Studies Centre, University of St Andrews, St Andrews, UK; 3 Glasgow Centre for Population Health, Glasgow, UK 10.1136/jech-2013-203126.31 Background Scotland experiences higher mortality than the rest of Western Europe with Glasgow experiencing higher mortality than the rest of Scotland. This excess persists even when control- ling for socio-economic status and levels of deprivation. More- over, available measures of deprivation have accounted for less of the Glasgow excess over last 30 years. Many theories have been put forward to account for these differences and a recent review identified seventeen hypotheses. One hypothesis, investigated here, was an artefact of inadequate control for deprivation. Methods Previous analyses of Glasgow’s ‘excess’ within Scotland have been based on comparison with nationally derived depriva- tion groupings (e.g. quintiles, deciles). Glasgow’s unique depriva- tion profile in comparison to the rest of Scotland renders such approaches problematic. Our approach, therefore, is as follows: small area geographies (datazones) were ranked by the Scottish Index of Multiple Deprivation (SIMD) income domain. Using a case control design datazones in Glasgow (cases) were matched with the closest non Glasgow datazone ranked above and below (controls). Geocoded mortality and population data were avail- able for all deaths by age and sex for 2000-2002. Age-standar- dised all-cause mortality rates were calculated for men and women between the ages of 0–64, and for all ages. Age-standar- dised mortality rates by sex were compared for cases and matched controls for each quintile of deprivation. Results Men in the 3 most deprived quintiles in Glasgow experi- enced higher levels of mortality compared to both the nearest non-Glasgow datazones and all non-Glasgow datazones within the quintile. Excess mortality was reduced for the 2 most afflu- ent quintiles. The pattern was broadly similar for women, with the excess reduced for the 3 most affluent quintiles. The age- standardised mean mortality rate for men aged 0–64 in Glasgow in the most deprived quintile was 769 per 100,000. The corre- sponding rate for all non-Glasgow datazones was 582 per 100,000. The mortality rate for the control group was 688 per 100,000, suggesting a substantially smaller excess. Using this case-control approach, the excess is reduced by 57% in the most deprived quintile. In the second most deprived quintile it is reduced by 15%. The corresponding results for women are reductions of 57% (most deprived quintile) and 47% (second most deprived quintile). Conclusion Using national population and mortality data, the excess mortality in Glasgow in the most deprived area is halved by means of a closer control for deprivation. Explanations are still required for the remaining excess suggesting it is only partly artefactual. OP32 SYMPTOMS, DIAGNOSIS AND TREATMENT IN SOCIO- ECONOMIC INEQUALITIES OF HEALTH 1 A C Hardcastle, 1 N Steel, 1 M O Bachmann, 2 D Melzer. 1 Norwich Medical School, University of East Anglia, Norwich, UK; 2 Department of Epidemiology and Public Health, Peninsular Medical School, Exeter, UK 10.1136/jech-2013-203126.32 Background Ill health and relative poverty are connected. This study aimed to determine first whether the worse health experi- enced by poorer participants was matched by appropriately greater receipt of healthcare, and second whether any inequal- ities in receipt occurred at the stage of diagnosis or treatment. Methods The English Longitudinal Study of Ageing is a cohort of participants aged 50 years or older. The relative distributions by wealth of symptoms, diagnosis and treatment of five common chronic conditions (angina, diabetes, depression, osteoarthritis, and cataract) were analysed in four waves of data collected from 2002 to 2010. Symptoms were defined for angina using the Rose Angina scale, diabetes using fasting HbA1c level, depression using the Centre for Epidemiologic Studies Depression Scale, osteoarthritis as self-reported pain and disability, and cataract as self-reported poor vision. Doctors’ diagnoses for all conditions were self-reported. Treatment was defined for angina as beta- blocker prescription, osteoarthritis and cataract as surgery, and diabetes and depression as receiving treatment described in qual- ity indicators. Binomial regression models tested variations between the hypothetically poorest and richest individuals for age and sex adjusted symptoms, diagnosis and treatment across the waves, using a slope index of inequality. Results Symptoms were commoner in poorer participants in all 5 conditions at all 4 timepoints, with ORs ranging from 2.5 to 7.0. In angina, depression and diabetes, receipt of diagnosis and treat- ment was similarly higher in poorer participants, with ORs rang- ing from 1.9 to 5.6. In osteoarthritis and cataract, receipt of diagnosis and treatment did not show substantial matching varia- tions by wealth, with ORs ranging from 0.8 to 1.9. For example, ORs for diabetes in 2008 were broadly similar for symptoms (2.5 [95% CI 1.5, 4.0], diagnosis (3.8 [3.0, 4.9]) and treatment (3.1 [2.4, 4.0]). In contrast, osteoarthritis ORs were substantially larger for symptoms (6.9 [5.2, 9.1]) than for diagnosis (1.4 [1.2, 1.7]) or treatment (0.8 [0.5, 1.3]). Conclusion Poorer participants were much more likely to have symptoms of osteoarthritis and cataract, but not much more likely to receive a diagnosis. The block in equitable receipt of healthcare was at the stage of diagnosis rather than treatment, and so interventions to reduce inequalities in osteoarthritis and cataract should focus on the diagnostic process. The same rela- tive inequalities in diagnosis were not seen in angina, depression and diabetes, which have all been the target of multiple quality improvement initiatives. These patterns remained consistent over 8 years. Sexual health OP33 HOW WELL DO VOLUNTEER WEB PANEL SURVEYS MEASURE SENSITIVE BEHAVIOURS IN THE GENERAL POPULATION, AND CAN THEY BE IMPROVED? A COMPARISON WITH THE THIRD BRITISH NATIONAL SURVEY OF SEXUAL ATTITUDES & LIFESTYLES (NATSAL3) 1,2 B Erens, 2 S Burkill, 2 A Copas, 3 M Couper, 3 F Conrad. 1 Health Services Research and Policy, London School of Hygiene and Tropical Medicine, London, UK; 2 Research Department of Infection and Population Health, University College London (UCL), London, UK; 3 Survey Research Center, University of Michigan, Ann Arbor, US 10.1136/jech-2013-203126.33 Background Surveys play an important role in providing public health data for researchers and policy-makers. Traditional inter- viewer-administered surveys are subject to declining response rates and increasing costs. With the spread of the internet among the general population, web surveys potentially provide a cost- effective alternative mode. Volunteer web panels are now widely used for market research and opinion polling, but less for Abstracts J Epidemiol Community Health 2013;67(Suppl 1):A1–A84 A17