International Journal of Pediatric Otorhinolaryngology (2008) 72, 1837—1843
www.elsevier.com/locate/ijporl
Pediatric retropharyngeal abscesses: A national perspective Lina Lander a, Sam Lu b, Rahul K. Shah c,* a
Department of Epidemiology, University of Nebraska Medical Center, Omaha, NE, United States University of Arkansas Medical School, Little Rock, AR, United States c Division of Otolaryngology, Children’s National Medical Center, George Washington University Medical School, 111 Michigan Avenue, NW, Washington, DC 20010, United States b
Received 30 June 2008; received in revised form 28 August 2008; accepted 2 September 2008 Available online 15 October 2008
KEYWORDS Resource utilization; Outcomes; Patient quality; Retropharyngeal abscesses; Kids’ Inpatient Database (KID); Healthcare Cost and Utilization Project (HCUP)
Summary Objectives: To determine the resource utilization and national variation in the management of pediatric retropharyngeal abscesses. Methods: The Kids’ Inpatient Database (KID) 2003 was analyzed. International Classification of Diseases, Ninth Revision code 478.24 was the inclusion criteria. Results: One thousand three hundred and twenty-one admissions with retropharyngeal abscess were sampled from the KID in 2003; there were no deaths. The mean age of patients was 5.1 years (S.D. 4.4 years); 63% were male. Of all admissions, 563 (43%) patients underwent surgical drainage of their infection; surgical patients had longer length of stays and total charges than patients managed medically. The average state spending per admission varied from $5126 (Utah) to $27,776 (California). There was seasonal variation in admissions with the highest percentage of admissions occurring in March (10.7%) and lowest in August (3.8%). Indicators of increased resource utilization included age (older patients), increased length of stay, nonelective admission, discharge quarter, and number of other diagnoses on record. There is a statistically significant decrease in the length of stay and total charges in patients admitted in the Midwest compared to other regions of the country. Conclusions: This study demonstrates national demographics and normative data on a commonly treated pediatric disease process, retropharyngeal space infections. The average demographic of such a patient is a 5-year-old male from an urban location admitted in a non-elective fashion via the emergency department. The mean total charges were $16,377; 90% of admissions had total charges less than $28,511. Patients who underwent surgical procedures had mean total charges of $22,013. There exists significant national variation in resource utilization for this commonly treated disease process. # 2008 Elsevier Ireland Ltd. All rights reserved.
* Corresponding author. Tel.: +1 202 476 3852; fax: +1 202 476 5038. E-mail addresses:
[email protected],
[email protected] (R.K. Shah). 0165-5876/$ — see front matter # 2008 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.ijporl.2008.09.001
1838
1. Introduction Retropharyngeal space infections are one of the more commonly treated disease processes at pediatric hospitals. Despite the prevalence of this disease process, little national data exists describing treatment variations and outcomes. Advances in imaging modalities as well as increasing potency of antibiotics, tempered by the emergence of drug-resistant organisms have changed the morbidity, mortality, and treatment algorithms associated with the management of retropharyngeal space infections and retropharyngeal abscesses (RPA) [1]. The nuances of RPA make these infections a challenge to manage. Effective and well-described strategies include solo or combination therapies which include intravenous antibiotics, needle aspiration, long-term intravenous antibiotics, and surgical drainage [2—6]. Given the variety of management options, the potential for morbidity and mortality, and the relatively high frequency of RPA in the realm of pediatrics, we hypothesize that there would be treatment variations in managing RPA; further we believe that from a national perspective, important trends and findings about resource utilization for this commonly managed disease could be identified. Examining a frequently managed disease process such as RPAs from a national perspective facilitates discussion about treatment trends and further enables individual practices and hospitals to benchmark themselves to this data. It is helpful to compare lengths of stay and mortality rates of patients admitted for RPA across states and institutions. National aggregate sampled data, such as those presented in this paper, however, are the first steps for introspective evaluation of one’s practice. Large, national database sets allow generalizations on the treatment of a specific disease while simultaneously facilitating the breakdown of the analysis into meaningful pieces. Our group has performed similar analyses to evaluate pediatric subglottic stenosis [7] and another common infectious process, pediatric mastoiditis [8]. Those analyses revealed significant national variations in the management of these disease processes.
2. Methods Institutional review board approval was obtained for this study. The source of the data is from the Kids’ Inpatient Database (KID), Healthcare Cost and Utilization Project (HCUP), Agency for Healthcare Research and Quality [9]. The KID 2003 data-
L. Lander et al. base serves as the only all-payer, pediatric database of inpatient hospital stays from the year 2003. Developed by Agency for Healthcare Research and Quality (AHRQ), the KID is one member of a family of databases in the Healthcare Cost and Utilization Project (HCUP) [9]. The KID provides researchers and physicians an objective tool in evaluating national trends in the management and resource utilization of pediatric diseases. The KID has been increasingly utilized as a research tool to establish normative data and assist in predicting parameters of high resource utilization [10—19]. The KID is an inpatient database consisting of a stratified random sample of 2,984,129 unweighted discharges from 3438 hospitals distributed across 36 states (Arizona, California, Colorado, Connecticut, Florida, Georgia, Hawaii, Illinois, Indiana, Iowa, Kansas, Kentucky, Maryland, Massachusetts, Michigan, Minnesota, Missouri, North Carolina, Nebraska, New Hampshire, New Jersey, Nevada, New York, Ohio, Oregon, Rhode Island, South Carolina, South Dakota, Tennessee, Texas, Utah, Virginia, Vermont, Washington, West Virginia, and Wisconsin). The sampled cohort includes pediatric discharges from community, non-rehabilitation hospitals in the United States in 2003. The American Hospital Association definition of community hospitals includes academic medical centers and specialty hospitals such as obstetrics/gynecology, ear nose throat, orthopedic, and pediatric hospitals. Excluded from the KID are federal hospitals, long-term hospitals, psychiatric hospitals, alcohol/ chemical dependency treatment facilities, and hospital units within institutions, such as prisons. Of the participating hospitals, the KID sampled 10% of uncomplicated, in-hospital births and 80% of complicated, in-hospital births and other pediatric cases from each hospital. Patients less than 21 years of age and discharged with a diagnosis of retropharyngeal abscess (RPA) — International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) code of 478.24 — were included in this study. Each admission included in the KID database contains up to 15 diagnoses. Inclusion into the study sample required the presence of RPA as one of the 15 available diagnoses. RPA, however, was not found in any patients as the eleventh through fifteenth diagnoses. Total hospital charges accrued during hospitalization were used as a surrogate for resource utilization. Rationale for this has been detailed extensively in previous studies [10,17,18]. Simple linear regression was fit to evaluate the effect of the potential predictors on increased total hospital
Pediatric retropharyngeal abscesses: A national perspective charges (Table 1). The p-value <0.05 indicated a significant predictor of increased total charges. Multivariate regression model was fit to determine which factors were associated with increased total charges in a multivariate model (Table 2).
1839
Patients with an RPA that underwent surgical drainage of the infection were also analyzed, inclusion criteria included the patient having ICD-9 procedure code 28.0 (incision and drainage of tonsil and peritonsillar structures, drainage (oral) (transcervical)
Table 1 Demographics of patients with RPA with p-value from univariate analysis (total charges was used as the baseline) Variable a
Mean
p-Value
Age (years), older patients have higher total charges
5.06
0.0379
Gender Male (n = 825, 62.5%) Female (n = 455, 34.4%)
$16,490 $16,170
Length of stay (days), longer stay has higher total charges Number of diagnoses Number of procedures performed during admission Mean total charges
4.27 2.63 1.21 $16,377
<0.0001 <0.0001 <0.0001 N/A
Source of admission Emergency room (n = 802, 61.2%) Another hospital (n = 117, 8.9%) Another facility (n = 15, 1.1%) Routine/birth/other (n = 364, 27.6%)
$16,700 $23,916 $29,677 $12,865
0.0047 — 0.0255 0.1269 0.0629
Admission type Elective admission (n = 140, 10.6%) Non-elective admission (n = 1157, 87.6%)
$9959 $17,188
Disposition of patient Routine (n = 1204, 91.1%) Short-term hospital (n = 52, 3.9%) Home health care services (n = 43, 3.3%)
$15,684 $11,091 $38,072
<0.0001 — 0.3101 <0.0001
Discharge quarter 1st (n = 438, 33.2%) 2nd (n = 352, 26.6%) 3rd (n = 200, 15.1%) 4th (n = 316, 23.9%)
$14,095 $16,682 $20,747 $16,434
0.1239 — 0.2678 0.0170 0.3312
Primary payer Medicaid (n = 450, 34.1%) Private insurance (n = 750, 56.8%) Self-pay (n = 52, 3.9%) Other (n = 49, 3.7%)
$17,815 $15,702 $14,290 $16,722
0.7047 — 0.2789 0.4621 0.8243
Patient location (urban/rural) Large metropolitan area (n = 794, 60.1%) Small metropolitan area (n = 364, 27.6%) Micropolitan area (n = 91, 6.9%) Non-urban (n = 53, 4.0%)
$18,251 $14,737 $10,913 $9139
0.0338 — 0.0890 0.0422 0.0491
Hospital location Rural (n = 66, 5.0%) Urban (n = 1182, 89.5%)
$7382 $17,060
NACHRI hospital type Not identified as a children’s hospital (n = 503, 38.1%) Children’s general hospital (n = 370, 28.0%) Children’s unit in a general hospital (n = 340, 25.7%)
$15,322 $16,025 $19,244
Hospital teaching status Teaching hospital (n = 268, 20.3%)
$13,820
0.8412
0.0136
0.0213
0.2313 — 0.7603 0.0971 0.1295
1840
L. Lander et al.
Table 1 (Continued )
Variable a
Mean
Non-teaching hospital (n = 980, 74.2%)
p-Value
$17,294
Median household income quartile $1—35,999 (n = 290, 22.0%) $36,000—44,999 (n = 301, 22.8%) $45,000—59,000 (n = 353, 26.7%) $60,000+ (n = 346, 26.2%)
$15,609 $13,670 $17,616 $16,232
0.3536 — 0.4039 0.3699 0.7819
Hospital region Northeast (n = 238, 18.0%) Midwest (n = 348, 26.3%) West (n = 300, 22.7%) South (n = 420, 31.8%)
$18,649 $10,718 $18,619 $18,176
0.0025 — 0.0038 0.9916 0.8577
Race White (n = 494, 37.4%) Black (n = 155, 11.7%) Hispanic (n = 147, 11.1%) Asian/Pacific Islander (n = 27, 2.0%) Other (n = 50, 3.8%)
$17,865 $16,944 $20,176 $22,716 $22,027
0.9170 — 0.7911 0.5146 0.5157 0.4576
Bold typeface with p-value <0.05 is considered to be a statistically significant predictor of increased total charges. a Note some variables do not total to 1321 as some data elements were missing from the KID.
of: parapharyngeal abscess, peritonsillar abscess, retropharyngeal abscess, tonsillar abscess).
3. Results There were 1321 admissions sampled from 36 states with a diagnosis of pediatric RPA in the KID 2003. There were no mortalities in the sampled admissions. The average total charges per admission was $16,377 (S.D. $32,637, range $859 to $728,351); 90% of admissions had charges less than $28,511. RPA admissions were found in all 36 states with 5 states (California, Texas, Florida, New York, and Ohio) accounting for over 40% of all admissions. There is significant variation in the number of admissions per state and costs per state as well as the months of admission (Figs. 1 and 2).
The mean presenting age at time of admission with an RPA was 5.1 years (S.D. 4.4). Males represented a disproportionate number of admissions compared to females, 64.6% versus 35.4%, respectively. A summary of pertinent variables and their association with total charges is presented in Table 1. Male patients had a slightly higher, but not statistically significant, mean charge than female patients. There were geographic variations in admissions and resource utilization with the South region having the highest percentage of RPA admissions (32%) and the Midwest having the lowest length of stay and total charges (Table 3). The length of stay was the highest in the South (5 days), but the highest total charges per RPA admission, however, occurred in the North. An analysis of patients that underwent surgical drainage for their infections is found in Table 4. There
Table 2 Indicators for increased resource utilization of RPA admissions in a multivariate analysis of admission characteristics of patients Variable Age Gender Race Elective type of admission Length of stay Type of admission Discharge quarter Hospital location Number of diagnoses on record Number of procedures on record Model R2 = 0.85, p < 0.0001.
Regression coefficient (95% CI) 286.1(41.6, 530.6) 1542.8 ( 3804.0, 717.3) 722.0 ( 1532.1, 88.1) 5828.6 ( 11672, 14.3) 6327.7 (6096.6, 6558.8) 3426.0 (845.6, 6006.5) 1067.5 (136.5, 1998.4) 2158.9 ( 6902.4, 2584.6) 875.7 ( 1502.0, 249.3) 531.9 ( 340.5, 1404.3)
p-Value
Partial R 2
0.0219 0.1806 0.0806 0.0506 <0.0001 0.0093 0.0247 0.3719 0.0062 0.2317
0.00707 0.00242 0.00411 0.00515 0.79589 0.00908 0.00679 0.00108 0.01006 0.00193
Pediatric retropharyngeal abscesses: A national perspective
1841
Fig. 1 Distribution of reported admissions of RPAs in KID by month. Total admissions for a respective month appear at the top of individual bar graphs.
Fig. 2 Total charges per admission by state in which the patient received care. The solid line graphically represents the overall mean total charge of $16,377 for all RPA cases. Total admissions for a respective state appear at the top of each bar graph (note: only states that had more than 10 admissions sampled are included in this graph).
were fewer surgical compared to non-surgical admissions; on average, total charges was almost twice as high for surgical admissions ( p < 0.0001) (Table 4). Further, surgical admissions involved longer lengths of stay (4.8 versus 3.9 days) ( p < 0.0001).
4. Discussion This series is the first to examine the variation and outcomes associated with pediatric retropharyngeal space infections on a national scale. Despite the prevalence of this disease process, normative, aggregate, and national data are lacking. In this series of patients, the average demographics of patients that develop RPAs include male children with an average age of 5 years. Most patients were
hospitalized for 4.3 days (S.D. 4.7) and did not have significant concomitant diagnoses (on average 2.6 diagnoses at time of admission). The majority of admissions occurred in urban, non-teaching hospitals and over 50% admissions were paid with private insurance. As to be expected, there is no correlation between race and increased total charges. Variables associated with increased total charges, or resource utilization, in the management of pediatric RPAs included: older patients, longer length of stay, number of diagnoses on admission, number of procedures performed during the hospitalization, patients transferred from another hospital, those necessitating home health care services, patients from large metropolitan areas, those admitted to urban hospitals, and those admitted regions of the United States other than the Midwest.
1842
L. Lander et al.
Table 3 Comparison of patient characteristics by admission region (n = 1321) Treatment Admissions, n (%) Length of stay (days), mean, S.D. Total charges, mean, S.D. Surgical procedure performed, n (%) Treated at teaching hospital, n (%) Treated at urban hospital, n (%) *
Northeast 238 (18%) 4.8 (3.4) 18648.8 (23000.7)
Midwest
West
349 (26%) 3.5 (2.7)
South
312 (24%) 3.9 (3.2)
10718.2 (8624.0)
18619.0 (25139.6)
422 (32%) 5.0 (7.0) 18176.1 (49693.0)
p-Value * <0.0001 <0.0001 0.0025
93 (7%)
165 (12%)
137 (10%)
168 (13%)
0.1169
202 (15%)
254 (19%)
227 (17%)
311 (24%)
<0.0001
228 (17%)
280 (21%)
294 (22%)
394 (30%)
0.2518
p-Values as determined by chi-square test for nominal variables and analysis of variance for continuous variables.
Table 4 Comparison of surgical and non-surgical RPA treatments (n = 1321) Treatment
Non-surgical
Surgical drainage
p-Value *
Admissions, n (%) Length of stay (days), mean, S.D. Total charges, mean, S.D. Treated at teaching hospital, n (%)
758 (57%) 3.9 (4.2) $12,198 ($24,645) 554 (42%)
563 (43%) 4.8 (5.4) $22,013 ($40,368) 440 (33%)
<0.0001 0.0002 <0.0001 0.0060
*
p-Values as determined by chi-square test for nominal variables and analysis of variance for continuous variables.
It was of interest that one region of the country would have significantly lower total charges than the other three (Table 3). Further analysis of patients admitted in the Midwest compared to the rest of the country revealed that these patients stayed a statistically significantly shorter time in the hospital and accrued approximately $8000 less in total charges. The rate of surgical admissions and admissions to teaching hospitals in the Midwest was approximately similar to the other regions of the country. The reason for the large difference between the Midwest and other regions deserves further investigation, however it is beyond the scope of the KID analysis. Several additional demographic characteristics identified in this national sampling of admissions were surprising, including an average age of 5.1 years, which is older than previously reported [2—4]. The disproportionate number of admissions of male compared to female patients (64.5% versus 34.5%, respectively), however, mirrors past findings [3—5]. A study by Schweinfurth found no fluctuations in RPA hospitalizations as a function of the seasons [6]. We observed markedly higher incidences of RPAs during the winter months (defined as December, January and February) and spring months (March, April and May) versus the summer (June, July and August) and fall (September, October and November) months (Fig. 1). Approximately 33.2% (n = 399) of admissions occurred in the winter followed closely by 31.2% (n = 375) in the spring. This is rather intuitive
as most children do have upper respiratory infections or other viral infections during the colder months. The majority of patients were admitted from emergency rooms; however, patients admitted from ‘‘another hospital’’ had higher total charges relative to other admissions sources. It may be that patients transferred from another hospital are sicker (hence the rational for the transfer of care) and eventually need more interventions. Perhaps better triage or proper initial identification of a hospital suitable in treating RPAs may alleviate excessive inter-hospital transfers and increased economic-burdens for this disease process. An overwhelming number of hospital admissions resulted from non-elective admissions (n = 1157, 87.6%) at a mean total charges of $17,188 versus elective admissions (n = 140, 10.6%) with mean total charges of $9959. Patients who become ill to the point where treatment and admission becomes urgent will have a more complicated hospital course than an elective admission. This was shown in our analysis with the difference in costs between elective and non-elective admissions. Primary care, emergency room, and triage physicians need to be cognizant of the standard of care for treating deep neck space infections; perhaps applications of algorithms such as suggested by Kirse and Roberson may prove to be cost-efficient in addition to optimizing patient care [1]. Comparison of patients that underwent surgical drainage of the RPA compared to non-surgical
Pediatric retropharyngeal abscesses: A national perspective intervention revealed a longer length of stay and charges for surgical patients. A statistically significant proportion of patients underwent non-surgical treatment at teaching institutions compared to those that were operated upon at teaching hospitals; it may be that these institutions are more apt to try conservative measures prior to surgical drainage. The limitations of using a national database should be highlighted. First, the data is an average of admissions and it is expected that an individual hospital and physician will have slightly different means for the variables presented. Further, not all states are included in the KID, which somewhat limits the scope of the database and an ability to make generalizations.
5. Conclusions This study demonstrated national demographics and normative data on a commonly treated pediatric disease process, retropharyngeal space infections. The average demographic of such a patient is a 5year-old male from an urban location admitted in a non-elective fashion via the emergency department. In a multivariate analysis, significant indicators of total charges include older patients, longer length of stay, discharge quarter, and number of diagnoses on record. There is significant variation between states and regions of the United States, with the Midwest having a statistically lower total charges and length of stay than other areas of the country. Data from this study will enable individual hospitals and physicians to compare themselves in an attempt to provide high quality and cost-effective care on a macrolevel. Furthermore, periodic studies of RPAs on a national level will allow identification of emerging changes in the demographics of this infectious process.
References [1] D.J. Kirse, D.W. Roberson, Surgical management of retropharyngeal space infections in children, Laryngoscope 111 (August (8)) (2001) 1413—1422. [2] B. Dodds, A.J. Maniglia, Peritonsillar and neck abscesses in the pediatric age group, Laryngoscope 116 (June (6)) (1988) 887—889.
1843
[3] J.M. Coticchia, G.S. Getnick, R.D. Yun, J.E. Arnold, Age-, site-, and time-specific differences in pediatric deep neck abscesses, Arch. Otolaryngol. Head Neck Surg. 130 (February (2)) (2004) 201—207. [4] J.W. Thompson, P. Reddix, S.R. Cohen, Retropharyngeal abscess in children: a retrospective and historical analysis, Laryngoscope (June (6 Pt. 1)) (1998) 589—592. [5] S.S. Lee, R.H. Schwartz, R.S. Bahadori, Retropharyngeal abscess: epiglottitis of the new millennium, J. Pediatr. 138 (March (3)) (2001) 435—437. [6] J.M. Schweinfurth, Demographics of pediatric head and neck infections in a tertiary care hospital, Laryngoscope 116 (June (6)) (2006) 887—889. [7] R.K. Shah, L. Lander, S.S. Choi, G.H. Zalzal, Resource utilization in the management of subglottic stenosis, Otolaryngol. Head Neck Surg. 138 (February (2)) (2008) 233—241. [8] J.L. Acevedo, L. Lander, U.K. Shah, R.K. Shah, Existence of Important Variations in the United States in the Treatment of Pediatric Mastoiditis, Arch. Otolaryngol. Head Neck Surg., in press. [9] HCUP Kids’ Inpatient Database (KID), Healthcare Cost and Utilization Project (HCUP), Agency for Healthcare Research and Quality, Rockville, MD, 2003, www.hcup-us.ahrq.gov/ kidoverview.jsp, accessed March 15, 2007. [10] J.A. Connor, K. Gauvreau, K.J. Jenkins, Factors associated with increased resource utilization for congenital heart disease, Pediatrics 116 (September (3)) (2005) 689—695. [11] J.G. Berry, T.A. Lieu, P.W. Forbes, D.A. Goldmann, Hospital volumes for common pediatric specialty operations, Arch. Pediatr. Adolesc. Med. 161 (January (1)) (2007) 38—43. [12] E.R. Kokoska, T.M. Bird, J.M. Robbins, S.D. Smith, J.M. Corsi, B.T. Campbell, Racial disparities in the management of appendicitis, J. Surg. Res. 137 (January (1)) (2007) 83—88. [13] G.H. Cosper, M.S. Hamann, A. Stiles, D.K. Nakayama, Hospital characteristics affect outcomes for common pediatric surgical conditions, Am. Surg. 72 (August (8)) (2006) 739—745. [14] M.A. Malek, A.T. Curns, R.C. Holman, T.K. Fischer, J.S. Bresee, R.I. Glass, et al., Diarrhea- and rotavirus-associated hospitalizations among children less than 5 years of age: United States, 1997 and 2000, Pediatrics 117 (June (6)) (2006) 1887—1892. [15] J.R. Meurer, E.M. Kuhn, G. Varghese, J.S. Yauck, P.M. Layde, Charges for childhood asthma by hospital characteristics, Pediatrics 102 (December (6)) (1998) E70. [16] D.S. Smink, S.J. Fishman, K. Kleinman, J.A. Finkelstein, Effects of race, insurance status and hospital volume on perforated appendicitis in children, Pediatrics 115 (April (4)) (2005) 920—925. [17] A.J. Schneier, B.J. Shields, B.A. Hostetler, H. Xiang, G.A. Smith, Incidence of pediatric traumatic brain injury and associated hospital resource utilization in the United States, Pediatrics 118 (August (2)) (2006) 483—492. [18] D. Merenstein, B. Egleston, M. Diener-West, Lengths of stay and costs associated with children’s hospitals, Pediatrics 115 (April (4)) (2005) 839—844. [19] S.L. Guthery, C. Hutchings, M. Dean, C. Hoff, National estimates of hospital utilization by children with gastrointestinal disorders: analysis of the 1997 Kids’ Inpatient Database, J. Pediatr. 144 (May (5)) (2004) 589—594.
Available online at www.sciencedirect.com