Dataset Information

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

ABSTRACT: Health surveys are an important resource for monitoring population health, but selective nonresponse may impede valid inference. This study aimed to assess nonresponse bias in a population-sampled health survey in Scotland, with a focus on alcohol-related outcomes. Nonresponse bias was assessed by examining whether rates of alcohol-related harm (i.e., hospitalization or death) and all-cause mortality among respondents to the Scottish Health Surveys (from 1995 to 2010) were equivalent to those in the general population, and whether the extent of any bias varied according to sociodemographic attributes or over time. Data from consenting respondents (aged 20-64 years) to 6 Scottish Health Surveys were confidentially linked to death and hospitalization records and compared with general population counterparts. Directly age-standardized incidence rates of alcohol-related harm and all-cause mortality were lower among Scottish Health Survey respondents compared with the general population. For all years combined, the survey-to-population rate ratios were 0.69 (95% confidence interval: 0.61, 0.76) for the incidence of alcohol-related harm and 0.89 (95% confidence interval: 0.83, 0.96) for all-cause mortality. Bias was more pronounced among persons residing in more deprived areas; limited evidence was found for regional or temporal variation. This suggests that corresponding underestimation of population rates of alcohol consumption is likely to be socially patterned.

SUBMITTER: Gorman E

PROVIDER: S-EPMC4207717 | biostudies-other | 2014 Nov

REPOSITORIES: biostudies-other

ACCESS DATA

Publications

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

Gorman Emma E Leyland Alastair H AH McCartney Gerry G White Ian R IR Katikireddi Srinivasa Vittal SV Rutherford Lisa L Graham Lesley L Gray Linsay L

American journal of epidemiology 20140916 9

Health surveys are an important resource for monitoring population health, but selective nonresponse may impede valid inference. This study aimed to assess nonresponse bias in a population-sampled health survey in Scotland, with a focus on alcohol-related outcomes. Nonresponse bias was assessed by examining whether rates of alcohol-related harm (i.e., hospitalization or death) and all-cause mortality among respondents to the Scottish Health Surveys (from 1995 to 2010) were equivalent to those in ...[more]

PMID: 25227767

Similar Datasets

Project description:BACKGROUND:QUALICOPC is an international survey of primary care performance. QUALICOPC data have been used in several studies, yet the representativeness of the Canadian QUALICOPC survey is unknown, potentially limiting the generalizability of findings. This study examined the representativeness of QUALICOPC physician and patient respondents in Ontario using health administrative data. METHODS:This representativeness study linked QUALICOPC physician and patient respondents in Ontario to health administrative databases at the Institute for Clinical Evaluative Sciences. Physician respondents were compared to other physicians in their practice group and all Ontario primary care physicians on demographic and practice characteristics. Patient respondents were compared to other patients rostered to their primary care physicians, patients rostered to their physicians' practice groups, and a random sample of Ontario residents on sociodemographic characteristics, morbidity, and health care utilization. Standardized differences were calculated to compare the distribution of characteristics across cohorts. RESULTS:QUALICOPC physician respondents included a higher proportion of younger, female physicians and Canadian medical graduates compared to other Ontario primary care physicians. A higher proportion of physician respondents practiced in Family Health Team models, compared to the provincial proportion for primary care physicians. QUALICOPC patient respondents were more likely to be older and female, with significantly higher levels of morbidity and health care utilization, compared with the other patient groups examined. However, when looking at the QUALICOPC physicians' whole rosters, rather than just the patient survey respondents, the practice profiles were similar to those of the other physicians in their practice groups and Ontario patients in general. CONCLUSIONS:Comparisons revealed some differences in responding physicians' demographic and practice characteristics, as well as differences in responding patients' characteristics compared to the other patient groups tested, which may have resulted from the visit-based sampling strategy. Ontario QUALICOPC physicians had similar practice profiles as compared to non-participating physicians, providing some evidence that the participating practices are representative of other non-participating practices, and patients selected by visit-based sampling may also be representative of visiting patients in other practices. Those using QUALICOPC data should understand this limited representativeness when generalizing results, and consider the potential for bias in their analyses.

Project description:IntroductionThe aims of this program of research are to use linked health and law enforcement data to describe individuals presenting to emergency and inpatient healthcare services with an acute alcohol harm or problematic alcohol use; measure their health service utilisation and law enforcement engagement; and quantify morbidity, mortality, offending and incarceration.Methods and analysisWe will assemble a retrospective cohort of people presenting to emergency departments and/or admitted to hospitals between 1 January 2005 and 31 December 2014 in New South Wales, Australia with a diagnosis denoting an acute alcohol harm or problematic alcohol use. We will link these data with records from other healthcare services (eg, community-based mental healthcare data, cancer registry), mortality, offending and incarceration data sets. The four overarching areas for analysis comprise: (1) describing the characteristics of the cohort at their first point of contact with emergency and inpatient hospital services in the study period with a diagnosis indicating an acute alcohol harm and/or problematic alcohol use; (2) quantifying health service utilisation and law enforcement engagement; (3) quantifying rates of mortality, morbidity, offending and incarceration; and (4) assessing predictors (eg, age, sex) of mortality, morbidity, offending and incarceration among this cohort.Ethics and disseminationEthics approval has been provided by the New South Wales Population and Health Services Research Ethics Committee. We will report our findings in accordance with the REporting of studies Conducted using Observational Routinely collected health Data (RECORD) statement and Guidelines for Accurate and Transparent Health Estimates Reporting (GATHER) where appropriate. We will publish data in tabular, aggregate forms only. We will not disclose individual results. We will disseminate project findings at scientific conferences and in peer-reviewed journals. We will aim to present findings to relevant stakeholders (eg, addiction medicine and emergency medicine specialists, policy makers) to maximise translational impact of research findings.

Project description:To develop a multivariate method for quantifying the population representativeness across related clinical studies and a computational method for identifying and characterizing underrepresented subgroups in clinical studies.We extended a published metric named Generalizability Index for Study Traits (GIST) to include multiple study traits for quantifying the population representativeness of a set of related studies by assuming the independence and equal importance among all study traits. On this basis, we compared the effectiveness of GIST and multivariate GIST (mGIST) qualitatively. We further developed an algorithm called "Multivariate Underrepresented Subgroup Identification" (MAGIC) for constructing optimal combinations of distinct value intervals of multiple traits to define underrepresented subgroups in a set of related studies. Using Type 2 diabetes mellitus (T2DM) as an example, we identified and extracted frequently used quantitative eligibility criteria variables in a set of clinical studies. We profiled the T2DM target population using the National Health and Nutrition Examination Survey (NHANES) data.According to the mGIST scores for four example variables, i.e., age, HbA1c, BMI, and gender, the included observational T2DM studies had superior population representativeness than the interventional T2DM studies. For the interventional T2DM studies, Phase I trials had better population representativeness than Phase III trials. People at least 65years old with HbA1c value between 5.7% and 7.2% were particularly underrepresented in the included T2DM trials. These results confirmed well-known knowledge and demonstrated the effectiveness of our methods in population representativeness assessment.mGIST is effective at quantifying population representativeness of related clinical studies using multiple numeric study traits. MAGIC identifies underrepresented subgroups in clinical studies. Both data-driven methods can be used to improve the transparency of design bias in participation selection at the research community level.

Project description:BackgroundHistorically, counting influenza recorded in administrative health outcome databases has been considered insufficient to estimate influenza attributable morbidity and mortality in populations. We used database record linkage to evaluate whether modern databases have similar limitations.MethodsPerson-level records were linked across databases of laboratory notified influenza, emergency department (ED) presentations, hospital admissions and death registrations, from the population (?6.9 million) of New South Wales (NSW), Australia, 2005 to 2008.ResultsThere were 2568 virologically diagnosed influenza infections notified. Among those, 25% of 40 who died, 49% of 1451 with a hospital admission and 7% of 1742 with an ED presentation had influenza recorded on the respective database record. Compared with persons aged ?65 years and residents of regional and remote areas, respectively, children and residents of major cities were more likely to have influenza coded on their admission record. Compared with older persons and admitted patients, respectively, working age persons and non-admitted persons were more likely to have influenza coded on their ED record. On both ED and admission records, persons with influenza type A infection were more likely than those with type B infection to have influenza coded. Among death registrations, hospital admissions and ED presentations with influenza recorded as a cause of illness, 15%, 28% and 1.4%, respectively, also had laboratory notified influenza. Time trends in counts of influenza recorded on the ED, admission and death databases reflected the trend in counts of virologically diagnosed influenza.ConclusionsA minority of the death, hospital admission and ED records for persons with a virologically diagnosed influenza infection identified influenza as a cause of illness. Few database records with influenza recorded as a cause had laboratory confirmation. The databases have limited value for estimating incidence of influenza outcomes, but can be used for monitoring variation in incidence over time.

Dataset Information

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

Publications

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

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure