national health surveys should include education.

Migrant's access to preventive health services in five EU countries.

BACKGROUND: Preventive health services (PHSs) form part of primary healthcare with the aim of screening to prevent disease. Migrants show significant differences in lifestyle, health beliefs and risk factors compared with the native populations. This can have a significant impact on migrants' access to health systems and participation in prevention programmes. Even in countries with widely accessible healthcare systems, migrants' access to PHSs may be difficult. The aim of the study was to compare access to preventive health services between migrants and native populations in five European Union (EU) countries. METHODS: Information from Health Interview Surveys of Belgium, Italy, Malta, Portugal and Spain were used to analyse access to mammography, Pap smear tests, colorectal cancer screening and flu vaccination among migrants. The comparative risk of not accessing PHSs was calculated using a mixed-effects multilevel model, adjusting for potential confounding factors (sex, education and the presence of disability). Migrant status was defined according to citizenship, with a distinction made between EU and non-EU countries. RESULTS: Migrants, in particular those from non-EU countries, were found to have poorer access to PHSs. The overall risk of not reporting a screening test or a flu vaccination ranged from a minimum of 1.8 times (colorectal cancer screening), to a high of 4.4 times (flu vaccination) for migrants. The comparison among the five EU countries included in the study showed similarities, with particularly limited access recorded in Italy and in Belgium for non-EU migrants. CONCLUSIONS: The findings of this study are in accordance with evidence from the scientific literature. Poor organization of health services, in Italy, and lack of targeted health policies in Belgium may explain these findings. PHSs should be responsive to patient diversity, probably more so than other health services. There is a need for diversity-oriented, migrant-sensitive prevention. Policies oriented to removing impediments to migrants' access to preventive interventions are crucial, to encourage more positive action for those facing the risk of intersectional discrimination.

Contribution of chronic diseases to the disability burden in a population 15 years and older, Belgium, 1997-2008.

BACKGROUND: Age-associated disability reduces quality of life in older populations and leads to wide-range implications for social and health policy. The identification of diseases that contribute to the disability burden is crucial to the development of prevention and intervention strategies to reduce disability. In this study, we assessed the contribution of chronic diseases to the prevalence of disability in Belgium. METHODS: Data from 35,837 individuals aged 15 years or older who participated in the 1997, 2001, 2004, or 2008 Belgian Health Interview Surveys were used. Disability was defined as difficulties in doing at least one of six activities of daily living (transfer in and out of bed, transfer in and out of chair, dressing, washing hands and face, feeding, and going to the toilet) and/or mobility limitations (ability to walk without stopping less than 200 m). Multiple additive regression models were fitted separately for men and women to estimate the age-specific background disability rate (experienced by everyone, independent of the presence of specific diseases) and disease-specific disability rates (disability rate in subjects who reported selected chronic diseases). RESULTS: Musculoskeletal, cardiovascular, and respiratory diseases were the main contributors to the disability burden in Belgium. Musculoskeletal diseases were the most prevalent diseases in men and women in all age groups. Neurological diseases and stroke were the most disabling diseases, i.e. caused the highest level of disability among the diseased individuals, in all age groups for men and women, respectively. Back pain was the main cause of disability in men aged 15 to 64 years, while heart attack was the major contributor to the disability prevalence in men aged 65 or older. Likewise, arthritis was the main cause of disability among women across all age groups. Depression was also an important contributor in young subjects (15-54 years). Cancer was not an important contributor to the disability prevalence in Belgium. CONCLUSIONS: To reduce the burden of disability in Belgium, interventions should target musculoskeletal, cardiovascular and respiratory diseases especially among elderly. Furthermore, attention should also be given to depression in young individuals.

Using multiple measures to assess changes in social inequalities for breast cancer screening.

OBJECTIVES: To identify changes in social inequalities for mammograms uptake in Belgium over the period 1997-2008 using multiple indices, and to assess the contribution of the national breast cancer screening programme in these changes. METHODS: Data were obtained from four waves of the Belgian Health Interview Survey. The socio-economic position was defined by the educational level. Inequalities were measured both with pairwise measures comparing extreme educational groups (prevalence difference and prevalence ratio), and with indices measuring the total inequality impact at population level: the Population Attributable Fraction (PAF), the Relative Index of Inequality (RII) and the Slope Index of Inequality (SII). RESULTS: All indices show a substantial decrease in inequalities in mammographic uptake between 1997 and 2008. For the indices of total impact (PAF, RII, SII), the change occurred between the first two waves (1997 and 2001) and stabilized afterwards, while for pairwise indices the evolution continued over the whole period. CONCLUSION: Using multiple indices of inequality is necessary for a more complete understanding of the changes: total impact inequality indices should always complement simple pairwise measures. The inequalities in mammograms uptake, as measured with total impact indices, only decreased before the start of the national screening programme.

Using multiple measures of inequalities to study the time trends in social inequalities in smoking.

BACKGROUND: The time trends in social inequalities in smoking have been examined in a number of international publications; however, these studies have rarely used multiple measures of health inequalities simultaneously. Also the analytical approach used often did not account, as recommended, for the changes in the relative distribution of social groups and the changes in the absolute level of the health outcome within social groups. METHODS: Data from four successive waves of the Belgian Health Interview Survey (1997, 2001, 2004, 2008) were used to study the time trends in educational inequalities in daily smoking for those aged between 15 and 74 years. We estimated two measures of relative inequalities: the OR and the relative index of inequality; and two measures of absolute inequalities: the population attributable fraction and the slope index of inequality. Three of these measures (relative index of inequality, population attributable fraction, slope index of inequality) account for the change in the relative size of the social groups over time. RESULTS: The four measures of inequality were consistent in showing significant inequalities among educational groups. The time trends, however, were less consistent. Measures of trends in relative inequalities witnessed a small linear increase. However, no substantial over time change was observed with the measures of absolute inequalities. CONCLUSION: The time trends in social inequalities in smoking varied according to the measure of inequality used. This study confirms the importance of using multiple measures of inequalities to understand and monitor social inequalities in smoking.

Socio-economic differences in participation of households in a Belgian national health survey.

BACKGROUND: Socio-economic inequalities in health survey participation can jeopardize the extrapolation of the survey findings to the total population. Earlier research, based on aggregated data, showed that in Belgium less-educated people with poor health were less likely to participate in a health survey. In this article, the association by socio-economic status and household non-response in a health survey is examined. METHODS: A linkage between the Belgian Health Survey 2001 with Census 2001 enabled us to evaluate the participation by socio-economic status. RESULTS: We observed that the socio-economic position was a determinant of health survey participation: participation rate was significantly lower in households with a lower socio-economic profile. CONCLUSION: Socio-economic inequalities in participation can introduce a bias in the health survey findings. Strategies targeting improvement of the participation of lower socio-economic groups need to be considered.

International differences in self-reported health measures in 33 major metropolitan areas in Europe.

BACKGROUND: The increasing concentration of populations into large conurbations in recent decades has not been matched by international health assessments, which remain largely focused at the country level. We aimed to demonstrate the use of routine survey data to compare the health of large metropolitan centres across Europe and determine the extent to which differences are due to socio-economic factors. METHODS: Multilevel modelling of health survey data on 126,853 individuals from 33 metropolitan areas in the UK, Republic of Ireland, Sweden, Norway, Finland, Spain, Belgium and Germany compared general health, longstanding illness, acute sickness, psychological distress and obesity with the average for all areas, accounting for education and social class. RESULTS: We found some areas (Greater Glasgow; Greater Manchester, Cheshire and Merseyside; Northumberland, Tyne and Wear and South Yorkshire) had significantly higher levels of poor health. Other areas (West Flanders and Antwerp) had better than average health. Differences in individual socio-economic circumstances did not explain findings. With a few exceptions, acute sickness levels did not vary. CONCLUSION: Health tended to be worse in metropolitan areas in the north and west of the UK and the central belt and south east of Germany, and more favourable in Sweden and north west Belgium, even accounting for socio-economic composition of local populations. This study demonstrated that combining national health survey data covering different areas is viable but not without technical difficulties. Future comparisons between European regions should be made using standardized sampling, recruitment and data collection protocols, allowing proper monitoring of health inequalities.

Multiple risk behaviour: increasing socio-economic gap over time?

BACKGROUND: Unhealthy behaviours often occur in combination. In this study the relationship between education and lifestyle, defined as a cluster of risk behaviours, has been analysed with the purpose to assess socio-economic changes in multiple risk behaviour over time. METHODS: Cross-sectional data from the Belgian Health Interview Surveys 1997, 2001 and 2004 were analysed. This study is restricted to persons aged ≥ 15 years with information on those health behaviours and education (n = 7431, n = 8142 and n = 7459, respectively). A lifestyle index was created based on the sum of the four unhealthy behaviours: smokers vs. non-smokers, risky versus non-risky alcohol use, sedentaryness vs. physically active and poor vs. healthy diet. The lifestyle index was dichotomized as low (0-2) vs. high (3-4). For the assessment of socio-economic inequalities in multiple risk behaviour, summary measures as Odds Ratio (OR) and Relative Index of Inequality (RII) were calculated using logistic regression, stratified by sex. RESULTS: Of the adult population, 7.5% combined three to four unhealthy behaviours. Lower educated men are the most at risk. Besides, the OR among men significantly increased from 1.6 in 2001 to 3.4 in 2004 (P = 0.029). The increase of the OR among women was less pronounced. The RII, on the other hand, did not show any gradient, neither for men nor for women. CONCLUSION: Multiple risk behaviour is more common among lower educated people. An increasing polarization in socio-economic inequalities is assessed from 2001 to 2004 among men. Therefore, health promotion programmes should focus on the lower socio-economic classes and target risk behaviours simultaneously.

The Belgian health examination survey: objectives, design and methods.

BACKGROUND: In 2018 the first Belgian Health Examination Survey (BELHES) took place. The target group included all Belgian residents aged 18 years and older. The BELHES was organized as a second stage of the sixth Belgian Health Interview Survey (BHIS). This paper describes the study design, recruitment method and the methodological choices that were made in the BELHES. METHODS: After a pilot period during the first quarter of the BHIS fieldwork, eligible BHIS participants were invited to participate in the BELHES until a predefined number (n = 1100) was reached. To obtain the required sample size, 4918 eligible BHIS participants had to be contacted. Data were collected at the participant's home by trained nurses. The data collection included: 1) a short set of questions through a face-to-face interview, 2) a clinical examination consisting of the measurement of height, weight, waist circumference, blood pressure and for people aged 50 years and older handgrip strength and 3) a collection of blood and urine samples. The BELHES followed as much as possible the guidelines provided in the framework of the European Health Examination Survey (EHES) initiative. Finally 1184 individuals participated in the BELHES, resulting in a participation rate of 24.1%. Results for all the core BELHES measurements were obtained for more than 90% of the participants. CONCLUSION: It is feasible to organize a health examination survey as a second stage of the BHIS. The first successfully organized BELHES provides useful information to support Belgian health decision-makers and health professionals. As the BELHES followed EHES recommendations to a large extent, the results can be compared with those from similar surveys in other EU (European Union) member states.