2015 NBR BEST PAPER BY A PHYSICIAN

The Effect of Burn Center Volume on Mortality in

a Pediatric Population: An Analysis of the National
Burn Repository
Erica I. Hodgman, MD, Melody R. Saeman, MD, Madhu Subramanian, MD, and
Steven E. Wolf, MD

Downloaded from https://academic.oup.com/jbcr/article/37/1/32/4582076 by guest on 27 April 2024

The effect of burn center volume on mortality has been demonstrated in adults. The
authors sought to evaluate whether such a relationship existed in burned children.
The National Burn Repository, a voluntary registry sponsored by the American Burn
Association, was queried for all data points on patients aged 18 years or less and treated
from 2002 to 2011. Facilities were divided into quartiles based on average annual burn
volume. Demographics and clinical characteristics were compared across groups, and
univariate and multivariate logistic regressions were performed to evaluate relationships
between facility volume, patient characteristics, and mortality. The authors analyzed 38,234
patients admitted to 88 unique facilities. Children under age 4 years or with larger burns
were more likely to be managed at high-volume and very high–volume centers (57.12 and
53.41%, respectively). Overall mortality was low (0.85%). Comparing mortality across
quartiles demonstrated improved unadjusted mortality rates at the low- and high-volume
centers compared with the medium-volume and very high–volume centers although
univariate logistic regression did not find a significant relationship. However, multivariate
analysis identified burn center volume as a significant predictor of decreased mortality
after controlling for patient characteristics including age, mechanism of injury, burn size,
and presence of inhalation injury. Mortality among pediatric burn patients is low and
was primarily related to patient and injury characteristics, such as burn size, inhalation
injury, and burn cause. Average annual admission rate had a significant but small effect on
mortality when injury characteristics were considered. (J Burn Care Res 2016;37:32–37)

The American Burn Association (ABA) estimates proportion sustaining scalds and/or intentional inju-
that severe burns are responsible for 40,000 annual ries, and the approach to resuscitation.2,3
admissions, of which approximately 30% are chil- Physicians, patient advocacy groups, policy-
dren.1,2 Pediatric burn patients differ from adults in makers, and insurance providers alike are placing
both the pattern of injuries, with a significantly larger increased emphasis on quality improvement initia-
tives in all disciplines of medicine. These initiatives
aim to reduce the mortality and morbidity while
From the Division of Burns, Trauma, and Critical Care, Depart-
ment of General Surgery, University of Texas, Southwestern increasing the patient satisfaction and cost effec-
Medical Center, Dallas. tiveness of care. Leading burn surgeons have long
Funding was received from the National Institutes of Health
(T32GM008593).
recognized the need for establishing and maintain-
Address correspondence to Steven E. Wolf, Division of Burns, ing a high standard of care for the specialized needs
Trauma, and Critical Care, Department of General Surgery, of burn patients to optimize outcomes.4 The ABA
University of Texas, Southwestern Medical Center, 5323 Harry
Hines Boulevard, Dallas, Texas 75390. Email: steven.wolf@ incorporated these standards into the criteria used
utsouthwestern.edu. for the Burn Center Verification process. Although
Copyright © 2015 by the American Burn Association. This is an open-
access article distributed under the terms of the Creative Commons
these guidelines were recently revised, the language
Attribution-Non Commercial-No Derivatives License 4.0 (CCBY- pertaining specifically to verification as a pediatric
NC-ND), where it is permissible to download and share the work burn center remains somewhat vague. Currently,
provided it is properly cited. The work cannot be changed in any way
or used commercially. verified pediatric centers are required to have a
1559-047X/2015 child life specialist available and be able to “dem-
DOI: 10.1097/BCR.0000000000000274 onstrate facilities, protocols, and personnel specific
32
Journal of Burn Care & Research
Volume 37, Number 1 Hodgman et al  33

to the care of critically injured pediatric patients.”5 average annual patient volume. All analyses were
The result is a wide variation in the types of centers completed using R (R Development Core Team,
that care for children, from adult centers to general Vienna, Austria). Descriptive analyses comparing
children’s hospitals, to one of only six verified pedi- demographics, clinical characteristics, and out-
atric burn centers in North America.6 Several stud- comes across the quartiles were completed using
ies have demonstrated that burn center volume does χ2 test, χ2 test for trend, Fisher’s exact test, and
have an effect on mortality.7–9 We therefore sought analysis of variance tests where appropriate. Back-
to evaluate whether patient volume impacted mor- ward and forward stepwise logistic regression was
tality in a pediatric burn population. performed to evaluate the relationships between
facility volume, patient characteristics, and mor-
tality using the rms package in R.10 We selected
METHODS
variables to include in the final model based on the

Downloaded from https://academic.oup.com/jbcr/article/37/1/32/4582076 by guest on 27 April 2024

The National Burn Repository (NBR) is a voluntary significance of each predictor (P < .05) and com-
registry sponsored by the ABA, which is updated annu- pared model iterations based on Akaike’s Informa-
ally.2 This computerized database contains informa- tion Criteria and the r2 values as evidence of the
tion on patient age, sex, race, cause of injury, percent, model’s ability to predict mortality.
and degree of %TBSA burned, presence of inhalational
injury, and pre-existing medical conditions. Verified RESULTS
burn centers are required to contribute to the NBR,
but many other nonverified facilities contribute data as A total of 38,234 patients were admitted to 88 facili-
well; the NBR does not contain information regarding ties. Patient age had a unimodal distribution, with
center verification status and identifies contributing 51.0% of children less than age 4 years (interquartile
centers only by a unique code. For the purposes of this range, 1.3–11 years). Only 325 deaths were reported
study, NBR version 8.0 was queried for all data points, in this population, yielding an overall mortality rate
yielding data on patients treated from 2002 to 2011. of 0.85%. The most common reported cause of death
We restricted our analysis to initial visits for all unique was multisystem organ failure (n = 41) although the
patients aged 0 to 18 years, yielding a total of 38,234 cause of death was omitted or unknown for more
records. Patient age, mechanism of injury, %TBSA than half of the entries. Scalds remain the most com-
burned, presence of inhalational injury, and outcomes mon cause of injury, representing more than 43% of
including mortality were collected. Data within this the database entries, followed by flame (23%) and
set were internally validated by removing duplicated contact (14%) burns.
entries identified using unique patient codes, remov- Only 3.8% of patients had a burn size greater than
ing patients identified as readmissions, and cross-ref- 30% TBSA; mortality in this group was significantly
erencing categories with free text submitted by the higher than patients with a burn size less than 30%
treating facility as part of the database to complete (12.7 vs 0.3%, P < .001). Presence of inhalation
missing values and standardize classification of burn injury was reported in 1072 patients (2.8%), was typ-
mechanism where possible. Institutional review board ically associated with flame burns (69.7% of all inha-
approval was not required since the analysis was done lations), and was also predictive of mortality (12.1
on deidentified data. vs 0.5%, P < .001). Age less than 4 years was associ-
We calculated the facility average annual burn ated with a lower rate of mortality among children
volume by summing the number of burn patients with small (0–29.9% TBSA) burn size and a higher
at each facility and dividing by the number of years rate among children with medium size (30–59.5%
for which the facility submitted data to the NBR. TBSA) injuries (Table 1).
Centers were then separated into quartiles (low, Facilities were grouped into quartiles (low,
medium, high, and very high) based on calculated medium, high, and very high) based on annual

Table 1. Mortality rate stratified by age and burn size

0–4 Yr >4 Yr

Burn Size n Mortality Rate, % n Mortality Rate, % P

Small (0–29.9% TBSA) 18,929 0.25 16,424 0.47 .0006

Medium (30–59.9% TBSA) 426 9.60 716 4.32 .0006
Large (>60% TBSA) 117 29.06 295 27.36 .82
 Journal of Burn Care & Research
34  Hodgman et al January/February 2016

Table 2. Patient characteristics by burn center volume

Quartile

Low Medium High Very High P

n 738 3748 8857 24,891

Average annual admissions (range) 1–15.5 17.8–35.7 35.8–87.3 90.3–386.7
Age (mean ± SD) 12.52 ± 6.34 7.67 ± 6.44 5.67 ± 5.53 5.93 ± 5.59 <.001
Age category (%) <.001
 0–4 yr 19.1 44.53 57.12 53.41
 >4 yr 80.89 55.47 42.88 46.59
Sex (% male) 71.41 65.82 62.79 63.11 <.001
Mechanism (%) <.001

Downloaded from https://academic.oup.com/jbcr/article/37/1/32/4582076 by guest on 27 April 2024

 Scald 27.73 46.38 56.26 47.11
 Flame 53.41 33.11 21.93 26.15
 Contact 9.19 9.32 14.59 17.38
 Grease 3.49 5.74 2.5 3.05
 Chemical 1.74 1.38 1.11 1.14
 Other 4.44 4.06 3.6 5.17
Inhalational injury (%) 4.81 3.96 3.16 3.08 .008
ICU stay (%) 49.05 31.37 44.09 20.83 <.001
Ventilated (%) 13.51 8.09 5.59 8.61 <.001
Burn size (%TBSA) (%) <.001
 0–9.9 73.05 75.52 78.31 78.46
 10–19.9 14.09 17.42 16.32 12.93
 20–29.9 5.47 3.97 3.16 3.6
 30–39.9 3.28 1.28 0.99 1.82
 40–49.9 2.33 0.78 0.51 1.04
 50–59.9 0.82 0.33 0.2 0.75
 60–69.9 0.41 0.22 0.2 0.56
 70–79.9 0.27 0.14 0.13 0.34
 80–89.9 0 0.19 0.09 0.27
 90–100 0.27 0.14 0.09 0.23
Disposition (%) <.001
 Home 91.1 92.49 86.93 94.66
 Transfer 3.71 1.74 9.02 1.29
 Rehabilitation/skilled nursing 3.86 1.98 1.5 1.64
facility
 Discharged to alternate caregiver 0.45 2.65 2.04 1.39
Mortality (%) 0.678 1.04 0.49 0.95 .0004

ICU, intensive care unit.

average number of admissions. Table 2 compares medium-volume centers were significantly more
basic demographic data as well as cause and size of likely to be transferred (3.71 and 9.02%) to another
burn, presence of inhalation injury, and mortality hospital or service on discharge than children at
between the quartiles. Significant differences were medium-volume and very high–volume centers
found between groups in nearly all patient char- (1.74 and 1.29%, respectively, P < .001). Mortality
acteristics examined, including patient age, sex, was low overall, but significantly lower rates were
%TBSA burned, and inhalation injury (P < .001). seen at centers in the low- and high-volume centers
Low- and medium-volume centers cared predomi- on this univariate analysis.
nantly for patients aged 4 years and older (80.89 As expected, children with larger burns were
and 55.47%, respectively), while a slight majority treated at the higher-volume centers (Figure 1).
of children were less than 4 years old at the high As burn size increased, the proportion of children
and very high centers (57.12 and 53.41%, respec- managed at low- and high-volume centers decreased
tively). A greater proportion of flame burns were (P = .002 and P < .001, respectively); similarly,
seen at low-volume centers. Children at low- and as burn size increased, the proportion of children
Journal of Burn Care & Research
Volume 37, Number 1 Hodgman et al  35

90.00%
mortality among pediatric burn patients. Patient
80.00%
and injury characteristics, including age, burn size,
70.00%
presence of inhalation injury, and cause of injury
Low Volume
60.00%
remain the most influential predictors of mortality.
50.00%
Medium After adjusting for these factors, the annual admis-
40.00% Volume
sion rate has a very small but nevertheless statisti-
30.00% High Volume
cally significant effect on observed mortality. The
20.00%
Very High lowest unadjusted mortality rates are found at the
10.00% Volume
low- and high-volume centers. The highest-volume
0.00%
centers are more likely to care for younger patients
and patients with larger injuries, and when patient
and injury characteristics are considered, the lowest

Downloaded from https://academic.oup.com/jbcr/article/37/1/32/4582076 by guest on 27 April 2024

Burn Size (% TBSA)
mortality was found in the higher-volume centers.
Figure 1. Proportion of children receiving care by center These findings are concordant with the bulk of the
volume and burn size. literature on pediatric burns and severe burns in
general.
managed at the very high–volume centers increased Conceptually, it is expected that higher mortal-
(P < .001). We found no significant changes in ity rates will be found among younger children
the medium-volume centers. When stratified by with immature immune, pulmonary, cardiovascular,
burn size and facility volume without adjustment and neurologic systems although this issue remains
for other factors such as mechanism of injury or somewhat controversial in the literature.10 While
presence of inhalation injury, we found no sig- several single-institution retrospective studies failed
nificant differences in mortality between quartiles to find an increased risk of death among young chil-
(Table 3). Unadjusted univariate logistic regression dren, a previous NBR review demonstrated worse
did not find center volume alone, either as a con- outcomes in children under the age of 4 years similar
tinuous variable or when divided into a categorical to what was found here.11–13 While a single-institu-
variable based on quartiles, to be a significant pre- tion study has the advantage of eliminating reporting
dictor of mortality. errors inherent to a national database review, it also
Multivariate logistic regression was used to esti- likely lacks the power to detect smaller differences, as
mate the risk of death for burn center volume, the overall mortality rate among children with severe
adjusting for age, presence of inhalational injury, burns is quite low. We also found no difference in
and %TBSA burned. This yielded a model with a mortality in the largest burns (>60% TBSA) between
moderate r2 (.35) and high C-statistic (.93). We the age groups, but we did find a marked difference
found the evidence for systematic overfitting and in those with burns between 30 and 60% TBSA, with
underestimation of mortality at higher probabilities a 122% increase in mortality rate in those less than
of mortality, but correction using bootstrapping did 4 years old.
not alter the C-statistic. Table 4 demonstrates the Prior studies of adult burn populations dem-
mortality odds as predicted by this model. Interest- onstrate a similar nonlinear relationship between
ingly, “other” causes of injury are associated with patient volume and outcomes, with the lowest
an increased risk of mortality; unfortunately, this adjusted mortality rate seen at the medium-volume
is difficult to interpret as this category represents centers.7,8,14 Both Light et al8 and Hranjec et al7
a highly heterogeneous population with desqua- postulate that the slightly worse outcomes seen
mating skin diseases and/or infections, degloving at the highest-volume centers could be related to
injuries, radiation injuries, and any other injury the the higher volumes “overwhelming” the system
treating center described as “other nonburn.” A although the retrospective nature of these stud-
subset analysis of children under the age of 4 years ies precludes additional investigation of this the-
yielded similar results, with burn center volume ory. We demonstrate here that the effect is likely
again demonstrating a small but significant impact a referral bias in that the most severely injured are
on mortality (results not published). cared for in the higher-volume centers, which have
more experience and/or resources available for the
DISCUSSION management of pediatric patients. Indeed, transfer
rates are higher at the low- and high-volume cen-
Before adjusting for other variables, facility aver- ters. In addition, the larger proportions of younger
age annual admission rate does have an impact on children or children with injuries classified as
 Journal of Burn Care & Research
36  Hodgman et al January/February 2016

Table 3. Mortality stratified by burn center volume and burn size

Annual Burn Center Admission Volume (Quartile)

Burn Size (%TBSA) Low, % Medium, % High, % Very High, % P

0–9.9 0.00 0.52 0.18 0.38 .37

10–19.9 0.00 0.16 0.28 0.29 .50
20–29.9 0.00 2.10 0.73 1.04 .79
30–39.9 8.33 0.00 5.81 4.61 .91
40–49.9 0.00 14.29 9.09 5.62 .57
50–59.9 0.00 33.33 0.00 10.61 .60
60–69.9 0.00 37.50 23.53 14.81 .31
70–79.9 50.00 20.00 27.27 13.58 .13

Downloaded from https://academic.oup.com/jbcr/article/37/1/32/4582076 by guest on 27 April 2024

80–89.9 — 42.86 37.50 43.75 .86
90–100 100.00 60.00 62.50 50.00 .18

“other” treated at the highest-volume centers may skilled nursing or rehabilitation facilities at higher-
also partially explain the slightly worse mortality volume centers.
rates seen. The conclusions of this study are limited by the fact
Palmieri et al9 recently published their analysis that the data contained within the NBR vary widely
of a pediatric subset from an older version of the in terms of the quality and consistency of reporting
NBR. In that study, the authors eliminated centers by contributing centers. Taylor et al15 addressed this
that did not treat a child under the age of 10 years issue to the best of their ability in their recent valida-
and then divided the remaining centers into five tion of the NBR, but missing data across important
groups based on predefined median yearly admis- data fields remain an issue. Taylor et al also identi-
sion rate, resulting in only seven centers being fied a number of patients who were readmitted for
included in the extra-large-volume and six cen- additional care but submitted to the NBR under a
ters in the large-volume groups. The authors did new identification code. Although we attempted to
find a linear improvement in mortality as volume remove all duplicate patients, it is still likely that our
increased. The use of quartiles instead of estimated population nonetheless contains some of these. In
division into groups in our study likely accounts addition, systematic differences in the reporting of
in part for the lack of a clear linear relationship of covariate factors used in the multivariate analysis (eg,
volume to mortality. a center’s routine omission of data or underestima-
For the purpose of this study, we assumed that tion or overestimation of burn injury size) could
average admission rate would act as a surrogate alter the apparent effect of facility volume.
marker for the institutional expertise accrued
in the course of caring for a large number of Table 4. Multivariate predictors of mortality
young patients. Although we did not find a lin-
β Coefficient OR (95% CI)
ear improvement in mortality with increasing burn
center volume, mortality is not the only marker of Intercept −13.84 OR (95% CI)
quality burn care. Facilities with significant pedi-
Burn center annual 0.23 1.26 (1.03–1.50)
atric experience have access to physiatrists, psy- admission volume
chiatrists, social workers, and case managers with quartile
significant pediatric experience. These resources Age −0.015 0.99 (0.97–1.01)
may lead to improvements in other outcome mea- Burn size 4.66 105.69 (66.91–167.00)
sures, such as complication rates, quality of life, Inhalational injury 1.8 6.08 (4.39–8.41)
functional recovery, and overall patient/family Mechanism
satisfaction. We believe that future studies evaluat- (reference = scald)
ing the impact of institutional expertise in caring  Flame 0.52 1.68 (1.04–2.71)
for pediatric patients on these outcomes are war-  Contact −2.2 0.11 (0.02–8.1)
ranted. Such studies will provide additional insights  Grease 0.78 2.17 (0.76–6.21)
 Chemical −6.14 0 (0–1.5 × 1014)
into the value of pediatric-specific resources and
 Other 1.76 5.82 (3.77–8.98)
may guide future quality initiatives. This may be
demonstrated by the lower rates of discharge to CI, confidence interval; OR, odds ratio.
Journal of Burn Care & Research
Volume 37, Number 1 Hodgman et al  37

REFERENCES burn center volume on the mortality of burned adults—an

analysis of the data in the National Burn Repository. J Burn
1. American Burn Association. Burn Incidence and Treatment Care Res 2009;30:776–82.
in the United States: 2013 Fact Sheet. 2013; available from 9. Palmieri TL, Taylor S, Lawless M, Curri T, Sen S, Greenhalgh
http://www.ameriburn.org/resources_factsheet.php; 14 DG. Burn center volume makes a difference for burned chil-
Jan. 2015. dren. Pediatr Crit Care Med 2015;16:319–24.
2. American Burn Association. 2014 NBR Annual Report; 10. Taylor SL, Lawless M, Curri T, Sen S, Greenhalgh DG,
2014. Palmieri TL. Predicting mortality from burns: the need for
3. Abeyasundara SL, Rajan V, Lam L, Harvey JG, Holland AJ. age-group specific models. Burns 2014;40:1106–15.
The changing pattern of pediatric burns. J Burn Care Res 11. Thombs BD, Singh VA, Milner SM. Children under 4 years
2011;32:178–84. are at greater risk of mortality following acute burn injury:
4. Brigham PA, Dimick AR. The evolution of burn care facili- evidence from a national sample of 12,902 pediatric admis-
ties in the United States. J Burn Care Res 2008;29:248–56. sions. Shock 2006;26:348–52.
5. American Burn Association. American Burn Association— 12. Sheridan RL, Weber JM, Schnitzer JJ, Schulz JT, Ryan CM,
Burn Center Verification Program Verification Criteria. 2015; Tompkins RG. Young age is not a predictor of mortality in

Downloaded from https://academic.oup.com/jbcr/article/37/1/32/4582076 by guest on 27 April 2024

available from http://www.ameriburn.org/Verification/ burns. Pediatr Crit Care Med 2001;2:223–4.
CriterionDeficiencies.pdf. Accessed February 4, 2015. 13. Tompkins RG, Remensnyder JP, Burke JF, et al. Significant
6. American Burn Association. Burn Center Verification. 2014; reductions in mortality for children with burn injuries
available from http://www.ameriburn.org/verification_ver- through the use of prompt eschar excision. Ann Surg
ifiedcenters.php; 4 Feb. 2015. 1988;208:577–85.
7. Hranjec T, Turrentine FE, Stukenborg G, Young JS, Sawyer 14. Allorto NL, Oosthuizen GV, Clarke DL, Muckart DJ. The
RG, Calland JF. Burn-center quality improvement: are burn spectrum and outcome of burns at a regional hospital in
outcomes dependent on admitting facilities and is there a South Africa. Burns 2009;35:1004–8.
volume-outcome “sweet-spot”? Am Surg 2012;78:559–66. 15. Taylor SL, Lee D, Nagler T, Lawless MB, Curri T, Palmieri
8. Light TD, Latenser BA, Kealey GP, Wibbenmeyer LA, TL. A validity review of the National Burn Repository.
Rosenthal GE, Sarrazin MV. The effect of burn center and J Burn Care Res 2013;34:274–80.