Scottish Longitudinal Study
Development & Support Unit
A comparative study of the relationship between deprivation and health status in Northern Ireland and Scotland
Dermot O'Reilly (Queens University, Belfast)
Gemma Catney (Queens University, Belfast)
Michael Rosato (Queens University, Belfast)
Chris Dibben (University of St Andrews)
Gillian Raab (University of St Andrews)
Frank Popham (University of St Andrews)
Paul Burton (University of Leicester)
The aim of the proposed study is two-fold (a) to investigate the relationship between deprivation and health in Northern Ireland and Scotland using the Northern Ireland Longitudinal Study (NILS) and the Scottish Longitudinal Study (SLS) and in doing so (b) to test a solution to the problem of conducting combining analysis on unlinked datasets. The specific research questions are …
- How strong is the relationship between deprivation and mortality (various causes) in Scotland and in Northern Ireland?
- Does the relationship between deprivation and mortality differ between the two countries?
- Does Scotland still have higher mortality than Northern Ireland after account for differences in deprivation and other factors?
- Does Glasgow still retain a higher relative mortality when Belfast is used as the comparator?
- Is the relationship between denomination and mortality the same in both countries?
- Can DataSHIELD offer a viable method for undertaking joint analysis of data held in different Longitudinal Studies?
High levels of deprivation and poor health have been found in both Northern Ireland and Scotland. For Scotland comparative research with England and Wales has suggested a “Scottish Effect” for mortality meaning that Scotland’s excess mortality can no longer be fully explained by its higher rate of deprivation. This research has been extended to the city level and it has been shown that a “Glasgow effect” exists in comparison to the mortality experience of similarly deprived Liverpool and Manchester. The current study aims to extend previous analyses by undertaking a comparison of the mortality experience of the Scottish and Northern Irish populations at an individual level and which simultaneously controls for variations in levels of deprivation and other potentially important factors in more detail than has been possible before.
Combined analysis of NILS and the SLS is needed to meet this aim however previous research has highlighted the difficulties of this given the important confidentiality arrangements of the datasets.1 However, recently a new method (called DataSHIELD) of conducting joint analysis without combining datasets has been demonstrated.2 This involves iterating statistical models in stages by passing non-disclosive model summary statistics between the studies. This procedure can be achieved within the existing confidentiality and security frameworks of the studies with model iterations being checked for disclosure before release. Importantly it does not require any linking of the datasets or abstracts from the datasets or any sharing of data between the studies. This datashield approach has been approved by NISRA for use in the NILS.
- Young H, Grundy E, O’Reilly D, Boyle P. Self-rated health and mortality in the UK: results from the first comparative analysis of the England and Wales, Scotland, and Northern Ireland Longitudinal Studies. Population Trends 2010;(139):11-36.
- Wolfson M, Wallace S, Masca N, Rowe G, Sheehan N, Ferretti V, LaFlamme P, Tobin M, Macleod J, Little J, Fortier I, Knoppers B, Burton P. DataSHIELD: resolving a conflict in contemporary bioscience—performing a pooled analysis of individual-level data without sharing the data. Int. J. Epidemiol. 2010; 39: 1372-1382.