Objectives: 

To investigate if the performance of Pediatric Index of Mortality 3 is improved by including imputed values for the Pao₂/Fio₂ ratio where measurements of Pao₂ or Fio₂ are missing.

Design: 

A prospective observational study.

Setting: 

A bi-national pediatric intensive care registry.

Patients: 

The records of 37,983 admissions of children less than 16 years old admitted to 19 ICUs.

Interventions: 

None.

Measurements and Main Results: 

Seven published equations describing an association between Pao₂/Fio₂ and oxygen saturation measured by pulse oximetry (Spo₂)/Fio₂ were used to derive an alternative variable d100 × Fio₂/Pao₂ for the Pediatric Index of Mortality 3 variable 100 × Fio₂/Pao₂. Six equations exclude Spo₂/Fio₂ values if Spo₂ is greater than 96–98%. 100 × Fio₂/Pao₂ was missing in 72% of patient records primarily due to missing Pao₂, d100 × Fio₂/Pao₂ was missing in 71% of patient records if values of Spo₂greater than 97% were excluded or in 17% of patient records if all measurements of Spo₂ were included. Univariable analysis supported the inclusion of Spo₂ values greater than 97%. Compared to the standard Pediatric Index of Mortality 3 model, two alternative models imputing 100 × Fio₂/Pao₂ from d100 × Fio₂/Pao₂ only if 100 × Fio₂/Pao₂ was missing, or using d100 × Fio₂/Pao₂ values exclusively, resulted in a small but statistically significant improvements in discrimination of Pediatric Index of Mortality 3 (area under the receiver operator curve 0.9068 [0. 8965–0. 9171]; 0.9083 [0.8981–0.9184]; 0.9087 [0.8987–0.9188], respectively).

Conclusions: 

Imputation of the Pao₂/Fio₂ ratio in cases where arterial sampling was not performed resulted in a large reduction in the rate of missing data if all values of Spo₂ were included. The imputation technique improved the discrimination of Pediatric Index of Mortality 3; however, the magnitude of the increment in overall model performance was small. A possible benefit of the approach is reducing the potential for bias resulting from variation in practice for invasive monitoring of oxygenation.

中文翻译：

---

从SpO2 / FIO2注入PaO2 / FIO2对小儿死亡率指标性能的影响3。

目标： 

为了研究的性能儿科死亡率3的指数是通过包括用于PA插补值提高Ò ₂ / F io的₂比，其中帕的测量Ò ₂或F io的₂缺失。

设计： 

前瞻性观察研究。

设置： 

两国的儿科重症监护中心。

耐心： 

19所重症监护病房收治了37,983名16岁以下儿童的入学记录。

干预措施： 

没有。

测量和主要结果： 

七发表了描述帕之间的关联方程Ò ₂ / F IO ₂和氧饱和度通过测量脉搏血氧仪（SP Ò ₂）/ F io的₂被用于导出一个替代可变D100×F io的₂ / PA ø ₂为儿科指数死亡率3变量100×F io ₂ / Pa o _2的估计。如果Sp o ₂大于96–98％，则六个方程式会排除Sp o ₂ / F io ₂值。100×F io ₂/ PA Ó ₂中的患者记录72％缺少这主要是由于缺少帕Ò ₂，D100×F IO ₂ / PA ö ₂缺失在患者记录71％，如果SP的值ö ₂大于97％被排除或在病人的病历17％，如果SP的所有测量Ø ₂都包括在内。单变量分析支持包含大于97％的Sp o ₂值。与标准的儿科死亡率指数3模型相比，两个替代模型从d100×F io推算出100×F io ₂ / Pa o ₂₂ / PA ö ₂仅当100×F io的₂ / PA ö ₂是缺失，或使用D100×F io的₂ / PA ö _2个值排他地，导致歧视的小，但统计学上显著改进儿科死亡率3的指数（接收器操作员曲线下方的区域分别为0.9068 [0. 8965-0。9171]； 0.9083 [0.8981-0.9184]； 0.9087 [0.8987-0.9188]。

结论： 

PA的插补ö ₂ / F io的₂比在情况下，动脉取样不进行导致如果SP的所有值的缺失数据的速率大的减速Ò ₂都包括在内。归因技术改善了对儿童死亡率指数3的区分；但是，总体模型性能的增加幅度很小。该方法的可能好处是减少了由于有创监测氧合的实践中的变化而导致产生偏差的可能性。