 |
|
 |
|
 |
 |
|
 |
 |
 |
 |
 |
 |
 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
To the Editor. —
I read with interest the article by Han et al1 but question the validity of adjusting their outcome measures by the Pediatric Risk of Mortality (PRISM) severity-of-illness measure. Adjustments are made to take into consideration confounding relationships such as the severity of illness. Unfortunately, the severity-of-illness measure used for adjusting in this study was measured at an illogical point in the sequence of events. The severity of illness will determine if resuscitation is necessary, the extent to which resuscitation measures need to be implemented, and the outcome of the resuscitation. Thus, the appropriate time to measure the severity of illness is at presentation to the community hospital. Because the PRISM score is measured within 24 hours of admission to the pediatric intensive care unit (PICU) and is a result of both the characteristics of the patient’s illness and the resuscitative efforts of the community hospital physicians, it is not appropriate to use it for adjusting the outcome data.
To illustrate this point, I have organized the variables presented in the article’s 3 tables into the sequence: input, process, output, outcome, and impact categories (Table 1). Inputs into the resuscitation process and medical care would be the characteristics of the patient and the illness. Inputs not presented would be duration of illness, prior medical intervention, and severity of illness on presentation to the community hospital. The primary process under study was the resuscitative effort of the community hospital physician. Characteristics of the transport process also were included in the analysis. The output of these 2 processes was a determination of whether the process was consistent with American College of Critical Care Medicine Pediatric Advanced Life Support (ACCM-PALS) guidelines. The outcome of the processes was the duration of shock, persistence/reversal of shock on arrival of the transport team, and the PRISM score determined within 24 hours of arrival at the PICU. The impact of the inputs and process (and subsequent PICU care) was survival/mortality.
|
REFERENCE
,
,,
Pediatrics
Department of Critical Care Medicine/Transport
In Reply. —
We agree entirely with Dr Hyslop’s point of view but are limited by the absence of an established risk of severity score for patients outside of the intensive care unit (ICU). The Pediatric Risk of Mortality (PRISM) score is validated for use only in patients in the ICU who have survived the first 12 to 24 hours. Four of the patients did not survive the first 24 hours (2 died before transport team arrival, 1 died before the transport team arrived back at the children’s hospital, and 1 died in the first 12 hours). Therefore, we performed our severity-of-illness adjusted analysis without including these patients.
We have subsequently performed the analyses that Dr Hyslop requested. The results were essentially the same when evaluating all patients without adjustment for severity of illness: survival odds ratio (OR) for shock reversed, 13.690 (95% confidence interval [CI], 1.741–107.65); survival OR for use of American College of Critical Care Medicine Pediatric Advanced Life Support (ACCM-PALS) guidelines, 7.500 (CI, 1.63–34.52); mortality OR for duration (per hour) of persistent shock, 2.43 (CI, 1.42–4.18); and mortality OR for delay (per hour) in resuscitation consistent with ACCM-PALS guidelines, 1.53 (CI, 1.149–2.041).
As Dr Hyslop requested, we also have evaluated all patients adjusting for the "pseudo-PRISM" score attained from community hospital patient notes. We did not perform this technique in the original article; we are not certain that it is valid because hypotension is one of the physiologic variables present in the PRISM score. Nevertheless, the results were similar, although less robust: survival OR for shock reversed, 5.71 (CI, 0.669–48.72); survival OR for use of ACCM-PALS guidelines, 4.97 (CI, 0.913–27.06); mortality OR for duration (per hour) of persistent shock, 1.83 (CI, 1.05–3.21); and mortality OR for delay (per hour) in resuscitation consistent with ACCM-PALS guidelines, 1.34 (CI, 0.96–1.88). A possible explanation for decreased significance may be that the pseudo-PRISM as an adjuster is already describing hypotension.
In conclusion, we used presently accepted severity-of-illness criteria for adjustment. However, we fully agree that the limitations of such an analysis are present, as Dr Hyslop describes. We await the validation of new scores that do not include prolonged capillary refill or hypotension to answer Dr Hyslop’s critique adequately. Strictly interpreted, our adjusted data apply only to survival after the first 24 hours of ICU stay.
This article has been cited by other articles:
 |
|
 |
|
  R. H. Palmer, S. Ezhuthachan, C. Newman, M. J. Maisels, and M. A. Testa Hyperbilirubinemia Benchmarking Pediatrics, September 1, 2004; 114(3): 902 - 904. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
 |
||
|
||
Home | My Pediatrics | Journal Information | Current Issue | Past Issues | Subscriptions & Services | Contact Us Click here for faster international access © 2004 American Academy of Pediatrics. All rights reserved. |
||
|
||
