Presentation on theme: "Linking lives through time www.lscs.ac.uk SCOTTISH LONGITUDINAL STUDY (SLS)"— Presentation transcript:
linking lives through time www.lscs.ac.uk SCOTTISH LONGITUDINAL STUDY (SLS)
linking lives through time www.lscs.ac.uk Why do we need the SLS? overall mortality rates are higher than the rest of Britain fertility rates are lower than the rest of Britain population ageing is a significant problem for the future in Scotland specific causes of death such as lung cancer and heart disease rank among the worst in Europe rates of household deprivation are higher than in the rest of Britain These issues, and the inter-relationships between them, are examples of what can be explored using the SLS
linking lives through time www.lscs.ac.uk Muller-Nordhorn, J. et al. Eur Heart J 2008 0:ehm604v1-11; doi:10.1093/eurheartj/ehm604 Age-standardized mortality from ischaemic heart disease in European regions (men; age group 45-74 years; year 2000)
linking lives through time www.lscs.ac.uk What is the SLS? SLS is a replica of the LS (Longitudinal Study of England and Wales), which started in 1971 The SLS is a large-scale, anonymised linkage study designed to capture 5.5% of the Scottish population Sample based on 20 semi-random birthdates It actually contains information on a 5.3% sample It is built using data available from… –Census data (initial sample from 1991, then 2001) –Vital Events data (births, deaths, marriages etc…) –National Health Service Central Register (NHSCR) data (migration into or out of Scotland) –NHS data (cancer registrations, hospital episodes and maternity data) –Education data
linking lives through time www.lscs.ac.uk Study structure SLS database 1991 original sample: 270,385 selected from 1991 Census plus household members Entries since 1991: Births 53k, Immigrants 31k Exits of sample members from SLS: deaths 48k, emigrants 20k Events to sample members since their entry to SLS: Births, widow(er)hoods, marriages, stillbirths, infant mortality (139k in total) 2001 Census sample 265,104 sample members found at 2001 plus household members Coming soon: School Census records, inc attendance/exclusions, attainment, qualifications ISD health data Project specific extracts of: Hospital admissions Cancer registrations Maternity data
SLS is one of three UK LS LS England and Wales (LS) 1% sampe Northern Ireland LS (NILS) 28% sample Northern Ireland mortality study (NIMS) 100%
linking lives through time www.lscs.ac.uk SLS project examples
linking lives through time www.lscs.ac.uk SLS project - Example 1 Assessing socio-economic inequalities in mortality at the Scottish national level, incorporating a comparison between mortality in Scotland and England Frank Popham – University of St Andrews Paul Boyle - University of St Andrews SLS project no. 2009_004 –www.lscs.ac.uk/sls/projects/2009_004.htmwww.lscs.ac.uk/sls/projects/2009_004.htm
linking lives through time www.lscs.ac.uk Is there a “Scottish effect” for mortality? Scottish effect = excess Scottish mortality over England after deprivation differences Found in unlinked ecological data but what about linked data joining SLS and ONS LS? Popham, F and Boyle, P: Journal of Public Health doi:10.1093/pubmed/fdr023
linking lives through time www.lscs.ac.uk Source: Mark Hillary What about those born in Scotland now living in England?
linking lives through time www.lscs.ac.uk Source: Scottish Longitudinal Study and ONS Longitudinal Study Impact on relative mortality rate 2001-2007 of adjusting for socio-economic differences
linking lives through time www.lscs.ac.uk Is there a “Scottish effect” for mortality? Scottish effect = excess Scottish mortality over England after deprivation differences Found in unlinked ecological data but what about linked data joining SLS and ONS LS? Additionally account for country of birth as Scots born have high mortality generally Confirms there appears to be a Scottish effect associated with born Scottish However, unmeasured confounding still possible explanation Socio-economic circumstances in childhood important? Popham, F and Boyle, P: Journal of Public Health doi:10.1093/pubmed/fdr023
linking lives through time www.lscs.ac.uk SLS project - Example 2 Does being widowed increase the risk of death? Paul Boyle - University of St Andrews Zhiqiang Feng - University of St Andrews Gillian Raab - University of St Andrews SLS project no. 2008_006 –www.lscs.ac.uk/sls/projects/2008_006.htmwww.lscs.ac.uk/sls/projects/2008_006.htm
linking lives through time www.lscs.ac.uk Research Overview Background Does widowhood increases mortality risk? Although commonly observed, this ‘widowhood effect’ could be due to selection effects as married couples share various characteristics related to the risk of death. The widowhood effect by different causes of spousal death; some are correlated with these shared characteristics, while others are not. Methods Using data from the SLS outcomes for men and women by different causes of death of their spouse, controlling for a range of individual- and household- level characteristics were compared.
linking lives through time www.lscs.ac.uk Effect of widowhood on the mortality hazard ratio in a Cox model, compared with the effect of age at the 1991 census, with age 50 as the baseline, adjusted for socio- economic, LLTI and ethnicity.
linking lives through time www.lscs.ac.uk Results The widowhood effect is greater than has been found from other recent studies, especially for older women. The risk is highest shortly after widowhood, but remains significant for over ten years. These broad results hold regardless of the cause of death of the spouse. Conclusions The analysis of the widowhood effect used three methods of classifying the spouse’s cause of death in an attempt to control for potential selection effects. The results are highly consistent and suggest that this is a causal effect, rather than a result of selection. Boyle, P, Feng, Z. & Raab, G, (2011) Does widowhood increase mortality risk? Comparing different causes of spousal death to test for selection effects. Epidemiology, 22 1-5
linking lives through time www.lscs.ac.uk SLS Future developments
linking lives through time www.lscs.ac.uk School Education data – 3 datasets: Pilot linkage exercise carried out using 2007 & 2008 data. 2009 linkage to be carried out summer 2011. Names not available so linkage carried out on DOB, sex and postcode only - >80% linkage success Data to be provided annually, and successive linkages will increase proportion of linked pupils (since each has up to 13 chances of being linked) - School Census data - SQA attainment data - Attendance/ absence & exclusions data available
linking lives through time www.lscs.ac.uk 2011 Census – Questions added 5 new questions included: national identity ability in spoken English languages other than English used at home long term health conditions month/year of arrival into the UK (for people not born in the UK)
linking lives through time www.lscs.ac.uk Summary: strengths of the SLS Sample size much larger than most surveys BHPS started with ~10,000 people in GB SLS has ±270 thousand traced members + ±505 thousand household members in Scotland (1991) This means that it’s possible to analyse relatively small sub populations Covers several fields of interest – demography, economic and social issues, health, housing, mortality, migration, fertility etc Can be linked with detailed geographical information Census compulsory so attrition rates low and linkage rates high Includes those in communal establishments Ability to link hospital episodes data to socio-economic characteristics Potential to link to other datasets in the future – (eg 2011 Census, HESA)
linking lives through time www.lscs.ac.uk Summary: weaknesses of the SLS Restricted range of variables Smoking Income Census information only collected every decade Not possible to return to the sample to ask extra questions The data are highly confidential Immigration and emigration data not very reliable We rely on census data - will there be a 2021 Census? Difficult dataset to analyse, both technically, and logistically!
linking lives through time www.lscs.ac.uk The SLS team – contact details Based in Edinburgh: Based in LSCS, St Andrews: Joan Nolan Database Administrator 0131 314 4598 email@example.com Zengyi Huang Programmer 0131 314 4253 firstname.lastname@example.org Claire Boag SLS Manager 0131 314 4310 Claire.Boag@gro-scotland.gsi.gov.uk Gillian Raab Statistician 0131 314 4599 email@example.com Christine Lamb Chris Dibben Administrator SLS Director SLS Support Officers Zhiqiang Feng Peteke Feijten Plus: Frank Popham & Lee Williamson Interested in using the SLS? Contact: firstname.lastname@example.org
Living circumstances Young women no births. Picture for young men looks very similar see next slide Over ages 15 to 30 gradual move away from parents to live alone or with husband or partner Others is an intermediate state
linking lives through time www.lscs.ac.uk Living circumstances at 2001 Young men no births. Picture very similar to that for young women Teen mums Teen dads
linking lives through time www.lscs.ac.uk Economic position at 2001