Adjustment for survey non-representativeness using record-linkage: refined estimates of alcohol consumption by deprivation in Scotland.
Ontology highlight
ABSTRACT: Background and aims
Analytical approaches to addressing survey non-participation bias typically use only demographic information to improve estimates. We applied a novel methodology which uses health information from data linkage to adjust for non-representativeness. We illustrate the method by presenting adjusted alcohol consumption estimates for Scotland.Design
Data on consenting respondents to the Scottish Health Surveys (SHeSs) 1995-2010 were linked confidentially to routinely collected hospital admission and mortality records. Synthetic observations representing non-respondents were created using general population data. Multiple imputation was performed to compute adjusted alcohol estimates given a range of assumptions about the missing data. Adjusted estimates of mean weekly consumption were additionally calibrated to per-capita alcohol sales data.Setting
Scotland.Participants
13 936 male and 18 021 female respondents to the SHeSs 1995-2010, aged 20-64 years.Measurements
Weekly alcohol consumption, non-, binge- and problem-drinking.Findings
Initial adjustment for non-response resulted in estimates of mean weekly consumption that were elevated by up to 17.8% [26.5 units (18.6-34.4)] compared with corrections based solely on socio-demographic data [22.5 (17.7-27.3)]; other drinking behaviour estimates were little changed. Under more extreme assumptions the overall difference was up to 53%, and calibrating to sales estimates resulted in up to 88% difference. Increases were especially pronounced among males in deprived areas.Conclusions
The use of routinely collected health data to reduce bias arising from survey non-response resulted in higher alcohol consumption estimates among working-age males in Scotland, with less impact for females. This new method of bias reduction can be generalized to other surveys to improve estimates of alternative harmful behaviours.
SUBMITTER: Gorman E
PROVIDER: S-EPMC5467727 | biostudies-literature |
REPOSITORIES: biostudies-literature
Similar Datasets
Project description:Reliable estimates of health-related behaviours, such as levels of alcohol consumption in the population, are required to formulate and evaluate policies. National surveys provide such data; validity depends on generalisability, but this is threatened by declining response levels. Attempts to address bias arising from non-response are typically limited to survey weights based on sociodemographic characteristics, which do not capture differential health and related behaviours within categories. This project aims to explore and address non-response bias in health surveys with a focus on alcohol consumption. The Scottish Health Surveys (SHeS) aim to provide estimates representative of the Scottish population living in private households. Survey data of consenting participants (92% of the achieved sample) have been record-linked to routine hospital admission (Scottish Morbidity Records (SMR)) and mortality (from National Records of Scotland (NRS)) data for surveys conducted in 1995, 1998, 2003, 2008, 2009 and 2010 (total adult sample size around 40 000), with maximum follow-up of 16 years. Also available are census information and SMR/NRS data for the general population. Comparisons of alcohol-related mortality and hospital admission rates in the linked SHeS-SMR/NRS with those in the general population will be made. Survey data will be augmented by quantification of differences to refine alcohol consumption estimates through the application of multiple imputation or inverse probability weighting. The resulting corrected estimates of population alcohol consumption will enable superior policy evaluation. An advanced weighting procedure will be developed for wider use. Ethics approval for SHeS has been given by the National Health Service (NHS) Multi-Centre Research Ethics Committee and use of linked data has been approved by the Privacy Advisory Committee to the Board of NHS National Services Scotland and Registrar General. Funding has been granted by the MRC. The outputs will include four or five public health and statistical methodological international journal and conference papers. Public health. Addiction: health policy; mental health.
| S-EPMC3612815 | biostudies-literature
Project description:BackgroundWe examine major causes of death amongst persons in contact with drug-treatment services across Scotland during April 1996-March 2006, hereafter Scottish Drug Misuse Database (SDMD) cohort.MethodsDrug-treatment records were linked to national registers of deaths and hepatitis C virus (HCV) diagnoses. For eras 1996/97-2000/01 and 2001/02-2005/06, we calculated cause-specific death-rates and standardised mortality ratios (SMRs) using age-, sex- and calendar-rates of the general Scottish population. Major causes of death were identified by high SMRs (>5 across eras) or rates (>50 per 100,000 person-years in either era), and their time-specific influences characterised by proportional hazards analyses.ResultsThe SDMD cohort comprised 69,456 individuals, 350,315 person-years and 2590 deaths. The overall SMR reduced from 6.4 (95% CI: 6.0-6.9) to 4.8 (95% CI: 4.6-5.0) between eras. We identified five major causes of death: drug-related (1383 deaths), homicide (118) and infectious diseases (90) with high SMRs; suicide (269) and digestive system disease (168) with high rates. HCV diagnosis marked individuals with at least double the risk of cause-specific mortality, including adjusted hazard ratio (HR) for no HCV diagnosis of 0.46 (95% CI: 0.41-0.53) for drug-related deaths (DRDs) and 0.15 (95% CI: 0.10-0.22) for death from digestive system disease. Increased DRD risk at older age (>34 years) appeared specific to HCV-diagnosed individuals (interaction: χ₁²=7.7, p=0.01). Alcohol misuse increased HRs: for DRD (1.76, 95% CI: 1.50-2.06), suicide (1.88, 95% CI: 1.35-2.60), deaths from digestive system disease (3.19, 95% CI: 2.21-4.60) and non-major causes (1.87, 95% CI: 1.49-2.35). Stimulant misuse increased suicide risk: adjusted HR 1.91 (95% CI: 1.43-2.54).ConclusionsDrug-users in Scotland are exposed to variously increased mortality risks. HCV-diagnosed individuals are particularly vulnerable, and may need additional support.
| S-EPMC3271367 | biostudies-literature
Project description:BackgroundA cancer diagnosis during childhood greatly disrupts the lives of those affected, causing physical and psychological challenges. We aim to investigate educational outcomes among schoolchildren with a previous cancer diagnosis compared to their peers.MethodsIndividual records from four national education databases and three national health databases were linked to construct a cohort of all singleton schoolchildren born in Scotland attending Scottish local-authority schools between 2009-2013. Pupils previously diagnosed with any cancer, haematological cancers, and central nervous system (CNS) cancers, were compared to their unaffected peers with respect to five educational outcomes: special educational need (SEN), absenteeism, school exclusion, academic attainment, and unemployment. Analyses were adjusted for sociodemographic and maternity factors and chronic conditions.ResultsOf 766,217 pupils, 1,313 (0.17%) had a previous cancer diagnosis. Children with any cancer had increased odds of SEN (OR 3.26, 95% CI 2.86-3.71), absenteeism (IRR 1.82, 95% CI 1.70-1.94), and low attainment (OR 2.15, 95% CI 1.52-3.03) compared to their peers. Similar findings were observed for haematological (SEN OR 2.62, 95% CI 2.12-3.24; absenteeism IRR 2.04, 95% CI 1.85-2.25; low attainment OR 2.17, 95% CI 1.31-3.61) and CNS (SEN OR 6.44, 95% CI 4.91-8.46; absenteeism IRR 1.75, 95% CI 1.51-2.04; low attainment OR 3.33, 95% CI 1.52-7.30) cancers. Lower exclusions were observed among children with any cancer (IRR 0.51, 95% CI 0.31-0.83) and CNS cancer (IRR 0.20, 95% CI 0.06-0.61). No associations were observed with unemployment.ConclusionsThis study highlights the wider impacts of childhood cancer on educational outcomes. These children need to be supported, as poor educational outcomes can further impact later health.
| S-EPMC10370705 | biostudies-literature
Project description:Health surveys are an important resource for monitoring population health, but selective nonresponse may impede valid inference. This study aimed to assess nonresponse bias in a population-sampled health survey in Scotland, with a focus on alcohol-related outcomes. Nonresponse bias was assessed by examining whether rates of alcohol-related harm (i.e., hospitalization or death) and all-cause mortality among respondents to the Scottish Health Surveys (from 1995 to 2010) were equivalent to those in the general population, and whether the extent of any bias varied according to sociodemographic attributes or over time. Data from consenting respondents (aged 20-64 years) to 6 Scottish Health Surveys were confidentially linked to death and hospitalization records and compared with general population counterparts. Directly age-standardized incidence rates of alcohol-related harm and all-cause mortality were lower among Scottish Health Survey respondents compared with the general population. For all years combined, the survey-to-population rate ratios were 0.69 (95% confidence interval: 0.61, 0.76) for the incidence of alcohol-related harm and 0.89 (95% confidence interval: 0.83, 0.96) for all-cause mortality. Bias was more pronounced among persons residing in more deprived areas; limited evidence was found for regional or temporal variation. This suggests that corresponding underestimation of population rates of alcohol consumption is likely to be socially patterned.
| S-EPMC4207717 | biostudies-other
Project description:BackgroundCOVID-19 has overwhelmed health services across the world; its global death toll has exceeded 5.3 million and continues to grow. There have been almost 15 million cases of COVID-19 in the UK. The need for rapid accurate identification, appropriate clinical care and decision making, remains a priority for UK ambulance service. To support identification and conveyance decisions of patients presenting with COVID-19 symptoms the Scottish Ambulance Service introduced the revised Medical Priority Dispatch System Protocol 36, enhanced physician led decision support and prehospital clinical guidelines. This study aimed to characterise the impact of these changes on the pathways and outcomes of people attended by the SAS) with potential COVID-19.MethodsA retrospective record linkage cohort study using National Data collected from NHS Scotland over a 5 month period (April-August 2020).ResultsThe SAS responded to 214,082 emergency calls during the study time period. The positive predictive value of the Protocol 36 to identify potentially COVID-19 positive patients was low (17%). Approximately 60% of those identified by Protocol 36 as potentially COVID-19 positive were conveyed. The relationship between conveyance and mortality differed between Protocol 36 Covid-19 positive calls and those that were not. In those identified by Protocol 36 as Covid-19 negative, 30 day mortality was higher in those not conveyed (not conveyed 9.2%; conveyed 6.6%) but in the Protocol 36 Covid-19 positive calls, mortality was higher in those conveyed (not conveyed 4.3% conveyed 8.8%). Thirty-day mortality rates of those with COVID-19 diagnosed through virology was between 28.8 and 30.2%.ConclusionThe low positive predictive value (17%) of Protocol 36 in identifying potential COVID-19 in patients emphasises the importance of ambulance clinicians approaching each call as involving COVID-19, reinforcing the importance of adhering to existing policy and continued use of PPE at all calls. The non-conveyance rate of people that were categorised as COVID-19 negative was higher than in the preceding year in the same service. The reasons for the higher rates of non-conveyance and the relationship between non conveyance rates and death at 3 and 30 days post index call are unknown and would benefit from further study.
| S-EPMC8795941 | biostudies-literature
Project description:BackgroundLooked after children are defined as children who are in the care of their local authority. Previous studies have reported that looked after children have poorer mental and physical health, increased behavioural problems, and increased self-harm and mortality compared to peers. They also experience poorer educational outcomes, yet population-wide research into the latter is lacking, particularly in the United Kingdom. Education and health share a bidirectional relationship; therefore, it is important to dually investigate both outcomes. Our study aimed to compare educational and health outcomes for looked after children with peers, adjusting for sociodemographic, maternity, and comorbidity confounders.Methods and findingsLinkage of 9 Scotland-wide databases, covering dispensed prescriptions, hospital admissions, maternity records, death certificates, annual pupil census, examinations, school absences/exclusions, unemployment, and looked after children provided retrospective data on 715,111 children attending Scottish schools between 2009 and 2012 (13,898 [1.9%] looked after). Compared to peers, 13,898 (1.9%) looked after children were more likely to be absent (adjusted incidence rate ratio [AIRR] 1.27, 95% confidence interval [CI] 1.24 to 1.30) and excluded (AIRR 4.09, 95% CI 3.86 to 4.33) from school, have special educational need (SEN; adjusted odds ratio [AOR] 3.48, 95% CI 3.35 to 3.62) and neurodevelopmental multimorbidity (AOR 2.45, 95% CI 2.34 to 2.57), achieve the lowest level of academic attainment (AOR 5.92, 95% CI 5.17 to 6.78), and be unemployed after leaving school (AOR 2.12, 95% CI 1.96 to 2.29). They were more likely to require treatment for epilepsy (AOR 1.50, 95% CI 1.27 to 1.78), attention deficit hyperactivity disorder (ADHD; AOR 3.01, 95% CI 2.76 to 3.27), and depression (AOR 1.90, 95% CI 1.62 to 2.22), be hospitalised overall (adjusted hazard ratio [AHR] 1.23, 95% CI 1.19 to 1.28) for injury (AHR 1.80, 95% CI 1.69 to 1.91) and self-harm (AHR 5.19, 95% CI 4.66 to 5.78), and die prematurely (AHR 3.21, 95% CI 2.16 to 4.77). Compared to children looked after at home, children looked after away from home had less absenteeism (AIRR 0.35, 95% CI 0.33 to 0.36), less exclusion (AIRR 0.63, 95% CI 0.56 to 0.71), less unemployment (AOR 0.53, 95% CI 0.46 to 0.62), and better attainment (AIRR 0.31, 95% CI 0.23 to 0.40). Therefore, among those in care, being cared for away from home appeared to be a protective factor resulting in better educational outcomes. The main limitations of this study were lack of data on local authority care preschool or before 2009, total time spent in care, and age of first contact with social care.ConclusionsLooked after children had poorer health and educational outcomes than peers independent of increased neurodevelopmental conditions and SEN. Further work is required to understand whether poorer outcomes relate to reasons for entering care, including maltreatment and adverse childhood events, neurodevelopmental vulnerabilities, or characteristics of the care system.
| S-EPMC8589203 | biostudies-literature
Project description:Studies involving the use of probabilistic record linkage are becoming increasingly common. However, the methods underpinning probabilistic record linkage are not widely taught or understood, and therefore these studies can appear to be a 'black box' research tool. In this article, we aim to describe the process of probabilistic record linkage through a simple exemplar. We first introduce the concept of deterministic linkage and contrast this with probabilistic linkage. We illustrate each step of the process using a simple exemplar and describe the data structure required to perform a probabilistic linkage. We describe the process of calculating and interpreting matched weights and how to convert matched weights into posterior probabilities of a match using Bayes theorem. We conclude this article with a brief discussion of some of the computational demands of record linkage, how you might assess the quality of your linkage algorithm, and how epidemiologists can maximize the value of their record-linked research using robust record linkage methods.
| S-EPMC5005943 | biostudies-literature
Project description:BackgroundThe goals of our study are to determine the most appropriate model for alcohol consumption as an exposure for burden of disease, to analyze the effect of the chosen alcohol consumption distribution on the estimation of the alcohol Population- Attributable Fractions (PAFs), and to characterize the chosen alcohol consumption distribution by exploring if there is a global relationship within the distribution.MethodsTo identify the best model, the Log-Normal, Gamma, and Weibull prevalence distributions were examined using data from 41 surveys from Gender, Alcohol and Culture: An International Study (GENACIS) and from the European Comparative Alcohol Study. To assess the effect of these distributions on the estimated alcohol PAFs, we calculated the alcohol PAF for diabetes, breast cancer, and pancreatitis using the three above-named distributions and using the more traditional approach based on categories. The relationship between the mean and the standard deviation from the Gamma distribution was estimated using data from 851 datasets for 66 countries from GENACIS and from the STEPwise approach to Surveillance from the World Health Organization.ResultsThe Log-Normal distribution provided a poor fit for the survey data, with Gamma and Weibull distributions providing better fits. Additionally, our analyses showed that there were no marked differences for the alcohol PAF estimates based on the Gamma or Weibull distributions compared to PAFs based on categorical alcohol consumption estimates. The standard deviation of the alcohol distribution was highly dependent on the mean, with a unit increase in alcohol consumption associated with a unit increase in the mean of 1.258 (95% CI: 1.223 to 1.293) (R2 = 0.9207) for women and 1.171 (95% CI: 1.144 to 1.197) (R2 = 0. 9474) for men.ConclusionsAlthough the Gamma distribution and the Weibull distribution provided similar results, the Gamma distribution is recommended to model alcohol consumption from population surveys due to its fit, flexibility, and the ease with which it can be modified. The results showed that a large degree of variance of the standard deviation of the alcohol consumption Gamma distribution was explained by the mean alcohol consumption, allowing for alcohol consumption to be modeled through a Gamma distribution using only average consumption.
| S-EPMC3352241 | biostudies-literature
Project description:Abstract:Heavy consumption of alcohol has shown to be associated with sleep disturbances among adult and elderly people in high income settings. So far, the relationship between alcohol drinking and sleeping pattern has not been studied in an African setting. Therefore, in this study we investigated whether alcohol consumption has any influence on sleeping hours among adult men and women in Ghana. Methods:Data for this survey were extracted from Ghana Demographic and Health Survey (GDHS 2008). GDHS is the only cross-sectional survey conducted on men and women aged above 15 years that collected information on variables such as sleeping hours and alcohol consumption. The analysis was controlled for various demographic, socioeconomic, household level factors, and smoking. Results:Prevalence of sleeping 1-3 h, 4-6 h, and >7 h was respectively 1.5% (1.2-2.0), 14.1% (12-16.5), and 84.4% (82.1-86.4), and that of alcohol use was 26.9% (23.4-30.6). In the multivariable regression analysis, compared with non-drinkers, those reported drinking had significantly lower odds of sleeping for at least 7 h. In the adjusted model, drinkers had 0.8 times (adjusted OR = 0.803, (95% CI [0.690-0.935])) lower odds of sleeping for at least 7 h. The odds for sleeping 4-6 h were not statistically significant. In the stratified analysis, the odds of sleeping for at least seven were comparatively lower among women (adjusted OR = 0.657, (95% CI [0.509-0.849]) then among men (adjusted OR = 0.867, (95% CI [0.740-0.965]). Conclusion:Men and women who reported consuming alcohol had significantly lower odds of getting adequate sleep (>7 h). The sleep-disrupting effect of alcohol appeared to be more prominent among women than among men. Currently there is not sufficient evidence on alcohol consumption and sleep disorder among Ghanaian population or any other country in the region. Further studies are required to understand sleeping patterns and the burden of alcohol drinking in this population to design intervention programs.
| S-EPMC6183558 | biostudies-other
Project description:Background and aimsWithin the context of Scotland's policy change to implement a minimum unit price (MUP) per unit of alcohol sold in licensed premises, this study used an N-of-1 design to assess between-person differences in the psychological and social factors associated with daily alcohol consumption.Design and settingA mixed-methods approach combined N-of-1 observational studies, comprising daily surveys followed by qualitative social network interviews (not reported here). Peer researchers with lived experience of substance use were involved in the study design and fieldwork was conducted in towns and rural areas in the East of Scotland.Participants/casesTwenty-five adults with current or recent history of alcohol dependence recruited for three 12-week waves: 11 in wave 1 (pre-MUP), 11 in wave 2 (pre- and post-MUP) and three in wave 3 (post MUP).MeasurementsGender, age, alcohol and other drug use history. Daily surveys for 12 weeks captured information about factors in the last 24 hours, e.g. amount and type of alcohol consumed, stress, social contact.FindingsEach participant was in the daily survey for a mean of 64 days [standard deviation (SD) = 42; median = 59], with a response rate of 48%; 15 participants provided sufficient data for analysis. Factors related to daily alcohol consumption differed between individuals. Models suggested that some individuals with high initial consumption reduced drinking after MUP, but explanatory factors differed, e.g. changing motivation was important for some, while alcohol availability was important for others.ConclusionsAdapting N-of-1 methods for an observational study uncovered differences in alcohol consumption change before and after minimum unit pricing implementation in Scotland, evidence of individual differences in the factors relating to alcohol consumption patterns and some evidence that post-MUP consumption changes may be related to changing psychosocial factors.
| S-EPMC8248017 | biostudies-literature