Presentation is loading. Please wait.

Presentation is loading. Please wait.

Comparing Results from the England and Wales, Scotland and Northern Ireland Longitudinal Studies: Health and Mortality as a case study Census Microdata.

Similar presentations


Presentation on theme: "Comparing Results from the England and Wales, Scotland and Northern Ireland Longitudinal Studies: Health and Mortality as a case study Census Microdata."— Presentation transcript:

1 Comparing Results from the England and Wales, Scotland and Northern Ireland Longitudinal Studies: Health and Mortality as a case study Census Microdata Conference 1 st September 2008 Harriet Young, Emily Grundy London School of Hygiene and Tropical Medicine Paul Boyle University of St Andrews Dermot OReilly Queens University Belfast

2 Background Mortality patterns differ between constituent countries of the UK – Scotland has highest mortality Mortality patterns differ between constituent countries of the UK – Scotland has highest mortality Standardised mortality rates by Local Authority across the UK Std. Rate per 100,000 population From Rosato and OReilly 2004

3 Self-reported health patterns also vary between countries – high levels of poor self-reported health in Northern Ireland and Wales Self-reported health patterns also vary between countries – high levels of poor self-reported health in Northern Ireland and Wales Std. Rate per 100,000 population From Rosato and OReilly 2004 Standardised rates of poor general health by Local Authority across the UK

4 Research Aims 1. Analyse socio-economic and socio- demographic variation in reporting poor health in 2001 in England and Wales, Scotland and Northern Ireland 1. Analyse socio-economic and socio- demographic variation in reporting poor health in 2001 in England and Wales, Scotland and Northern Ireland 2. Analyse associations between reporting of health status in 2001 and subsequent mortality (taking account of socio-economic and socio-demographic factors) 2. Analyse associations between reporting of health status in 2001 and subsequent mortality (taking account of socio-economic and socio-demographic factors)

5 Datasets ONS Longitudinal Study – England and Wales (ONS LS) Scottish Longitudinal Study (SLS) Northern Ireland Longitudinal Study (NILS) Record linkage studies Record linkage studies Samples drawn from census data, based on a number of birthdays in the year Samples drawn from census data, based on a number of birthdays in the year After initial census starting point, subsequent census points linked in After initial census starting point, subsequent census points linked in Vital events data linked in (birth to sample mother, death of spouse, death & other) Vital events data linked in (birth to sample mother, death of spouse, death & other) Samples maintained by addition of new births and immigrants Samples maintained by addition of new births and immigrants

6 Datasets Census data from % of population in sample # of birthdays selected Sample size in 2001 ONS LS Approx 500,000

7 Datasets Census data from % of population in sample # of birthdays selected Sample size in 2001 ONS LS Approx 500,000 SLS 1991, Approx 274,000

8 Datasets Census data from % of population in sample # of birthdays selected Sample size in 2001 ONS LS Approx 500,000 SLS 1991, Approx 274,000 NIILS Approx 500,000

9 Example timeline for an individual in the ONS LS

10 Developmental Aim To demonstrate how parallel and if possible combined analyses of the 3 datasets can give UK wide results To demonstrate how parallel and if possible combined analyses of the 3 datasets can give UK wide results Consider issues that could undermine cross- country analysis: address and document Consider issues that could undermine cross- country analysis: address and document Negotiate procedures facilitating UK wide analysis and a user guide for others Negotiate procedures facilitating UK wide analysis and a user guide for others

11 Confidentiality Access to individual level data only in relevant Statistical Office safe setting. Access to individual level data only in relevant Statistical Office safe setting. Outputs from safe setting governed by disclosure guidelines. Outputs from safe setting governed by disclosure guidelines. Variations in data access between the studies reflect Variations in data access between the studies reflect differences in sampling fractions differences in sampling fractions legal differences. legal differences.

12 Analysis strategy 1. Parallel analysis using individual level data in relevant safe setting 1. Parallel analysis using individual level data in relevant safe setting 2. Parallel analysis using aggregated datasets (cleared to researcher if possible) 2. Parallel analysis using aggregated datasets (cleared to researcher if possible) 3. Combined analysis 3. Combined analysis Combined individual level data not possible Combined individual level data not possible combined aggregated data combined aggregated data

13 Sample population Population aged 35 and over Population aged 35 and over Present at 2001 Census Present at 2001 Census Follow up to end 2005 Follow up to end 2005 Analyses separate by gender Analyses separate by gender Analyses separate by broad age group: 35-54, 55-74, 75+ Analyses separate by broad age group: 35-54, 55-74, 75+ This presentation: individual country analysis for England and Wales, and Northern Ireland This presentation: individual country analysis for England and Wales, and Northern Ireland

14 Outcome variables Self-rated health Self-rated health Over the last twelve months would you say your health has on the whole been: Over the last twelve months would you say your health has on the whole been: Good ? Fairly good ? Not good ? Good ? Fairly good ? Not good ? Limiting long term illness Limiting long term illness Do you have any long term illness, health problem or disability which limits your daily activities or the work you can do ? (include problems which are due to old age) Do you have any long term illness, health problem or disability which limits your daily activities or the work you can do ? (include problems which are due to old age) Yes, No Yes, No Mortality Mortality 2001 Census to end Census to end 2005

15 Explanatory Variables Self rated health and limiting long term illness (for Aim #2) Self rated health and limiting long term illness (for Aim #2) Age, gender, marital status Age, gender, marital status Education, NS-SEC Education, NS-SEC Housing tenure, car access Housing tenure, car access Region (for ONS LS only) Region (for ONS LS only)

16 1. Variation in health reporting

17 Odds ratios from logistic regression showing age and gender differences in reporting a limiting long term illness by country, ONS LS and NILS, * p<0.05 ** p<0.01 *** p<0.001 Models controlled for age group and gender only EnglandWales Northern Ireland OR P value OR OR Gender (reference: men) Women1.03*** *** Age group (reference: 35-54) ***3.59***3.21*** ***8.79***8.68***

18 Odds ratios from logistic regression showing the association of socio-economic factors and limiting long term illness by country for men aged ONS LS and NILS, England & Wales Northern Ireland Odds Ratio P value Odds Ratio P value Housing tenure (reference: owner occupier) Private rent 1.51***2.03*** Social rent 2.48***2.40*** Education (reference: upper secondary or degree level) Lower secry 1.06*1.36*** None1.48***1.99*** Car access (reference : has car) No car 1.90***1.40*** * p<0.05 ** p<0.01 *** p<0.001 Models controlled for age, marital status, tenure, car access, education, social class, region for England and Wales

19 2. Association of morbidity and mortality Crude models controlled for age Adjusted models controlled for age, marital status, housing tenure, car access, education, social class, and region for England and Wales

20 Crude models controlled for age Adjusted models controlled for age, marital status, housing tenure, car access, education, social class, and region for England and Wales

21

22 Discussion Results to date indicate: Are differences between England, Wales and Northern Ireland in gender and age effects on reporting poor health. Are differences between England, Wales and Northern Ireland in gender and age effects on reporting poor health. In all areas, reporting a limiting long term illness and reporting poor self-rated health or an illness are associated with mortality. In all areas, reporting a limiting long term illness and reporting poor self-rated health or an illness are associated with mortality.

23 Limitations: dataset comparability Variable differences e.g. education Variable differences e.g. education Differences in question order in Census form between Scotland and other countries Differences in question order in Census form between Scotland and other countries Differences in methods of event data collection e.g. on emigrations: NILS, ONS LS Differences in methods of event data collection e.g. on emigrations: NILS, ONS LS Small sample size for Wales Small sample size for Wales

24 Future activities Analysis of data from the SLS Analysis of data from the SLS Creation of aggregated (weighted count) datasets for each country Creation of aggregated (weighted count) datasets for each country Appending aggregated datasets and carrying out combined country analyses Appending aggregated datasets and carrying out combined country analyses Development of tools to aid others to carry out analyses on these datasets in future Development of tools to aid others to carry out analyses on these datasets in future

25 ONS Longitudinal Study ONS Longitudinal Study Northern Ireland Longitudinal Study Northern Ireland Longitudinal Study Scottish Longitudinal Study Scottish Longitudinal Study For more information:


Download ppt "Comparing Results from the England and Wales, Scotland and Northern Ireland Longitudinal Studies: Health and Mortality as a case study Census Microdata."

Similar presentations


Ads by Google