Satu Helakorpi1, Pia Mäkelä2, Ansku Holstila3, Antti Uutela4, Erkki Vartiainen5. 1. 1 Department of Lifestyle and Participation, National Institute for Health and Welfare (THL) satu.helakorpi@thl.fi. 2. 2 Department of Alcohol, Drugs and Addiction, THL. 3. 3 Hjelt Institute, University of Helsinki. 4. 1 Department of Lifestyle and Participation, National Institute for Health and Welfare (THL). 5. 4 Division of Welfare and Health Promotion, THL.
Abstract
BACKGROUND: Prevalence estimates may be biased if the characteristics of respondents differ from those of non-respondents in surveys. In this study, we used a follow-up telephone interview of initial non-respondents to examine the differences--in terms of self-rated health and health behaviours--to initial postal respondents and to assess improvements in prevalence estimates. METHODS: Following a postal questionnaire survey using a random sample (n = 5000) of the Finnish working-age population with a response rate of 57% (n = 2826), a follow-up telephone survey was performed based on 1261 non-respondents (response rate 56%, n = 708) in 2010. Prevalence of smoking, alcohol use, body mass index, physical activity, self-rated fitness, dietary habits and self-rated health were calculated for the survey population with and without a telephone interview. Logistic regression models were used to examine differences in health behaviours and health between the initial postal questionnaire respondents and follow-up telephone interview respondents. RESULTS: The total response rate increased from 57% to 71% when the telephone respondents were included. The telephone survey indicated that both male and female telephone respondents were more often smokers, and female telephone respondents were more often heavy episodic drinkers and less often reported poor self-rated fitness than postal respondents. Nonetheless, the prevalence rates of outcome variables did not change significantly when telephone respondents were included. CONCLUSION: The response rate of surveys can be increased by using a telephone survey in follow-up contacts with non-respondents. As non-respondents differ from respondents, this contributes to an improvement--although small--in internal validity.
BACKGROUND: Prevalence estimates may be biased if the characteristics of respondents differ from those of non-respondents in surveys. In this study, we used a follow-up telephone interview of initial non-respondents to examine the differences--in terms of self-rated health and health behaviours--to initial postal respondents and to assess improvements in prevalence estimates. METHODS: Following a postal questionnaire survey using a random sample (n = 5000) of the Finnish working-age population with a response rate of 57% (n = 2826), a follow-up telephone survey was performed based on 1261 non-respondents (response rate 56%, n = 708) in 2010. Prevalence of smoking, alcohol use, body mass index, physical activity, self-rated fitness, dietary habits and self-rated health were calculated for the survey population with and without a telephone interview. Logistic regression models were used to examine differences in health behaviours and health between the initial postal questionnaire respondents and follow-up telephone interview respondents. RESULTS: The total response rate increased from 57% to 71% when the telephone respondents were included. The telephone survey indicated that both male and female telephone respondents were more often smokers, and female telephone respondents were more often heavy episodic drinkers and less often reported poor self-rated fitness than postal respondents. Nonetheless, the prevalence rates of outcome variables did not change significantly when telephone respondents were included. CONCLUSION: The response rate of surveys can be increased by using a telephone survey in follow-up contacts with non-respondents. As non-respondents differ from respondents, this contributes to an improvement--although small--in internal validity.
Authors: Heidi Amalie Rosendahl Jensen; Cathrine Juel Lau; Michael Davidsen; Helene Birgitte Feveile; Anne Illemann Christensen; Ola Ekholm Journal: Eur J Public Health Date: 2022-06-01 Impact factor: 4.424
Authors: Benjamin Clarsen; Jens Christoffer Skogen; Thomas Sevenius Nilsen; Leif Edvard Aarø Journal: BMC Public Health Date: 2021-04-15 Impact factor: 3.295