Kestin I G
Derriford Hospital, Plymouth.
Br J Anaesth. 1995 Dec;75(6):805-9. doi: 10.1093/bja/75.6.805.
Cusum analysis is a statistical technique to distinguish deviations from an acceptable failure rate. The progress of anaesthetic trainees learning four practical procedures (obstetric extradural anaesthesia, spinal anaesthesia, central venous cannulation and arterial cannulation) was monitored from their first attempt using cusum analysis. Suitable acceptable and unacceptable failure rates for each procedure were chosen by consultant anaesthetists. For obstetric extradural anaesthesia, four trainees eventually achieved acceptable failure rates (5%) and the number of attempts required to demonstrate this statistically ranged from 29 to 185; three trainees had an unacceptable failure rate (10%) and five trainees had inconclusive records. For spinal anaesthesia, two trainees achieved an acceptable failure rate (10%) and the number of attempts required to demonstrate this statistically ranged from 39 to 67; two trainees had an unacceptable failure rate (20%) and four trainees had inconclusive records. One trainee demonstrated statistically an acceptable failure rate in arterial cannulation (20%) after 14 attempts and four trainees had inconclusive records. Two records of central venous cannulation were inconclusive. Some records showed variable failure rates which were sometimes associated with lack of practice or a change in technique. Cusum analysis can be used to monitor training in practical procedures and as a continuous audit of quality of clinical practice.
累积和分析是一种用于区分与可接受故障率偏差的统计技术。使用累积和分析对麻醉实习医生学习四种实际操作（产科硬膜外麻醉、脊髓麻醉、中心静脉置管和动脉置管）的进展情况从他们的首次尝试开始进行监测。每种操作合适的可接受和不可接受故障率由麻醉顾问医生选定。对于产科硬膜外麻醉，4名实习医生最终达到了可接受的故障率（5%），从统计学上证明这一点所需的尝试次数在29次至185次之间；3名实习医生的故障率不可接受（10%），5名实习医生的记录不明确。对于脊髓麻醉，2名实习医生达到了可接受的故障率（10%），从统计学上证明这一点所需的尝试次数在39次至67次之间；2名实习医生的故障率不可接受（20%），4名实习医生的记录不明确。1名实习医生在进行14次尝试后从统计学上证明动脉置管的故障率可接受（20%），4名实习医生的记录不明确。中心静脉置管的两份记录不明确。一些记录显示故障率存在变化，有时与缺乏练习或技术改变有关。累积和分析可用于监测实际操作的培训，并作为对临床实践质量的持续审核。
