Selection bias was reduced by recontacting nonparticipants

Abstract

Objective: One of the main goals of health examination surveys (HES) is to provide unbiased estimates of health indicators at the population level. We demonstrate how multiple imputation methods may help to reduce the selection bias if partial data on some non-participants are collected. Study Design and Setting: In the FINRISK 2007 study, a populationbased health study conducted in Finland, a random sample of 10 000 men and women aged 25-74 were invited to participate. The study included a questionnaire data collection and a health examination. 6255 individuals participated in the study. Out of 3745 non-participants, 473 returned a simplified questionnaire after a recontact. Both the participants and the nonparticipants were followed up for death and hospitalizations. The follow-up data allowed to check the assumptions on the missing data mechanism and tailored multiple imputation methods were used to handle the missing data. Results: Non-participation is a strong predictor for mortality in the fiveyear follow-up. However, the recontact response does not predict mortality or morbidity among the non-participants when adjusted for age and sex. The M A N U S C R I P T A C C E P T E D ACCEPTED MANUSCRIPT 2 result suggests that the recontact respondents can be used as proxy for all non-participants. A comparison of raw estimates and estimates adjusted for selection bias reveals clear differences in the estimated population prevalences of smoking and heavy alcohol usage. Conclusion: All efforts to collect data on non-participants are likely to be useful even if the response rate for the recontact remains low. Statistical analysis of the recontact respondents provides an indication of the extent of the selection bias, even in studies where follow-up data are not available to check the assumptions.Keywords: survey; non-response; bias; missing data; multiple imputation What is new?:• Register data on cause specific mortality and morbidity can be used to check the assumptions on the missing data mechanism.• In the FINRISK 2007 study, we found that while the participants and the non-participants clearly differ, the non-participants with and without a recontact response have similar mortality and morbidity when adjusted for age and sex.• We propose a multiple imputation approach to handle data missing not at random. The approach is applicable when data on a non-response questionnaire are available.

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

“…Due to these differences, health indicators estimated from the participants only are likely to suffer from selection bias. For example, recent results indicate that the true prevalence of smoking in Finland may be five percentage points higher than the prevalence based on the participants only [10].…”

“…Re-contact data were collected in FINRISK 2007 and 2012 surveys where the main survey questionnaire was re-sent to non-participants. Although the response rates for the re-contact were very low, 13% and 14%, respectively, the re-contact data turned out to be useful in reducing selection bias [10,36]. It was found that when adjusted for age and sex, the re-contacted respondents tend to resemble the non-participants with respect to their mortality and hospitalizations.…”

“…Textual descriptions often miss important details. In the NoPaHES project, we used graphical models to visualize study design and the assumptions on the missing data mechanism together with the assumed causal model [10,20,27,28].…”

“…Unfortunately, it is practically impossible in public health monitoring and research to obtain survey data that are truly representative for the population. There is a lot of evidence that even if the invited sample is a representative random sample, the actual participants differ from the population with respect to their health [1][2][3][4][5][6][7][8][9][10][11], health behaviours [10,[12][13][14] and socio-demographic status [14][15][16]. Due to these differences, health indicators estimated from the participants only are likely to suffer from selection bias.…”

Bayesian models for data missing not at random in health examination surveys

“…The decision about participation have been found to depend on the risk factors, such as smoking (Shahar et al, 1996), either directly or via a common cause such as health awareness. This may be deduced from the fact that the non-participants have a higher risk of death (Jousilahti et al, 2005;Harald et al, 2007;Karvanen et al, 2016). This dependence causes missing data to be classified as missing not at random (MNAR) (Rubin, 1976).…”