Other Age CPR prior to admission Mechanical ventilation

Non-elective surgery

Patient's age at last birthday in years 0.03057

Within 24 h prior to ICU admission 0.56995 CPR includes chest compression, defibrillation, or cardiac massage; record as 'yes' regardless of where the CPR was administered

Patient is on a ventilator at the time of ICU admission or immediately thereafter 0.79105

This question should be answered 'yes' except for (1) elective surgery (i.e. surgery scheduled at least 24 h in advance) patients and (2) 1.19098 preoperative Swan-Ganz insertion if an elective surgery patient

In order to test the overall quality of the finished model, discrimination and calibration were tested using the developmental set. The importance of performing and reporting both calibration and discrimination cannot be overstated; it is entirely possible for a model to perform well in one of these calculations and poorly in the other, resulting in a poor model. In the case of MPM0 the results were an area under the ROC of 0.837 and a Hosmer-Lemeshow goodness-of-fit test showing good calibration with p = 0.623. Finally, the model was applied to and computed for all the cases in the validation set. Again, there was excellent discrimination (ROC value, 0.824) and calibration (p = 0.327). The result of the MPM II calculation again computes a direct estimate of the probability of mortality for the patient using 15 variables.

A similar process was performed using data available for the 10 357 patients in the development sample still alive at 24 h. The resulting model contains 13 variables: five values already collected in the admission model, two re-evaluated at 24 h, and six new variables ( Table...?). This model (MPM24) was tested and demonstrated excellent discrimination (ROC value, 0.844) and good calibration ( p = 0.764) in the developmental sample, and again for those cases in the validation group who survived at 24 h (ROC value, 0.836; p = 0.231 for goodness-of-fit test).

Table 2 Variables in MPM24 with definitions and estimated coefficients


Definition b

Table 2 Variables in MPM24 with definitions and estimated coefficients



Variables at admisison


As in MPM0



As in MPM0



As in MPM0

Metastatic neoplasm

Non-elective surgery

Variables reassessed at

24 h

Coma or deep stupor

Mechanical ventilation

New variables


Confirmed infection

As in MPM0 As in MPM0

Partial pressure of oxygen

Prothrombin time > 3s above standard

Urine output < 150 ml in 8 h

Vasoactive drug therapy

At 24 h, whether or not due to drug overdose. Determination is made at the 24-h mark. For patients on a paralyzing muscle relaxant, awakening from anesthesia, or heavily sedated, use best judgment of the level of consciousness prior to sedation. The definition for this variable is the same as for coma/deep stupor in MPM0, but overdose patients still in coma or deep stupor at 24 h are not an exception to the definition

Determination applies at any time during the first 24 h

>176.8 mmol/l (2.0 mg/dl): determination applies at any time during the first 24 h

Confirmed admission infection or confirmed new infection. Determination applies at any time during the first 24 h. 'Yes' only if cultures, Gram stains, or radiographs confirm a suspected admission infection or new infection that developed in the first 24 h, or there is evidence of gross purulence. Laboratory confirmation (including verbal or fax-type confirmation) must be obtained by the 24-h mark for 'Confirmed' to be checked. If used to confirm presence of infection, radiological or other imaging evidence must be strongly suggestive

Determination applies at any time during the first 24 h. Po2 values in the operating room and recovery room should not be considered. Does not count transient drop during tracheal suctioning. A sustained drop in pulse oximetry < 90% is acceptable evidence of low Po2. Record as 'yes' regardless of Fio2

Determination applies at any time during the first 24 h

Determination applies to any 8-h period during the first 24 h

Continuous IV, defined as for 1 h or more Determination applies at any time during the first 24 h Vasoactive drugs include:

1.161G9 G.834G4



G.72283 G.49742



G.82286 G.71628

1. Any dose of metaraminol (Aramine), epinephrine (adrenaline chloride), norepinephrine (Levophed), dobutamine (Dobutrex), phenylephrine (Neo-Synephrine), amrinone (Inocor)

2. Naloxone (Narcan) if used to elevate blood pressure

3. Dopamine (Intropin) only if > 5 mg/kg/min

Subsequent to the original publication of the MPM models, models for 48 h (MPM48) and 72 h (MPM72) were developed and published (L.e.m.,e.s..h,o.w...,et.,al 1994). These models use the same variables, and except for b0 (or constant term), the same bk values as MPM24. MPM48 and MPM72 have their own unique b0 terms (b0 = -5.39153

Because of the ease of collection, the acceptance and use of MPM 0 and MPM24 have been significant. Unlike APACHE III, the MPM II probability of mortality does not depend on a single diagnosis because individual diagnoses, other than those included as part of the model itself, did not seem be significantly related to mortality. This is a significant simplifying factor since the necessity of designating a single overriding diagnosis within the first 24 h of an ICU admission is fraught with obvious difficulties. It is estimated that the collection time for any of the MPM II models is less than 2 min per patient (with some sites reporting less than 1 min) when the data are collected concurrently.

Sleep Apnea

Sleep Apnea

Have You Been Told Over And Over Again That You Snore A Lot, But You Choose To Ignore It? Have you been experiencing lack of sleep at night and find yourself waking up in the wee hours of the morning to find yourself gasping for air?

Get My Free Ebook

Post a comment