The dramatic improvement in outcomes of very low birth weight infants over the last 30 years has been associated with many innovations in care. We have explored the CUSUM technique as a form of continuous quality control to ascertain the effects on survival to 28 days postnatal age and severe intraventricular hemorrhage (ivh) in extremely low birth weight infants following the introduction of a management protocol. Methods: All infants of less than or equal to 26 weeks gestation born between January 1996 and August 1997 were entered. The protocol was introduced in December 1996, and consists of a series of computerized admission orders which mandate: Aquaphor skin care, early amino acids before 24 h, prophylactic low dose indomethacin, prophylactic surfactant in the delivery room, reduced heparin concentrations in lines, and suggest permissive hypercapnia and earlier extubation. We collected data on birth weight, gestation, survival to 28 days, and survival without an ivh of greater than grade 3. Data were gathered prospectively and then entered into a spreadsheet designed to perform the CUSUM analysis. Results: 57 infants of less than or equal to 1000g were studied, 38 were treated prior to introduction of the protocol and 19 after. Prior to introduction of the protocol the CUSUM function was consistent with a null hypothesis that survival without major ivh was not different from 60%, and did not show any trend to improvement. The CUSUM function representing survival without major ivh crossed the line representing the alternate hypothesis (improvement in outcome to 80%) between the 10th and 11th infant treated after introduction of the protocol, p<0.05. This occurred even though the infants studied after introduction of the protocol were of lower birth weight 701 (SD 175)g vs. 773 (SD 147)g. Removing the infants who died before 24 hours of age reinforced the discrepancy between pre and post protocol periods. All infants who died before 24 h of age and all but one who died after the first 24 h did so following discontinuation of care because of major intraventricular hemorrhage. Conclusion: Instituting protocolized care for the very preterm infant was associated with a significant improvement in survival to 28 days without severe ivh. The CUSUM technique is applicable to relatively homogeneous populations where short term outcomes can be determined and facilitates assessment of new therapeutic approaches.