Background: Health services research is a well-articulated research methodology and can be a powerful vehicle to implement sustainable health service reform. This paper presents a summary of a five-year collaborative program between stakeholders and researchers that led to sustainable improvements in the maternity services for remote-dwelling Aboriginal women and their infants in the Top End (TE) of Australia. Methods. A mixed-methods health services research program of work was designed, using a participatory approach. The study area consisted of two large remote Aboriginal communities in the Top End of Australia and the hospital in the regional centre (RC) that provided birth and tertiary care for these communities. The stakeholders included consumers, midwives, doctors, nurses, Aboriginal Health Workers (AHW), managers, policy makers and support staff. Data were sourced from: hospital and health centre records; perinatal data sets and costing data sets; observations of maternal and infant health service delivery and parenting styles; formal and informal interviews with providers and women and focus groups. Studies examined: indicator sets that identify best care, the impact of quality of care and remoteness on health outcomes, discrepancies in the birth counts in a range of different data sets and ethnographic studies of 'out of hospital' or health centre birth and parenting. A new model of maternity care was introduced by the health service aiming to improve care following the findings of our research. Some of these improvements introduced during the five-year research program of research were evaluated. Results: Cost effective improvements were made to the acceptability, quality and outcomes of maternity care. However, our synthesis identified system-wide problems that still account for poor quality of infant services, specifically, unacceptable standards of infant care and parent support, no apparent relationship between volume and acuity of presentations and staff numbers with the required skills for providing care for infants, and an 'outpatient' model of care. Services were also characterised by absent Aboriginal leadership and inadequate coordination between remote and tertiary services that is essential to improve quality of care and reduce 'system-introduced' risk. Conclusion: Evidence-informed redesign of maternity services and delivery of care has improved clinical effectiveness and quality for women. However, more work is needed to address substandard care provided for infants and their parents.