Randomized intervention trial on preventive home visits to older people: Baseline and follow-up characteristics of participants and non-participants

Aims: In controlled intervention studies, a selective non-response or refusal to participate at baseline may bias measurable effects of the intervention. The aim of this study was to compare mortality and nursing home admission among older persons who accepted (participants) and older person who declined (non-participants) to join a controlled feasibility trial, and to describe and evaluate defined subgroups of non-participants. Methods: Prospective controlled three-year intervention study (1999-2001) in 34 Danish municipalities with five-year follow-up. Randomization and intervention (education of municipality employees) was done at municipality level. In total 5,788 home-dwelling 75- and 80-year-olds living in these municipalities were invited to participate in the study. Written consent was obtained from 4,060 persons (participation rate 71%). Results: During five-year follow-up non-participants had a higher mortality rate (survival analysis risk ratio RR=1.5, 95% CI=13-1.7,p<0.0001) and a higher rate of nursing home admissions (RR=1.7, 95% CI=1.32. 1, p < 0.0001) compared with participants. Subgroups of non-participants describing themselves as "too ill" and persons "not reached" had a significantly higher mortality rate and risk of admission to nursing home than participants, whereas the subgroups of non-participants describing themselves as "too healthy" and having "another reason for refusal" did not differ from the participants. There was no difference in mortality rates between non-participants living in intervention municipalities compared with non-participants living in control municipalities. Conclusions: Mortality and nursing home admissions were higher among non-participants. Selection participation bias was of no clinical importance since subgroups of non-participants eligible for the intervention did not differ from the participants.