The Breastfeeding Attrition Prediction Tool (BAPT) is based on the theory of planned behavior, which explains behavior as a function of attitudes, subjective norm, and perceived control. The BAPT subscales are positive and negative attitudes toward breastfeeding, family and professional expectations, and perceived ability to be successful. The purpose of this study was to test the reliability and validity of the BAPT among 269 women who planned to breastfeed for at least 8 weeks. Subjects completed the BAPT in the hospital and during a telephone interview at 8 weeks. Based on factor analysis, several items were deleted. Using discriminant function analysis, the modified BAPT was an effective predictor of 78% of women who stopped breastfeeding before 8 weeks and 68% of those who were still breastfeeding. There is potential for the modified BAPT to be an adjunct for the clinician in identifying women at risk for early cessation of breastfeeding.