ABSTRACT
The federal Healthy Start program began 20 years ago, yet outcome evaluations lack sufficient rigor to draw conclusions on program impact. We evaluated the impact of the Healthy Start program on birth outcomes, prenatal care, and public services utilization. Birth record data for the St. Louis Healthy Start Program (SLHS) and non-SLHS controls (matched using a propensity score technique) were assessed for differences. Propensity score matching techniques matched SLHS to non-SLHS clients on potentially confounding variables for births from years 2006 to 2008. Traditional multivariable logistic regression on the full, unmatched sample was also conducted for comparison. Matching eliminated any prior statistical differences between groups on covariates. 168 controls and 84 SLHS participants remained in the final matched analysis group. Both analysis techniques were similar on all outcomes, revealing significant group differences for low birth weight (matched OR = 0.28, p = 0.023) and prematurity (matched OR = 0.25, p = 0.012) but not for prenatal care (matched OR = 0.76, p = 0.414), or public services utilization (matched OR = 3.31, p = 0.121). Early results for this Healthy Start project are positive in key areas directly impacting infant mortality. However, continued analysis of this program for sustained impact in these areas and ultimately, a reduction in infant mortality is needed. Additionally, more rigorous experimental and quasi-experimental evaluation designs are needed to assess the impact of other Healthy Start programs around the country.
