Background: During the 1990s, researchers at the Navrongo Health Research Centre in northern Ghana developeda highly successful community health program. The keystone of the Navrongo approach was the deployment of nurses termed community health officers to village locations. A trial showed that, compared to areas relying on existing services alone, the approach reduced child mortality by half, maternal mortality by 40%, and fertility by nearly a birth -- from a total fertility rate of 5.5 in only five years. In 2000, the government of Ghana launched a national program called Community-based Health Planning and Services (CHPS) to scale up the Navrongo model. However, CHPS scale-up has been slow in districts located outside of the Upper East Region, where the "Navrongo Experiment" was first carried out. This paper describes the Ghana Essential Health Intervention Project (GEHIP), a plausibility trial of strategies for strengthening CHPS, especially in the areas of maternal and newborn health, and generating the political will to scale up the program with strategies that are faithful to the original design. Description of the intervention: GEHIP improves the CHPS model by 1) extending the range and quality of services for newborns; 2) training community volunteers to conduct the World Health Organization service regimen known as integrated management of childhood illness (IMCI); 3) simplifying the collection of health management information and ensuring its use for decision making; 4) enabling community health nurses to manage emergencies, particularly obstetric complications and refer cases without delay; 5) adding $0.85 per capita annually to district budgets and marshalling grassroots political commitment to financing CHPS implementation; and 6) strengthening CHPS leadership at all levels of the system. Evaluation design: GEHIP impact is assessed by conducting baseline and endline survey research and computing the Heckman "difference in difference" test for under-5 mortality in three intervention districts relative to four comparison districts for core indicators of health status and survival rates. To elucidate results, hierarchical child survival hazard models will be estimated that incorporate measures of health system strength as survival determinants, adjusting for the potentially confounding effects of parental and household characteristics. Qualitative systems appraisal procedures will be used to monitor and explain GEHIP implementation innovations, constraints, and progress. Discussion: By demonstrating practical means of strengthening a real-world health system while monitoring costs and assessing maternal and child survival impact, GEHIP is expected to contribute to national health policy, planning, and resource allocation that will be needed to accelerate progress with the Millennium Development Goals.