Introduction Evidences on whether poor self-reported health (SRH) predicts subsequent mortality across different socio-economic groups are inconsistent. This study assesses whether education modifies how poor SRH influences mortality among older people in Indonesia.

Methods A cohort of 11 753 men and women aged 50 years and over was recruited in the INDEPTH/WHO Study on Adult Health and Global Ageing (SAGE) in Purworejo Health and Demographic Surveillance (HDSS) site in 2007. SRH was measured using the single global SRH question with 5-point response scales (very good, good, moderate=moderately good SRH; bad, very bad=poor SRH). The baseline data were linked to the HDSS mortality data in 2010. HR for mortality was calculated for poor SRH using Cox proportional hazard regression after adjustment for age, education levels, age, marital status, living area, history of chronic diseases, and presence of disabilities.

Results During follow-up (median duration=37 months), 1199 deaths (10.2%) and 1.9% lost to follow-up were identified. Poor SRH increased the mortality risk in men (HR 3.59, 95% CI 1.96 to 6.57) and women (HR 3.16, 1.12 to 8.90). Education levels were not associated with mortality risk. The association between poor SRH and mortality did not differ across education groups, neither in men nor in women. Presence of disabilities, history of chronic diseases, and living alone increased the mortality risk.

Conclusion Poor SRH predicts mortality among older population in Indonesia. Education does not modify the association between poor SRH and mortality. Health promotion in the general population is important for the reduction of the mortality risk among older people.