External validation of Acute Physiology and Chronic Health Evaluation IV in Dutch intensive care units and comparison with Acute Physiology and Chronic Health Evaluation II and Simplified Acute Physiology Score II

Caution when using prognostic models: A prospective comparison of 3 recent prognostic models

“…SAPS3 was published in 2005, and Italian data from 2007 showed it was not calibrated to that population [15]. APACHE IV was published in 2006, and Dutch data from 2006-2009 showed it was not calibrated to that population [17]. We arrived basically at the same conclusion and also showed poor calibration for MPM 0 -III.…”

“…There was a 5-to 7-year interval between the development of these models and our study. This could partly explain our results, as demonstrated by other authors in external validations of SAPS 3 and APACHE IV [14,15,17]. However, if this is the case, then how often should a prognostic model be updated?…”

“…Interestingly, this trend was seen in SAPS 3 validation studies, with calibrations found to be better in smaller samples [10][11][12][13]16] and worse in larger ones [14,15], as what happened in our Table 3 Performance of the scores by type of admission (medical/surgical) and reason for admission 423.e5 Caution when using prognostic scores cohort. The Dutch study [17] also had a larger sample than the North American study that showed a good calibration for APACHE IV [18].…”

“…SAPS 3 was developed in a heterogeneous population from various countries in Europe, North, South and Central Americas and Australasia, which may have contributed to improve its external validity. Some studies have validated these new versions in other populations [10][11][12][13][14][15][16][17][18][19], but to the best of our knowledge, they have never been compared with each other before. The objective of this study is to assess external validity of the 3 scores and to compare their performance in a Brazilian population.…”

Utility of using the Therapeutic Intervention Scoring System-28 to predict mortality in intensive care unit

“…[4][5][6] In our daily clinical routine, we use APACHE II and SOFA scores to determine mortality, and the TISS-28 score is used by nurses in the ICU. The TISS-28 calculates the amount of nursing time spent on direct patient care and does not require any laboratory findings.…”