The application of risk-adjusted control charts using the Paediatric Index of Mortality 2 for monitoring paediatric intensive care performance in Australia and New Zealand

Abstract

Objective: To assess the applicability of risk-adjusted sequential control charts using the Paediatric Index of Mortality version 2 for monitoring of the quality of paediatric intensive care.

Design: Observational study.

Setting: A registry of patient admissions to paediatric intensive care units (PICUs) in Australia and New Zealand.

Patients: A total of 10,710 patients admitted to eight PICUs during a 24-month period.

Measurements and results: A series of risk-adjusted control charts was created for each PICU. Modified sequential probability ratio tests were used to test the hypothesis that the PICUs being monitored were 'out of control', where loss of control was arbitrarily defined as the odds of death exceeding twice the odds of dying as estimated by PIM2. In 24 months of monitoring, there was one alarm signal, suggesting the odds of deaths had doubled, and there was one signal, in another PICU, suggesting the odds of death had halved.

Conclusions: The major advantage of risk-adjusted sequential control charts is that the technique allows unit performance to be monitored continuously over time, rather than intermittently, with the aim of rapidly detecting a change in performance as soon as possible after it occurs. This technique is suitable for continuously screening for a change in outcome within a PICU over time and complements other methods of monitoring the quality of paediatric intensive care.