Health behaviours depend upon preferences and individuals’ assessments of the long-term costs and benefits to themselves and others. We investigate family spillovers by examining couple concordance in uptake of bowel cancer screening using the English Longitudinal Study of Ageing. We exploit the sharp age cut-off for screening programme eligibility at age 60 and use polynomial regression to study how initial screening participation depends upon partner’s uptake. We find that individuals with screened older partners are 42 percentage points more likely to screen than individuals with non-screened older partners, whose probability of screening is similar to comparable singles. Additionally, eligible but non-screened individuals update their screening behaviour if their younger partners participate in screening once they became eligible. We find heterogeneity in this concordance that is consistent with learning among partners being the most likely mechanism driving couple concordance. The findings have important implications for the optimal design of population-wide screening campaigns.