Original Article
A Web-based survey among adults aged 40–54 years was time effective and yielded stable response patterns

https://doi.org/10.1016/j.jclinepi.2018.08.021Get rights and content

Abstract

Objective

We want to present information about response patterns obtained by Web-based survey in a large-scale epidemiological study.

Study Design and Setting

Within the RAPS (Risk Adapted Prevention Strategies for colorectal cancer [CRC]) study, we invited 160,000 randomly selected persons aged 40–54 years in three large German cities from 2015 to 2016 to complete a Web-based questionnaire on CRC risk factors and screening (97 items, average time for completion 15 minutes). Invitation letters and up to two reminder letters were sent to each individual.

Results

A total of 21.4% of women and 18.0% of men completed the questionnaire. Overall cumulative response rates were 7.5%, 14.3%, and 19.6% after the initial invitation letter, and the first and second reminder, respectively, with prevalence of and associations of key epidemiological parameters (such as family history of cancer, previous colonoscopy, etc.) being remarkably stable across waves of responses. For example, the sex and age distribution of the sample did not change with additional answers gained from additional letters.

Conclusion

Web-based questionnaires are feasible, cost-effective, and time effective in the setting of large-scale epidemiological studies. Although response patterns were remarkably stable over several rounds of reminders with substantially increasing cumulative response rates, future research should address possibilities to further enhance response rates.

Introduction

Epidemiological studies need large number of participants and reasonable response rates. In population-based studies, initial contacts and first responses have traditionally most commonly been made by written letters or phone. In the Internet era, these media are more and more replaced by online communication, but Web-based surveys have not yet been widely used in large-scale epidemiological studies [1].

Generally, online questionnaires seem to have a number of advantages compared to paper-based alternatives. For example, they have previously been found to be returned quicker than postal questionnaires [2], [3]. Another advantage of Web-based questionnaires compared to paper-based surveys is the possibility for researchers to immediately adjust questions if necessary without having to discard already printed copies. Furthermore, preliminary results can normally be easily obtained [4]. Although online questionnaires were thought to yield more measurement error than traditional data collection [4], [5], more recent research indicates that the contrary seems to be true [1]. A further advantage is that built-in plausibility checks may immediately alert participants of missing or inconsistent or implausible responses, which may increase data quality and completeness and substantially reduce the load of time-consuming contacts with participants and data cleaning. Importantly, there is evidence that individuals responding to Web-based surveys are comparable to those responding to traditional modes of data collection regarding a large variety of important characteristics (e.g., age, gender, income, education, and health status) [6], [7], [8].

Yet very little is known about response rates and response patterns that may be achieved with this approach in times of almost complete Internet coverage. Here, we report on response rates and response patterns and their relationship to observed prevalences and associations in a Web-based survey on family history (FH) of colorectal cancer (CRC) and related factors relevant to CRC screening for which 160,000 men and women aged 40–54 years in three large German cities were invited.

Section snippets

Methods

This survey was part I of the RAPS (Risk Adapted Prevention Strategies for CRC) study, a multiphase multicenter study conducted in the framework of the German Cancer Consortium. The full description of the RAPS study has been published elsewhere [9]. Briefly, in part I of this study, we aimed for recruiting 30,000 men and women aged 40–54 years in the catchment areas of three German study centers located in the cities of Dresden, Munich, and Stuttgart (approximately 540,000, 1,450,000, and

Results

Table 1 shows the characteristics of the random samples retrieved from the population registries, in comparison to the general population. Overall, the obtained data sets represented very good random samples of the underlying population with regard to their sex and age distribution. A total of 28.3% of all eligible persons in the population of the three cities were invited to take part in the study. This proportion varied from approximately 18% in Munich to approximately 46% in Stuttgart.

Discussion

The present study gives an overview on the response rate of an online questionnaire used in the RAPS study, a large population-based epidemiological study. One invitation letter and up to two subsequent reminder letters resulted in a cumulative response rate of approximately 20% in a population-based setting, with higher response rates among women and residents from a city located in Eastern Germany compared to two cities located in Western Germany. These response rates were achieved without

Acknowledgments

The authors especially thank all study participants. The authors also thank Drs. Utz Benscheid, Isabel Lerch, Volker Herrmann, Bernd Rasemann, Katrin Kollaske, Ulrike Seifert, Miriam Hofmann, Helga Lex, Caroline Tetzner, and Bernd Borstel for their excellent technical assistance and their contributions in data collection, monitoring, and documentation and thank Kai Ravesloot for the help with programming the online questionnaire.

References (26)

  • R. West et al.

    The ATTEMPT cohort: a multi-national longitudinal study of predictors, patterns and consequences of smoking cessation; introduction and evaluation of internet recruitment and data collection methods

    Addiction

    (2006)
  • K. Tikk et al.

    Study protocol of the RaPS study: novel risk adapted prevention strategies for people with a family history of colorectal cancer

    BMC Cancer

    (2018)
  • S. McCluskey et al.

    Increasing response rates to lifestyle surveys: a pragmatic evidence review

    Perspect Public Health

    (2011)
  • Cited by (6)

    • Prevalence of a First-Degree Relative With Colorectal Cancer and Uptake of Screening Among Persons 40 to 54 Years Old

      2020, Clinical Gastroenterology and Hepatology
      Citation Excerpt :

      Of those who received the invitation letter, 30,766 people participated in the online questionnaire. After exclusion of incomplete (n = 1814) or highly implausible (n = 3) questionnaires and data from participants with discrepancies between self-reported sex and/or age with the obtained registry offices data (n = 238), a total of 28,711 participants could be analyzed (response rate as previously published was 19.6%12). Study characteristics of the remaining 28,711 participants stratified by FH are shown in Table 1.

    Funding: This work was supported by the German Cancer Consortium (DKTK). J.R. and M.S. in part were supported by the Robert Bosch Stiftung, Stuttgart. The funding sources had no involvement in the study design; the collection, analysis, and interpretation of data; in the writing of the report; and in the decision to submit the article for publication.

    Conflict of interest: None.

    Transcript profiling: None.

    Writing assistance: None.

    1

    Authors share the first authorship and contributed equally to the preparation of the manuscript.

    View full text