PURPOSE: Continuous monitoring tools can be used to monitor surgical outcomes over time. We illustrate the use of CUmulative SUM (CUSUM) charts in monitoring outcomes of Kasai portoenterostomy for treatment of biliary atresia at a supraregional unit. METHODS: Data on 57 consecutive infants who underwent a Kasai portoenterostomy performed by a single surgeon between June 1994 and June 2006 were collected. A procedure was defined as successful if clearance of jaundice (plasma bilirubin level <20 micromol/l) was achieved within 6 months of surgery. We applied cumulative observed-minus-expected, sequential probability ratio test (SPRT), and zero-resetting SPRT CUSUM charts and compared the results with those of standard aggregate data analyses. An expected failure rate of 43.0%, based on the national average failure rate, was used. RESULTS: The failure rate observed after 57 operations was 29.8%. The zero-resetting SPRT chart indicated a lower-than-expected failure rate earlier than did the aggregate data analyses and any of the other continuous monitoring techniques. CONCLUSIONS: The CUSUM chart method provides ongoing feedback that can be used for continuous monitoring of the outcome of a procedure to ensure that standards of care are maintained. Its use as a routine monitoring tool in pediatric surgery deserves wider recognition.
PURPOSE: Continuous monitoring tools can be used to monitor surgical outcomes over time. We illustrate the use of CUmulative SUM (CUSUM) charts in monitoring outcomes of Kasai portoenterostomy for treatment of biliary atresia at a supraregional unit. METHODS: Data on 57 consecutive infants who underwent a Kasai portoenterostomy performed by a single surgeon between June 1994 and June 2006 were collected. A procedure was defined as successful if clearance of jaundice (plasma bilirubin level <20 micromol/l) was achieved within 6 months of surgery. We applied cumulative observed-minus-expected, sequential probability ratio test (SPRT), and zero-resetting SPRT CUSUM charts and compared the results with those of standard aggregate data analyses. An expected failure rate of 43.0%, based on the national average failure rate, was used. RESULTS: The failure rate observed after 57 operations was 29.8%. The zero-resetting SPRT chart indicated a lower-than-expected failure rate earlier than did the aggregate data analyses and any of the other continuous monitoring techniques. CONCLUSIONS: The CUSUM chart method provides ongoing feedback that can be used for continuous monitoring of the outcome of a procedure to ensure that standards of care are maintained. Its use as a routine monitoring tool in pediatric surgery deserves wider recognition.