A Positive Deviance (PD) Hearth intervention is a home and neighborhood-based nutrition program for children who are at risk for protein-energy malnutrition in a low resource community. The intervention uses the ‘Positive Deviance’ approach to identify those behaviors practiced by the mothers or caretakers of well-nourished children from poor families and transfers such positive practices to other mothers who are equally disadvantaged economically. Positive Deviance Hearth intervention is designed to treat malnourished children, enable the families to sustain their rehabilitation at home on their own and to prevent malnutrition in younger siblings. However, PD Hearth intervention monitoring system in Migori only assesses a program’s ability to treat, one of the three PD Hearth objectives. Thus, there was need for impact evaluation to measure outcomes of the PD Hearth intervention to sustain rehabilitation and prevent malnutrition in younger siblings. The objectives of the study were to determine the level to which PD Hearth enables families to sustain rehabilitation at home on their own and to identify the practices which influence PD Hearth outcomes. The study was designed as a pipeline quasi-experimental and mixed method was used to collect data and perform statistical analyses. Single stage cluster sampling was used to identify 53 and 54 children on the intervention and comparison group in five communities. Weight measurements of the children on the intervention aged 6 to 59 months at the entry, exit and graduation stages were retrieved from Kenya Medical Research Institute Family AIDS Care and Education Services programme activities reports. Anthropometry (height measurements) for the children on the intervention and comparison children was taken. Caregivers filled in a questionnaire, assisted by the researchers as necessary. At entry, 18.9% children on the intervention had moderate underweight while 43.4% had mild underweight. At current status though, 3.8% and 34.0% had moderate and mild underweight respectively. The regression model predicted that Weight-for-Height (WAZ) of the children on the intervention at current status lied on 51.5 percentile, thus, normal for underweight. Increased feeding frequency made the largest contribution to weight gain than other caregiver practices. Therefore, the Migori County government in collaboration with the Ministry of Health needs to scale up PD Hearth intervention to reverse cases of Moderate Acute Malnutrition (MAM) and prevent Severe Acute Malnutrition (SAM) in the County.