search text   Advanced Search

This Article Abstract Full Text (PDF) P3Rs: Submit a response P3Rs: View responses Alert me when this article is cited Alert me when P3Rs are posted Alert me if a correction is posted Citation Map Services E-mail this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Add to My File Cabinet Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via CrossRef Citing Articles via ISI Web of Science (3) Citing Articles via Google Scholar Google Scholar Articles by Donoghue, A. J. Search for Related Content PubMed PubMed Citation Articles by Donoghue, A. J. Related Collections Emergency Medicine

Effect of Hospital Characteristics on Outcomes From Pediatric Cardiopulmonary Resuscitation: A Report From the National Registry of Cardiopulmonary Resuscitation

^a Divisions of Emergency Medicine
^b Critical Care Medicine, Children's Hospital of Philadelphia, Philadelphia, Pennsylvania
^c Department of Biostatistics and Epidemiology, University of Pennsylvania School of Medicine, Philadelphia, Pennsylvania

	ABSTRACT

TOP ABSTRACT METHODS RESULTS DISCUSSION REFERENCES

 
OBJECTIVE. Cardiac arrest is uncommon among pediatric patients. Prehospital data demonstrate differences in care processes between children and adults receiving cardiopulmonary resuscitation and advanced life support. We sought to evaluate whether children receiving in-hospital cardiopulmonary resuscitation would attain superior 24-hour survival in hospitals with a higher level of pediatric physician staffing, greater intensity of pediatric care services, and higher pediatric patient volume.

METHODS. A retrospective cohort of 778 hospital inpatients aged <18 years receiving cardiopulmonary resuscitation was identified from the National Registry of Cardiopulmonary Resuscitation from January 2000 to December 2002. Data on hospital pediatric facilities were obtained via telephone survey. Univariate analyses comparing 24-hour survivors and nonsurvivors were conducted using Wilcoxon rank-sum testing for continuous variables and ² analysis for dichotomous variables. Multivariate regression analysis was done to examine hospital characteristics as independent predictors of 24-hour survival.

RESULTS. Complete data were available for 677 patients. Univariate analyses showed an association between several pediatric-specific facility characteristics and 24-hour survival. After accounting for indicators of pre-event clinical condition and monitoring, multivariate analysis showed improved 24-hour survival in hospitals staffed by pediatric residents and surgeons and pediatric residents, surgeons, and fellows than for hospitals with no pediatric physician staffing or pediatric surgeons alone. Measures of available facilities and patient volume were not associated with improved outcome.

CONCLUSIONS. Improved 24-hour survival for children receiving in-hospital cardiopulmonary resuscitation is associated with the presence of pediatric residents and fellows.

Key Words: cardiopulmonary resuscitation • hospital performance • cardiac arrest

Abbreviations: CPR—cardiopulmonary resuscitation • NRCPR—National Registry of Cardiopulmonary Resuscitation • NoPeds—no pediatric housestaff or surgeons • PedSurg—pediatric surgeons only • PedRes—pediatric surgeons and residents • PedFellow—pediatric surgeons—residents—and pediatric emergency medicine and/or pediatric critical care medicine fellows • OR—odds ratio • CI—confidence interval • ED—emergency department

Cardiac arrest is an uncommon phenomenon in children. Differences in physiology and etiology between adults and children have led to the creation of separate treatment algorithms for cardiac arrest in children and adults. Research in pediatric cardiac arrest and cardiopulmonary resuscitation (CPR) has not yielded consistent evidence that the most commonly used processes of care in pediatric CPR improve clinical outcomes. A recent collective review by Young and Seidel¹ reported an overall survival rate of 13% among out-of-hospital and in-hospital arrests combined and a high prevalence of neurocognitive morbidity among survivors. Outcomes from inpatient pediatric CPR are better, with 24-hour survival rates ranging from 33% to 37% and less neurologic morbidity than in patients receiving out-of-hospital CPR.^2–4

Although numerous studies have demonstrated significant differences in processes of care for children receiving CPR and advanced life support when compared with adults in the prehospital setting,^5–7 little is known about how processes of care or hospital resources impact survival. Differences in processes of care and clinical outcomes for children by hospital characteristics have been investigated in other clinical contexts, such as abdominal trauma, intussusception, and procedural sedation and analgesia.^8–10 A number of these studies have looked specifically at differences between processes of care in children's hospitals and hospitals that care for adults, as well as children.

This study was designed to examine the effect of hospital and physician staffing characteristics on 24-hour survival after in-hospital CPR in children. Specifically, we examined whether a higher level of pediatric physician staffing, increased volume of pediatric patients, or increased pediatric specialization were associated with improved 24-hour survival after CPR while controlling for likely clinical confounders such as age, location of arrest, underlying diagnosis, and intensity of pre-event clinical support.

	METHODS

TOP ABSTRACT METHODS RESULTS DISCUSSION REFERENCES

 
A retrospective cohort of all patients aged <18 years who required in-hospital CPR were selected from the National Registry of Cardiopulmonary Resuscitation (NRCPR) during the 36-month period between January 1, 2000, and December 31, 2002. The NRCPR is a multisite database sponsored by the American Heart Association in which hospitals voluntarily participate. Participating hospitals pay an annual fee for data support and report generation. Data abstractors at participating hospitals are trained and certified before beginning data collection. Data quality assurance measures include data completeness reports, random reabstraction of data with error rate reports, and intrinsic data quality checks in database software. During our study period, 258 hospitals in 49 states were participating in the NRCPR. No standard indices of rural or urban hospital status are used.

Patients enrolled in the database are those who experience a resuscitation event, defined as "cardiopulmonary arrest requiring chest compressions and/or defibrillation of ventricular fibrillation or pulseless ventricular tachycardia," and for whom "a hospital-wide (eg, for general inpatient area) or unit-based emergency response by facility personnel" is prompted. Patients for whom resuscitation efforts were altered because of palliative care or preexisting do-not-attempt-resuscitation orders were excluded. Neonates who received CPR at the time of delivery or in the NICU were also excluded. An additional subset of patients were identified that seemed consistent with cases of sudden infant death syndrome with out-of-hospital cardiopulmonary arrest; these children were <1 year of age, underwent CPR in the emergency department, had a presenting rhythm of asystole, had no preexisting medical conditions, and did not survive. Thirteen patients in all met these criteria. Despite the fact that the attempted resuscitation of these patients began in the hospital and not in the prehospital phase, these children were deemed by investigator consensus to have likely suffered out-of-hospital arrests and, as such, were excluded.

Data included in the analysis for each patient were age, location of the event within the hospital, illness category, whether or not the patient was pulseless at onset of CPR, and details of preexisting clinical interventions when the event occurred. For pre-event clinical interventions, a separate categorical variable was generated as shown in Table 1. A list of hospital characteristics pertinent to pediatric care were generated by the investigators and is shown in Table 2. Values for these variables were obtained by telephone interviews and were self-reported by hospital personnel designated as contacts for the NRCPR database. Data elements not collected by this method were obtained by alternate hospital contacts, from NRCPR facility data, or from the National Resident Matching Program Web site (www.nrmp.org).

The outcome of interest was survival at 24-hours after CPR event. This was determined by the presence of return of spontaneous circulation in the absence of external chest compressions and a postevent length of stay of 1 day. This outcome likely reflects the effects of both resuscitation and postresuscitation interventions on short-term clinical outcome. Patients for whom postevent length of stay was missing from the database (n = 7) were excluded. Patients for whom postevent length of stay was <1 day and whose discharge disposition was coded as "discharged alive" (n = 27) were also excluded. On subsequent examination, these 27 patients were children for whom CPR was initiated at their enrolling hospital, achieved return of spontaneous circulation, and were transferred to other facilities. These patients, whereas not included in the main analysis, were examined separately to determine the potential for bias resulting from their exclusion (see below).

Descriptive statistics were performed on all of the variables; measures of spread in continuous variables were expressed as medians and ranges. Univariate analyses were performed between survivors and nonsurvivors by ² analysis for all of the hospital characteristic variables and for several clinical variables that were potential sources of confounding (age category, shockable versus nonshockable rhythm, level of pre-event clinical intervention, event location, and illness category). Multivariate regression analysis was performed to examine the effect of hospital characteristics on 24-hour survival while controlling for clinical confounders, with relationships between predictor and 24-hour survival expressed as odds ratios (ORs) with 95% confidence intervals (CIs). Generalized estimating equations were used to account for clustering by hospital.

Pairwise comparisons for event locations and diagnostic categories demonstrated significant differences in 24-hour survival for events that occurred in the emergency department ([ED] worse survival compared with all other locations), patients whose diagnostic category was trauma (worse survival compared with all other categories), and patients whose diagnostic category was surgical (better survival compared with all other categories). The distribution of event locations and diagnostic categories also differed by hospital category. Therefore, these variables were coded as ED versus other for event location and trauma versus surgical versus other for diagnostic category and were included in the multivariate model as potential clinical confounders. In addition, the categorical variable for pre-event intervention level was included in the model as a surrogate measure of illness severity and clinical status. The final model was the result of a stepwise multivariate regression where the initial model included a complete set of all hospital characteristics. Characteristics of which the ß coefficients yielded P values <0.2 were also considered for inclusion in the final model as possible independent predictors of survival. All of the potential clinical confounders remained in our final model irrespective of the level of significance in the preliminary analysis with the exception of shockable rhythm, which had no association with our exposure or outcome of interest on univariate analysis.

All of the statistical analyses were performed by using Stata 8.0 software (Stata Corp, College Station, TX). All aspects of this study were reviewed and approved by the Children's Hospital of Philadelphia Institutional Review Board and the Science Advisory Board of the American Heart Association NRCPR.

	RESULTS

TOP ABSTRACT METHODS RESULTS DISCUSSION REFERENCES

 
A total of 23 535 events were recorded during the study period. Of these, 778 (3.3%) occurred in children <18 years. Among the 778 pediatric events, 34 had either incomplete or no data on survival at 24 hours after event; 13 were apparent sudden infant death syndrome or delivery room events; and 54 occurred in hospitals with incomplete facility data. Thus, 677 events were included in the final analysis (see Figure 1).

View larger version (20K): [in this window] [in a new window]   FIGURE 1 Flow diagram for patient inclusion.

	DISCUSSION

TOP ABSTRACT METHODS RESULTS DISCUSSION REFERENCES

 
Our study suggests that 24-hour survival for children requiring inpatient CPR is associated with the presence of pediatric physicians when compared with hospitals without pediatric personnel. The presence of pediatric residents was associated with the greatest improvement in 24-hour survival. We did not find any significant association on multivariate analysis between the presence or size of specialized care areas, such as ICUs, and improved 24-hour survival. We also found no association on multivariate analysis between measures of patient volume and 24-hour survival.

Our findings with respect to survival after CPR extend those of previously published studies on pediatric inpatients. Emerging epidemiological data on pediatric inpatient CPR has suggested that clinical outcomes are superior to those from prehospital CPR and that greater understanding of the epidemiology of inpatient CPR in children is necessary for improving treatment and outcomes further. Studies from Brazil,² Finland,³ and the United States⁴ have examined outcomes from inpatient pediatric CPR while using consensus Utstein definitions; 24-hour survival rates from these studies ranged from 33% to 37%. These studies, however, have been predominantly conducted in single centers and have not attempted to examine relationships between characteristics of different hospitals and patient outcomes. Our overall 24-hour survival rate of 51% is, in fact, higher than these studies. Given the demonstrated association between a higher level of pediatric staffing and improved outcome, the fact that a large majority of our patients received CPR in hospitals staffed by pediatric residents and/or fellows may account for this elevated survival rate.

The association between improved care processes and outcomes for children in hospitals with greater levels of pediatric specialization has been demonstrated in previous studies of children with blunt splenic injury, intussusception, and ED procedural sedation and analgesia.^8–10 These studies have suggested that variation in pediatric volumes may have an effect on both practice patterns and outcomes. Studies of PICUs have demonstrated relationships between the presence of pediatric critical care fellows and improved survival,¹¹ as well as patient volume and improved survival.¹² In our study, we identified pediatric-trained physician staffing as a predictor of improved outcome, whereas other measures of patient volume were not associated with improved survival after CPR.

The majority of the children in our data set were inpatients at hospitals with pediatric housestaff and tertiary pediatric services. Children cared for in pediatric hospitals have greater complexity and severity of underlying illness and are likely at greater risk for clinical events requiring CPR. Children at such hospitals had improved odds of survival after CPR in our study, despite these differences. Resuscitation training in pediatric hospitals likely confers a greater familiarity with developmental physiology, equipment, medication indications and dosing, and other processes of care essential for critically ill children. Our data demonstrate that the presence of physicians with such pediatric-specific training is associated with improved short-term outcomes after CPR.

Survival at 24 hours after cardiac arrest is a discreet, measurable outcome of clinical significance that reflects resuscitation and postresuscitation interventions more completely than even shorter-term outcomes, such as return of spontaneous circulation. Longer-term outcomes, such as survival to hospital discharge and neurologically intact survival, although of great clinical importance, were not examined in this study for a number of reasons. First, underlying medical conditions are likely to have a greater effect on long-term outcomes after CPR, and the distribution of patients with such conditions was irregular across hospital types. Second, with a longer time horizon to our outcome of interest, factors related to regionalization of pediatric tertiary care and interfacility transport are likely to be of significance and were beyond the scope of the source of data used. Third, practices related to end-of-life and palliative care in children are more likely to emerge for children who survive such an event. These practices are common in pediatric intensive care settings¹³ but vary significantly by hospital and patient characteristics.^14,15

Several limitations to our study should be acknowledged. The NRCPR database collects data on all patients requiring CPR, but the vast majority of patients enrolled are adults. The data collection form reflects this fact in that the list of diagnoses and clinical interventions are replete with conditions common to adults, such as myocardial infarction, stroke, and so forth. In our analysis, the diagnostic category had no significant confounding effect on our exposures of interest. It is probable that severity of illness for children is incompletely captured by the data elements available from the NRCPR. In our present model, the categorical variable reflecting the pre-event level of clinical intervention suggests that patients have optimal survival likelihood when patients are in an ICU but do not require assisted ventilation or hemodynamic support. We believe that this variable characterizes the prearrest severity of illness in an accurate way, and our results remain significant while controlling for this factor.

Children's hospitals with tertiary services inevitably have higher caseloads of ill children than general centers. A greater knowledge and experience with children with specific complex diseases may allow patient care teams to anticipate and prevent complications to a greater extent that facilities without the same level of experience. Patients benefiting from this phenomenon would not be enrolled in the database, despite being part of the population at risk, and, as such, this may result in an underestimate of the effect of tertiary pediatric facilities. In addition, subjects enrolled from such facilities may, therefore, represent a group of patients with higher illness severity. This may result in a biased distribution of illness severity among hospitals, with the children in tertiary centers being more severely ill and possibly less likely to survive. Therefore, our results may underestimate the true association between variables common in pediatric hospitals with 24-hour survival.

The centers contributing data to the NRCPR during our study period comprise 10% of all hospitals in the United States, and although they represent a wide variety of hospital sizes and settings, the database does constitute a convenience sample. Enrollment in the NRCPR is voluntary, and the possibility exists of bias based on differences in the catchment areas, regionalization patterns, or other unknown hospital-level factors. In our study, the use of a population-based generalized estimating equation attempted to control for clustering effects by hospital, but additional covariates that may be associated with the hospital to which a child was admitted are difficult to measure. Although the use of propensity scores may allow an estimate of this, the absence of significant collinearity between our measured variables for hospital characteristics makes it difficult to construct these scores in our data set.

The efficacy and indications for CPR in a patient with a pulsatile rhythm but gross hypoperfusion is not clear. Current pediatric advanced life support guidelines state that bradycardia in the presence of poor perfusion is an indication for chest compressions but also acknowledge that an absolute heart rate at which CPR should be initiated has not been identified by current data.¹⁶ In our data set, 289 patients (42%) had CPR initiated while a pulse was present. Although the case definition for the NRCPR is cardiopulmonary arrest requiring chest compressions or defibrillation, it is possible that some of these patients who were not pulseless at the onset of the systematic response received chest compressions when they were not indicated. Because presence of a pulse was univariately associated with both hospital type and 24-hour survival, it was necessary to exclude is as both a confounder and a source of effect modification in our multivariate analyses. Although our data did not, on multivariate analysis, demonstrate an association between 24-hour survival and the presence of a pulse at the initiation of CPR, this group of children bears further study to delineate patient characteristics and clinical interventions that affect their outcomes.

Differences in 24-hour survival from cardiac arrest based on availability of pediatric personnel have potential implications with respect to hospital physician and nursing staffing, mandatory training and certification of caretakers of hospitalized children, and baseline preparedness for critical situations of hospitals that admit children. Future research is required to identify specific processes of care or clinical interventions that are associated with improved short- and longer-term outcomes for infants, children, and adolescents who suffer cardiac arrest, toward which improvements in readiness and knowledge can be directed.

	ACKNOWLEDGMENTS

 
This study was funded in part by National Institutes of Health grant T32 HD 043021 and the American Heart Association Emergency Cardiovascular Care Committee, National Registry of Cardiopulmonary Resuscitation.

We thank Robert A. Berg, MD, for ongoing review and comments; Scott Carey for help with data acquisition and database administration; Stanley Dunn for extensive contributions to data collection and management; and Mark A. Helfaer, MD, and Kathy Shaw, MD, MSCE, for mentorship and administrative support.

	FOOTNOTES

 
Accepted Apr 11, 2006.

Address correspondence to Aaron J. Donoghue, MD, MSCE, Divisions of Emergency Medicine and Critical Care Medicine, Children's Hospital of Philadelphia, 34th Street and Civic Center Boulevard. E-mail: donoghue{at}email.chop.edu

The authors have indicated they have no financial relationships relevant to this article to disclose.

	REFERENCES

TOP ABSTRACT METHODS RESULTS DISCUSSION REFERENCES

 

1. Young KD, Seidel JS. Pediatric cardiopulmonary resuscitation: a collective review. Ann Emerg Med. 1999;33 :195 –205[CrossRef][ISI][Medline]
2. Reis AG, Nadkarni V, Perondi MB, Grisi S, Berg RA. A prospective investigation into the epidemiology of in-hospital pediatric cardiopulmonary resuscitation using the international Utstein reporting style. Pediatrics. 2002;109 :200 –209[Abstract/Free Full Text]
3. Suominen P, Olkkola KT, Voipio V, Korpela R, Palo R, Rasanen J. Utstein style reporting of in-hospital paediatric cardiopulmonary resuscitation. Resuscitation. 2000;45 :17 –25[CrossRef][ISI][Medline]
4. Torres A, Jr, Pickert CB, Firestone J, Walker WM, Fiser DH. Long-term functional outcome of inpatient pediatric cardiopulmonary resuscitation. Pediatr Emerg Care. 1997;13 :369 –373[ISI][Medline]
5. Aijian P, Tsai A, Knopp R, Kallsen GW. Endotracheal intubation of pediatric patients by paramedics. Ann Emerg Med. 1989;18 :489 –494[CrossRef][ISI][Medline]
6. Kumar VR, Bachman DT, Kiskaddon RT. Children and adults in cardiopulmonary arrest: are advanced life support guidelines followed in the prehospital setting? Ann Emerg Med. 1997;29 :743 –747[CrossRef][ISI][Medline]
7. Losek JD, Hennes H, Glaeser P, Hendley G, Nelson DB. Prehospital care of the pulseless, nonbreathing pediatric patient. Am J Emerg Med. 1987;5 :370 –374[CrossRef][ISI][Medline]
8. Davis DH, Localio AR, Stafford PW, Helfaer MA, Durbin DR. Trends in operative management of pediatric splenic injury in a regional trauma system. Pediatrics. 2005;115 :89 –94[Abstract/Free Full Text]
9. Krauss B, Zurakowski D. Sedation patterns in pediatric and general community hospital emergency departments. Pediatr Emerg Care. 1998;14 :99 –103[ISI][Medline]
10. Bratton SL, Haberkern CM, Waldhausen JH, Sawin RS, Allison JW. Intussusception: hospital size and risk of surgery. Pediatrics. 2001;107 :299 –303[Abstract/Free Full Text]
11. Pollack MM, Patel KM, Ruttimann E. Pediatric critical care training programs have a positive effect on pediatric intensive care mortality. Crit Care Med. 1997;25 :1637 –1642[CrossRef][ISI][Medline]
12. Marcin JP, Song J, Leigh JP. The impact of pediatric intensive care unit volume on mortality: a hierarchical instrumental variable analysis. Pediatr Crit Care Med. 2005;6 :136 –141[CrossRef][Medline]
13. Vernon DD, Dean JM, Timmons OD, Banner W Jr, Allen-Webb EM. Modes of death in the pediatric intensive care unit: withdrawal and limitation of supportive care. Crit Care Med. 1993;21 :1798 –1802[ISI][Medline]
14. Gilban S, Kumar D, de Caprariis PJ, Olivieri F, Ho K. Pediatric AIDS and advanced directives: a three-year prospective study in New York State. AIDS Patient Care STDS. 1996;10 :168 –170[ISI][Medline]
15. Lantos JD, Berger AC, Zucker AR. Do-not-resuscitate orders in a children's hospital. Crit Care Med. 1993;21 :52 –55[ISI][Medline]
16. Part 11: Pediatric Basic Life Support. Circulation. 2005;112 :IV156 –IV166

This article has been cited by other articles:

				D. W. Jones, E. D. Peterson, R. O. Bonow, F. A. Masoudi, G. C. Fonarow, S. C. Smith Jr, P. Solis, M. Girgus, P. C. Hinton, A. Leonard, et al. Translating Research Into Practice for Healthcare Providers: The American Heart Association's Strategy for Building Healthier Lives, Free of Cardiovascular Diseases and Stroke Circulation, August 5, 2008; 118(6): 687 - 696. [Abstract] [Full Text] [PDF]

				R. R. Thiagarajan, P. C. Laussen, P. T. Rycus, R. H. Bartlett, and S. L. Bratton Extracorporeal Membrane Oxygenation to Aid Cardiopulmonary Resuscitation in Infants and Children Circulation, October 9, 2007; 116(15): 1693 - 1700. [Abstract] [Full Text] [PDF]

				D. C. Goff Jr, L. Brass, L. T. Braun, J. B. Croft, J. D. Flesch, F. G.R. Fowkes, Y. Hong, V. Howard, S. Huston, S. F. Jencks, et al. Essential Features of a Surveillance System to Support the Prevention and Management of Heart Disease and Stroke: A Scientific Statement From the American Heart Association Councils on Epidemiology and Prevention, Stroke, and Cardiovascular Nursing and the Interdisciplinary Working Groups on Quality of Care and Outcomes Research and Atherosclerotic Peripheral Vascular Disease Circulation, January 2, 2007; 115(1): 127 - 155. [Full Text] [PDF]

P³Rs:

This Article Abstract Full Text (PDF) P3Rs: Submit a response P3Rs: View responses Alert me when this article is cited Alert me when P3Rs are posted Alert me if a correction is posted Citation Map Services E-mail this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Add to My File Cabinet Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via CrossRef Citing Articles via ISI Web of Science (3) Citing Articles via Google Scholar Google Scholar Articles by Donoghue, A. J. Search for Related Content PubMed PubMed Citation Articles by Donoghue, A. J. Related Collections Emergency Medicine

	Home | My Pediatrics | Journal Information | Current Issue | Past Issues | Subscriptions & Services | Contact Us Click here for faster international access © 2006 American Academy of Pediatrics. All rights reserved.