OBJECTIVES: Women in poverty may benefit from avoiding closely spaced pregnancies. This study sought to identify predictive factors that could identify women at risk for closely spaced pregnancies. METHODS: We studied 20,028 women receiving welfare (cash assistance) from Washington State. Using Cox proportional hazards methods, we estimated the effects of individual- and community-level variables on time from an index birth until a subsequent pregnancy (between June 1992 and December 1999). Prediction models developed in a random half of our data were validated in the other half. Receiver operator characteristic plots appropriate for proportional hazards models were calculated to compare the sensitivity and specificity of each model. RESULTS: At 5 years of follow-up, the most predictive model contained just individual-level variables (age, education, race, marital status, number of prior pregnancies); the area under the receiver operator characteristic curve was 0.66 (.62-.69). The addition of community-level variables (percent in poverty, with a high school degree or higher, Black, Hispanic, in an urban area; female unemployment rate; income inequality) added little predictive ability. Differences were found between women with different individual- and community-level characteristics, but the results suggest that these factors are not strong predictors of pregnancy spacing. CONCLUSIONS: Individual- and community-level characteristics are associated with interpregnancy intervals; however, we found little evidence that the selected variables predicted pregnancy interval in a useful manner.
OBJECTIVES:Women in poverty may benefit from avoiding closely spaced pregnancies. This study sought to identify predictive factors that could identify women at risk for closely spaced pregnancies. METHODS: We studied 20,028 women receiving welfare (cash assistance) from Washington State. Using Cox proportional hazards methods, we estimated the effects of individual- and community-level variables on time from an index birth until a subsequent pregnancy (between June 1992 and December 1999). Prediction models developed in a random half of our data were validated in the other half. Receiver operator characteristic plots appropriate for proportional hazards models were calculated to compare the sensitivity and specificity of each model. RESULTS: At 5 years of follow-up, the most predictive model contained just individual-level variables (age, education, race, marital status, number of prior pregnancies); the area under the receiver operator characteristic curve was 0.66 (.62-.69). The addition of community-level variables (percent in poverty, with a high school degree or higher, Black, Hispanic, in an urban area; female unemployment rate; income inequality) added little predictive ability. Differences were found between women with different individual- and community-level characteristics, but the results suggest that these factors are not strong predictors of pregnancy spacing. CONCLUSIONS: Individual- and community-level characteristics are associated with interpregnancy intervals; however, we found little evidence that the selected variables predicted pregnancy interval in a useful manner.
Authors: R Gold; Frederick A Connell; Patrick Heagerty; Stephen Bezruchka; Robert Davis; Mary Lawrence Cawthon Journal: Soc Sci Med Date: 2004-09 Impact factor: 4.634
Authors: Julie M Petersen; Malcolm Barrett; Katherine A Ahrens; Eleanor J Murray; Allison S Bryant; Carol J Hogue; Sunni L Mumford; Salini Gadupudi; Matthew P Fox; Ludovic Trinquart Journal: Res Synth Methods Date: 2022-01-05 Impact factor: 9.308