An infographic is available for this article at: https://links.lww.com/MPG/C767.
What Is Known
- Patients with Hirschsprung disease (HD) remain at risk for long term gastrointestinal morbidity despite advances in the diagnosis and management.
- Pull-through procedure for HD is associated with high readmission rates with an associated economic burden.
What Is New
- Cost of HD-related hospitalization shows increasing trend despite stable rate of admissions and decreasing length of hospital stay.
- Median zip code household income in lowest quartile is associated with higher odds of hospital mortality and increased length of stay.
Hirschsprung disease (HD) is a congenital condition characterized by the absence of ganglion cells in the myenteric and submucosal plexuses of the distal intestine, which results in lack of peristalsis and functional intestinal obstruction. The aganglionosis involves the rectum or rectosigmoid in most cases, but it can extend for varying lengths, and in 5–10% of cases can involve the entire colon or even a significant amount of the small intestine. The incidence of this disease is approximately 1 in 5000 live births and is approximately three times more common in males (1,2). HD is a multifactorial disease caused by both genetic and environmental factors, and HD may also be part of a syndrome, most commonly Trisomy 21 (3).
The principal purpose of surgical management of HD is to remove aganglionic bowel and reconstruct the intestinal tract by connecting the normally innervated bowel just above the anus so that normal sphincter function is preserved. Over the last decades, surgical management has changed from original three-stage approach to recent introduction of minimally invasive one-stage procedures, which appears to be safe, and facilitates early feeding and discharge (4–6). Although surgical treatment definitively removes the histologically defined pathological intestinal segment, many patients with HD continue to experience bowel dysfunction, obstructive symptoms and recurrent enterocolitis requiring frequent hospitalizations (7–15).
Huang et al (16) had shown that HD discharges, associated demographics, and numbers of pull-through procedures remained stable from 1997 to 2006. There is a paucity in the literature regarding cost components associated with care of hospitalized patients with the primary diagnosis of HD. An information update regarding inpatient care utilization for HD is needed to better understand current HD disease burden and inpatient care practice. The primary objective of this study was to analyze cost of HD-related admissions in the United States between 2009 and 2014 using the National Inpatient Sample (NIS) database. Secondary objectives were to update national and regional trends in the epidemiology of HD related hospital utilization, and to examine changing demographics, hospital mortality and length of hospital stay (LOS) in these HD patients.
METHODS
Data Source
We obtained the NIS databases for the years 2009 to 2014 for this study with diagnoses and procedures reported using the International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) coding system. The NIS is the largest all-payer inpatient care database in the United States and is maintained as part of the Healthcare Cost and Utilization Project (HCUP) by the Agency for Healthcare Research and Quality (AHRQ) (17). NIS samples 20% of the discharges from US nonfederal community hospitals and contains >7 million hospitalizations annually. With the established weights in NIS, this sample could be weighted to represent the standardized US population and more than 35 million hospitalizations each year. The available data include demographic variables (such as age, sex, race/ethnicity, and median income for ZIP code), discharge disposition, primary and secondary diagnoses (up to 25 for years 2009–2013 and up to 30 for 2014), primary and secondary procedures (up to 15), primary insurance payers, total hospital charges, and LOS.
STUDY POPULATION AND DESIGN
We included all patients who were 18 years or younger with an ICD-9-CM code of 751.3 for HD among the first 10 diagnosis. Patients whose discharge disposition was transfer to another acute hospital was excluded from analysis on the assumption that these transferred patients were represented in the national sample because the receiving hospitals of these transfers were included in the sampling scheme. This exclusion avoided over-counting HD hospitalizations. We studied patient-level and hospital-level characteristics. Patient level characteristics included gender, race, quartiles of zip code median household income, primary payer (Medicare/Medicaid, private insurance, other); hospital-level characteristics included hospital location and teaching status (rural, urban non-teaching and urban teaching), and hospital bed size. We obtained national estimates of the total number of discharges with an HD diagnosis in each year in relation to these characteristics. We also reported estimates of common discharge diagnoses, coexisting congenital anomalies, hospital mortality and LOS per year. We identified patients with a pull-through procedure using the following ICD-9-CM codes: 4592, 4593, 4594, 4595, 4840, 4841, 4842, 4843, 4849, and 4865 (Table 1, Supplemental Digital Content, https://links.lww.com/MPG/C768). We used all 15 procedure codes to identify pull-through procedures and tracked the number of these pull-through procedures a patient received. We also identified gastrointestinal and other procedures including endoscopic biopsy of rectum (ICD-9-CM 4824), endoscopic biopsy of large intestine (4525), open biopsy of rectum (4825), open biopsy of large intestine (4526), colostomy (4610) and central venous catheter placement with guidance (3897). To identify common discharge diagnoses we used the first five diagnosis codes. To identify congenital anomalies, we queried the following ICD-9-CM codes: anencephalus and similar anomalies (740), spina bifida (741), other congenital anomalies of nervous system (742), bulbus cordis anomalies and anomalies of cardiac septal closure (745), other congenital anomalies of heart (746), cleft palate and cleft lip (749), congenital anomalies of genital organs (752), congenital anomalies of urinary system (753), or chromosomal anomalies including trisomy 21 (758). A binary indicator variable of any congenital anomalies being diagnosed during the admission was used for the analysis. The NIS contains data on total charges for each hospital discharge. Charges are the initial list prices a hospital sets for individual items and services it provides while costs represent the actual expenses incurred during a hospitalization. To analyze how hospital charges translate into actual inpatient cost, we utilized HCUP cost-to-charge ratio (CCR) files. Costs were calculated by multiplying the hospital charges to the CCR. To adjust inflation, we downloaded the medical care expenditure (MCE) indices from the U.S. Department of Commerce Bureau of Economic Analysis website (18,19). We divided the sub-index for gastrointestinal (GI)-track congenital anomalies into the inpatient cost to derive inflation-adjusted cost. We calculated incidence of HD using the number of pull-through procedures as a surrogate marker for the number of newly diagnosed HD patients in the United States each year. These weighted estimates for each year were then divided by the number of live births for that particular year and for each sub-group, obtained from the Centers for Disease Control’s Online database (20).
Statistical Analysis
All analyses were performed using R (R Core Team, 2019). National weighted estimates were calculated by applying the NIS survey design (weights, stratum and hospital cluster) to the sample dataset using the Survey package in R (21). A modified Cochrane–Armitage (C–A) trend test (22) was used to account for hospital-level clustering in trend tests for dichotomous outcomes. The test extends the C–A trend test by using an additional multiplier of the cluster size and intra-class correlation to inflate the binomial variance. We wrote our own R codes of the trend test which are available upon request. P < 0.05 was considered statistically significant. We used generalized linear mixed-effect (GLME) models in R package lmer to analyze the relationship between an outcome measure and a set of determinants that include patient demographics, hospital characteristics, and clinical procedures. Three main outcomes were hospital mortality, LOS and inflation-adjusted inpatient cost of hospitalization. The GLME models permit the use of random effects aligned with clusters (hospitals) and facilitate the analyses of trend and disparities in a unified framework while adjusting for heterogeneity in patient clinical presentation.
RESULTS
Hospitalizations
From 2009 to 2014, there were 4643 discharges with HD among patients 18 years of age or younger. Of these, 322 ended in a transfer to another acute care facility. National estimates showed HD-discharges relative to annual total pediatric inpatient volume ranged from 2796 (95% confidence internal [CI] 1946–3646) in 2009 to 3925 (95% CI 3242–4608) in 2014 (Table 1). Accordingly, HD-discharge incidence per 100k pediatric inpatient volume ranged from 46 in 2009 to 69 in 2014, demonstrating no significant change (P = 0.27). The overall incidence of HD was 1.78/10,000 live births, with the highest value at 2.42/10,000 live births in 2012 (data not shown).
TABLE 1. -
National sampled and estimated discharges (95% CI) related to HD and outcomes
| Pediatric inpatient samples |
2009 |
2010 |
2011 |
2012 |
2013 |
2014 |
P
|
| Annual volume |
1,269,515 |
1,268,134 |
1,202,498 |
1,168,168 |
1,138,463 |
1,131,655 |
|
| HD (incidence × 100k–1) |
589 (46) |
802 (63) |
653 (54) |
823 (70) |
715 (63) |
803 (71) |
|
| HD in first 10 diagnosis codes (incidence ×100k–1) |
578 (46) |
794 (63) |
645 (54) |
814 (70) |
705 (62) |
785 (69) |
|
| National estimates |
2009 |
2010 |
2011 |
2012 |
2013 |
2014 |
P
|
| HD discharges |
2851 (1984–3718) |
3944 (2822–5065) |
2985 (1939–4031) |
4115 (3418–4812) |
3575 (2992–41,581) |
4015 (3316–4714) |
0.25 |
| HD in first 10 diagnosis codes |
2796 (1946–3646) |
3904 (2785–5023) |
2946 (1911–3981) |
4070 (3380–4760) |
3525 (2949–4101) |
3925 (3242–4608) |
0.27 |
| Pull through procedures (%) |
22.6 (18.6–26.5) |
22.1 (18.7–25.5) |
22.9 (19.1–26.7) |
22.9 (20.0–25.7) |
23.5 (20.3–26.8) |
21.9 (19.0–24.8) |
0.88 |
| LOS (days): median, (IQR) |
5 (3–13) |
5 (3–10) |
5 (3–11) |
5 (2–10) |
5 (2–9) |
5 (3–9) |
|
| Deaths in hospital (%) |
0.9% (0.2–1.5) |
0.7% (0.0–1.0) |
0.4% (0.0–0.9) |
0.9% (0.1–1.6) |
0.6% (0.0–1.2) |
0.4% (0.0–0.8) |
|
| Cost ($): median (IQR) |
9923 (4267–24,666) |
9398 (4081–20,865) |
10,792 (4966–26,251) |
10,044 (4567–23,917) |
9783 (5027–25,026) |
11,800 (4903–26,288) |
|
CI = confidence interval; HD = Hirschsprung disease; IQR = interquartile range; LOS = length of stay.
Cost of Hospitalization
Cost showed a significant increasing trend of $1137 (SE $326) per year (P = 0.0005) despite our LOS model confirmed a decreasing trend in LOS for HD-related admissions (Table 2). Each additional pull-through procedure was associated with $6538 ($937) extra cost (P < 0.001); each additional gastrointestinal and other procedures was associated with $5584 ($214) extra cost (P < 0.001), and for neonates such procedure was linked to further $2368 ($253) extra cost (P < 0.001). Whereas increasing severity of function loss was associated with increasing cost (Table 2), the higher cost for African American and Native American patients (P = 0.080) and higher cost of West region compared with Northeast region (P = 0.053) were not statistically significant.
TABLE 2. -
Generalized mixed-effect regression models of mortality, length of stay, and inflation-adjusted cost in association with clinical and demographic factors among HD discharges
| Covariate |
Mortality, OR (95% CI) |
P-value |
LOS, days (SE) |
P value |
Cost (SE) |
P value |
| Year (YOY-trend) |
0.94 (0.84–1.07) |
0.352 |
–0.23 (0.11) |
0.036 |
1137 (326) |
5.37E−4 |
| Age |
|
|
|
|
|
|
| Neonates (<29 days) |
Ref |
|
Ref |
|
|
|
| 29 days to <1 y |
0.55 (0.35–0.85) |
0.007 |
–4.10 (0.53) |
8.48E–15 |
|
|
| 1–18 y |
0.12 (0.06–0.23) |
<1.0E–08 |
–4.69 (0.49) |
<2.0E–16 |
|
|
| No of pull-through procedures |
0.47 (0.29–0.75) |
0.002 |
1.13 (0.36) |
0.002 |
6538 (937) |
3.45E–12 |
| No of other procedures |
1.31 (1.27–1.36) |
<2.0E–16 |
2.29 (0.07) |
<2.0E–16 |
5584 (214) |
<2.0E–16 |
| No. of other procedures for neonate* |
|
|
|
|
2368 (253) |
<2.0E–16 |
| Congenital anomalies |
3.07 (2.01–4.67) |
<1.0E–07 |
|
|
|
|
| Sex |
|
|
|
|
|
|
| Male |
Ref |
|
|
|
|
|
| Female |
0.41 (0.24–0.69) |
0.001 |
|
|
|
|
| Zip code median household income |
|
|
|
|
|
|
| 1st quartile |
Ref |
|
Ref |
|
|
|
| 2nd quartile |
0.25 (0.14–0.46) |
5.12E–06 |
–0.77 (0.44) |
0.081 |
|
|
| 3rd quartile |
0.40 (0.24–0.68) |
0.004 |
–1.07 (0.45) |
0.016 |
|
|
| 4th quartile |
0.30 (0.17–0.53) |
0.001 |
–1.51 (0.48) |
0.002 |
|
|
| Severity |
|
|
|
|
|
|
| Minor loss of function |
|
|
Ref |
|
Ref |
|
| Moderate loss |
|
|
0.82 (0.42) |
0.053 |
964 (1146) |
0.400 |
| Major loss |
|
|
3.85 (0.50) |
1.28E–14 |
5885 (1349) |
1.33E–5 |
| Extreme loss |
|
|
12.47 (0.78) |
<2.0E–16 |
33007 (2173) |
<2.0E–16 |
| Hospital birth |
|
|
|
|
|
|
| No |
|
|
Ref |
|
Ref |
|
| Yes |
|
|
3.76 (0.82) |
4.74E–6 |
–5563 (2105) |
0.008 |
| Race |
|
|
|
|
|
|
| Non-AA or Non-NA |
|
|
Ref |
|
Ref |
|
| AA or NA |
|
|
1.12 (0.44) |
0.012 |
3382 (1934) |
0.080 |
| region |
|
|
|
|
|
|
| Northeast |
|
|
|
|
Ref |
|
| Midwest |
|
|
|
|
527 (2683) |
0.846 |
| South |
|
|
|
|
2424 (2565) |
0.357 |
| West |
|
|
|
|
5605 (2738) |
0.053 |
AA = African American; CI = confidence interval; HD = Hirschsprung disease; LOS = length of stay; NA = Native American; OR = odds ratio; SE = standard error; YOY = year over year.
Demographics and Disparities
Table 3 demonstrates trends in demographic data. Neonates, patients of 29 days to <1 year in age, and patients 1–18 years accounted for 24%, 23%, and 49% of all HD-discharges in this national sample. There was no temporal trend in HD-discharges for neonates (P = 0.50) or 29 days to <1 year age group (P = 0.85) as a fraction of total HD-discharges. Medicaid patients accounted for 51% of the HD-related discharges, and this percentage increased from 49% in 2009 to 54% in 2014 (P = 0.02). A sensitivity analysis using 2011–2014 data suggested a significant increasing trend of Medicaid patients that coincided with the period of full implementation of Affordable Care Act. About 31% of the HD-discharges were from communities whose zip code median household income was in the lowest quartile compared with 20% from communities whose zip code median household income was in the highest quartile. In 2014, the lowest quartile of median household incomes was $39,999 or lower. African American, Hispanics and Native Americans accounted for about 17%, 13%, and 0.4% of the total HD-discharges. Noteworthy is that the percentage of African American HD-discharges increased from 12% in 2009 to 20% in 2014 (P = 0.002). Because NIS does not report race and ethnicity individually, separate analysis of race and ethnicity was not feasible.
TABLE 3. -
National estimates (95% CI) of HD patient demographics and hospital characteristics
| Demographics |
2009 |
2010 |
2011 |
2012 |
2013 |
2014 |
P
|
| Age |
| 0–28 days |
581 (378–783) |
710 (483–937) |
643 (411–875) |
955 (765–1145) |
675 (526–824) |
795 (624–966) |
0.500 |
| 0–<1 y* |
1380 (927–1832) |
1878 (1345–2411) |
1471 (972–1969) |
2050 (1660–2440) |
1680 (1374–1986) |
1820 (1461–2179) |
0.850 |
| 1–18 y |
1416 (959–1873) |
2026 (1380–2672) |
1475 (874–2077) |
2020 (1642–2398) |
1845 (1491–2199) |
2105 (1702–2508) |
|
| Gender |
| Female |
679 (428–929) |
1088 (724–1452) |
785 (512–1058) |
1060 (832–1288) |
955 (762–1148) |
1105 (868–1342) |
0.689 |
| Male |
2108 (1479–2736) |
2812 (1997–3627) |
2161 (1375–2946) |
3010 (2492–3528) |
2570 (2126–3014) |
2820 (2322–3318) |
|
| Payer |
| Medicaid |
1374 (877–1872) |
2041 (1422–2659) |
1514 (911–2118) |
1940 (1581–2299) |
1880 (1556–2204) |
2100 (1719–2481) |
0.015 |
| Private |
1181 (764–1598) |
1642 (1081–2203) |
1219 (756–1681) |
1820 (1434–2206) |
1430 (1135–1725) |
1545 (1210–1880) |
0.441 |
| Others |
241 (152–331) |
221 (96–345) |
213 (112–314) |
310 (172–457) |
215 (101–329) |
280 (117–447) |
0.0004 |
| Quartile of zip code median household income |
| 1st Quartile |
662 (408–915) |
1216 (823–1610) |
941 (480–1401) |
1280 (979–1581) |
1105 (878–1332) |
1135 (871–1399) |
0.483 |
| 2nd Quartile |
688 (449–928) |
964 (607–1321) |
746 (457–1035) |
980 (777–1183) |
800 (627–973) |
1015 (798–1232) |
0.650 |
| 3rd Quartile |
841 (492–1190) |
976 (575–1378) |
723 (466–980) |
925 (701–1149) |
850 (652–1048) |
865 (660–1070) |
0.060 |
| 4th Quartile |
535 (309–762) |
654 (377–931) |
469 (259–679) |
790 (603–977) |
725 (556–894) |
830 (621–1039) |
0.010 |
| Race |
| White |
1442 (924–1959) |
1881 (1174–2589) |
1235 (698–1772) |
2150 (1681–2619) |
1645 (1285–2005) |
1945 (1519–2371) |
0.311 |
| Black |
343 (203–483) |
616 (366–867) |
407 (225–588) |
670 (506–834) |
690 (514–866) |
770 (585–955) |
0.002 |
| Hispanic |
416 (184–649) |
586 (323–848) |
375 (208–542) |
460 (333–587) |
485 (358–612) |
515 (367–663) |
0.055 |
| Other |
224 (93–356) |
497 (251–743) |
236 (59–414) |
410 (235–590) |
330 (179–485) |
335 (180–494) |
0.995 |
| Region |
| Northeast |
610 (147–1073) |
775 (231–1319) |
407 (8–506) |
660 (369–951) |
585 (373–797) |
670 (393–947) |
0.163 |
| Midwest |
403 (131–675) |
743 (215–1271) |
872 (175–1568) |
1040 (673–1407) |
875 (543–1207) |
980 (600–1360) |
0.118 |
| South |
777 (431–1124) |
1400 (832–1968) |
1175 (572–1777) |
1655 (1205–2105) |
1350 (995–1705) |
1455 (1044–1866) |
0.133 |
| West |
1006 (446–1566) |
985 (390–1581) |
493 (237–748) |
715 (484–946) |
715 (489–941) |
820 (545–1095) |
0.102 |
| Teaching status and location |
| Rural |
56 (18–95) |
98 (0–197) |
49 (15–82) |
45 (16–74) |
40 (12–68) |
40 (0–92) |
0.030 |
| Urban. non–teaching |
287 (146–428) |
244 (91–398) |
137 (69–204) |
290 (195–385) |
245 (145–345) |
125 (67–183) |
0.269 |
| Urban, teaching |
2373 (1535–3210) |
3472 (2366–4577) |
2673 (1640–3707) |
3735 (3052–4418) |
3240 (2673–3807) |
3760 (3082–4438) |
0.020 |
| Hospital Bed size |
| Small |
257 (0–528) |
157 (0–356) |
221 (0–456) |
515 (277–753) |
360 (189–531) |
405 (217–593) |
0.239 |
| Medium |
580 (54–1105) |
952 (394–1510) |
626 (113–1139) |
1060 (665–1455) |
955 (616–1294) |
1095 (674–1516) |
0.227 |
| Large |
1880 (1170–2491) |
2705 (1754–3657) |
2011 (1143–2880) |
2495 (1982–3008) |
2210 (1776–2644) |
2425 (1922–2928) |
0.206 |
| Hospital ownership |
| Government (nonfederal) |
302 (57–547) |
482 (40–923) |
307 (19–595) |
470 (195–745) |
360 (165–555) |
440 (220–660) |
0.994 |
| Private,non–profit |
2115 (1311–2921) |
3136 (2078–4193) |
2423 (1416–3432) |
3400 (2776–4024) |
2920 (2394–3446) |
3280 (2644–3916) |
0.272 |
| Private, invest-own |
299 (24–573) |
197 (0–443) |
127 (0–271) |
200 (98–302) |
245 (113–377) |
205 (93–317) |
0.340 |
CI = confidence interval; HD = Hirschsprung disease.
Pull-through Procedures
About 23% of the HD-discharges had at least one pull-through procedure, according to both sampled data and national estimate. The estimated percentage of patients receiving pull-through procedures each year remained stable from 2009 to 2014 (P = 0.88, Table 1). Among HD patients receiving any pull-through procedures, 91% received one and 9% received two (only one patient received three). The percentage of patients receiving a pull-through procedure in neonatal age group increased from 33.0% in 2009 to 36.5% in 2014 (P = 0.003). Among those with any pull-through procedures, laparoscopic pull-through resection of rectum increased from 16.5% in 2009 to 31.4% in 2014 (P = 0.0001), showing a significant increasing trend.
Congenital Anomalies and Discharge Diagnoses
Chromosomal anomalies including trisomy 21 were the most common diagnoses of HD discharges, ranging between 10% and 15% of the total estimated discharges over the years, followed by bulbus cordis anomalies and anomalies of cardiac septal closure (5–10%), other cardiac anomalies (1.5–3.5%), and central nervous system anomalies (1.4–3.3%). Other associated anomalies included cleft palate and lip, genital organs anomalies, and urinary system anomalies making up <1% of discharges.
Gastroenteritis, functional disorders including constipation, enterocolitis and intestinal obstruction were the most common discharge diagnoses (Table 4). The estimated percentage of HD patients with these common discharge diagnoses showed a statistically significant trend from 2009 to 2014 in 29 days to <1 year age group (P = 0.002) as well as in the 1–18 year age group (P = 0.049). There was a statistically significant increasing trend of functional digestive disorders among the HD discharges, from 9.5% (95% CI 7.1–12.0) in 2009 to 14.6% (95% CI 12.0–17.3) in 2014 (P = 0.001) (Table 4).
TABLE 4. -
National estimates (95% CI) of percentage of patients with leading diagnoses among all HD discharges by year
| Diagnosis (ICD-9-CM code) |
2009 |
2010 |
2011 |
2012 |
2013 |
2014 |
Trend test
P value |
| Disorders of electrolyte and acid-base balance (276) |
19.6% (16.4–22.9) |
16.1% (13.0–19.7) |
15.0% (12.3–17.7) |
17.3% (14.0–20.7) |
17.4% (14.5–20.4) |
16.4% (13.4–19.4) |
0.513 |
| Intestinal obstruction without mention of hernia (560) |
11.6% (8.2–15.0) |
12.4% (9.5–15.2) |
11.4% (8.8–14.0) |
9.7% (7.5–11.9) |
9.8% (7.6–12.0) |
11.0% (8.7–13.2) |
0.101 |
| Functional digestive disorders not elsewhere classified (564) |
9.5% (7.1–12.0) |
11.3% (8.3–14.3) |
12.2% (9.1–15.3) |
13.0% (10.3–15.8) |
12.9% (10.3–15.5) |
14.6% (12.0–17.3) |
0.001 |
| Other and ill-defined conditions originating in the perinatal period (779) |
8.6% (6.3–10.9) |
8.7% (6.7–10.6) |
10.6% (8.0–13.2) |
10.2% (7.9–12.5) |
7.1% (5.0–9.1) |
7.4% (5.5–9.2) |
0.158 |
| Other and unspecified noninfectious gastroenteritis and colitis (558) |
10.0% (7.3–12.7) |
9.5% (6.6–12.4) |
8.1% (5.9–10.2) |
10.3% (8.0–12.6) |
10.4% (7.9–12.8) |
13.0% (10.3–15.6) |
0.328 |
| Symptoms involving digestive system (787) |
5.5% (3.7–7.2) |
5.9% (4.4–7.5) |
7.4% (5.7–9.2) |
9.1% (6.9–11.3) |
7.8% (5.8–9.8) |
9.8% (7.5–12.1) |
0.156 |
CI = confidence interval; HD = Hirschsprung disease; ICD-9-CM = International Classification of Diseases, Ninth Revision, Clinical Modification.
Length of Stay
The median LOS was 5 days (Table 1). LOS demonstrated a decreasing trend of 0.23 days per year (P = 0.04) in the mixed-effects regression analysis (Table 2). LOS for neonates was on average 4.1 and 4.7 days longer than for 29 days to <1 year of age and for ≥1 year, respectively (P = <0.0001). Each pull-through procedure was associated with an extra 1.1 day LOS (P = 0.002), and each gastrointestinal and other procedures was associated with an additional 2.3 days LOS (P = <0.0001). Higher zip code median household income was associated with up to 1.5 day reduced LOS (P = 0.004). LOS increased with functional severity, by up to 13 days for those with extreme loss of function compared with minor loss. Finally, we note that Native American and African American combined had an extra 1.1 days LOS than the rest of the patient sample (P = 0.01).
Hospital Mortality
Hospital mortality remained stable between 0.4% and 0.9% (P = 0.598). The mortality model confirmed no temporal trend in the risk of death during a HD-related inpatient encounter (Table 2). However, neonates had the highest odds of death, 1.7 (95% CI 1.1–2.6) and 7.7 (95% CI 3.8–16.7) times that of 29 days to <1 year and >1 year groups, respectively. Each gastrointestinal and other procedures were associated with 31% (95% CI 26–36) increase in the odds of mortality. Presence of any other congenital anomalies increased the odds of mortality (OR = 3.1, 95% CI 2.0–4.7). Finally, patients who lived in a community where its zip code median income was in the lowest quartile among all communities had an odds of death 2.5–3.8 times that of higher median income communities.
DISCUSSION
In the last few decades, the early diagnosis and appropriate pre- and postoperative management of HD, including medical and surgical treatments, have undergone various changes leading to improved postoperative outcomes (4–6); however, there remains a paucity of studies regarding hospital utilization and national trends. Our large nationwide study is an updated analysis of the inpatient care utilization and epidemiology of HD in the United States from 2009 to 2014. We found a statistically significant increase in inflation-adjusted hospital cost of HD-related hospitalization over the study period despite a decreasing trend in LOS. The Bureau of Economic Analysis of the Department of Commerce reported the medical care expenditure inflation rates of 19.4%, 38.1%, 11.4%, 13.4%, 9.0% between 2010 and 2014 compared with the 2009 value for GI-track congenital anomalies (19). The analysis of the cost without adjusting for inflation yielded slightly higher costs. There may be variability in care amongst hospitals in the United States. Protocol driven care of patients with HD is not common in pediatric surgery training centers, leading to great variability in care between institutions, as well as among faculty within single programs. Development of clinical pathways and protocols may help to standardize care and reduce hospital cost.
HD has a favorable prognosis in the developed world with the reported survival of >95% (23–25) and the results of the present study are consistent with the survival outcomes in these prior studies. Among patients with unfavorable outcomes, neonates, presence of congenital anomalies, and patients from communities in the lowest quartile of zip code medium incomes have a significantly higher risk of mortality. Studies have shown that Hirschsprung’s associated enterocolitis (HAEC) is the most common cause of death in patients with HD with the mortality ranges between 5% and 50%, depending on patient’s age and type of infection (23,26). Application of preventative strategies combined with early aggressive treatment of HAEC may improve outcomes and health care utilization for these vulnerable populations.
The results from this study are consistent with several previous single-center studies (26–29). There are only a few previous large database studies of HD hospital utilization to which we can compare our findings (16,30–32). Huang et al (16) demonstrated that the percentage of patients receiving pull-through procedures had remained stable from 1997 to 2006, and the frequency of constipation and enterocolitis had trended upward based on data from KIDS Inpatient Database. Pruitt and coworkers conducted a retrospective cohort study using the Pediatric Health Information System Database (PHIS) and identified a 7% rate of recurrent post-operative enterocolitis (30). Quiroz et al (31) showed pull-through procedure for HD is associated with high readmissions and economic burden using the Nationwide Readmissions Database. The most common diagnoses for readmission were gastrointestinal disorders (46%) and infections (39%) in their study population.
This study depends entirely on the accuracy of administrative data coding and, therefore, has some limitations. First, there is a potential for bias due to inaccuracies and incompleteness of reporting and coding. Second, the NIS database does not identify longitudinal follow-up visits, which could have provided more information regarding longer-term postoperative outcomes. Third, we acknowledge that our analysis was limited to data from 2009 to 2014. On October 1, 2015, the USA transitioned from the ICD-9-CM to the Tenth Revision (ICD-10-CM) for clinical coding in medical settings. The ICD-9-CM contains just under 17,000 codes whereas ICD-10-CM contains approximately 155,000 codes, and the codes contain much more details compared to ICD-9-CM (33). The NIS database for 2015 includes a mixture of ICD-9-CM and ICD-10-CM diagnosis and procedure codes, and databases from 2016 to the most recent version in 2018 include ICD-10-CM diagnosis and procedure codes. We included databases from 2009 to 2014 to avoid billing and administrative errors possibly occurring during transition as well as uncertainties in combining ICD-9-CM and ICD-10-CM for HD. The NIS database is released annually, approximately 18 to 22 months following the end of a calendar year. Newer versions of the NIS database can be used for future research to study trends and changes in HD healthcare utilization over time.
CONCLUSIONS
Our large nationwide study is an updated analysis of the inpatient care utilization and epidemiology of HD in the United States from 2009 to 2014. We demonstrated increasing trends in the cost of hospitalization, decreasing trends in the LOS and no statistically significant changes in the hospital mortality. HD-related hospitalization rates were stable and the ratio of HD patients receiving pull-through procedures increased among neonates. Future studies and development of protocols to standardize patient care could improve outcomes and decrease healthcare spending.
REFERENCES
1. Granstrom AL, Svenningsson E, Hagel E, et al. Maternal risk factors and perinatal characteristics for Hirschsprung disease. Pediatrics. 2016;138:e2015460.
2. Bradnock TJ, Knight M, Kenny S, et al. Hirschsprung’s disease in the UK and Ireland: incidence and anomalies. Arch Dis Child. 2017;102:722–7.
3. Ikeda K, Goto S. Additional anomalies in Hirschsprung’s disease: an analysis based on the nationwide survey in Japan. Z Kinderchir. 1986;41:279–81.
4. Langer JC, Durrant AC, Torre L, et al. One-stage transanal Soave pullthrough for Hirschsprung disease. Ann Surg. 2003;238:569–76.
5. Torre-Mondragon LD, Ortega-Salgado JA. Transanal endorectal pull-through for Hirschsprung’s disease. J Pediatr Surg. 1998;33:1283–6.
6. Georgeson KE, Fuenfer MM, Hardin WD. Primary laparoscopic pull-through for Hirschsprung’s disease in infants and children. J Pediatr Surg. 1995;30:1017–22.
7. Stensrud KJ, Emblem R, Bjornland K. Functional outcome after operation for Hirschsprung disease—transanal vs transabdominal approach. J Pediatr Surg. 2010;45:1640–4.
8. Langer JC. Persistent obstructive symptoms after surgery for Hirschsprung disease: development of a diagnostic and therapeutic algorithm. J Pediatr Surg. 2004;39:1458–62.
9. Granstrom AL, Husberg B, Nordenskjold A, et al. Laparoscopic assisted pull-through for Hirschsprung’s disease, a prospective repeated evaluation of functional outcome. J Pediatr Surg. 2013;48:2536–9.
10. Jarvi K, Laitakari EM, Koivusalo A, et al. Bowel function and gastrointestinal quality of life among adults operated for Hirschsprung disease during childhood: a population based study. Ann Surg. 2010;252:977–81.
11. Ieiri S, Nakatsuji T, Akiyoshi J, et al. Long-term outcomes and the quality of life of Hirschsprung disease in adolescents who have reached 18 years or older—a 47-year single-institute experience. J Pediatr Surg. 2010;45:2398–402.
12. Heikkinen M, Rintala RJ, Louhimo I. Bowel function and quality of life in adult patients with operated Hirschsprung’s disease. Pediatr Surg Int. 1995;10:342–4.
13. Conway SJ, Craigie RJ, Cooper LH, et al. Early adult outcome of the Duhamel procedure for left-sided Hirschsprung disease—a prospective serial assessment study. J Pediatr Surg. 2007;42:1429–32.
14. Granstrom AL, Danielson J, Husberg B, et al. Adult outcomes after surgery for Hirschsprung’s disease: evaluation of bowel function and quality of life. J Pediatr Surg. 2015;50:1865–9.
15. Neuvonen MI, Kyrklund K, Rintala RJ, et al. Bowel function and quality of life after transanal endorectal pull-through for Hirschsprung disease: controlled outcomes up to adulthood. Ann Surg. 2017;265:622–9.
16. Huang EY, Tolley EA, Blakely ML, et al. Changes in hospital utilization and management of Hirschsprung disease: analysis using the kids’ inpatient database. Ann Surg. 2013;257:371–5.
17. Healthcare Cost and Utilization Project (HCUP). Overview of the national (nationwide) inpatient sample (NIS). Available at:
https://www.hcup-us.ahrq.gov/nisoverview.jsp. Accessed September 15, 2020.
18. Dunn A, Grosse SD, Zuvekas SH. Adjusting health expenditures for inflation: a review of measures for health services research in the United States. Health Serv Res. 2018;53:175–96.
19. Bureau of Economic Analysis, U.S. Department of Commerce. Available at:
https://www.bea.gov/data/special-topics/health-care. Accessed at November 12, 2021.
20. CDC NCHS Natality public use data 2000-2014, on CDC Wonder Online Database. Available at:
https://wonder.cdc.gov/wonder/help/natality.html. Accessed September 15, 2020.
21. Lumley T. Survey: analysis of complex survey samples. R package version 4.1. 2019. Available at:
http://r-survey.r-forge.r-project.org/survey. Accessed September 15, 2020.
22. Zhu Y, Fung K. Statistical Methods in Developmental Toxicity Risk Assessment. Marcel Dekker; 1996:413–46.
23. Granstrom AL, Wester T. Mortality in Swedish patients with Hirschsprung disease. Pediatr Surg Int. 2017;33:1177–81.
24. Downey EC, Hughes E, Putnam AR, et al. Hirschsprung disease in the premature newborn: a population-based study and 40-year single center experience. J Pediatr Surg. 2015;50:123–5.
25. Anderson JE, Vanover MA, Saadai P, et al. Epidemiology of Hirschsprung disease in California from 1995 to 2013. Pediatr Surg Int. 2018;34:1299–303.
26. Menezes M, Puri P. Long-term outcome of patients with enterocolitis complicating Hirschsprung’s disease. Pediatr Surg Int. 2006;22:316–8.
27. Rescorla FJ, Morrison AM, Engles D, et al. Hirschsprung’s disease: evaluation of mortality and long-term function in 260 cases. JAMA Surg. 1992;127:934–42.
28. Moore SW, Albertyn R, Cywes S. Clinical outcome and long-term quality of life after surgical correction of Hirschsprung’s disease. J Pediatr Surg. 1996;31:1496–502.
29. Wildhaber BE, Teitelbaum DH, Coran AG. Total colonic Hirschsprung’s disease: a 28-year experience. J Pediatr Surg. 2005;40:203–6.
30. Pruitt LCC, Skarda DE, Rollins MD, et al. Hirschsprung-associated enterocolitis in children treated at US children’s hospitals. J Pediatr Surg. 2020;55:535–40.
31. Quiroz HJ, Perez EA, Franklin KN, et al. Pull-through procedure in children with Hirschsprung disease: a nationwide analysis on postoperative outcomes. J Pediatr Surg. 2020;55:899–903.
32. Taguchi T, Obata S, Ieiri S. Current status of Hirschsprung’s disease: based on a nationwide survey of Japan. Pediatr Surg Int. 2017;33:497–504.
33. Topaz M, Shafran-Topaz L, Bowles KH. ICD-9 to ICD-10: evolution, revolution, and current debates in the United States. Perspect Health Inf Manag. 2013;10:1d.
