Background:
A consensus is building that improving the health of women prior to pregnancy is necessary to improve poor birth outcomes reflected in health disparities, but few studies demonstrate that interventions designed to improve women's health prior to pregnancy do improve birth outcomes.

Objectives:
This study provides promising results showing that pre and interconceptional case management reduces poor birth outcomes and the factors contributing to poor birth outcomes.

Methods:
An evaluation that included prospective and retrospective study designs was used to assess the impact of case management interventions targeting high risk women before they become pregnant (again). The prospective study intervention group consisted of Magnolia clients who were recently enrolled in case management and a comparison group of a similar high risk population. The prospective study relied on pre and post intervention primary data collection for self efficacy, perceived stress, social support, high risk behavior and goals and future orientation. The retrospective study intervention group consisted of closed cases of Magnolia clients who received case management services for at least 90 days (2000-2005) and a randomly selected comparison group from a high risk population. The retrospective study relied on secondary analysis of health care data that was linked to surveillance and vital statistics data reflecting birth outcomes and factors contributing to poor birth outcomes.


Results:
The program was very successful in recruiting exceptionally high risk populations. Preliminary findings revealed statistically significant reductions in STDs (p<.05) and low-birth-weights (p=.06) among the participating women. Highly meaningful reductions in infant mortality (over 40/1000) were observed, but reductions were not statistically significant due to the small number of live births. Prospective study results using the scales were promising but statistically inconclusive.

Conclusion and implications for practice:
Results were promising, but study size and single location present challenges to generalizing results. The results show that this approach should continue to be investigated with more extensive studies accompanied by continued refinement of the intervention for fidelity, replicability and generalizability.