Bias due to non-participation and heterogenous sub-groups in population surveys

Factors Influencing Community Participation in Internet Interventions Compared With Research Trials: Observational Study in a Nationally Representative Adult Cohort

BACKGROUND

Digital mental health (DMH) programs can be effective in treating and preventing mental health problems. However, community engagement with these programs can be poor. Understanding the barriers and enablers of DMH program use may assist in identifying ways to increase the uptake of these programs, which have the potential to provide broad-scale prevention and treatment in the community.

OBJECTIVE

In this study, we aimed to identify and compare factors that may influence participation in DMH programs in practice and research trials, identify any respondent characteristics that are associated with these factors, and assess the relationship between intentions to use DMH programs and actual uptake.

METHODS

Australian adults aged ≥18 years were recruited from market research panels to participate in the study. The sample was representative of the Australian adult population based on age, gender, and location. Participants completed a cross-sectional web-based survey assessing demographic characteristics, mental health symptom measures, attitudes and use of DMH programs in practice and in research studies, and the factors influencing their use in both settings.

RESULTS

Across both research and practice, trust in the organization delivering the service or trial was the top-ranked factor influencing participation, followed by anonymity or privacy and adequate information. There was little variation in rankings across demographic groups, including intentions to use DMH programs or mental health status. Intentions to use DMH programs were a strong predictor of both current (odds ratio 2.50, 99% CI 1.41-4.43; P<.001) and past (odds ratio 2.98, 99% CI 1.71-5.19; P<.001) use behaviors.

CONCLUSIONS

Efforts to increase the uptake of DMH programs or participation in research trials should focus on clearly communicating the following to users: the legitimacy of the organization delivering the program, security and use of participant data, and effectiveness of DMH programs.

Visual acuity and causes of central visual loss in the adult Danish population 2020-2022. Results from the FORSYN study

PURPOSE

Knowledge of visual health in the population is necessary for designing and implementing measures to handle visual handicap. The purpose of the FORSYN (Forekomst af synshandicap og synshjaelpemidler i Danmark) project was to study visual health in the Danish population 2020-2022 after implementation of the recent advances in the management of choroidal and retinal vascular disease. The present study reports visual acuity and causes of central visual loss from this study.

METHODS

A population-representative sample of 10 350 citizens living within 40 kilometres from Aarhus University Hospital were invited to answer a questionnaire about quality of life related to vision, measurement of visual acuity and a non-mydriatic examination of the eyes. The data were corrected for selection bias on the basis of demographic and socioeconomic factors so that the results could be projected to represent the adult Danish population.

RESULTS

Population-adjusted visual acuity in ETDRS letters (mean ± SD) differed significantly (p < 0.0001) among the worse eye (84.1 ± 0.25), the better eye (88.4 ± 0.11) and binocularly (89.2 ± 0.15). Social blindness affected 0.22% (95% CI: 0.14%-0.33%) of the population and was in none of the studied persons due to exudative age-related macular degeneration (AMD) or diabetic retinopathy. The most frequent causes of visual loss were atrophic AMD, neuro-ophthalmic disorders and other chorioretinal diseases.

CONCLUSIONS

Recent advances in the therapy of chorioretinal vascular diseases have been paralleled with a reduction in central vision loss secondary to exudative AMD and diabetic retinopathy in Denmark. The demographic development can be expected to increase the demand for treatments of vision-threatening diseases that mainly affect older persons.

Representativeness in population-based studies of older adults: five waves of cross-sectional examinations in the Gothenburg H70 Birth Cohort Study

OBJECTIVES

To describe representativeness in the Gothenburg H70 1930 Birth Cohort Study.

DESIGN

Repeated cross-sectional examinations of a population-based study.

SETTING

Gothenburg, Sweden.

PARTICIPANTS

All residents of Gothenburg, Sweden, born on specific birth dates in 1930 were invited to a comprehensive health examination at ages 70, 75, 79, 85 and 88. The number of participants at each examination was 524 at age 70, 767 at age 75, 580 at age 79, 416 at age 85, and 258 at age 88.

PRIMARY OUTCOME MEASURES

We compared register data on sociodemographic characteristics and hospital discharge diagnoses between participants and (1) refusals, (2) all same-aged individuals in Gothenburg and (3) all same-aged individuals in Sweden. We also compared mortality rates between participants and refusals.

RESULTS

Refusal rate increased with age. At two or more examination waves, participants compared with refusals had higher educational level, more often had osteoarthritis, had lower mortality rates, had lower prevalence of neuropsychiatric, alcohol-related and cardiovascular disorders, and were more often married. At two examination waves, participants compared with same-aged individuals in Gothenburg had higher education and were more often born in Sweden. At two examination waves or more, participants compared with same-aged individuals in Sweden had higher education, had higher average income, less often had ischaemic heart disease, were less often born in Sweden and were more often divorced.

CONCLUSIONS

Participants were more similar to the target population in Gothenburg than to refusals and same-aged individuals in Sweden. Our study shows the importance of having different comparison groups when assessing representativeness of population studies, which is important in evaluating generalisability of results. The study also contributes unique and up-to-date knowledge about participation bias in these high age groups.

Risk of subarachnoid haemorrhages and aneurysms in Danish People Living With HIV, a nationwide cohort study

OBJECTIVE

It remains unclear whether people with HIV (PWH) have increased risk of aneurysms. We aimed to investigate if the risk of subarachnoid haemorrhage, cerebral aneurysm, aortic aneurysm and other arterial aneurysms and dissections is increased in PWH compared with the general population.

DESIGN

We performed a nationwide population-based matched cohort study.

METHODS

We compared PWH with age-matched and sex-matched comparison cohort members to calculate incidence rate ratios (IRR) of subarachnoid haemorrhage, cerebral aneurysm, aortic aneurysm and other arterial aneurysms and dissections as well as surgery for these conditions.

RESULTS

We included all PWH, who were Danish residents and treated at a HIV healthcare centre between January 1995 and February 2018 (n = 6935) and an age-matched and sex-matched comparison cohort (n = 55 480). PWH had increased risk of subarachnoid haemorrhage (IRR 1.7, 95% CI, 1.1-2.6), but no increased risk of surgery for this condition. PWH had no increased risk of cerebral aneurysm or aortic aneurysm or surgery for these conditions. The risk of other arterial aneurysms and dissections was increased in PWH (IRR 2.0, 95% CI, 1.4-2.9), but this only applied for PWH who reported intravenous substance use as route of HIV transmission (IRR 18.4, 95% CI, 9.3-36.6), and not for PWH without reported injection drug use (IRR 1.2, 95% CI, 0.73-1.7).

CONCLUSION

PWH were not at an increased risk of cerebral, aortic or other arterial aneurysms and dissections. Although PWH were at an increased risk of subarachnoid haemorrhage, subarachnoid haemorrhages were rare among PWH.

Revisiting the continuum of resistance model in the digital age: a comparison of early and delayed respondents to the Norwegian counties public health survey

Background

The continuum of resistance model’s premise is that delayed respondents to a survey are more similar to non-respondents than early respondents are. For decades, survey researchers have applied this model in attempts to evaluate and adjust for non-response bias. Despite a recent resurgence in the model’s popularity, its value has only been assessed in one large online population health survey.

Methods

Respondents to the Norwegian Counties Public Health Survey in Hordaland, Norway, were divided into three groups: those who responded within 7 days of the initial email/SMS invitation (wave 1, n = 6950); those who responded after 8 to 14 days and 1 reminder (wave 2, n = 4950); and those who responded after 15 or more days and 2 reminders (wave 3, n = 4045). Logistic regression analyses were used to compare respondents’ age, sex and educational level between waves, as well as the prevalence of poor general health, life dissatisfaction, mental distress, chronic health problems, weekly alcohol consumption, monthly binge drinking, daily smoking, physical activity, low social support and receipt of a disability pension.

Results

The overall response to the survey was 41.5%. Respondents in wave 1 were more likely to be older, female and more highly educated than those in waves 2 and 3. However, there were no substantial differences between waves for any health outcomes, with a maximal prevalence difference of 2.6% for weekly alcohol consumption (wave 1: 21.3%, wave 3: 18.7%).

Conclusions

There appeared to be a mild continuum of resistance for demographic variables. However, this was not reflected in health and related outcomes, which were uniformly similar across waves. The continuum of resistance model is unlikely to be useful to adjust for nonresponse bias in large online surveys of population health.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12889-021-10764-2.

Sickness absenteeism, work performance, and healthcare use due to respiratory infections for shift and non-shift workers

This study aimed to compare sickness absenteeism, work performance, and healthcare use due to respiratory infections, as well as general sickness absenteeism and work performance between shift and non-shift workers. In this study, 589 shift and non-shift workers employed in hospitals were included. For 6 months, participants kept a daily record of their influenza-like illness/acute respiratory infection (ILI/ARI) symptoms using a diary application. After an episode of ILI/ARI symptoms ended, participants (n = 531) were questioned about their sickness absenteeism (occurrence and duration in hours), work performance (on a 10 point scale), and healthcare use during the ILI/ARI episode. At the end of the 6 months follow-up, participants (n = 498) were also asked about general sickness absenteeism and work performance in the past 4 weeks. Mixed-model and regression analyses were used to compare absenteeism, work performance, and healthcare use between shift and non-shift workers. No differences were found in sickness absenteeism [Odds Ratio (OR) = 1.00 (95%‒Confidence Interval (CI): 0.61‒1.64)] and work performance [Regression coefficient (B) = -0.19 (95%‒CI: -0.65‒0.26)] due to ILI/ARI between shift and non-shift workers. In addition, healthcare use due to ILI/ARI was similar between shift and non-shift workers. Furthermore, similar general sickness absenteeism rates and work performance levels were found between shift and non-shift workers. As this is the first study that examined the associations with shift work due to ILI/ARI, further studies are needed to confirm our findings.

Practice of reporting social characteristics when describing representativeness of epidemiological cohort studies - A rationale for an intersectional perspective

Representativeness has been defined as the degree of similarity of a study population compared to an external population. To characterize a study population, both health-related and social or demographic features should be considered according to current guidelines. However, little guidance is given on how to describe social complexity of study populations when aiming to conclude on representativeness. We argue that sociological concepts should inform characterizations of study populations in order to increase credibility of conclusions on representativeness. The concept of intersectionality suggests to conceptualize social location as a combination of characteristics such as sex/gender and ethnicity instead of focusing on each feature independently. To contextualize advantages of integrating the concept of intersectionality when investigating representativeness, we reviewed publications that described the baseline population of selected epidemiological cohort studies. Information on the applied methods to characterize the study population was extracted, as well as reported social characteristics. Nearly all reviewed studies reported descriptive statistics of the baseline population and response proportions. In most publications, study populations were characterized according to place of residence, age and sex/gender while other social characteristics were reported irregularly. Differential patterns of representativeness were revealed in analyses that stratified social characteristics by sex/gender or age. Furthermore, the included studies did not explicitly state the theoretical approach that underlay their description of the study population. Intersectionality might be particularly fruitful when applied to descriptions of representativeness, because this concept provides an understanding of social location that has been developed based on situated experiences of people at the intersection of multiple axes of social power relations. An intersectional perspective, hence, contributes to approximate social complexity of study populations and might contribute to increase validity of conclusions on representativeness of population-based studies.

Short- and Longer-Term Benefits of Temporary Alcohol Abstinence During ‘Dry January’ Are Not Also Observed Among Adult Drinkers in the General Population: Prospective Cohort Study

Aims

The alcohol abstinence challenge ‘Dry January’ continues to grow, but there is a lack of knowledge of how Dry January participants compare to the general population. There is also a need to determine whether benefits experienced by Dry January participants are unique to that group or are also observed among other people.

Methods

We conducted a prospective cohort study using online questionnaires in early January, February and August 2019. We compared 1192 Dry January participants and 1549 adult drinkers who did not attempt to abstain from alcohol. Key outcomes were self-rated physical health, psychological well-being (Warwick–Edinburgh Mental Well-Being Scale), control over drinking (Drink Refusal Self-Efficacy Scale (DRSE)) and alcohol intake (Alcohol Use Disorders Identification Test–Consumption (AUDIT-C) subscale). Baseline differences in demographic and alcohol consumption variables were included as covariates in between-group analyses.

Results

Dry January participants had higher SES, poorer well-being, higher AUDIT-C scores and less control over their drinking than the general population. Beneficial changes in health, WEMWBS, DRSE and AUDIT-C observed among people completing Dry January were not observed among other adult drinkers.

Conclusions

Dry January appears to attract people who are heavier drinkers than the general population and who are more concerned about their alcohol intake. Completion of Dry January is associated with short- and longer-term benefits to well-being that are not observed in the general population.

Assessment of participation biases for a confidential non-anonymous adolescent study: A based population study

A prospective study often receives a low participation rate that may alter the results quality. This study assessed the participation bias for a confidential non-anonymous adolescent survey among 1559 middle-school adolescents from north-eastern France (mean age 13.5 ± 1.3). They completed an anonymous questionnaire gathering demographic/socioeconomic features as well as school, behavior and health-related difficulties, and adolescent's assent to participate with perceived parents' consent (APC) if they were contacted for a confidential non-anonymous survey at home. Such a survey received an APC of 60%. The logistic model including all socioeconomic factors and school, behavior and health-related difficulties showed that the adolescents with APC were less often male (adjusted odds ratio = 0.77, p = 0.014), non-European immigrant (0.48, p = 0.016), living with a single parent (0.72, p = 0.046), in manual-worker families (0.69, p = 0.007), had less often low parents' education (0.70, p = 0.002), body-mass-index measurement refusal (0.60, p = 0.010), no regular physical/sports activity (0.70, p = 0.035), poor social relationships (0.73, p = 0.046) and poor living environment (0.63, p = 0.007). The percentage of subjects with APC steadily decreased with the number of these criteria: from 74% for 0 criterion to 19% for 6-8 criteria. Because of these possible strong participation biases the construction of adolescent cohorts and the results interpretation should be made with prudence.

Perceived discrimination In Primary Healthcare in Europe: evidence from the cross-sectional QUALICOPC study

Recent figures show that discrimination in healthcare is still persistent in the European Union. Research has confirmed these results but focused mainly on the outcomes of perceived discrimination. Studies that take into account socioeconomic determinants of discrimination limit themselves to either ethnicity, income or education. This article explores the influence of several socioeconomic indicators (e.g. gender, age, income, education and ethnicity) on perceived discrimination in 30 European countries. Data from the QUALICOPC study were used. These data were collected between October 2011 and December 2013 in the participating countries. In total, 7183 GPs (general practitioners) and 61932 patients participated in the study, which had an average response rate of 74.1%. Data collection was co-ordinated by NIVEL (Dutch Institute for Research of Health Care). Bivariate binomial logistic regressions were used to estimate the impact of each socioeconomic indicator on perceived discrimination. Multivariate logistic regressions were used to estimate the unique effect of each indicator. Results indicate that in Europe, overall 7% of the respondents felt discriminated, ranging between 1.4% and 12.8% at the country level. With regard to socioeconomic determinants in perceived discrimination, income and age are both important indicators, with lower income groups and younger people having a higher chance to feel discriminated. In addition, we find significant influences of education, gender, age and ethnicity in several countries. In most countries, higher educated people, older people, women and the indigenous population appeared to feel less discriminated. In conclusion, perceived discrimination in healthcare is reported in almost all European countries, but there is large variation between European countries. A high prevalence of perceived discrimination within a country also does not imply a correlation between socioeconomic indicators and perceived discrimination.

Registry-based analysis of participator representativeness: a source of concern for sickness absence research?

OBJECTIVES

Selective participation can bias results in epidemiological surveys. The importance of health status is often suggested as a possible explanation for non-participation but few empirical studies exist. In a population-based study, explicitly focused on sickness absence, health and work, we examined whether a history of high levels of sickness absence was associated with non-participation.

DESIGN

The study is based on data from official sickness absence registers from participants, non-participants and the total target population of the baseline survey of the Health Assets Project (HAP).

SETTING

HAP is a population-based cohort study in the Västra Götaland region in South Western Sweden.

PARTICIPANTS

HAP included a random population cohort (n=7984) and 2 cohorts with recent sickness absence (employees (n=6140) and non-employees (n=990)), extracted from the same overall general working-age population.

PRIMARY OUTCOME MEASURES

We examined differences in participation rates between cohorts (2008), and differences in previous sickness absence (2001-2008) between participants (individual-level data) and non-participants or the target population (group-level data) within cohorts.

RESULTS

Participants had statistically significant less registered sickness absence in the past than non-participants and the target population for some, but not all, of the years analysed. Yet these differences were not of substantial size. Other factors than sickness absence were more important in explaining differences in participation, whereby participants were more likely to be women, older, born in Nordic countries, married and have higher incomes than non-participants.

CONCLUSIONS

Although specifically addressing sickness absence, having such experience did not add any substantial layer to selective participation in the present survey. Detailed measures are needed to gain a better understanding for health selection in health-related surveys such as those addressing sickness absence, for instance in order to discriminate between selection due to ability or motivation for participation.

Interprofessional clinical training improves self-efficacy of health care students

BACKGROUND

Interprofessional collaboration potentially enhances patient safety and satisfaction, and reduces tensions and conflicts among health professionals. However, health professionals often lack sufficient knowledge of other professional roles and competences to engage in interprofessional teamwork. The aim of this study was to assess the impact of an interprofessional training programme on students' perceived self-efficacy.

METHODS

A quasi-experimental study with an intervention group (239 students) and a control group (405 students). The intervention was an interprofessional clinical study (ICS) unit including students from nursing, medicine, physiotherapy, occupational therapy, laboratory technology and radiography. Data on students' perceived self-efficacy were collected through web-based questionnaires. Aspects of self-efficacy measured were: (1) collaboration with other professions in planning goals and actions for patients; (2) collaboration with other professions for rehabilitation; (3) identifying the functions of other professions and (4) assessing and describing patients' needs and problems.

RESULTS

All scores of perceived self-efficacy for the ICS group improved over time although one score change was non-significant (p = 0.08). After adjustment for baseline differences and the score change for the control group, the ICS group's self-efficacy score gain remained statistically significant.

CONCLUSION

The study showed that interprofessional training improved students' perception of self-efficacy more than traditional clinical training.

The Parental Dental Concerns Scale (PDCS): its development and initial psychometric properties

OBJECTIVES

To assess the validity and reliability of a Parental Dental Concerns Scale (PDCS) to identify parents unable to access preventive dental care for their children.

METHODS

Two studies were conducted. In Study One, a purposive convenience sample of 399 Scottish parents answered questions on going to the dentist, family life and demographics. Parents were retested eight weeks later. In Study Two, 574 Scottish parents participating in a preventive oral health programme were posted the same questionnaire. Information on child dental attendance was gained from dental records. Data were analysed using exploratory (EFA) and confirmatory (CFA) factor analysis. Internal consistency and test-retest correlations provided reliability estimates. Validity was assessed with confirmatory factor analysis, correlations and independent t-tests.

RESULTS

EFA indicated that the PDCS had a four factor structure, supported by a subsequent CFA. The PDCS and its four subscales had good internal consistency, concurrent validity and test-retest reliability. Further work is required to confirm the scale's predictive validity in discriminating between children and parents who did and did not attend the dental practice.

CONCLUSIONS

The PDCS is a reliable scale, which demonstrates good construct validity. Further testing is required to confirm its predictive validity.

Difficulties with telephone-based surveys on alcohol consumption in high-income countries: the Canadian example

Accurate information concerning alcohol consumption level and patterns is vital to formulating public health policy. The objective of this paper is to critically assess the extent to which survey design, response rate and alcohol consumption coverage obtained in random digit dialling, telephone-based surveys impact on conclusions about alcohol consumption and its patterns in the general population. Our analysis will be based on the Canadian Alcohol and Drug Use Monitoring Survey (CADUMS) 2008, a national survey intended to be representative of the general population. The conclusions of this paper are as follows: (1) ignoring people who are homeless, institutionalized and/or do not have a home phone may lead to an underestimation of the prevalence of alcohol consumption and related problems; (2) weighting of observations to population demographics may lead to a increase in the design effect, does not necessarily address the underlying selection bias, and may lead to overly influential observations; and (3) the accurate characterization of alcohol consumption patterns obtained by triangulating the data with the adult per capita consumption estimate is essential for comparative analyses and intervention planning especially when the alcohol coverage rate is low like in the CADUMS with 34%.

A Comparison of Early and Late Respondents in a Twin–Family Survey Study

Differences between early (within 30 days) and late (after 30 days) respondents in a survey study were analyzed in twins and siblings registered with the Netherlands Twin Register. We compared early and late respondents on personality traits, health, lifestyle, and demographic variables. The odds of being a late respondent were significantly higher for men (OR 1.14), alcohol use on a daily/weekly basis (OR 1.20), having a relationship (OR 1.40), higher score on experience seeking scale (OR 1.02), and criticizing the questionnaire as too long (OR 1.27). The odds of being a late respondent were significantly lower for nontwin subjects (OR 0.71), regular cycling (OR 0.83), and judging the questionnaire to be fun (OR 0.80). There were no significant interactions with sex. To examine to what extent early and late response is influenced by genetic factors, twin and sibling data of 5040 subjects were analyzed. The best model includes genetic factors (31%), shared environmental influences (36%), and unique environmental influences (43%) on variation in response time.

The effect of smoking on response and drug survival in rheumatoid arthritis patients treated with their first anti-TNF drug

OBJECTIVES

Smoking has been associated with higher disease activity and poor response to anti-tumour necrosis factor (anti-TNF) therapy in patients with rheumatoid arthritis (RA). We wanted to study the effect of smoking on response to therapy, disease activity measures, and drug survival in RA patients starting their first anti-TNF drug.

METHODS

In 2005, RA patients in a voluntary rheumatology biologics register in Southern Sweden answered a questionnaire that included smoking habits. The primary endpoint comprised the European League Against Rheumatism (EULAR) response criteria at 3, 6, and 12 months. Secondary endpoints were the Simplified Disease Activity Index (SDAI), Clinical Disease Activity Index (CDAI) response criteria, and drug survival.

RESULTS

Between 1999 and 2005, 23% of RA patients (216/934) in Southern Sweden were current smokers at the start of anti-TNF therapy. Smoking did not influence disease activity at baseline. Heavy smokers had the poorest drug survival. Current smoking was a negative predictive factor for EULAR response at the 3-month follow-up [odds ratio (OR) 0.53, 95% confidence interval (CI) 0.32-0.87, p = 0.012], and for SDAI response at 3 months (OR 0.45, 95% CI 0.27-0.77, p = 0.003) and 6 months (OR 0.47, 95% CI 0.25-0.88, p = 0.02). A pack-year history of 11-20 was a negative predictive factor for SDAI response at 12 months (OR 0.30, 95% CI 0.13-0.70, p = 0.005). Smokers had higher visual analogue scale (VAS) global scores, C-reactive protein (CRP) levels, and erythrocyte sedimentation rate (ESR) at 3 months.

CONCLUSION

Current smoking was predictive of poor response to anti-TNF treatment for up to 12 months and heavy smokers had the poorest drug survival.

Pain-related fear, perceived harmfulness of activities, and functional limitations in complex regional pain syndrome type I

Numerous studies have shown that pain-related fear is one of the strongest predictors of pain disability in patients with chronic musculoskeletal pain, and there is evidence that the reduction of pain-related fear through an exposure treatment can be associated with restoration of functional abilities in patients with complex regional pain syndrome type I (CRPS-I). These findings suggest that pain-related fear may be associated with functional limitations in neuropathic pain as well. The aim of the current study was to test whether the debilitating role of pain-related fear generalizes to patients with CRPS-I. The results of 2 studies are presented. Study I includes a sample of patients with early CRPS-I referred to an outpatient pain clinic. In Study II, patients with chronic CRPS who are members of a patients' association were invited to participate. The results show that in early CRPS-I, pain severity but not fear of movement/(re)injury as measured with the Tampa Scale for Kinesiophobia was related to functional limitations. In patients with chronic CRPS-I, however, perceived harmfulness of activities as measured with the pictorial assessment method significantly predicted functional limitations beyond and above the contribution of pain severity. Not fear of movement/(re)injury in general, but the perceived harmfulness of activities appears a key factor that might be addressed more systematically in the clinical assessment of patients with CRPS-I. These results support the idea that pain-related fear might be a promising concept in the understanding of pain disability in patients with neuropathic pain.

PERSPECTIVE

This is the first study showing that perceived harmfulness of activities contribute to the functional limitations in CRPS-I. The current findings may help clinicians customizing cognitive-behavioral treatments for patients with chronic neuropathic pain.

Non-participation in the second wave of the PART study on mental disorder and its effects on risk estimates

BACKGROUND

In epidemiological studies, analyses are needed to investigate the consequences of non-response.

AIMS

To analyse the consequences of attrition in the second wave of the population-based PART study, which was performed three years after the first wave.

METHODS

Potential determinants for non-participation obtained from population registers and the first wave were analyzed. The relationships between potential determinants and reduced well-being or depressive mood in the first wave questionnaire were investigated separately for participants and non-participants in the second wave. Samples of respondents to the second wave questionnaire with reduced or not reduced well-being were summoned for interview regarding determinants of distress and disorder. The occurrence of potential determinants was compared between participants and non-participants in both groups.

RESULTS

Low income, low education, non-Nordic origin, not being married and previous psychiatric diagnosis were associated with lower participation rates. These variables were similarly related to depressive mood and low psychological well-being in the first wave among participants and non-participants in the second wave. Potential determinants were not or only weakly related to participation in the interview groups.

CONCLUSION

Although the true prevalence of distress and disorder is underestimated, the true associations between potential determinants and the outcomes seem reasonably well reproduced.

Differences in longitudinal disease and treatment characteristics of patients with rheumatoid arthritis replying and not replying to a postal questionnaire. Experience from a biologics register in southern Sweden

OBJECTIVE

Studies on patients not answering postal questionnaires are scarce. We assessed the demographics and longitudinal disease and treatment characteristics of patients with rheumatoid arthritis (RA) in a Swedish biologics register who replied and who did not reply to a postal questionnaire.

METHODS

In the South Swedish Arthritis Treatment Group register, we have detailed disease severity characteristics at baseline and at followup for rheumatology patients taking biologic drugs. In 2005 a questionnaire on smoking, comorbidities, education, and ethnicity was sent to 1234 RA patients who had started their first biologic drug.

RESULTS

In total, 989 subjects (80%) answered the questionnaire. The 245 (20%) who did not answer generally had more severe RA [higher Disease Activity Score, worse Health Assessment Questionnaire score, higher visual analog scale scores for general health and pain at baseline and at followup, and stopped the drug treatment more frequently (72% vs 53%; p=0.0001)]. There were no statistically significant differences in gender and disease duration between those who replied and those who did not reply, but in general the patients who did not reply were younger.

CONCLUSION

Patients with RA in a Swedish biologics register not replying to a postal questionnaire had more severe RA and stopped biological drug treatment more frequently. Thus a detailed analysis of prospectively collected data can clarify selection bias introduced by subjects who do not answer a postal questionnaire, which may influence the validity and interpretation of results from postal survey studies.

Non-participants in a general population study of men, with special reference to social and alcoholic problems

The bias introduced by non-participation in a study depends on the size and the composition of the non-participant group. Out of 10,000 men invited to a screening examination in a large primary prevention trial in Göteborg, Sweden, 25% did not come to the examination. The non-participants could be shown to be registered by the Board of Social Welfare for social problems and alcohol abuse to a greater extent than the participants in the study. The annual mortality among the non-participants was about twice that of the participants during the entire follow-up period of 11.8 years. The incidence of non-fatal myocardial infarction was not significantly higher in non-participants. Coronary death, on the other hand, was significantly more common among those not attending the examination (3.5 vs. 7.6%). Participants registered for alcoholic problems had coronary death rates approaching those of the non-participants. Sudden coronary death accounted for most of the excess coronary mortality. Most of the excess mortality in the non-participant group was from other causes than cardiovascular diseases and cancer but even so, about one third of the excess deaths could be attributed to coronary heart disease. Possible explanations of this excess coronary mortality include that non-participants may smoke more, more frequently have alcoholic problems and that non-participation may reflect an unwillingness to seek medical care even in the event of illness.

Non-respondents in a post-myocardial infarction trial: characteristics and reasons for refusal

We surveyed the 270 survivors of acute myocardial infarction who refused to participate in the Warfarin Re-Infarction Study (WARIS). Information on medical variables were derived from registration forms completed by hospital staff upon discharge, whereas data on a variety of health conditions and reasons for refusal were gathered by mailed questionnaires, 178 (66%) of which were returned. Some disparities were found when comparing non-respondents and participants, the former showing more potential bad risk factors. The diversities between participants and non-respondents are of yet unknown prognostic importance. However, the presence of such differences imply that information on characteristics of non-respondents in clinical trials is desirable in terms of generalizability of the trial results. Reasons stated for non-participation reflect poor motivation, low mobility and saturation with focusing on disease. A slight co-variation between social status and reasons for refusal was noted.

Non-response to baseline, non-response to follow-up and mortality in the Whitehall II cohort

BACKGROUND

Little is known about the associations between non-response to follow-up surveys and mortality, or differences in these associations by socioeconomic position in studies with repeat data collections.

METHODS

The Whitehall II study of socioeconomic inequalities in health provided response status from five data collection surveys; Phase 1 (1985-88, n = 10 308), Phase 5 (1997-99, n = 6533), and all-cause mortality to 2006. Odd-numbered phases included a medical examination in addition to a questionnaire.

RESULTS

Non-response to baseline and to follow-up phases that included a medical examination was associated with a doubling of the mortality hazard in analyses adjusted for age and sex. Compared with complete responders, responders who missed one or more phases, but completed the last possible phase before they died, had a 38% excess risk of mortality. However, those who missed one or more phases including the last possible phase before death had an excess risk of 127%. There was no evidence that these associations differed by socioeconomic position.

CONCLUSION

In studies with repeat data collections, non-response to follow-up is associated with the same doubling of the mortality risk as non-response to baseline; an association that is not modified by socioeconomic position.

Genome wide association for substance dependence: convergent results from epidemiologic and research volunteer samples

Background

Dependences on addictive substances are substantially-heritable complex disorders whose molecular genetic bases have been partially elucidated by studies that have largely focused on research volunteers, including those recruited in Baltimore. Maryland. Subjects recruited from the Baltimore site of the Epidemiological Catchment Area (ECA) study provide a potentially-useful comparison group for possible confounding features that might arise from selecting research volunteer samples of substance dependent and control individuals. We now report novel SNP (single nucleotide polymorphism) genome wide association (GWA) results for vulnerability to substance dependence in ECA participants, who were initially ascertained as members of a probability sample from Baltimore, and compare the results to those from ethnically-matched Baltimore research volunteers.

Results

We identify substantial overlap between the home address zip codes reported by members of these two samples. We find overlapping clusters of SNPs whose allele frequencies differ with nominal significance between substance dependent vs control individuals in both samples. These overlapping clusters of nominally-positive SNPs identify 172 genes in ways that are never found by chance in Monte Carlo simulation studies. Comparison with data from human expressed sequence tags suggests that these genes are expressed in brain, especially in hippocampus and amygdala, to extents that are greater than chance.

Conclusion

The convergent results from these probability sample and research volunteer sample datasets support prior genome wide association results. They fail to support the idea that large portions of the molecular genetic results for vulnerability to substance dependence derive from factors that are limited to research volunteers.

A pilot study of the effect of providing daily free fruit to primary-school children in Auckland, New Zealand

OBJECTIVE

To determine the uptake of a free fruit provision to low-decile primary-school children by quantitatively assessing changes in fruit intake.

DESIGN

A randomised controlled trial using a paired, cluster randomisation.

SETTING

Twenty low-decile primary schools (schools attended by the most deprived children) in Auckland, New Zealand.

SUBJECTS

In total 2032 children, aged 7-11 years, provided data on at least one occasion.

INTERVENTION

Ten pairs of low-decile primary schools matched by roll size and location were randomly allocated to control (no free fruit) or intervention (free fruit) for a school term. Dietary assessments using the 24 h recall methodology were made at baseline, on the last week of the intervention and 6 weeks post-intervention.

RESULTS

Fruit intakes in this cohort were lower than the national average with over 40 % reporting no fruit intake at baseline and did not differ between groups. After the free fruit period the intervention group increased school fruit intakes by 0.39 pieces/school d from baseline (P < or = 0.001) and the proportion of children consuming no fruit reduced to 22 %. This increase, however, was not sustained and fruit intakes fell below baseline levels at 6 weeks post-intervention. Control subjects did not significantly alter their fruit intakes throughout the study.

CONCLUSIONS

Improving exposure and accessibility to fruits at school increases fruit intakes of low socio-economic group children, particularly those who do not normally eat fruit. The present pilot study demonstrates some possible negative effects of short-term free fruit interventions, but is informative for developing and evaluating sustained fruit intervention programmes.

Participant characteristics that influence consent for genetic research in a population-based survey: the Baltimore epidemiologic catchment area follow-up

BACKGROUND

The purpose of this study is to investigate the sociodemographic and health characteristics associated with the willingness to donate a DNA sample, and consent to testing and long-term storage of that sample, among participants in a longitudinal community-based survey.

SAMPLE

Eighty-three percent of the 1,071 participants interviewed in 2004/5 agreed to donate a biological specimen (blood or buccal).

RESULTS

Age was consistently inversely associated with the willingness to allow genetic testing (OR 0.97; p < 0.05), but was unrelated to the willingness to donate or allow storage. There was no association between race and the consent to donate a specimen, but Blacks were less likely to consent to DNA storage for future research as compared with members of other racial groups (OR 0.50; p < 0.01). Four conditions were listed on the consent form as relevant to the genes targeted for assay. Participants with a family history of 1 or more of these conditions were more likely to donate than those without (OR 1.68; p < 0.01). Participants with a personal history of 1 of the 4 conditions listed were not more or less likely to donate, allow testing or allow storage than respondents without such a history.

CONCLUSIONS

Sociodemographic characteristics were unrelated to the willingness to donate a biological sample. Age, but not race, sex or education, was related to consent to genetic testing. Race, but not age, sex or education, was related to consent to storage. A family history of health conditions listed as relevant to the assays being requested was related to the willingness to donate. Factors that affect the willingness to donate a biological sample in an epidemiologic study are not the same as those associated with the willingness to allow genetic testing or storage of that sample for unspecified future research.

Participation bias and its impact on the assembly of a genetic specimen repository for a myocardial infarction cohort

OBJECTIVE

To assess participation bias in the assembly of a specimen repository for genetic studies and to examine the association of participation with outcome within the Olmsted County myocardial infarction (MI) cohort.

PARTICIPANTS AND METHODS

From January 1, 1979, to May 31, 2006, 3081 persons had MI in Olmsted County, MN. Face-to-face contact was used to recruit patients who were hospitalized for an acute event. Persons who had had an MI before establishment of this repository were contacted by mail. At initial contact, we sought consent to use blood samples for genetic studies. Persons who refused were contacted by mail and were asked to consent to the use of stored tissue samples. For deceased subjects, stored tissue was collected when available.

RESULTS

Of the 3081 persons in the Olmsted County MI cohort, 1994 participated in the study; 1007 (50.5%) blood and 987 (49.5%) tissue specimens were provided. Participants were more likely to be younger men with hypertension, comorbidities, and non-ST-segment elevation MI (all, P<.05). Participants who provided blood specimens were more likely to have non-ST-segment elevation MI and lower Killip class than those who provided tissue. After adjustment for age, sex, hypertension, ST-segment elevation, Killip class, and comorbidities, participation was not associated with outcome. Participants who provided blood specimens were less likely to have heart failure (hazard ratio, 0.49; 95% confidence interval, 0.40-0.59; P<.01) or to die (hazard ratio, 0.16; 95% confidence interval, 0.12-0.21; P<.01) than those who provided tissue.

CONCLUSION

A variety of sources can be used to assemble community specimen repositories. Baseline characteristics differed between participants and nonparticipants and, among participants, by specimen source. Participants who provided blood specimens had better outcomes than those who provided tissue specimens. No survival advantage was observed for participants after combining blood and tissue specimens.

Study design, exposure variables, and socioeconomic determinants of participation in Diet, Cancer and Health: a population-based prospective cohort study of 57,053 men and women in Denmark

AIMS

Diet is considered an important aspect of lifestyle related to cancer development. To contribute further knowledge within this field a Danish prospective cohort study "Diet, Cancer and Health" has been initiated. The aims of this paper are to give a description of the study design, measurement procedures, and differences between participants and non-participants with special reference to socioeconomic characteristics.

METHODS

A total of 160,725 individuals 50-64 years of age living in Copenhagen or Aarhus were invited to participate. Information concerning diet and other lifestyle factors was obtained from 57,053 participants using questionnaires and interviews. Anthropometric measurements were taken and biological material collected. In addition, detailed (selected) socioeconomic information on all invited persons including 103,671 non-participants was obtained from statistical registers in Statistics Denmark.

RESULTS

Differences were seen between participants and non-participants on a number of socioeconomic factors. The highest participation in relation to education was found among participants with higher education, with a significant tendency to be highest in the second highest level of higher education (3-4 years). Married people were more likely to participate than persons living alone or cohabiting.

CONCLUSION

Results from the prospective cohort study "Diet, Cancer and Health" support the general assumption that lower socioeconomic groups are underrepresented in epidemiological studies.

Personality, health and lifestyle in a questionnaire family study: a comparison between highly cooperative and less cooperative families

The effect of nonresponse on health and lifestyle measures has received extensive study, showing at most relatively modest effects. Nonresponse bias with respect to personality has been less thoroughly investigated. The present study uses data from responding individuals as a proxy for the missing data of their nonresponding family members to examine the presence of nonresponse bias for personality traits and disorders as well as health and lifestyle traits. We looked at the Big Five personality traits, borderline personality disorder (BPD) features, attention-deficit/hyperactivity disorder, Anger, and several measures of health (Body Mass Index, migraine) and lifestyle (smoking, alcohol use). In general, outcomes tend to be slightly more favorable for individuals from highly cooperative families compared to individuals from less cooperative families. The only significant difference was found for BPD features (p = .001). However, the absolute difference in mean scores is very small, less than 1 point for a scale ranging from 0 to 72. In conclusion, survey data on personality, health and lifestyle are relatively unbiased with respect to nonresponse.

Non-participation and mortality in different socioeconomic groups: the FINRISK population surveys in 1972-92

BACKGROUND

Declining response rates pose a serious threat to the validity of estimates derived from epidemiological studies. If respondents and non-respondents differ systematically from each other, there can be a bias in the results of the study. A population-based cohort study was conducted to investigate disparities in socioeconomic structure between respondents and non-respondents and the contribution of these disparities to socioeconomic differences in total and cardiovascular mortality.

DESIGN

Data comprised 32,354 male and female participants and 4890 non-participants aged 35-74 years who belonged to the sample in one of the five FINRISK surveys in 1972, 1977, 1982, 1987 or 1992 in Finland. They were followed up for 9 years and 6 months.

RESULTS

It was found that the lower socioeconomic groups were over-represented among non-respondents both in men and women. When comparing the relative risk of death using the highest socioeconomic group of the participants as the reference group, it was found that although the socioeconomic gradient was similar for participants and non-participants-that is, lower groups had a higher risk of death-the risk was at a higher level among non-respondents.

CONCLUSIONS

Basing analysis on participants does not distort the relative risk of death associated with socioeconomic position. However, it does underestimate the absolute risk.

Assessing participation bias in a population-based study of measles-mumps-rubella vaccine immunity in children and adolescents aged 12-18

The purpose of our study was to determine whether specific characteristics were associated with study participation in a group of children residing in Olmsted County, MN. We compared 346 participants and 848 non-participants from a study examining associations between human leukocyte antigen gene variants and immunity following measles-mumps-rubella (MMR) vaccination by demographic characteristics, MMR vaccination history and length of time in the Olmsted County health care system. We also compared the frequency and reasons for health care visits between participants and non-participants by comparing diagnostic codes for all visits that had occurred between 1999 and 2001. Characteristics were compared using chi-square and t-tests, followed by multivariable logistic regression. Study participants were more likely to be white/Caucasian, to have received their first MMR vaccination at a younger age, and to have had more health care visits (especially for acute respiratory illnesses, vaccinations, or other acute conditions such as fainting and headaches) than study non-participants. These results suggest that frequent use of local health care systems by children and parents may increase comfort levels with local physicians and physician researchers, thereby improving participation rates in research studies among these populations. However, special efforts may be necessary to improve research participation among children who are infrequent users of the health care systems of interest.

"To whom do the results of this trial apply?" External validity of a randomized controlled trial involving 130 patients scheduled for primary total hip replacement

BACKGROUND

Although the randomized controlled trial (RCT) is regarded as the gold standard for evaluation of the effect of an intervention, its external validity has been questioned. RCTs cannot be expected to produce results that are directly relevant to all patients and all settings, but they should at least allow patients and clinicians to judge to whom trial results can reasonably be applied. We assessed the external validity of an RCT investigating the efficacy of a fast-track program after total hip replacement.

METHODS

130 patients were identified as potential participants.18 patients were excluded, 33 enrolled patients declined to participate, and 79 patients were enrolled and randomized. We studied the distribution of preoperative characteristics and postoperative clinical variables in these 3 groups.

RESULTS

A significant difference was found in both preoperative characteristics and clinical outcome variables. The non-consenters were older, less healthy, and needed more help from the home care system. Furthermore, they were hospitalized longer and were more often transferred to a rehabilitation ward.

INTERPRETATION

Our findings demonstrate the importance of patient inclusion criteria in RCTs. Moreover, they may account for the lack of reproducibility of RCT results in clinical practice dealing with fast-track programs.

Representativeness in population-based studies: a detailed description of non-response in a Danish cohort study

BACKGROUND

Decreasing rates of participation in population-based studies increasingly challenge the interpretation of study results, in both analytic and descriptive epidemiology. Consequently, estimates of possible differences between participants and non-participants are increasingly important for the interpretation of study results and generalization to the background population.

METHODS

An age-specific, population-based cohort of 1,198 individuals was examined at age 40, 45, 51, and 60. Participants were compared with non-participants and when possible also with the background population using a wide range of detailed information on somatic and mental health collected at each examination, including data from a clinical examination, biochemical measurements, questionnaires, interviews, and public registers.

RESULTS

Participation rates were higher than 80% at examinations at age 40, 45, and 51, but decreased to 65% at age 60. At the baseline investigation at age 40, analyses indicated that participants were representative of the cohort as well as the background population. However, the mortality rate was higher among non-participants in the succeeding 20 years. Among living cohort members at the 60-year examination, non-participants had lower socioeconomic status, higher hospitalization rate, and a worse overall health profile than participants.

CONCLUSIONS

The detailed data presented reinforce the contention that the health profile of non-participants is typically worse than that of participants. The results also indicate that while data from public registers give easily accessible information about non-participants, these crude proxy measures of health may not be enough to document representativeness.

Trends in educational inequalities in cardiovascular risk factors: a longitudinal study among 48,000 middle-aged Norwegian men and women

The aim of the study was to compare educational and gender specific trends in the classical cardiovascular risk factors body mass index (BMI), cholesterol, systolic blood pressure and smoking in a Norwegian population during a period when the coronary heart disease mortality had just reached its peak in the late 1970s and 1980s. We used The Norwegian Counties study: a longitudinal study with three screenings on the same individuals in Oppland, Sogn og Fjordane and Finnmark counties in Norway in the period 1974-1988. All residents aged 35-49 were invited and 48,422 participated (89%) in the baseline screening. To estimate the cardiovascular risk factor change in individuals over time longitudinal statistical methods were used. BMI, cholesterol and blood pressure levels increased with age, while the amount of daily smokers decreased. The higher the educational level the lower the level of BMI, blood pressure, smoking and cholesterol. This pattern persisted through the whole study period. In men, however, the educational gradient in cholesterol diminished in the last screening. Among women there was an increase in inequality in systolic blood pressure, and for smoking there was a steeper decrease for men than women across all educational groups. The educational differences in classical cardiovascular risk factors persisted, except for BMI, cholesterol and daily smokers in men which tended to decrease.

Health, quality of life, social network and use of health care: a comparison between those granted and those not granted disability pensions

PURPOSE

The aim was to compare self-rated health, health-related quality of life (HRQoL), social networks and health care utilisation of those granted disability pension (DP) and those not granted disability pension (nDP).

METHOD

Demographic data and medical diagnoses were obtained from the records of the social insurance office. Data concerning self-reported health, HRQoL social networks, and use of health care were collected by a postal questionnaire. The nDP group included all those not granted full DPs between 1999 and 2000 (n = 99). The DP group were a random sample of those granted full DPs, during the same period (n = 197).

RESULTS

The nDP group had more often multiple diagnoses, and lower self-reported health and HRQoL compared to those granted DP. In particular, their average scores were lower on the SF-36 scales social functioning, role limitations due to physical problems and mental health. The nDP group also had significantly smaller social networks.

CONCLUSIONS

Contrary to expectations, those not granted a disability pension do not seem to have better health, but rather to suffer from more sickness than those who were granted a disability pension.

Effect on trend estimates of the difference between survey respondents and non-respondents: results from 27 populations in the WHO MONICA Project

INTRODUCTION

In the World Health Organization (WHO) MONICA (multinational MONItoring of trends and determinants in CArdiovascular disease) Project considerable effort was made to obtain basic data on non-respondents to community based surveys of cardiovascular risk factors. The first purpose of this paper is to examine differences in socio-economic and health profiles among respondents and non-respondents. The second purpose is to investigate the effect of non-response on estimates of trends.

METHODS

Socio-economic and health profile between respondents and non-respondents in the WHO MONICA Project final survey were compared. The potential effect of non-response on the trend estimates between the initial survey and final survey approximately ten years later was investigated using both MONICA data and hypothetical data.

RESULTS

In most of the populations, non-respondents were more likely to be single, less well educated, and had poorer lifestyles and health profiles than respondents. As an example of the consequences, temporal trends in prevalence of daily smokers are shown to be overestimated in most populations if they were based only on data from respondents.

CONCLUSIONS

The socio-economic and health profiles of respondents and non-respondents differed fairly consistently across 27 populations. Hence, the estimators of population trends based on respondent data are likely to be biased. Declining response rates therefore pose a threat to the accuracy of estimates of risk factor trends in many countries.

Effect of sampling frames on response rates in the WHO MONICA risk factor surveys

Sample surveys are used to investigate occurrence and determinants of diseases in populations. Their reliability is influenced by quality of sampling frame and response rate. We investigated relationship between sampling frame type and response rates and assessed their impact on non-response bias, using data from the WHO MONICA Project, where 37 centres in 20 countries conducted sample surveys, employing the best locally available sampling frame. Sampling frames fell into three categories: Population registers (PR), electoral registers (ER), and health care registers (HR). Response rate (rrs) was factored into components reflecting quality of sampling frame (contact rate cr) and characterizing willingness of sample members to participate (enrolment rate er). The mean quality score for the sampling frames was 92% for PR, 87% for HR and 85% for ER; they contributed on average 23, 20, and 26% to the respective non-response rates. For all frame types and both sexes the lowest quality score occurred in the age group 35 - 44, suggesting a reduced ability to track migration of a highly mobile population group. The patterns in the age/sex distribution of er indicate at least for males in PR and females in HR a potential for non-response bias. Estimation of non-response bias through an abbreviated questionnaire failed because of low item response. We found that contact rate characterizes sampling frame quality. For all frame types it had a major influence on response rate. It is likely that low er and low cr cause different kind of bias, requiring different measures to minimize their effects.

Determinants of non-participation, and the effects of non-participation on potential cause-effect relationships, in the PART study on mental disorders

OBJECTIVE

The study had two objectives: (i) to analyse determinants of non-participation in a general population study of mental disorder in Stockholm, and (ii) to determine whether associations between mental disorder and some potential determinants were different among participants and non-participants in the study.

METHODS

The study was based on a questionnaire including potential risk factors for mental disorder and symptom scales. The study group was a random sample of the Stockholm County population aged 20-64 years (19,742 persons). Replies were obtained from 10,441 participants, i. e. the response rate was 53%. The symptom scales were used to identify a group with increased likelihood of mental disorder, screening-positive, and a group with low likelihood of mental disorder, screening-negative. Random samples of the screening-positive and screening-negative respondents were summoned for interview concerning psychiatric symptoms. The database of respondents and non-respondents was linked to several population registers.

RESULTS

The participation was higher in females, among older persons, married persons, among persons with higher income and education and among those born in the Nordic countries. There seemed to be little further risk indicator based selection to interview participation. The associations between in-patient psychiatric care and register variables were strong and similar among participants and non-participants.

CONCLUSIONS

The occurrence of mental disorders is likely to be underestimated in studies with this design and with substantial dropout rates. However, the study participants can most likely be a base for generalising risk indicators for, or social consequences of,mental disorder, to the general population.

Total and cause specific mortality among participants and non-participants of population based health surveys: a comprehensive follow up of 54 372 Finnish men and women

STUDY OBJECTIVE

To assess total and cause specific mortality among participants and non-participants of large population based health surveys.

DESIGN

A prospective follow up study. Baseline surveys were conducted in 1972, 1977, 1982, 1987, and 1992. Study end points were overall, cardiovascular, cancer and violent mortality, and deaths related to smoking and alcohol. Study cohorts were followed up until the end of 2000 through computerised record linkage. All analyses were adjusted for age.

SETTING

Finland.

PARTICIPANTS

Participants and non-participants of five population based risk factor surveys. The samples included 54 372 men and women aged 25 to 64 years at baseline.

MAIN RESULTS

The average participation rate was 81.7% among men and 87% among women. At eight year follow up, the non-participating men had twice and non-participating women 2.5-fold higher overall mortality than the participating men and women. Non-participants had also significantly higher cause specific mortality, except cancer and smoking related mortality among women. Relative differences in mortality were largest in violent and alcohol related deaths. Non-participants had considerably higher overall mortality than smoking participants, and their mortality was threefold compared with non-smoking participants.

CONCLUSIONS

Observed differences in mortality show that health behaviour and health status substantially differ between non-participants and participants. Low participation rate may considerably bias the results of population based health surveys.

Estimating non-response bias in family studies: application to mental health and lifestyle

Non-response to mailed surveys reduces the effective sample size and may introduce bias. Non-response has been studied by (1) comparison to available data in population based registers, (2) directly contacting non-respondents by telephone or single-item reply cards, and (3) longitudinal repetition of the survey. The goal of this paper was to propose an additional method to study non-response bias: when the variable of interest has a familial component, data from respondents can be used as proxy for the data from their non-responding family members. This approach was used with data on smoking, alcohol consumption, physical activity, coffee- and tea-use, education, body mass index, religion, burnout, life events, personality and mental health in large number of siblings and DZ twins registered with the Netherlands Twin Register. In addition, for smoking behavior, we also used the second strategy by sending a reply card. Results show that scores of members from less cooperative families or incomplete twin pairs tended to be more unfavorable than the scores from highly cooperative families or complete twin pairs. For example, family members from less cooperative families cycled less often and scored higher on anxious depression and neuroticism. For smoking, both the results of the reply card and the results of the additional method suggested a higher percentage smokers among the non-respondents but this was only significant with reply card method. In general, differences between highly/less cooperative families and complete/incomplete DZ twins were small. Results suggest that, even for studies with moderate response rates, data collected on health, personality and lifestyle are relatively unbiased.

The Oslo Health Study: The impact of self-selection in a large, population-based survey

Background

Research on health equity which mainly utilises population-based surveys, may be hampered by serious selection bias due to a considerable number of invitees declining to participate. Sufficient information from all the non-responders is rarely available to quantify this bias. Predictors of attendance, magnitude and direction of non-response bias in prevalence estimates and association measures, are investigated based on information from all 40 888 invitees to the Oslo Health Study.

Methods

The analyses were based on linkage between public registers in Statistics Norway and the Oslo Health Study, a population-based survey conducted in 2000/2001 inviting all citizens aged 30, 40, 45, 59–60 and 75–76 years. Attendance was 46%. Weighted analyses, logistic regression and sensitivity analyses are performed to evaluate possible selection bias.

Results

The response rate was positively associated with age, educational attendance, total income, female gender, married, born in a Western county, living in the outer city residential regions and not receiving disability benefit. However, self-rated health, smoking, BMI and mental health (HCSL) in the attendees differed only slightly from estimated prevalence values in the target population when weighted by the inverse of the probability of attendance.

Observed values differed only moderately provided that the non-attending individuals differed from those attending by no more than 50%. Even though persons receiving disability benefit had lower attendance, the associations between disability and education, residential region and marital status were found to be unbiased. The association between country of birth and disability benefit was somewhat more evident among attendees.

Conclusions

Self-selection according to sociodemographic variables had little impact on prevalence estimates. As indicated by disability benefit, unhealthy persons attended to a lesser degree than healthy individuals, but social inequality in health by different sociodemographic variables seemed unbiased. If anything we would expect an overestimation of the odds ratio of chronic disease among persons born in non-western countries.

Effect of a fruit and vegetable subscription in Danish schools

OBJECTIVE

To measure the effect of a school fruit and vegetable subscription on children's intake of fruit and vegetables after 5 weeks of intervention.

SETTING

Seven primary schools in Denmark.

DESIGN AND METHODS

Intervention schools (n=4) were offered a fruit and vegetable subscription comprising one piece per day. Control schools situated in another municipality were not offered the subscription. Intake of fruit and vegetables was measured at baseline and 5 weeks after the start of the subscription. Two methods were used for dietary assessment: a pre-coded 24-hour recall form including total food intake and a food-frequency questionnaire (FFQ) including only fruit and vegetables.

SUBJECTS

Children aged 6-10 years (n=804 from intervention schools and n=689 from control schools). Response rate in the dietary assessment was 31%.

RESULTS

At intervention schools 45% of the children enrolled in the subscription. After 5 weeks of intervention, both subscribers and non-subscribers had increased their intake of fruit by 0.4 (P=0.019) and 0.3 (P=0.008) pieces per school day, respectively, but no change was observed in vegetable intake. Total intake increased only for non-subscribers by 0.4 piece/school day (P=0.008) mainly due to the consistent increase in fruit intake. No change in intake was measured at control schools. Only the 24-hour recall questionnaire was sensitive enough to pick up the changes of the subscription, whereas the FFQ was not.

CONCLUSION

Five weeks with the subscription affected both subscribers and non-subscribers to increase intake of fruit. This may indicate that the subscription had an additional effect of stimulating parents of non-subscribers to supply their children with fruit. The results stress the importance of evaluating the effect of this type of programme, and the carefulness needed in designing the evaluation study.

Analysis of nonresponse bias in a population-based case-control study on lung cancer

The objective of this study was to identify characteristics of nonrespondents and late respondents in a population-based case-control study on lung cancer conducted in the city of Turin (Italy). Information about demographic and socioeconomic variables of 634 cases and 859 controls who responded to an interview, as well as of 154 cases and 154 controls who did not respond were obtained from census and the public register of Turin. The socioeconomic level of nonrespondents was high in cases but low in controls. Late respondent controls (i.e., individuals who responded after contact through their general practitioner) had socioeconomic characteristics comparable with those of nonrespondents, while they were similar to respondents with respect to demographic variables. The interview of late respondents halved, from 14 to 7%, the magnitude of the bias introduced by nonresponse on the estimate of the association between educational level and lung cancer. Nonresponse, associated with socioeconomic status, is an important potential source of bias in population-based case-control studies, which should always be considered and discussed. The direction and magnitude of the bias can be quantified. General practitioners may contribute to decrease nonresponse bias. Caution should be used in inferring characteristics of nonrespondents on the basis of those of late respondents.

Does non-responder bias have a significant effect on the results in a postal questionnaire study?

BACKGROUND AND AIM

In epidemiological questionnaire studies results can be influenced by non-responder bias. However, in respiratory epidemiology this has been analysed in very few recently published papers. The aim of our paper is to assess if the results found in our previous postal questionnaire study in an adult population in Northern Finland were biased by non-response.

METHODS

A random sample of 385 persons from the 1,284 non-responders in a previous postal questionnaire study was examined. The same questionnaire as in the original study was again mailed to these persons, and those still not answering were contacted by phone.

RESULTS

Totally 183 complete answers (48%) were collected. Lack of interest (56%) and forgetting to mail the response letter (22%) were the most common reasons to non-response. Typical non-responders were young men and current smokers who less frequently reported respiratory symptoms in exercise and asthma than the responders in the original study. Answers collected by phone gave for some questions higher prevalence rates than postal answers.

CONCLUSION

Firstly, in this population the response rate (83.6%) in the original study was high enough to provide reliable results for respiratory symptoms and diseases, only the prevalence of current smoking was biased by non-response. Secondly, the methods used for collecting responses in a non-response study may influence the results.

The Malmö Diet and Cancer Study: representativity, cancer incidence and mortality in participants and non-participants

In order to investigate potential selection bias in population-based cohort studies, participants (n = 28098) and non-participants (n = 40807) in the Malmö Diet and Cancer Study (MDCS) were compared with regard to cancer incidence and mortality. MDCS participants were also compared with participants in a mailed health survey with regard to subjective health, socio-demographic characteristics and lifestyle. Cancer incidence prior to recruitment was lower in non-participants, Cox proportional hazards analysis yielded a relative risk (RR) with a 95% confidence interval of 0.95 (0.90-1.00), compared with participants. During recruitment, cancer incidence was higher in non-participants, RR: 1.08 (1.01-1.17). Mortality was higher in non-participants both during, 3.55 (3.13-4.03), and following the recruitment period, 2.21 (2.03-2.41). The proportion reporting good health was higher in the MDCS than in the mailed health survey (where 74.6% participated), but the socio-demographic structure was similar. We conclude that mortality is higher in non-participants than in participants during recruitment and follow-up. It is also suggested that non-participants may have a lower cancer incidence prior to recruitment but a higher incidence during the recruitment period.