Population:
(1) 55 children from 6 months to 18 years
(2) Seen between 1974 and 1978.
(3) Most had a hematologic malignancy (50 of 55).
(4) Ventilatory support was begun when the FIO2 was > 50%.
Data collected at 36 hours:
(1) respiratory rate in breaths per minute
(2) heart rate in beats per minute
(3) temperature in °C
(4) pH (respiratory alkalosis at high respiratory rates)
(5) PaCO2 in mm Hg
(6) PaO2 in mm Hg
(7) FIO2 in percent (decreases with improvement)
discriminant function using data at 36 hours =
= (-8133.3) + (5.5 * (respiratory rate)) – (1.7 * (heart rate)) + (109 * (temperature)) + (504 * (pH)) + (4.8 * (PaCO2)) – (1.47 * (PaO2)) + (4.99 * (FIO2))
Interpretation:
• minimum score: < -100.0
• maximum score: > +100.0
• 100% of children with a discriminant score > +1 required ventilatory support.
• About 38% (9 out of 24) of children with a score from –1 to +1 required ventilatory support.
• About 7% (1 of 15) of children with a score -1 to –3 required ventilatory support.
Performance:
• 91% of children requiring ventilatory support were identified.
• The equation needs to be validated using larger groups of patients.
NOTE: Looking at Figure 3, all of the patients are tightly packed from –3.0 to +3.0. When I enter various data, the discriminant function values vary over a fairly large range. There may be a conversion factor that is being missed.