Interventions designed to assist informal caregivers who serve individuals at or near the end of life have predominantly focused on caregiving spouses. Can we define other caregiver subpopulations--by intensity of care provided--so as to enable better a) identification of caregiver needs and b) targeting of support to caregivers? The Health Omnibus Survey, an annual face-to-face survey in South Australia, collects health-related data from a representative sample of 4400 households. Piloted questions included in the 2001-2005 Health Omnibus surveys addressed death of a loved one, caregiving provided, impact of caregiving and caregiver characteristics. Of 18,224 respondents, 5302 reported a loved one's death due to terminal illness in the previous 5 years. In all, 502 (10%) provided daily care [5-7 days/week], 619 (12%) provided intermittent care [2-4 days/week] and 425 (8%) provided rare care. Active (daily plus intermittent) caregivers, compared with non-active (rare) caregivers, were more often women (63% vs 50%; P < 0.0001). Daily caregivers were distinguishable from intermittent; daily caregivers were more often widowed (95% vs 7%; P < 0.0001) and >or=60 years (80% vs 64%; P < 0.0001); intermittent caregivers were more commonly children/parents (35%), other relatives (33%), or friends (26%; P < 0.0001) and were better educated, more active in paid work and wealthier. Financial burden, experience at time of death, ability to move on after the death and need for grief support also differed by intensity of caregiving. Caregiver subpopulations can be defined according to intensity of caregiving with distinct demographic features helping to distinguish them.