To assess the impact of the introduction of kangaroo mother care (KMC) in hospitals using the Perinatal Problem Identification Programme (PPIP) in South Africa, a survey was conducted of the PPIP sentinel sites in South Africa requesting information on the practice of KMC in the hospital and if practised, when it had been initiated. Data on live births and the neonatal deaths of infants weighing between 1000 and 1999 g for each institution were obtained from the national PPIP database and, where applicable, divided into two periods, before and after the introduction of KMC. The practice of KMC and PPIP data could be combined for 40 of the hospitals that had responded to the survey. Of these, eight hospitals had not initiated KMC by January 2005, 21 had PPIP data for a period after KMC had commenced and 11 had PPIP data for periods before and after the introduction of KMC. The neonatal death rate (NNDR) for all hospitals with no KMC or before the introduction of KMC was 88.14/1000 live births, whereas the NNDR for hospitals with KMC or after the introduction of KMC was 71.43/1000 live births [relative risk (RR) 0.81; 95% confidence interval (CI) 0.72-0.91]. For the 11 hospitals that had reliable PPIP data for periods before and after the initiation of KMC, the NNDR was 87.72/1000 live births before KMC and 60.76/1000 live births after KMC had been introduced (RR 0.62; 95% CI 0.53-0.73). The large and significant reduction in the NNDR of neonates weighing between 1000 and 1999 g was associated with the introduction of KMC.