RESUMO
Recently, we validated a simple method for estimating peritoneal dialysis (PD) peritonitis rate. Despite good agreement between estimates and gold-standard measurements in two large dialysis registries, the International Society of Peritoneal Dialysis (ISPD) was hesitant to recommend adoption of the estimating equation. Their perception is that inaccuracies, as small as they are, might still be detrimental to clinical decision-making. In this study, we apply new analyses to the original validation data sets. We quantify agreement using standards from the International Organization for Standardization (ISO). We also identify a subset of centres with poorest performance of the estimating equation and qualitatively assess the potential for compromised clinical decision-making associated with its use. Inter-assay % coefficient of variation between estimates and measurements was 4.2% in the Australia and New Zealand Dialysis and Transplant Registry and 4.6% in Le Registre de Dialyse Péritonéale de Langue Française, easily meeting ISO requirements. Mandel's h values and Grubb's tests confirmed more outlying estimates compared to the measurements, while Mandel's k values and Cochran's C tests showed that identical precision by the two methods. Misclassification of centres as being above versus below the ISPD standard of 0.4 episodes/patient-year occurred only with rates close to the threshold, affecting approximately 3% of patient-years. In the 26 (out of 268) centres with poorest performance of the estimating equation, examination of the time series of their annual PD peritonitis rate estimates/measurements showed that using estimates would not detrimental to clinical decision-making. In conclusion, the estimating equation is sufficiently accurate for routine clinical use.