Assessing the representativeness of population-sampled health surveys through linkage to administrative data on alcohol-related outcomes

Trends and socioeconomic inequalities in the dental attendance of adult smokers in Scotland from 2009 to 2019, a repeated cross-sectional study

BACKGROUND

Smoking continues to be the single largest cause of preventable disease and death and a major contributor to health inequalities. Dental professionals are well placed to offer behavioural support in combination with pharmacotherapy to increase smoking cessation rates across the population. We aimed to assess the trends and socioeconomic inequalities in the dental attendance of adult smokers in Scotland from 2009 to 2019 and examine the potential population reach of dental settings for smoking cessation interventions.

METHODS

A secondary analysis was conducted of combined Scottish Health Surveys (SHeS) from 2009/11, 2013/15 and 2017/19. 'Recent' dental attendance (within the past two years) was the focus and descriptive analysis examined attendance of self-reported smokers compared to non-smokers and stratified by the area-based Scottish Index of Multiple Deprivation (SIMD) and individual socioeconomic measures (income, education, and occupation). Generalised linear models were used to model recent attendance in non-smokers relative to smokers adjusted by the socioeconomic measures, for each of the survey cohorts separately. Absolute differences and risk ratios were calculated with 95% Confidence Intervals (CI).

RESULTS

Recent dental attendance was generally high and increased in both smokers (70-76%) and non-smokers (84-87%) from 2009/11 to 2017/19 and increased across all SIMD groups. After adjustment for sociodemographic variables, the adjusted Risk Difference (aRD) for recent attendance between non-smokers and smokers was 8.9% (95% CI 4.6%, 13.2%) by 2017/19. Within smokers, recent attendance was 7-9% lower in those living in the most deprived areas compared to those living in the least deprived areas over the three surveys.

CONCLUSIONS

SHeS data from 2009 to 2019 demonstrated that a high and increasing proportion of smokers in the population attend the dentist, albeit slightly less frequently than non-smokers. There were large inequalities in the dental attendance of smokers, to a lesser extent in non-smokers, and these persisted over time. Dental settings provide a good potential opportunity to deliver population-level smoking cessation interventions, but smokers in the most deprived groups and older age groups may be harder to reach. Consideration should be given to ensure that these groups are given appropriate proportionate support to take up preventive interventions.

The impact of assortative mating, participation bias and socioeconomic status on the polygenic risk of behavioural and psychiatric traits

To investigate assortative mating (AM), participation bias and socioeconomic status (SES) with respect to the genetics of behavioural and psychiatric traits, we estimated AM signatures using gametic phase disequilibrium and within-spouses and within-siblings polygenic risk score correlation analyses, also performing a SES conditional analysis. The cross-method meta-analysis identified AM genetic signatures for multiple alcohol-related phenotypes, bipolar disorder, major depressive disorder, schizophrenia and Tourette syndrome. Here, after SES conditioning, we observed changes in the AM genetic signatures for maximum habitual alcohol intake, frequency of drinking alcohol and Tourette syndrome. We also observed significant gametic phase disequilibrium differences between UK Biobank mental health questionnaire responders versus non-responders for major depressive disorder and alcohol use disorder. These results highlight the impact of AM, participation bias and SES on the polygenic risk of behavioural and psychiatric traits, particularly in alcohol-related traits.

Recoverability and estimation of causal effects under typical multivariable missingness mechanisms

In the context of missing data, the identifiability or "recoverability" of the average causal effect (ACE) depends not only on the usual causal assumptions but also on missingness assumptions that can be depicted by adding variable-specific missingness indicators to causal diagrams, creating missingness directed acyclic graphs (m-DAGs). Previous research described canonical m-DAGs, representing typical multivariable missingness mechanisms in epidemiological studies, and examined mathematically the recoverability of the ACE in each case. However, this work assumed no effect modification and did not investigate methods for estimation across such scenarios. Here, we extend this research by determining the recoverability of the ACE in settings with effect modification and conducting a simulation study to evaluate the performance of widely used missing data methods when estimating the ACE using correctly specified g-computation. Methods assessed were complete case analysis (CCA) and various implementations of multiple imputation (MI) with varying degrees of compatibility with the outcome model used in g-computation. Simulations were based on an example from the Victorian Adolescent Health Cohort Study (VAHCS), where interest was in estimating the ACE of adolescent cannabis use on mental health in young adulthood. We found that the ACE is recoverable when no incomplete variable (exposure, outcome, or confounder) causes its own missingness, and nonrecoverable otherwise, in simplified versions of 10 canonical m-DAGs that excluded unmeasured common causes of missingness indicators. Despite this nonrecoverability, simulations showed that MI approaches that are compatible with the outcome model in g-computation may enable approximately unbiased estimation across all canonical m-DAGs considered, except when the outcome causes its own missingness or causes the missingness of a variable that causes its own missingness. In the latter settings, researchers may need to consider sensitivity analysis methods incorporating external information (e.g., delta-adjustment methods). The VAHCS case study illustrates the practical implications of these findings.

Trends in inequalities in childhood overweight and obesity prevalence: a repeat cross-sectional analysis of the Health Survey for England

OBJECTIVE

To examine trends in socio-economic and ethnic inequalities in childhood overweight and obesity in the England between 1995 and 2019 in survey data and to compare these to administrative data.

DESIGN

Observational repeated cross-sectional study using the Health Survey for England (HSE) and National Child Measurement Programme (NCMP).

OUTCOME

Age and sex standardised overweight, obesity and overweight including obesity.

ANALYSIS

Inequalities assessed by parental education, family structure, ethnicity (binary non-white vs white) and area-level Index of Multiple Deprivation. Estimates stratified by age and sex. Trends compared against NCMP data (age 4-5 and 10-11 years).

RESULTS

Prevalence of childhood overweight including obesity increased from 26.0% in 1995 to 31.7% in 2019, with the highest and fastest growing levels in those aged 11-15 years, rising from 29.7% to 38.0%. Despite a plateau in overall childhood obesity since 2004, differences between groups demonstrated widening inequalities over time. Inequalities widened by area-level deprivation, household educational attainment, household structure and ethnicity driven primarily by increased prevalence among socioeconomically disadvantaged children. For example, the gap between children from households with no qualifications versus degree-level qualifications increased from -1.1% to 13.2%, and the gap between single-parent households and couple households increased from 0.5% to 5.3%. HSE trends in prevalence of childhood overweight and obesity by deprivation quintile were consistent with those in NCMP.

CONCLUSION

Overall levels of child overweight and obesity increased between 1995 and 2004. Since then, increases in prevalence among less advantaged groups have driven widening of inequalities.

Linking health survey data with health insurance data: methodology, challenges, opportunities and recommendations for public health research. An experience from the HISlink project in Belgium

In recent years, the linkage of survey data to health administrative data has increased. This offers new opportunities for research into the use of health services and public health. Building on the HISlink use case, the linkage of Belgian Health Interview Survey (BHIS) data and Belgian Compulsory Health Insurance (BCHI) data, this paper provides an overview of the practical implementation of linking data, the outcomes in terms of a linked dataset and of the studies conducted as well as the lessons learned and recommendations for future links.Individual BHIS 2013 and 2018 data was linked to BCHI data using the national register number. The overall linkage rate was 92.3% and 94.2% for HISlink 2013 and HISlink 2018, respectively. Linked BHIS-BCHI data were used in validation studies (e.g. self-reported breast cancer screening; chronic diseases, polypharmacy), in policy-driven research (e.g., mediation effect of health literacy in the relationship between socioeconomic status and health related outcomes, and in longitudinal study (e.g. identifying predictors of nursing home admission among older BHIS participants). The linkage of both data sources combines their strengths but does not overcome all weaknesses.The availability of a national register number was an asset for HISlink. Policy-makers and researchers must take initiatives to find a better balance between the right to privacy of respondents and society's right to evidence-based information to improve health. Researchers should be aware that the procedures necessary to implement a link may have an impact on the timeliness of their research. Although some aspects of HISlink are specific to the Belgian context, we believe that some lessons learned are useful in an international context, especially for other European Union member states that collect similar data.

Gender-based harassment in Swedish workplaces and alcohol-related morbidity and mortality: A prospective cohort study

OBJECTIVE

The study investigated experiences of different types of work-related gender-based harassment (GBH), specifically sexual and gender harassment, as risk factors for alcohol-related morbidity and mortality (ARMM).

METHODS

Information about experiences of (i) sexual harassment (SH-I) and (ii) gender harassment (GH-I) from inside the organization and (iii) sexual harassment from a person external to the organization (SH-E) were obtained from the Swedish Work Environment Survey 1995-2013, a biannual cross-sectional survey, administered to a representative sample of the Swedish working population. The survey responses from 86 033 individuals were connected to multiple registers containing information about alcohol-related diagnoses, treatment, or cause of death. Cox proportional hazard models were fitted to assess hazard ratios (HR) of incident ARMM during a mean follow-up of eight (SH-I and GH-I) and ten (SH-E) years.

RESULTS

A higher prospective risk estimate of ARMM was found among participants who reported experiences of SH-E [HR 2.01, 95% confidence interval (CI) 1.61-2.52], GH-I (HR 1.33, CI 1.03-1.70), or SH-I (HR 2.37, CI 1.42-3.00). Additional analyses, distinguishing one-time from reoccurring harassment experiences, indicated a dose-response relationship for all three harassment types. Gender did not modify the associations. Under the assumption of causality, 9.3% (95% CI 5.4-13.1) of the risk of ARMM among Swedish women and 2.1% (95% CI 0.6-3.6) among Swedish men would be attributable to any of the three types of GBH included in this study.

CONCLUSIONS

Experiences of GBH in the work context may be a highly relevant factor in the etiology of ARMM.

Adjustment for survey non-participation using record linkage and multiple imputation: A validity assessment exercise using the Health 2000 survey

AIMS

It is becoming increasingly possible to obtain additional information about health survey participants, though not usually non-participants, via record linkage. We aimed to assess the validity of an assumption underpinning a method developed to mitigate non-participation bias. We use a survey in Finland where it is possible to link both participants and non-participants to administrative registers. Survey-derived alcohol consumption is used as the exemplar outcome.

METHODS

Data on participants (85.5%) and true non-participants of the Finnish Health 2000 survey (invited survey sample N=7167 aged 30-79 years) and a contemporaneous register-based population sample (N=496,079) were individually linked to alcohol-related hospitalisation and death records. Applying the methodology to create synthetic observations on non-participants, we created 'inferred samples' (participants and inferred non-participants). Relative differences (RDs) between the inferred sample and the invited survey sample were estimated overall and by education. Five per cent limits were used to define acceptable RDs.

RESULTS

Average weekly consumption estimates for men were 129 g and 131 g of alcohol in inferred and invited survey samples, respectively (RD -1.6%; 95% confidence interval (CI) -2.2 to -0.04%) and 35 g for women in both samples (RD -1.1%; 95% CI -2.4 to -0.8%). Estimates for men with secondary levels of education had the greatest RD (-2.4%; 95% CI -3.7 to -1.1%).

CONCLUSIONS

The sufficiently small RDs between inferred and invited survey samples support the assumption validity and use of our methodology for adjusting for non-participation. However, the presence of some significant differences means caution is required.

Effects of Substitution of Higher-Alcohol Products with Lower-Alcohol Products on Population-Level Alcohol Purchases: ARIMA Analyses of Spanish Household Data

In its action plan (2022−2030) to reduce the harmful use of alcohol, the WHO calls on economic operators to “substitute, whenever possible, higher-alcohol products with no-alcohol and lower-alcohol products in their overall product portfolios, with the goal of decreasing the overall levels of alcohol consumption in populations and consumer groups”. This paper investigates substitution at the level of the consumer, based on Spanish household purchase data. ARIMA modelling of market research data of 1.29 million alcohol purchases from 18,954 Spanish households is used to study the potential impact of lower-strength alcohol products and the impact of beer prices in reducing household purchases of grams of alcohol between the 2nd quarter of 2017 and 1st quarter of 2022. Reducing the alcohol strength of existing higher-strength beers and wines had a much greater associated impact on reducing the purchases of all grams of alcohol than the relatively small increases in purchases of no-alcohol beers (ABV ≤ 1.0%) and zero-alcohol wines (ABV = 0.0%). For beers, the relative price per gram of alcohol decreased with the increasing ABV of the beer. Increasing the price per gram of alcohol in beers with an ABV > 3.5%, adjusted for the ABV of the beer, was associated with much greater increases in purchases of no-alcohol beers (ABV ≤ 1.0%) and much greater decreases in purchases of all grams of alcohol than decreases in the price of no-alcohol beers or increases in the price of beers with an ABV > 3.5% unadjusted for ABV. Thus, a key to reducing purchases of grams of alcohol, which also results in increased purchases of no-alcohol beers, is to increase the price of higher strength beers (ABV > 3.5%) with the price per gram of alcohol increasing as the ABV of the product increases.

The Impact of Lower-Strength Alcohol Products on Alcohol Purchases by Spanish Households

In its action plan (2022-2030) to reduce the harmful use of alcohol, the WHO calls on economic operators to "substitute, whenever possible, higher-alcohol products with no-alcohol and lower-alcohol products in their overall product portfolios, with the goal of decreasing the overall levels of alcohol consumption in populations and consumer groups". In this paper, we investigate substitution at the level of the consumer based on Spanish household purchase data. ARIMA modelling of market research data of 1.29 million alcohol purchases from 18,954 Spanish households is used to study the potential impact of lower-strength alcohol products on reducing household purchases of grams of alcohol between the second quarter of 2017 and the first quarter of 2022. For households that recently bought either no-alcohol beer or wine (ABV ≤ 0.5%), the subsequent associated purchases of higher-strength beers and wines, respectively, and total grams of alcohol were reduced, the more so the higher the volume of initial purchases of beers and wines. The introduction of 20% ABV variants of same-branded 40% ABV whisky and gin during early 2021 was associated with reduced purchases of grams of alcohol within all spirits and of total grams of alcohol as a result of switching from other spirits products to the 20% variants, although not associated with reduced purchases of grams of alcohol within all variants of the studied same-branded whisky and gin; instead, an increase was observed in this category. With respect to Spanish household purchase data, the evidence behind the WHO's call for substitution is substantiated. Further research across different jurisdictions is needed to provide ongoing monitoring of the impact of potential substitution on consumer behavior and public health, including unintended consequences, with findings from research informing future alcohol policies at all levels.

A methodology for preprocessing structured big data in the behavioral sciences

The characteristics of big data, including high volume, increased variety, and velocity, pose special challenges for data analysis. As these characteristics generally preclude manual data inspection and processing, researchers must often use computational methodologies to deal with this type of data; techniques that may be unfamiliar to nonspecialists, including behavioral scientists. However, previous data analytics methodologies within the field of computer science, developed to handle the generic tasks of data collection, preprocessing, and analysis, can be appropriated for use in other disciplines. These methodologies involve a sequential pipeline of quality checks to prepare data sets for analysis and application. Building upon these methodologies, this paper describes the Big Data Quality & Statistical Assurance (BDQSA) model, applicable for researchers in the behavioral sciences. It involves a series of data preprocessing tasks, to achieve data understanding, as well as data screening, cleaning, and transformation. These are followed by a statistical quality phase, which includes extraction of the relevant data subset, type conversions, ensuring sample representativeness when appropriate, and assessing statistical assumptions. The resulting model thereby provides methodological guidance for the preprocessing of behavioral science big data, aimed at ensuring acceptable data quality before analysis is undertaken. Sample R code snippets demonstrating the application of this model are provided throughout the paper.

The impact of lower strength alcohol products on alcohol purchases: ARIMA analyses based on 4 million purchases by 69 803 households, 2015-2019

BACKGROUND

Lowering the strength of alcohol products could lead to less alcohol being bought and drunk. In its prevention White Paper, the UK Government aims to promote a significant increase in the availability of alcohol-free and low-alcohol products by 2025.

METHODS

Through descriptive analysis and ARIMA modelling of >4 million alcohol purchases from 69 803 British households, we study the potential impact of lower strength alcohol products in reducing household purchases of grams of alcohol over 2015-2019. Households are divided into predominantly beer, wine or spirits purchasers.

RESULTS

Over 5 years, there were decreases in purchases of grams of alcohol within beer amongst beer-purchasing households and increases in purchases of grams of alcohol within wine and spirits amongst, respectively, wine- and spirits-purchasing households. Almost all the changes were due to beer-purchasing households buying less regular strength beer, and wine and spirits-purchasing households buying, respectively, more regular strength wine and spirits, rather than increases in purchases of no- and low-alcohol products.

CONCLUSIONS

In general, lower strength alcohol products have not contributed to British households buying fewer grams of alcohol over the 5-year follow-up period during 2015-2019.

Educational Gradients in Drinking Amount and Heavy Episodic Drinking among Working-Age Men and Women in Spain

Alcohol-related harm decreases as socioeconomic position increases, although sometimes the opposite happens with alcohol intake. The objective was to know the educational gradient in monthly measures of drinking amount and heavy episodic drinking (HED) among people aged 25−64 years in Spain from 1997−2017. Such gradient was characterized with the relative percent change (PC) in drinking measures per year of education from generalized linear regression models after adjusting for age, year, region, marital status and immigration status. Among men, the PCs were significantly positive (p < 0.05) for prevalence of <21 g alcohol/day (2.9%) and 1−3 HED days (1.4%), and they were negative for prevalences of 21−40 g/day (−1.1%), >40 g/day (−6.0%) and ≥4 HED days (−3.2%), while among women they ranged from 3.6% to 5.7%. The gradient in prevalences of >40 g/day (men) and >20 g/day (women) was greatly attenuated after additionally adjusting for HED, while that of ≥4 HED days was only slightly attenuated after additionally adjusting for drinking amount. Among women, the gradients, especially in HED measures, seem steeper in 2009−2017 than in 1997−2007. Educational inequality remained after additional adjustment for income and occupation, although it decreased among women. These results can guide preventive interventions and help explain socioeconomic inequalities in alcohol-related harm.

Development of an algorithm to classify primary care electronic health records of alcohol consumption: experience using data linkage from UK Biobank and primary care electronic health data sources

OBJECTIVES

Develop a novel algorithm to categorise alcohol consumption using primary care electronic health records (EHRs) and asses its reliability by comparing this classification with self-reported alcohol consumption data obtained from the UK Biobank (UKB) cohort.

DESIGN

Cross-sectional study.

SETTING

The UKB, a population-based cohort with participants aged between 40 and 69 years recruited across the UK between 2006 and 2010.

PARTICIPANTS

UKB participants from Scotland with linked primary care data.

PRIMARY AND SECONDARY OUTCOME MEASURES

Create a rule-based multiclass algorithm to classify alcohol consumption reported by Scottish UKB participants and compare it with their classification using data present in primary care EHRs based on Read Codes. We evaluated agreement metrics (simple agreement and kappa statistic).

RESULTS

Among the Scottish UKB participants, 18 838 (69%) had at least one Read Code related to alcohol consumption and were used in the classification. The agreement of alcohol consumption categories between UKB and primary care data, including assessments within 5 years was 59.6%, and kappa was 0.23 (95% CI 0.21 to 0.24). Differences in classification between the two sources were statistically significant (p<0.001); More individuals were classified as 'sensible drinkers' and in lower alcohol consumption levels in primary care records compared with the UKB. Agreement improved slightly when using only numerical values (k=0.29; 95% CI 0.27 to 0.31) and decreased when using qualitative descriptors only (k=0.18;95% CI 0.16 to 0.20).

CONCLUSION

Our algorithm classifies alcohol consumption recorded in Primary Care EHRs into discrete meaningful categories. These results suggest that alcohol consumption may be underestimated in primary care EHRs. Using numerical values (alcohol units) may improve classification when compared with qualitative descriptors.

The COVID-19 alcohol paradox: British household purchases during 2020 compared with 2015-2019

British supermarket-panel data suggest no increases in overall sales and purchases of alcohol following COVID-19 lockdowns, yet survey and mortality data suggest otherwise. This paper attempts to unravel the paradox. Based on purchase data of 79,417 British households from Kantar Worldpanel, we undertake controlled interrupted time series analysis of the impact of COVID-19 confinement introduced on 23rd March 2020, and variably applied during 2020, compared to purchases during 2015 to 2019 as controls. We also undertook Poisson regression analyses to estimate if changes in purchases differed by household socio-demographic and economic factors. Excess off-trade household alcohol purchases (expressed as grams of ethanol) following the introduction of confinement, were 29.2% higher (95% CI = 25.8% to 32.5%) for the post-confinement months of 2020, being larger until mid-July 2020 (37.5%, 95%CI = 33.9 to 41.26%) when pubs re-opened with restrictions, and smaller (24.6%, 95%CI = 21.6 to 27.7) thereafter. During the time of complete pub closures, and fully adjusting for no on-trade purchases, household purchases of alcohol did not change when compared with the same time period during 2015-2019 (coefficient = -0.9%, 95%CI = -5.6 to 3.8). Excess purchases from 23rd March to 31st December 2020 varied by region of Great Britain, being higher in the north of England, and lower in Scotland and Wales. Excess purchases were greater in the most deprived households, compared with the least deprived households. Excess purchases increased substantially as the amount of alcohol normally purchased by a household increased, with the top one fifth of households that normally bought the most alcohol increasing their purchases more than 17 times than the bottom one fifth of households that bought the least alcohol. That the heaviest buyers of alcohol increased their purchases the most, with some independent impact of socio-economic disadvantage, might explain why reported alcohol problems and recent alcohol-related death rates might have increased. A conclusion of this is that alcohol policy to reduce high consumption of alcohol, and the availability of help and treatment to reduce alcohol consumption become more important during extraordinary times, such as COVID lockdowns.

Intended and unintended consequences of the implementation of minimum unit pricing of alcohol in Scotland: a natural experiment

Background

Scotland was the first country to implement minimum unit pricing for alcohol nationally. Minimum unit pricing aims to reduce alcohol-related harms and to narrow health inequalities. Minimum unit pricing sets a minimum retail price based on alcohol content, targeting products preferentially consumed by high-risk drinkers. This study comprised three components.

Objectives

This study comprised three components assessing alcohol consumption and alcohol-related attendances in emergency departments, investigating potential unintended effects of minimum unit pricing on alcohol source and drug use, and exploring changes in public attitudes, experiences and norms towards minimum unit pricing and alcohol use.

Design

We conducted a natural experiment study using repeated cross-sectional surveys comparing Scotland (intervention) and North England (control) areas. This involved comparing changes in Scotland following the introduction of minimum unit pricing with changes seen in the north of England over the same period. Difference-in-difference analyses compared intervention and control areas. Focus groups with young people and heavy drinkers, and interviews with professional stakeholders before and after minimum unit pricing implementation in Scotland allowed exploration of attitudes, experiences and behaviours, stakeholder perceptions and potential mechanisms of effect.

Setting

Four emergency departments in Scotland and North England (component 1), six sexual health clinics in Scotland and North England (component 2), and focus groups and interviews in Scotland (component 3).

Participants

Research nurses interviewed 23,455 adults in emergency departments, and 15,218 participants self-completed questionnaires in sexual health clinics. We interviewed 30 stakeholders and 105 individuals participated in focus groups.

Intervention

Minimum unit pricing sets a minimum retail price based on alcohol content, targeting products preferentially consumed by high-risk drinkers.

Results

The odds ratio for an alcohol-related emergency department attendance following minimum unit pricing was 1.14 (95% confidence interval 0.90 to 1.44; p = 0.272). In absolute terms, we estimated that minimum unit pricing was associated with 258 more alcohol-related emergency department visits (95% confidence interval –191 to 707) across Scotland than would have been the case had minimum unit pricing not been implemented. The odds ratio for illicit drug consumption following minimum unit pricing was 1.04 (95% confidence interval 0.88 to 1.24; p = 0.612). Concerns about harms, including crime and the use of other sources of alcohol, were generally not realised. Stakeholders and the public generally did not perceive price increases or changed consumption. A lack of understanding of the policy may have caused concerns about harms to dependent drinkers among participants from more deprived areas.

Limitations

The short interval between policy announcement and implementation left limited time for pre-intervention data collection.

Conclusions

Within the emergency departments, there was no evidence of a beneficial impact of minimum unit pricing. Implementation appeared to have been successful and there was no evidence of substitution from alcohol consumption to other drugs. Drinkers and stakeholders largely reported not noticing any change in price or consumption. The lack of effect observed in these settings in the short term, and the problem-free implementation, suggests that the price per unit set (£0.50) was acceptable, but may be too low. Our evaluation, which itself contains multiple components, is part of a wider programme co-ordinated by Public Health Scotland and the results should be understood in this wider context.

Future work

Repeated evaluation of similar policies in different contexts with varying prices would enable a fuller picture of the relationship between price and impacts.

Trial registration

Current Controlled Trials ISRCTN16039407.

Funding

This project was funded by the National Institute for Health Research (NIHR) Public Health Research programme and will be published in full in Public Health Research; Vol. 9, No. 11. See the NIHR Journals Library website for further project information.

Is Buying and Drinking Zero and Low Alcohol Beer a Higher Socio-Economic Phenomenon? Analysis of British Survey Data, 2015-2018 and Household Purchase Data 2015-2020

Zero and low alcohol products, particularly beer, are gaining consideration as a method to reduce consumption of ethanol. We do not know if this approach is likely to increase or decrease health inequalities. The aim of the study was to determine if the purchase and consumption of zero and low alcohol beers differs by demographic and socio-economic characteristics of consumers. Based on British household purchase data from 79,411 households and on British survey data of more than 104,635 adult (18+) respondents, we estimated the likelihood of buying and drinking zero (ABV = 0.0%) and low alcohol (ABV > 0.0% and ≤ 3.5%) beer by a range of socio-demographic characteristics. We found that buying and consuming zero alcohol beer is much more likely to occur in younger age groups, in more affluent households, and in those with higher social grades, with gaps in buying zero alcohol beer between households in higher and lower social grades widening between 2015 and 2020. Buying and drinking low alcohol beer had less consistent relationships with socio-demographic characteristics, but was strongly driven by households that normally buy and drink the most alcohol. Common to many health-related behaviours, it seems that it is the more affluent that lead the way in choosing zero or low alcohol products. Whilst the increased availability of zero and low alcohol products might be a useful tool to reduce overall ethanol consumption in the more socially advantageous part of society, it may be less beneficial for the rest of the population. Other evidence-based alcohol policy measures that lessen health inequalities, need to go hand-in-hand with those promoting the uptake of zero and low alcohol beer.

Observational study of the implementation of telephone advice nursing in Sweden: did callers follow recommendations and did the rate of healthcare visits change?

OBJECTIVES

Telephone advice nursing is introduced in many countries with one key aim being a reduction of avoidable healthcare visits. The aim of this study was to explore whether callers to a telephone advice nursing service followed the telenurses' recommendations, and whether there was a change in the level and trend of the rate of healthcare visits after the introduction of telephone advice nursing.

DESIGN

Observational study.

SETTING

Primary and secondary care in Jönköping Region, Sweden.

PARTICIPANTS

Telephone advice nursing calls, 6:00-23:00, 2014-2015 (n=185 994) and outpatient healthcare visits 2012-2015 (n=6 877 266).

PRIMARY OUTCOME

Proportion of callers who visited healthcare within the time period advised by the telenurse.

SECONDARY OUTCOME

Change in level or trend of the overall rate of healthcare visits per 1000 persons and 4-week period after the introduction of telephone advice nursing, with subgroup analysis for primary and secondary care.

RESULTS

77% of callers who were recommended either to visit healthcare within 24 hours or to 'wait and see' followed the recommendations. There was no significant change in level (-5.15; 95% CI -15.80 to 5.50; p=0.349) or trend (-0.24; 95% CI -0.86 to 0.38; p=0.448) of the overall rate of visits per 1000 persons and 4-week period after the introduction of telephone advice nursing. For the rate of primary care visits, an increase in level (8.01; 95% CI 6.36 to 9.66; p<0.001) and trend (1.28; 95% CI 1.17 to 1.39; p<0.001) were observed. For the rate of secondary care visits, a decrease in level (-8.77, 95% CI -14.41 to -3.13; p=0.004) and trend (-1.03, 95% CI -1.35 to -0.71; p<0.001) were observed.

CONCLUSIONS

The introduction of telephone advice nursing may have contributed to a shift in the rate of healthcare visits from secondary to primary care.

Standard multiple imputation of survey data didn't perform better than simple substitution in enhancing an administrative dataset: the example of self-rated health in England

Background

Health surveys provide a rich array of information but on relatively small numbers of individuals and evidence suggests that they are becoming less representative as response levels fall. Routinely collected administrative data offer more extensive population coverage but typically comprise fewer health topics. We explore whether data combination and multiple imputation of health variables from survey data is a simple and robust way of generating these variables in the general population.

Methods

We use the UK Integrated Household Survey and the English 2011 population census both of which included self-rated general health. Setting aside the census self-rated health data we multiply imputed self-rated health responses for the census using the survey data and compared these with the actual census results in 576 unique groups defined by age, sex, housing tenure and geographic region.

Results

Compared with original census data across the groups, multiply imputed proportions of bad or very bad self-rated health were not a markedly better fit than those simply derived from the survey proportions.

Conclusion

While multiple imputation may have the potential to augment population data with information from surveys, further testing and refinement is required.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12982-021-00099-z.

Alcohol-related Outcomes and All-cause Mortality in the Health 2000 Survey by Participation Status and Compared with the Finnish Population

BACKGROUND

In the context of declining levels of participation, understanding differences between participants and non-participants in health surveys is increasingly important for reliable measurement of health-related behaviors and their social differentials. This study compared participants and non-participants of the Finnish Health 2000 survey, and participants and a representative sample of the target population, in terms of alcohol-related harms (hospitalizations and deaths) and all-cause mortality.

METHODS

We individually linked 6,127 survey participants and 1,040 non-participants, aged 30-79, and a register-based population sample (n = 496,079) to 12 years of subsequent administrative hospital discharge and mortality data. We estimated age-standardized rates and rate ratios for each outcome for non-participants and the population sample relative to participants with and without sampling weights by sex and educational attainment.

RESULTS

Harms and mortality were higher in non-participants, relative to participants for both men (rate ratios = 1.5 [95% confidence interval = 1.2, 1.9] for harms; 1.6 [1.3, 2.0] for mortality) and women (2.7 [1.6, 4.4] harms; 1.7 [1.4, 2.0] mortality). Non-participation bias in harms estimates in women increased with education and in all-cause mortality overall. Age-adjusted comparisons between the population sample and sampling weighted participants were inconclusive for differences by sex; however, there were some large differences by educational attainment level.

CONCLUSIONS

Rates of harms and mortality in non-participants exceed those in participants. Weighted participants' rates reflected those in the population well by age and sex, but insufficiently by educational attainment. Despite relatively high participation levels (85%), social differentiating factors and levels of harm and mortality were underestimated in the participants.

Causal inference with observational data: the need for triangulation of evidence

The goal of much observational research is to identify risk factors that have a causal effect on health and social outcomes. However, observational data are subject to biases from confounding, selection and measurement, which can result in an underestimate or overestimate of the effect of interest. Various advanced statistical approaches exist that offer certain advantages in terms of addressing these potential biases. However, although these statistical approaches have different underlying statistical assumptions, in practice they cannot always completely remove key sources of bias; therefore, using design-based approaches to improve causal inference is also important. Here it is the design of the study that addresses the problem of potential bias - either by ensuring it is not present (under certain assumptions) or by comparing results across methods with different sources and direction of potential bias. The distinction between statistical and design-based approaches is not an absolute one, but it provides a framework for triangulation - the thoughtful application of multiple approaches (e.g. statistical and design based), each with their own strengths and weaknesses, and in particular sources and directions of bias. It is unlikely that any single method can provide a definite answer to a causal question, but the triangulation of evidence provided by different approaches can provide a stronger basis for causal inference. Triangulation can be considered part of wider efforts to improve the transparency and robustness of scientific research, and the wider scientific infrastructure and system of incentives.

Mental health and health behaviours before and during the initial phase of the COVID-19 lockdown: longitudinal analyses of the UK Household Longitudinal Study

BACKGROUND

There are concerns that COVID-19 mitigation measures, including the 'lockdown', may have unintended health consequences. We examined trends in mental health and health behaviours in the UK before and during the initial phase of the COVID-19 lockdown and differences across population subgroups.

METHODS

Repeated cross-sectional and longitudinal analysis of the UK Household Longitudinal Study, including representative samples of over 27,000 adults (aged 18+) interviewed in four survey waves between 2015 and 2020. A total of 9748 adults had complete data for longitudinal analyses. Outcomes included psychological distress (General Health Questionnaire-12), loneliness, current cigarette smoking, use of e-cigarettes and alcohol consumption. Cross-sectional prevalence estimates were calculated and multilevel Poisson regression assessed associations between time period and the outcomes of interest, as well as differential associations by age, gender, education level and ethnicity.

RESULTS

Psychological distress increased 1 month into lockdown with the prevalence rising from 19.4% (95% CI 18.7% to 20.1%) in 2017-2019 to 30.6% (95% CI 29.1% to 32.3%) in April 2020 (RR=1.3, 95% CI 1.2 to 1.4). Groups most adversely affected included women, young adults, people from an Asian background and those who were degree educated. Loneliness remained stable overall (RR=0.9, 95% CI 0.6 to 1.5). Smoking declined (RR=0.9, 95% CI=0.8,1.0) and the proportion of people drinking four or more times per week increased (RR=1.4, 95% CI 1.3 to 1.5), as did binge drinking (RR=1.5, 95% CI 1.3 to 1.7).

CONCLUSIONS

Psychological distress increased 1 month into lockdown, particularly among women and young adults. Smoking declined, but adverse alcohol use generally increased. Effective measures are required to mitigate negative impacts on health.

Mental health and health behaviours before and during the initial phase of the COVID-19 lockdown: longitudinal analyses of the UK Household Longitudinal Study

BACKGROUND

There are concerns that COVID-19 mitigation measures, including the 'lockdown', may have unintended health consequences. We examined trends in mental health and health behaviours in the UK before and during the initial phase of the COVID-19 lockdown and differences across population subgroups.

METHODS

Repeated cross-sectional and longitudinal analysis of the UK Household Longitudinal Study, including representative samples of over 27,000 adults (aged 18+) interviewed in four survey waves between 2015 and 2020. A total of 9748 adults had complete data for longitudinal analyses. Outcomes included psychological distress (General Health Questionnaire-12), loneliness, current cigarette smoking, use of e-cigarettes and alcohol consumption. Cross-sectional prevalence estimates were calculated and multilevel Poisson regression assessed associations between time period and the outcomes of interest, as well as differential associations by age, gender, education level and ethnicity.

RESULTS

Psychological distress increased 1 month into lockdown with the prevalence rising from 19.4% (95% CI 18.7% to 20.1%) in 2017-2019 to 30.6% (95% CI 29.1% to 32.3%) in April 2020 (RR=1.3, 95% CI 1.2 to 1.4). Groups most adversely affected included women, young adults, people from an Asian background and those who were degree educated. Loneliness remained stable overall (RR=0.9, 95% CI 0.6 to 1.5). Smoking declined (RR=0.9, 95% CI=0.8,1.0) and the proportion of people drinking four or more times per week increased (RR=1.4, 95% CI 1.3 to 1.5), as did binge drinking (RR=1.5, 95% CI 1.3 to 1.7).

CONCLUSIONS

Psychological distress increased 1 month into lockdown, particularly among women and young adults. Smoking declined, but adverse alcohol use generally increased. Effective measures are required to mitigate negative impacts on health.

Self-reported health without clinically measurable benefits among adult users of multivitamin and multimineral supplements: a cross-sectional study

OBJECTIVE

Multiple clinical trials fail to identify clinically measurable health benefits of daily multivitamin and multimineral (MVM) consumption in the general adult population. Understanding the determinants of widespread use of MVMs may guide efforts to better educate the public about effective nutritional practices. The objective of this study was to compare self-reported and clinically measurable health outcomes among MVM users and non-users in a large, nationally representative adult civilian non-institutionalised population in the USA surveyed on the use of complementary health practices.

DESIGN

Cross-sectional analysis of the effect of MVM consumption on self-reported overall health and clinically measurable health outcomes.

PARTICIPANTS

Adult MVM users and non-users from the 2012 National Health Interview Survey (n=21 603).

PRIMARY AND SECONDARY OUTCOME MEASURES

Five psychological, physical, and functional health outcomes: (1) self-rated health status, (2) needing help with routine needs, (3) history of 10 chronic diseases, (4) presence of 19 health conditions in the past 12 months, and (5) Kessler 6-Item (K6) Psychological Distress Scale to measure non-specific psychological distress in the past month.

RESULTS

Among 4933 adult MVM users and 16 670 adult non-users, MVM users self-reported 30% better overall health than non-users (adjusted OR 1.31; 95% CI 1.17 to 1.46; false discovery rate adjusted p<0.001). There were no differences between MVM users and non-users in history of 10 chronic diseases, number of present health conditions, severity of current psychological distress on the K6 Scale and rates of needing help with daily activities. No effect modification was observed after stratification by sex, education, and race.

CONCLUSIONS

MVM users self-reported better overall health despite no apparent differences in clinically measurable health outcomes. These results suggest that widespread use of multivitamins in adults may be a result of individuals' positive expectation that multivitamin use leads to better health outcomes or a self-selection bias in which MVM users intrinsically harbour more positive views regarding their health.

Drinking beer, wine or spirits - does it matter for inequalities in alcohol-related hospital admission? A record-linked longitudinal study in Wales

BACKGROUND

Alcohol-related harm has been found to be higher in disadvantaged groups, despite similar alcohol consumption to advantaged groups. This is known as the alcohol harm paradox. Beverage type is reportedly socioeconomically patterned but has not been included in longitudinal studies investigating record-linked alcohol consumption and harm. We aimed to investigate whether and to what extent consumption by beverage type, BMI, smoking and other factors explain inequalities in alcohol-related harm.

METHODS

11,038 respondents to the Welsh Health Survey answered questions on their health and lifestyle. Responses were record-linked to wholly attributable alcohol-related hospital admissions (ARHA) eight years before the survey month and until the end of 2016 within the Secure Anonymised Information Linkage (SAIL) Databank. We used survival analysis, specifically multi-level and multi-failure Cox mixed effects models, to calculate the hazard ratios of ARHA. In adjusted models we included the number of units consumed by beverage type and other factors, censoring for death or moving out of Wales.

RESULTS

People living in more deprived areas had a higher risk of admission (HR 1.75; 95% CI 1.23-2.48) compared to less deprived. Adjustment for the number of units by type of alcohol consumed only reduced the risk of ARHA for more deprived areas by 4% (HR 1.72; 95% CI 1.21-2.44), whilst adding smoking and BMI reduced these inequalities by 35.7% (HR 1.48; 95% CI 1.01-2.17). These social patterns were similar for individual-level social class, employment, housing tenure and highest qualification. Inequalities were further reduced by including either health status (16.6%) or mental health condition (5%). Unit increases of spirits drunk were positively associated with increasing risk of ARHA (HR 1.06; 95% CI 1.01-1.12), higher than for other drink types.

CONCLUSIONS

Although consumption by beverage type was socioeconomically patterned, it did not help explain inequalities in alcohol-related harm. Smoking and BMI explained around a third of inequalities, but lower socioeconomic groups had a persistently higher risk of (multiple) ARHA. Comorbidities also explained a further proportion of inequalities and need further investigation, including the contribution of specific conditions. The increased harms from consumption of stronger alcoholic beverages may inform public health policy.

Immediate impact of minimum unit pricing on alcohol purchases in Scotland: controlled interrupted time series analysis for 2015-18

OBJECTIVE

To assess the immediate impact of the introduction of minimum unit pricing in Scotland on household alcohol purchases.

DESIGN

Controlled interrupted time series analysis.

SETTING

Purchase data from Kantar Worldpanel's household shopping panel for 2015-18.

PARTICIPANTS

5325 Scottish households, 54 807 English households as controls, and 10 040 households in northern England to control for potential cross border effects.

INTERVENTIONS

Introduction of a minimum price of 50p (€0.55; $0.61) per UK unit (6.25p per gram) for the sale of alcohol in Scotland on 1 May 2018.

MAIN OUTCOME MEASURES

Price per gram of alcohol, number of grams of alcohol purchased from off-trade by households, and weekly household expenditure on alcohol.

RESULTS

The introduction of minimum unit pricing in Scotland was associated with an increase in purchase price of 0.64p per gram of alcohol (95% confidence interval 0.54 to 0.75), a reduction in weekly purchases of 9.5 g of alcohol per adult per household (5.1 to 13.9), and a non-significant increase in weekly expenditure on alcohol per household of 61p (-5 to 127). The increase in purchase price was higher in lower income households and in households that purchased the largest amount of alcohol. The reduction in purchased grams of alcohol was greater in lower income households and only occurred in the top fifth of households by income that purchased the greatest amount of alcohol, where the reduction was 15 g of alcohol per week (6 to 24). Changes in weekly expenditure were not systematically related to household income but increased with increasing household purchases.

CONCLUSIONS

In terms of immediate impact, the introduction of minimum unit pricing appears to have been successful in reducing the amount of alcohol purchased by households in Scotland. The action was targeted, in that reductions of purchased alcohol only occurred in the households that bought the most alcohol.

Recommendations for design and analysis of health examination surveys under selective non-participation

Background

The decreasing participation rates and selective non-participation peril the representativeness of health examination surveys (HESs).

Methods

Finnish HESs conducted in 1972-2012 are used to demonstrate that survey participation rates can be enhanced with well-planned recruitment procedures and auxiliary information about survey non-participants can be used to reduce selection bias.

Results

Experiments incorporated to pilot surveys and experience from previously conducted surveys lead to practical improvements. For example, SMS reminders were taken as a routine procedure to the Finnish HESs after testing their effect on a pilot study and finding them as a cost-effective way to increase participation rate especially among younger age groups. Auxiliary information about survey non-participants can be obtained from many sources: sampling frames, previous measurements in longitudinal setting, re-contacts and non-response questionnaires, and record linkage to administrative data sources. These data can be used in statistical modelling to adjust the population level estimates for the selection bias. Information on the characteristics of non-participants also helps to improve targeting the recruitment in the future.

Conclusion

All methods discussed and recommended are relatively easy to incorporate to any national HES in Europe except the record linkage of survey data from administrative data sources. This is not feasible in all European countries because of non-existence of registries, lack of an identifier needed for record linkage, or national data protection legislation which restricts the data use.

Correcting for non-participation bias in health surveys using record-linkage, synthetic observations and pattern mixture modelling

Surveys are key means of obtaining policy-relevant information not available from routine sources. Bias arising from non-participation is typically handled by applying weights derived from limited socio-demographic characteristics. This approach neither captures nor adjusts for differences in health and related behaviours between participants and non-participants within categories. We addressed non-participation bias in alcohol consumption estimates using novel methodology applied to 2003 Scottish Health Survey responses record-linked to prospective administrative data. Differences were identified in socio-demographic characteristics, alcohol-related harm (hospitalisation or mortality) and all-cause mortality between survey participants and, from unlinked administrative sources, the contemporaneous general population of Scotland. These were used to infer the number of non-participants within each subgroup defined by socio-demographics and health outcomes. Synthetic observations for non-participants were then generated, missing only alcohol consumption. Weekly alcohol consumption values among synthetic non-participants were multiply imputed under missing at random and missing not at random assumptions. Relative to estimates adjusted using previously derived weights, the obtained mean weekly alcohol intake estimates were up to 59% higher among men and 16% higher among women, depending on the assumptions imposed. This work demonstrates the universal value of multiple imputation-based methodological advancement incorporating administrative health data over routine weighting procedures.

Validation of non-participation bias methodology based on record-linked Finnish register-based health survey data: a protocol paper

INTRODUCTION

Decreasing participation levels in health surveys pose a threat to the validity of estimates intended to be representative of their target population. If participants and non-participants differ systematically, the results may be biased. The application of traditional non-response adjustment methods, such as weighting, can fail to correct for such biases, as estimates are typically based on the sociodemographic information available. Therefore, a dedicated methodology to infer on non-participants offers advancement by employing survey data linked to administrative health records, with reference to data on the general population. We aim to validate such a methodology in a register-based setting, where individual-level data on participants and non-participants are available, taking alcohol consumption estimation as the exemplar focus.

METHODS AND ANALYSIS

We made use of the selected sample of the Health 2000 survey conducted in Finland and a separate register-based sample of the contemporaneous population, with follow-up until 2012. Finland has nationally representative administrative and health registers available for individual-level record linkage to the Health 2000 survey participants and invited non-participants, and the population sample. By comparing the population sample and the participants, synthetic observations representing the non-participants may be generated, as per the developed methodology. We can compare the distribution of the synthetic non-participants with the true distribution from the register data. Multiple imputation was then used to estimate alcohol consumption based on both the actual and synthetic data for non-participants, and the estimates can be compared to evaluate the methodology's performance.

ETHICS AND DISSEMINATION

Ethical approval and access to the Health 2000 survey data and data from administrative and health registers have been given by the Health 2000 Scientific Advisory Board, Statistics Finland and the National Institute for Health and Welfare. The outputs will include two publications in public health and statistical methodology journals and conference presentations.

Canonical Causal Diagrams to Guide the Treatment of Missing Data in Epidemiologic Studies

With incomplete data, the “missing at random” (MAR) assumption is widely understood to enable unbiased estimation with appropriate methods. While the need to assess the plausibility of MAR and to perform sensitivity analyses considering “missing not at random” (MNAR) scenarios has been emphasized, the practical difficulty of these tasks is rarely acknowledged. With multivariable missingness, what MAR means is difficult to grasp, and in many MNAR scenarios unbiased estimation is possible using methods commonly associated with MAR. Directed acyclic graphs (DAGs) have been proposed as an alternative framework for specifying practically accessible assumptions beyond the MAR-MNAR dichotomy. However, there is currently no general algorithm for deciding how to handle the missing data given a specific DAG. Here we construct “canonical” DAGs capturing typical missingness mechanisms in epidemiologic studies with incomplete data on exposure, outcome, and confounding factors. For each DAG, we determine whether common target parameters are “recoverable,” meaning that they can be expressed as functions of the available data distribution and thus estimated consistently, or whether sensitivity analyses are necessary. We investigate the performance of available-case and multiple-imputation procedures. Using data from waves 1–3 of the Longitudinal Study of Australian Children (2004–2008), we illustrate how our findings can guide the treatment of missing data in point-exposure studies.

Mental health and the jilted generation: Using age-period-cohort analysis to assess differential trends in young people's mental health following the Great Recession and austerity in England

Those born in the United Kingdom post-1979 have been described as a ‘jilted generation’, materially disadvantaged by economic and social policy; however, it is unclear whether this resulted in their experiencing poorer mental health than previous cohorts. Following the 2008 recession, UK austerity reforms associated with worsening mental health also disproportionately impacted those of younger working-age. This study aimed to identify any historic cohort changes in population mental health, and whether austerity widened generational inequalities. Repeat cross-sectional data from the Health Survey for England (1991–2014) were used to calculate prevalence of psychopathology for those of younger and older working-age (16–30 and 31–64 years) and retirement-age (65+ years), measured by General Health Questionnaire-12 (GHQ) score ≥ 4 (caseness). Descriptive age-period-cohort analysis was performed for 15-year birth cohorts, including the jilted generation (born 1976-90). Logistic regression tested differences in outcome between groups.

Age-specific GHQ caseness between successive birth cohorts did not significantly change for men, and significantly improved between 2.8% (95% CI 0.1%–5.5%) and 4.4% (95% CI 2.2%–6.7%) for women. Secondary analysis adjusting for education partially explained this improvement. Following the recession, GHQ caseness worsened in men of younger and older working-age by 3.7% (95% CI 1.2%–6.2%) and 3.5% (95% CI 2.1%–5.0%) respectively before returning to baseline during austerity. All women experienced non-significant increases post-recession, but trends diverged during austerity with caseness worsening by 2.3% (95% CI 1.0%–3.6%) for older working-age women versus 3.7% (95% CI 1.3%–6.2%) for younger working-age women. Those of retirement-age experienced little change throughout. In summary, mental health has historically improved between successive cohorts, including for the jilted generation. However, the 2008 recession and subsequent austerity could be most impacting those of younger working-age, particularly women, to create a new cohort effect. Policymakers should consider the differential impact economic and social policy may have across society by age.

•

Mental health in England has consistently improved from generation to generation.

•

These generational improvements have been most marked for women.

•

Following the onset of austerity young women experienced worsening mental health.

•

Mental health of over 65s was relatively protected during recession and austerity.

•

Austerity policies have potential to widen generational inequality in mental health.

Assessment of demographic and perinatal predictors of non-response and impact of non-response on measures of association in a population-based case control study: findings from the Georgia Study to Explore Early Development

Background

Participation in epidemiologic studies has declined, raising concerns about selection bias. While estimates derived from epidemiologic studies have been shown to be robust under a wide range of scenarios, additional empiric study is needed. The Georgia Study to Explore Early Development (GA SEED), a population-based case-control study of risk factors for autism spectrum disorder (ASD), provided an opportunity to explore factors associated with non-participation and potential impacts of non-participation on association studies.

Methods

GA SEED recruited preschool-aged children residing in metropolitan-Atlanta during 2007-2012. Children with ASD were identified from multiple schools and healthcare providers serving children with disabilities; children from the general population (POP) were randomly sampled from birth records. Recruitment was via mailed invitation letter with follow-up phone calls. Eligibility criteria included birth and current residence in study area and an English-speaking caregiver. Many children identified for potential inclusion could not be contacted. We used data from birth certificates to examine demographic and perinatal factors associated with participation in GA SEED and completion of the data collection protocol. We also compared ASD-risk factor associations for the final sample of children who completed the study with the initial sample of all likely ASD and POP children invited to potentially participate in the study, had they been eligible. Finally, we derived post-stratification sampling weights for participants who completed the study and compared weighted and unweighted associations between ASD and two factors collected via post-enrollment maternal interview: infertility and reproductive stoppage.

Results

Maternal age and education were independently associated with participation in the POP group. Maternal education was independently associated with participation in the ASD group. Numerous other demographic and perinatal factors were not associated with participation. Moreover, unadjusted and adjusted odds ratios for associations between ASD and several demographic and perinatal factors were similar between the final sample of study completers and the total invited sample. Odds ratios for associations between ASD and infertility and reproductive stoppage were also similar in unweighted and weighted analyses of the study completion sample.

Conclusions

These findings suggest that effect estimates from SEED risk factor analyses, particularly those of non-demographic factors, are likely robust.

The socioeconomic profile of alcohol-attributable mortality in South Africa: a modelling study

BACKGROUND

Globally, illness and life expectancy follow a social gradient that puts people of lower socioeconomic status (SES) at higher risk of dying prematurely. Alcohol consumption has been shown to be a factor contributing to socioeconomic differences in mortality. However, little evidence is available from low- and middle-income countries. The objective of this study was to quantify mortality attributable to alcohol consumption in the adult (15+ years) general population of South Africa in 2015 by SES, age, and sex.

METHODS

A comparative risk assessment was performed using individual and aggregate data from South Africa and risk relations reported in the literature. Alcohol-attributable fractions (AAFs) and alcohol-attributable mortality rates were estimated for cause-specific mortality by SES, sex, and age. Monte Carlo simulation techniques were used to calculate 95% uncertainty intervals (UI).

RESULTS

Overall, approximately 62,300 (95% UI 27,000-103,000) adults died from alcohol-attributable causes in South Africa in 2015, with 60% of deaths occurring in people in the low and 15% in the high SES groups. Age-standardized, alcohol-attributable mortality rates per 100,000 adults were highest for the low SES group (727 deaths, 95% UI 354-1208 deaths) followed by the middle (377 deaths, 95% UI 165-687 deaths) and high SES groups (163 deaths, 95% UI 71-289 deaths). The socioeconomic differences were highest for mortality from infectious diseases. People of low SES had a lower prevalence of current alcohol use but heavier drinking patterns among current drinkers. Among men, AAFs were elevated at low and middle SES, particularly for the middle and higher age groups (35+). Among women, AAFs differed less across SES groups and, in the youngest age group (15-34), women of high SES had elevated AAFs.

CONCLUSIONS

Alcohol use contributed to vast socioeconomic differences in mortality. Where observed, elevated AAFs for people of low and middle SES arose from higher levels of consumption among current drinkers and not from the prevalence of current alcohol use per se. The findings can direct preventive measures and interventions on those at highest risk. Future research is needed to investigate socioeconomic differences in the risk functions relating alcohol use to cause-specific mortality.

Design and development of a complex narrative intervention delivered by text messages to reduce binge drinking among socially disadvantaged men

Background

Socially disadvantaged men are at high risk of suffering from alcohol-related harm. Disadvantaged groups are less likely to engage with health promotion. There is a need for interventions that reach large numbers at low cost and which promote high levels of engagement with the behaviour change process. The aim of this study was to design a theoretically and empirically based text message intervention to reduce binge drinking by socially disadvantaged men.

Results

Following MRC guidance, the intervention was developed in four stages. Stage 1 developed a detailed behaviour change strategy based on existing literature and theory from several areas. These included the psychological theory that would underpin the intervention, alcohol brief interventions, text message interventions, effective behaviour change techniques, narratives in behaviour change interventions and communication theory. In addition, formative research was carried out. A logic model was developed to depict the pathways between intervention inputs, processes and outcomes for behaviour change. Stage 2 created a narrative which illustrated and modelled key steps in the strategy. Stage 3 rendered the intervention into a series of text messages and ensured that appropriate behavioural change techniques were incorporated. Stage 4 revised the messages to ensure comprehensive coverage of the behaviour change strategy and coherence of the narrative. It also piloted the intervention and made final revisions to it.

Conclusions

The structured, systematic approach to design created a narrative intervention which had a strong theoretical and empirical basis. The use of a narrative helped make the intervention realistic and allowed key behaviour change techniques to be modelled by characters. The narrative was intended to promote engagement with the intervention. The intervention was rendered into a series of short text messages, and subsequent piloting showed they were acceptable in the target group. Delivery of an intervention by text message offers a low-cost, low-demand method that can reach large numbers of people. This approach provides a framework for the design of behaviour change interventions which could be used for interventions to tackle other health behaviours.

Electronic supplementary material

The online version of this article (10.1186/s40814-018-0298-0) contains supplementary material, which is available to authorized users.

Text message intervention to reduce frequency of binge drinking among disadvantaged men: the TRAM RCT

Background

Socially disadvantaged men are more likely to binge drink frequently and to experience high levels of alcohol-related harm.

Objectives

To test the effectiveness and cost-effectiveness of a text message intervention in reducing the frequency of binge drinking among disadvantaged men.

Study design

A four-centre, parallel-group, pragmatic, individually randomised controlled trial was conducted. Randomisation was carried out using a secure remote web-based system. It was stratified by participating centre and recruitment method and restricted using block sizes of randomly varying lengths.

Setting

The study was conducted in the community. Members of the public helped to develop the study methods.

Participants

Participants were men aged 25–44 years who had ≥ 2 episodes of binge drinking (> 8 units of alcohol in a single session) in the preceding 28 days. Men were recruited from areas of high deprivation.

Interventions

An empirically and theoretically based text message intervention was delivered by 112 interactive text messages over a 12-week period. The control group received an attentional control comprising 89 text messages on general health.

Primary outcome measure

The primary outcome measure was the proportion of men consuming > 8 units of alcohol on ≥ 3 occasions (in the previous 28 days) at 12 months post intervention.

Results

The recruitment target of 798 was exceeded and 825 men were randomised. Retention was high and similar in the intervention (84.9%) and control (86.5%) groups. Most men in the intervention group engaged enthusiastically with the text messages: almost all (92%) replied to text messages and over two-thirds (67%) replied more than 10 times. The intervention was estimated to have had a modest, statistically non-significant effect on the primary outcome at the 12-month follow-up [odds ratio 0.79, 95% confidence interval (CI) 0.57 to 1.08]. This corresponds to a net reduction of 5.7% in regular binge drinking. Five secondary outcomes showed small non-significant and inconsistent effects on alcohol consumption, with one suggesting a positive effect and four suggesting an adverse effect. Both the short- and the long-term cost per quality-adjusted life-year (QALY) analysis suggested that the brief intervention was dominated by a ‘do-nothing’ option. The intervention’s impacts on patterns of alcohol consumption, QALYs and downstream costs were inconsistent and uncertain.

Limitations

The study used an active control that, combined with the recruitment procedures and baseline assessments, could have biased the treatment effect towards the null. The measurement of alcohol consumption relied on self-reported drinking.

Conclusions

The trial has demonstrated that it is possible to recruit and retain large numbers of socially disadvantaged men in a research study. The text messages delivered a complex theoretically and empirically based intervention that fostered enthusiastic engagement with the key components of the behaviour change sequence. The intervention produced a modest, statistically non-significant effect on the primary outcome, with wide CIs. Further research is needed to reduce uncertainty about the treatment effect. The methods developed for this study provide a platform for the design and testing of interventions to reduce inequalities in health.

Future work

A future trial could reduce the uncertainty around the treatment effect of the intervention.

Trial registration

Current Controlled Trials ISRCTN07695192.

Funding

This study was funded by the National Institute for Health Research (NIHR) Public Health Research programme and will be published in full in Public Health Research; Vol. 6, No. 6. See NIHR Journals Library website for further information.

Assessing causal relationships using genetic proxies for exposures: an introduction to Mendelian randomization

BACKGROUND AND AIMS

Studying the consequences of addictive behaviours is challenging, with understanding causal relationships from observational data being particularly difficult. For example, people who smoke or drink excessively are often systematically different from those who do not, are less likely to participate in research and may misreport their behaviours when they do. Furthermore, the direction of causation between an addictive behaviour and outcome may be unclear. Mendelian randomization (MR) offers potential solutions to these problems.

METHODS

We describe MR's principles and the criteria under which it is valid. We identify challenges and potential solutions in its application (illustrated using two applied examples) and describe methodological extensions in its application.

RESULTS

MR is subject to certain assumptions, and requires the availability of appropriate genetic data, large sample sizes and careful design and conduct. However, it has already been applied successfully to the addiction literature. The relationship between alcohol consumption (proxied by a variant in the ADH1B gene) and cardiovascular risk has been investigated, finding that alcohol consumption increases risk, with no evidence of a cardioprotective effect at moderate consumption levels. In addition, heaviness of smoking (proxied by a variant in the CHRNA5-A3-B4 gene cluster) and risk of depression and schizophrenia have been investigated, with no evidence of a causal effect of smoking on depression but some evidence of a causal effect on schizophrenia.

CONCLUSIONS

Mendelian randomization analyses are already producing robust evidence for addiction-related practice and policy. As genetic variants associated with addictive behaviours are identified, the potential for Mendelian randomization analyses will grow. Methodological developments are also increasing its applicability.

How Healthy Are Survey Respondents Compared with the General Population?: Using Survey-linked Death Records to Compare Mortality Outcomes

BACKGROUND

National surveys are used to capture US health trends and set clinical guidelines, yet the sampling frame often includes those in noninstitutional households, potentially missing those most vulnerable for poor health. Declining response rates in national surveys also represent a challenge, and existing inputs to survey weights have limitations. We compared mortality rates between those who respond to surveys and the general population over time.

METHODS

Survey respondents from 20 waves of the National Health Interview Survey from 1990 to 2009 who have been linked to death records through 31 December 2011 were included. For each cohort in the survey, we estimated their mortality rates along with that cohort's mortality rate in the census population using vital statistics records, and differences were examined using Poisson models.

RESULTS

In all years, survey respondents had lower mortality rates compared with the general population when data were both weighted and unweighted. Among men, survey respondents in the weighted sample had 0.86 (95% confidence interval = 0.853, 0.868) times the mortality rate of the general population (among women, RR = 0.887; 95% confidence interval, 0.879, 0.895). Differences in mortality are evident along all points of the life course. Differences have remained relatively stable over time.

CONCLUSION

Survey respondents have lower death rates than the general US population, suggesting that they are a systematically healthier source population. Incorporating nonhousehold samples and revised weighting strategies to account for sample frame exclusion and nonresponse may allow for more rigorous estimation of the US population's health.

Recruitment methods for survey research: Findings from the Mid-South Clinical Data Research Network

PURPOSE

The objective of this study was to report survey response rates and demographic characteristics of eight recruitment approaches to determine acceptability and effectiveness of large-scale patient recruitment among various populations.

METHODS

We conducted a cross sectional analysis of survey data from two large cohorts. Patients were recruited from the Mid-South Clinical Data Research Network using clinic-based recruitment, research registries, and mail, phone, and email approaches. Response rates are reported as patients who consented for the survey divided by the number of eligible patients approached.

RESULTS

We contacted more than 90,000 patients and 13,197 patients completed surveys. Median age was 56.3years (IQR 40.9, 67.4). Racial/ethnic distribution was 84.1% White, non-Hispanic; 9.9% Black, non-Hispanic; 1.8% Hispanic; and 4.0% other, non-Hispanic. Face-to-face recruitment had the highest response rate of 94.3%, followed by participants who "opted-in" to a registry (76%). The lowest response rate was for unsolicited emails from the clinic (6.1%). Face-to-face recruitment enrolled a higher percentage of participants who self-identified as Black, non-Hispanic compared to other approaches (18.6% face-to-face vs. 8.4% for email).

CONCLUSIONS

Technology-enabled recruitment approaches such as registries and emails are effective for recruiting but may yield less racial/ethnic diversity compared to traditional, more time-intensive approaches.

Adjustment for survey non-representativeness using record-linkage: refined estimates of alcohol consumption by deprivation in Scotland

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.

Socioeconomic status as an effect modifier of alcohol consumption and harm: analysis of linked cohort data

Background

Alcohol-related mortality and morbidity are high in socioeconomically disadvantaged populations compared with individuals from advantaged areas. It is unclear if this increased harm reflects differences in alcohol consumption between these socioeconomic groups, reverse causation (ie, downward social selection for high-risk drinkers), or a greater risk of harm in individuals of low socioeconomic status compared with those of higher status after similar consumption. We aimed to investigate whether the harmful effects of alcohol differ by socioeconomic status, accounting for alcohol consumption and other health-related factors.

Methods

The Scottish Health Surveys are record-linked cross-sectional surveys representative of the adult population of Scotland. We obtained baseline demographics and data for alcohol consumption (units per week and binge drinking) from Scottish Health Surveys done in 1995, 1998, 2003, 2008, 2009, 2010, 2011, and 2012. We matched these data to records for deaths, admissions, and prescriptions. The primary outcome was alcohol-attributable admission or death. The relation between alcohol-attributable harm and socioeconomic status was investigated for four measures (education level, social class, household income, and area-based deprivation) using Cox proportional hazards models. The potential for alcohol consumption and other risk factors (including smoking and body-mass index [BMI]) mediating social patterning was explored in separate regression models. Reverse causation was tested by comparing change in area deprivation over time.

Findings

50 236 participants (21 777 men and 28 459 women) were included in the analytical sample, with 429 986 person-years of follow-up. Low socioeconomic status was associated consistently with strikingly raised alcohol-attributable harms, including after adjustment for weekly consumption, binge drinking, BMI, and smoking. Evidence was noted of effect modification; for example, relative to light drinkers living in advantaged areas, the risk of alcohol-attributable admission or death for excessive drinkers was increased (hazard ratio 6·12, 95% CI 4·45–8·41 in advantaged areas; and 10·22, 7·73–13·53 in deprived areas). We found little support for reverse causation.

Interpretation

Disadvantaged social groups have greater alcohol-attributable harms compared with individuals from advantaged areas for given levels of alcohol consumption, even after accounting for different drinking patterns, obesity, and smoking status at the individual level.

Funding

Medical Research Council, NHS Research Scotland, Scottish Government Chief Scientist Office.

Assessment of Non-Response Bias in Estimates of Alcohol Consumption: Applying the Continuum of Resistance Model in a General Population Survey in England

Background

Previous studies have shown heavier drinkers are less likely to respond to surveys and require extended efforts to recruit. This study applies the continuum of resistance model to explore how survey estimates of alcohol consumption may be affected by non-response bias in three consecutive years of a general population survey in England.

Methods

Using the Health Survey for England (HSE) survey years 2011–13, number of contact attempts (1–6 and 7+) were explored by socio-demographic and drinking characteristics. The odds of drinking more than various thresholds were modelled using logistic regression. Assuming that non-participants were similar to those who were difficult to contact (the continuum of resistance model), the effect of non-response on measures of drinking was investigated.

Results

In the fully-adjusted regression model, women who required 7+ calls were significantly more likely to drink more than the UK Government’s recommended daily limit (OR 1.19, 95% CI 1.06–1.33, P = 0.003) and to engage in heavy episodic drinking (OR 1.23, 95% CI 1.07–1.42, P = 0.004), however this was not significant in men in the fully-adjusted model. When the continuum of resistance model was applied, there was an increase in average weekly alcohol consumption of 1.8 units among men (a 12.6% relative increase), and an increase of 1.5 units among women (a 20.5% relative increase). There was also an increase in the prevalence of heavy episodic drinking of 2.5% among men (an 12.0% relative increase) and of 2.0% among women (a 15.8% relative increase), although other measures of drinking were less affected.

Conclusion

Overall alcohol consumption and the prevalence of heavy episodic drinking were higher among HSE participants who required more extended efforts to contact. The continuum of resistance model suggests non-response bias does affect survey estimates of alcohol consumption.

A cross-sectional investigation of public attitudes toward dementia in Bristol and South Gloucestershire using the approaches to dementia questionnaire

BACKGROUND

To date, surveys of attitudes toward dementia have largely been conducted using unvalidated materials or have focused on healthcare professionals supporting people affected by dementia. The aim of this study was to carry out a survey of public attitudes toward people affected by dementia in Bristol and South Gloucestershire.

METHODS

A survey was carried out using a modified version of the Approaches to Dementia Questionnaire (ADQ). Data from people living outside the area, and people who were working with people affected by dementia were omitted from the analysis. Responses from the remaining 794 ADQ questionnaires were weighted to correct for under-represented age, gender, and ethnic groups.

RESULTS

Younger people held more positive attitudes toward dementia than older people. Individuals who identified themselves as White held more positive attitudes than non-White individuals. Individuals with personal experience of dementia held more positive attitudes than those with no experience of dementia. When considering age differences, gender played a role, with younger men having more positive scores than other groups.

CONCLUSIONS

This is one of the first surveys of public attitudes to dementia to use a validated questionnaire such as the ADQ. The study provides a baseline of attitudes toward dementia for the Bristol and South Gloucestershire areas, against which we will be able to compare changes over time. This is important due to the emphasis in public health campaigns on improving attitudes toward dementia.

Social relationships and cognitive decline: a systematic review and meta-analysis of longitudinal cohort studies

BACKGROUND

Although poor social relationships are assumed to contribute to cognitive decline, meta-analytic approaches have not been applied. Individual study results are mixed and difficult to interpret due to heterogeneity in measures of social relationships. We conducted a systematic review and meta-analysis to investigate the relation between poor social relationships and cognitive decline.

METHODS

MEDLINE, Embase and PsycINFO were searched for longitudinal cohort studies examining various aspects of social relationships and cognitive decline in the general population. Odds ratios (ORs) with 95% confidence intervals (CIs) were pooled using random effects meta-analysis. Sources of heterogeneity were explored and likelihood of publication bias was assessed. We stratified analyses according to three aspects of social relationships: structural, functional and a combination of these.

RESULTS

We identified 43 articles. Poor social relationships predicted cognitive decline; for structural (19 studies): pooled OR: 1.08 (95% CI: 1.05-1.11); functional (8 studies): pooled OR: 1.15 (95% CI: 1.00-1.32); and combined measures (7 studies): pooled OR: 1.12 (95% CI: 1.01-1.24). Meta-regression and subgroup analyses showed that the heterogeneity could be explained by the type of social relationship measurement and methodological quality of included studies.

CONCLUSIONS

Despite heterogeneity in study design and measures, our meta-analyses show that multiple aspects of social relationships are associated with cognitive decline. As evidence for publication bias was found, the association might be overestimated and should therefore be interpreted with caution. Future studies are needed to better define the mechanisms underlying these associations. Potential causality of this prognostic association should be examined in future randomized controlled studies.

Estimating average alcohol consumption in the population using multiple sources: the case of Spain

Background

National estimates on per capita alcohol consumption are provided regularly by various sources and may have validity problems, so corrections are needed for monitoring and assessment purposes. Our objectives were to compare different alcohol availability estimates for Spain, to build the best estimate (actual consumption), characterize its time trend during 2001–2011, and quantify the extent to which other estimates (coverage) approximated actual consumption.

Methods

Estimates were: alcohol availability from the Spanish Tax Agency (Tax Agency availability), World Health Organization (WHO availability) and other international agencies, self-reported purchases from the Spanish Food Consumption Panel, and self-reported consumption from population surveys. Analyses included calculating: between-agency discrepancy in availability, multisource availability (correcting Tax Agency availability by underestimation of wine and cider), actual consumption (adjusting multisource availability by unrecorded alcohol consumption/purchases and alcohol losses), and coverage of selected estimates. Sensitivity analyses were undertaken. Time trends were characterized by joinpoint regression.

Results

Between-agency discrepancy in alcohol availability remained high in 2011, mainly because of wine and spirits, although some decrease was observed during the study period.

The actual consumption was 9.5 l of pure alcohol/person-year in 2011, decreasing 2.3 % annually, mainly due to wine and spirits. 2011 coverage of WHO availability, Tax Agency availability, self-reported purchases, and self-reported consumption was 99.5, 99.5, 66.3, and 28.0 %, respectively, generally with downward trends (last three estimates, especially self-reported consumption). The multisource availability overestimated actual consumption by 12.3 %, mainly due to tourism imbalance.

Conclusions

Spanish estimates of per capita alcohol consumption show considerable weaknesses. Using uncorrected estimates, especially self-reported consumption, for monitoring or other purposes is misleading. To obtain conservative estimates of alcohol-attributable disease burden or heavy drinking prevalence, self-reported consumption should be shifted upwards by more than 85 % (91 % in 2011) of Tax Agency or WHO availability figures. The weaknesses identified can probably also be found worldwide, thus much empirical work remains to be done to improve estimates of per capita alcohol consumption.

Electronic supplementary material

The online version of this article (doi:10.1186/s12963-016-0090-4) contains supplementary material, which is available to authorized users.

Measuring Alcohol Consumption in Population Surveys: A Review of International Guidelines and Comparison with Surveys in England

AIMS

To review the international guidelines and recommendations on survey instruments for measurement of alcohol consumption in population surveys and to examine how national surveys in England meet the core recommendations.

METHODS

A systematic search for international guidelines for measuring alcohol consumption in population surveys was undertaken. The common core recommendations for alcohol consumption measures and survey instruments were identified. Alcohol consumption questions in national surveys in England were compared with these recommendations for specific years and over time since 2000.

RESULTS

Four sets of international guidelines and three core alcohol consumption measures (alcohol consumption status, average volume of consumption, frequency and volume of heavy episodic drinking) with another optional measure (drinking context) were identified. English national surveys have been inconsistent over time in including questions that provide information on average volume of consumption but have not included questions on another essential alcohol consumption measure, frequency of heavy episodic drinking. Instead, they have used questions that focus only on maximum volume of alcohol consumed on any day in the previous week.

CONCLUSIONS

International guidelines provide consistent recommendations for measuring alcohol consumption in population surveys. These recommendations have not been consistently applied in English national surveys, and this has contributed to the inadequacy of survey measurements for monitoring vital aspects of alcohol consumption in England over recent years.

The importance of post hoc approaches for overcoming non-response and attrition bias in population-sampled studies

Population-based health studies are critical resources for monitoring population health and related factors such as substance use, but reliable inference can be compromised in various ways. Non-response and attrition are major methodological problems which reduce power and can hamper the generalizability of findings if individuals who participate and who remain in a study differ systematically from those who do not. In this issue of SPPE, McCabe et al. studied participants of the 2001-2002 National Epidemiologic Survey on Alcohol and Related Conditions, comparing attrition in Wave 2 across participants with different patterns of substance use at Wave 1. The implications of differential follow-up and further possibilities for addressing selective participation are discussed.

What is wrong with non-respondents? Alcohol-, drug- and smoking-related mortality and morbidity in a 12-year follow-up study of respondents and non-respondents in the Danish Health and Morbidity Survey

AIM

Response rates in health surveys have diminished over the last two decades, making it difficult to obtain reliable information on health and health-related risk factors in different population groups. This study compared cause-specific mortality and morbidity among survey respondents and different types of non-respondents to estimate alcohol-, drug- and smoking-related mortality and morbidity among non-respondents.

DESIGN

Prospective follow-up study of respondents and non-respondents in two cross-sectional health surveys.

SETTING

Denmark.

PARTICIPANTS

A total sample of 39 540 Danish citizens aged 16 years or older.

MEASUREMENTS

Register-based information on cause-specific mortality and morbidity at the individual level was obtained for respondents (n = 28 072) and different types of non-respondents (refusals n = 8954; illness/disabled n = 731, uncontactable n = 1593). Cox proportional hazards models were used to examine differences in alcohol-, drug- and smoking-related mortality and morbidity, respectively, in a 12-year follow-up period.

FINDINGS

Overall, non-response was associated with a significantly increased hazard ratio (HR) of 1.56 [95% confidence interval (CI) = 1.36-1.78] for alcohol-related morbidity, 1.88 (95% CI = 1.38-2.57) for alcohol-related mortality, 1.55 (95% CI = 1.27-1.88) for drug-related morbidity, 3.04 (95% CI = 1.57-5.89) for drug-related mortality and 1.15 (95% CI = 1.03-1.29) for smoking-related morbidity. The hazard ratio for smoking-related mortality also tended to be higher among non-respondents compared with respondents, although no significant association was evident (HR = 1.14; 95% CI = 0.95-1.36). Uncontactable and ill/disabled non-respondents generally had a higher hazard ratio of alcohol-, drug- and smoking-related mortality and morbidity compared with refusal non-respondents.

CONCLUSION

Health survey non-respondents in Denmark have an increased hazard ratio of alcohol-, drug- and smoking-related mortality and morbidity compared with respondents, which may indicate more unfavourable health behaviours among non-respondents.