Although there are numerous studies of depression and its linkages with non-communicable diseases (NCDs), most rely on a single cross-section or a single wave of the National Income Dynamics Study (SANIDS) for South Africa, which does not allow for incorporation of individual unobservable effects. Such effects are potentially significant as it is frequently observed that there is considerable variation in depressive symptoms even when an old person suffers from common NCDs. We use correlated random effects probit model on the first 5 waves of SA-NIDS panel data collected every two years between 2008-2016/17 to examine thereverseassociation from Depression to selected NCDs, controlling for socio-economic and demographic characteristics. The analysis yields useful insights into the complex relationships between NCDs and depression. Policy options that focus on biological and behavioural links in the co-occurrence of NCDs and depression are examined. Of particular importance is integration of depression and NCD care in primary health care with a view to increasing prevention, screening, self-management, treatment and rehabilitation in order to achieve equitable, efficient and quality health services in South Africa.