Validation of pediatric index of mortality 2 (PIM2) in a single pediatric intensive care unit of Argentina

Pediatr Crit Care Med. 2007 Jan;8(1):54-7. doi: 10.1097/01.pcc.0000256619.78382.93.


Objective: Pediatric Index of Mortality 2 (PIM2) is an up-to-date mortality prediction model in the public domain that has not yet been widely validated. We aimed to evaluate this score in the population of patients admitted to our pediatric intensive care unit.

Design: Prospective cohort study.

Setting: Multidisciplinary pediatric intensive care unit in a general university hospital in Buenos Aires, Argentina.

Patients: All consecutive patients admitted between January 1, 2004, and December 31, 2005.

Interventions: None.

Measurements and main results: There were 1,574 patients included in the study. We observed 41 (2.6%) deaths, and PIM2 estimated 48.1 (3.06) deaths. Discrimination assessed by the area under the receiver operating characteristic curve was 0.9 (95% confidence interval, 0.89-0.92). Calibration across five conventional mortality risk intervals assessed by the Hosmer-Lemeshow goodness-of-fit test showed chi5 = 12.2 (p = .0348). The standardized mortality ratio for the whole population was 0.85 (95% confidence interval, 0.6-1.1).

Conclusions: PIM2 showed an adequate discrimination between death and survival and a poor calibration assessed by the Hosmer-Lemeshow goodness-of-fit test. The standardized mortality ratio and clinical analysis of the Hosmer-Lemeshow table make us consider that PIM2 reasonably predicted the outcome of our patients.

Publication types

  • Comparative Study
  • Evaluation Study

MeSH terms

  • Adolescent
  • Argentina
  • Chi-Square Distribution
  • Child
  • Child Mortality*
  • Child, Preschool
  • Cohort Studies
  • Confidence Intervals
  • Female
  • Hospital Mortality*
  • Humans
  • Infant
  • Infant Mortality*
  • Intensive Care Units, Pediatric*
  • Length of Stay
  • Male
  • Outcome Assessment, Health Care
  • Probability
  • Prognosis
  • Prospective Studies
  • Respiration, Artificial
  • Survival Analysis
  • Time Factors