http://informahealthcare.com/jmf ISSN: 1476-7058 (print), 1476-4954 (electronic) J Matern Fetal Neonatal Med, Early Online: 1–7 ! 2014 Informa UK Ltd. DOI: 10.3109/14767058.2014.927428
ORIGINAL ARTICLE
Factor affecting length of stay in late preterm infants: an US national database study J Matern Fetal Neonatal Med Downloaded from informahealthcare.com by University of Maastricht on 07/05/14 For personal use only.
Hany Aly1, Heather Hoffman2, Mohamed El-Dib1, Lujain Said2, and Mohamed Mohamed1 1
Department of Neonatology, The George Washington University and Children’s National Medical Center, Washington, DC, USA and 2Department of Epidemiology and Biostatistics, The George Washington University, School of Public Health and Health Services, Washington, DC, USA Abstract
Keywords
Objectives: Late preterm infants are the fastest growing segment of the premature infant population in the United States. However, it is not known if demographic and clinical factors can impact the length of hospital stay (LOS) in this population. The objectives of this study are to determine the following: (a) factors associated with a LOS43 d and (b) whether there is any difference in risks between infants born at 33–34 versus 35–36 weeks. Methods: Utilizing the Nationwide Inpatient Sample Database, a de-identified dataset produced by the Healthcare Cost and Utilization Project, analysis of 81 913 infants born at 33–36 weeks from 2007 to 2008 was conducted. LOS outcome was defined as 3 and 43 d. Bivariable and multivariable logistic regression was used to evaluate predictors of LOS among this population. Results: Only 42.7% of infants were discharged home within three days. Factors associated with a LOS43 d included gestational age of 535 weeks (RR ¼ 1.63; CI: 1.58–1.68), birth weight of 52500 g (RR ¼ 1.36; CI: 1.33–1.39), male sex (RR ¼ 1.06; CI: 1.05–1.07), delivery via C-section (RR ¼ 1.46; CI: 1.41–1.51) and multiple gestation (RR ¼ 1.08; 95% CI: 1.06–1.09). Other significant factors included race, birth region, primary insurance payer and clinical complications. In the adjusted interaction model, these variables have more impact on longer LOS in the 35–36 weeks group (p50.0001). Conclusion: Birth region in addition to gestational age, birth weight, gender, mode of delivery, multiple gestation and primary insurance payer affect LOS in late preterm infants. These variables are more critical for the 35–36 week population.
HCUP, hospitalization, neonates, NIS, premature
Introduction Premature birth is a determining factor of neonatal morbidity and mortality. Approximately 450 000 infants in the United States are born preterm, accounting for about 11.5% of live births [1]. Of these, 70% are late preterm births; approximately, 320 000 live births annually [1]. Late preterm infants, defined as births between 34 weeks and 0/7 d through 36 weeks and 6/7 d [2], are the fasting growing segment of singleton preterm births in the United States [3]. From 1990 to 2005, the birth rate of late preterm infants in the United States increased by 25% [4]. However, from 2006 to 2012, the percentage of infants born late preterm has declined by 11% [1]. Factors associated with late preterm births include spontaneous labor, preterm premature rupture of membranes and elected deliveries [4]. Late preterm delivery is associated with increased mortality and morbidities when compared to full term infants. In 2004, the infant mortality rate among late preterm infants
Address for correspondence: Hany Aly, MD, 900 23rd Street, NW, Suite G-2092, Room G-132, Washington, DC 20037. Tel: 202-715-5236. Fax: 202-715-5354. E-mail: [email protected]
History Received 28 January 2014 Revised 23 April 2014 Accepted 20 May 2014 Published online 19 June 2014
was three times higher than among full-term infants (7.3 versus 2.4 infant deaths per 1000 live births); however, studies conducted after 2004 have shown a significant decrease in mortality with each weekly increase of gestational age [5–8]. Late preterm infants are more likely to develop respiratory distress, temperature instability, hypoglycemia, kernicterus, apnea, seizures and feeding problems when compared to fullterm infants. They are more likely to be admitted to neonatal intensive care units, to be re-admitted and to have worse neurodevelopmental outcome [9,10]. Consequently, these risk factors have a significant impact on the length of hospital stay (LOS). Previous studies showed that LOS increases with decreasing gestational age and birth weight, and increases the risk of infants being exposed to adverse environmental stimuli, such as healthcare-associated infections and medical complications. For these reasons, LOS is an important indicator that reflects quality of care and efficient resource utilization and is a goal for hospitals to reduce it without compromising the quality of care [11,12]. LOS, among various outcomes and morbidities, has been studied in the very preterm infant population [11,13,14]. Yet, large studies that assess several of these factors in relation to the LOS of late preterm infants are lacking [5]. In addition, these studies
2
H. Aly et al.
do not consist of a national, population-based cohort, but rather were conducted in a single unit or a cohort of hospitals within one locality with a relatively small sample size [11]. This study aims to assess the relationship between demographic and clinical factors in relation to the LOS among a national, population-based cohort of infants born late preterm. We are also interested to identify the factors that influence the LOS of the late preterm infant at different gestation.
J Matern Fetal Neonatal Med, Early Online: 1–7
admission date from the discharge date. LOS was defined as a binary outcome of 3 d and 43 d for this study. This cutoff was used based on the LOS distribution of our dataset with a median LOS of four days (IQR 8.00), which confirms previous studies [21]. We included infants who had a LOS of zero days in our analyses. These infants only made up 0.11% of the study population, but were deemed important to since they did not fit any of the exclusion criteria. Independent variables
J Matern Fetal Neonatal Med Downloaded from informahealthcare.com by University of Maastricht on 07/05/14 For personal use only.
Patients and methods This study was approved by the George Washington University Institutional Review Board, Office of Human Research. Data were extracted through the Nationwide Inpatient Sample Database (NIS), a de-identified dataset produced by the Healthcare Cost and Utilization Project (HCUP), which is sponsored by the Agency for Healthcare Research and Quality. Data were obtained for years 2007 across 40 states and 2008 across 42 states [15]. HCUP is the largest healthcare database in the United States, reproduced from an all-payer national database collected annually from millions of inpatient hospitalization records across the United States [16]. These datasets reflect a national stratified sample of about 1000 hospitals from across the United States with various care levels (primary, tertiary), type of insurance (public, private) and academic (university, general) settings [16–19]. Datasets include more than 100 data elements for each hospital stay and represent a 20% sample of all hospital admissions during any given year for patients of all ages [16,20].
Patient selection and identification Patients included are those between 33 and 36 weeks GA. The NIS dataset groups infants into two week increments, thus we categorized late preterm infants into two groups: 33–34 and 35–36 weeks of gestation. In order to avoid duplicate inclusion, we excluded infants who were admitted from another hospital, other healthcare facility including long-term care and court/law enforcement. In addition, we excluded late preterm infants who were not discharged home, meaning they were discharged to a shortterm hospital, skilled nursing facility, intermediate care facility or another type of facility. If LOS information was missing or invalid from the HCUP NIS dataset, cases were excluded. We also excluded late preterm infants born with conditions that would have likely increased the LOS, such as infants born at a birth weight of 51000 g and central nervous system anomalies, lung anomalies, congenital heart disease, congenital diaphragmatic hernia, abdominal wall defects, multiple congenital anomalies and trisomies 13, 18 and 21. Late preterm infants who died during hospitalization were also excluded from the analysis as this could have impacted the results by decreasing the LOS.
Studied variables Outcome variable The LOS was calculated as the total number of inpatient hospital stay days until the first discharge by subtracting the
A total of 11 independent variables were analyzed in relation to LOS. The independent variables in this study fall within two categories: demographic and clinical characteristics of the infant and mother. Demographic characteristics tested in the analyses included gestational age, sex, race, primary expected payer and geographical birth region. Race was broken down into four categories: White (reference group), Black, Hispanic and Other (which includes Asian or Pacific Islanders, Native Americans, other racial groups and infants with an unknown race). Primary expected payer represents the primary payer that bears the major financial responsibility for the infants’ cost of hospitalization and was categorized into four groups: Private insurance (Blue Cross, commercial carriers, PPOs and HMOs), Medicaid (reference group), Uninsured (the infant’s hospitalization was self-paid or there was no charge) and Other (Medicare, Worker’s Compensation, CHAMPUS, CHAMPVA, Title V and other government programs). Geographical birth region represents the region in the United States in which infants were given birth to. To define birth region in our study, hospital state postal codes in the dataset were re-defined according to the four US Census Bureau regions: Northeast (reference group), South, West and Midwest. Clinical characteristics included birth weight (52500 g or 2500 g), maternal mode of delivery, maternal gestation type (singleton versus multiple), respiratory distress syndrome (RDS), sepsis and necrotizing enterocolitis (NEC). Statistical analysis methods Univariable analysis was performed to obtain descriptive statistics of the late preterm infant population. Chi-square test was used to examine the association between gestational age and each predictor for descriptive statistics. Bivariable analyses were used to examine the association between gestational age and each predictor independently against LOS. Multivariable logistic regression was performed to assess the relationship between all predictor variables on LOS among the overall late preterm infant population and to test for the interaction of gestational age with each predictor on the LOS outcome, controlling for all demographic and clinical factors. Unadjusted and adjusted relative risks, 95% confidence intervals and p values were calculated to compare differences. A log-link, generalized linear model with a Modified Poisson distribution was utilized for bivariable and multivariable logistic regression models. Compound symmetry was used in the logistic regression models as the correlation structure. The HCUP hospital identifier was applied as a cluster variable to account for differences that
Length of stay in late preterm infants
J Matern Fetal Neonatal Med Downloaded from informahealthcare.com by University of Maastricht on 07/05/14 For personal use only.
DOI: 10.3109/14767058.2014.927428
3
Figure 1. Distribution of length of stay days of late preterm infants.
may exist among and between hospital entities and among the types of patients admitted to these hospitals based on the effects of the factors in the study. Statistical analyses were performed using SAS version 9.3 for Windows (SAS Institute Inc., Cary, NC).
Results The data included 95 399 late preterm infants in years 2007 and 2008. Of these, 414 infants died during hospitalization and were excluded from the analysis. Additional records were excluded if the infants met the exclusion criteria as previously stated. Thus, a total of 81 913 late preterm infants were included in the final analysis of this study. As previously mentioned, the median LOS for the entire late preterm infant population was four days, and about 99% of infants in the study had a LOS of no more than 32 days, as shown in Figure 1. Little variation existed among infants between hospitals based on the intraclass correlations (ICC). Most of the variation occurred among infants within hospitals; thus, infants within a hospital were not more alike one another than infants between hospitals (ICC in the adjusted multivariable model with all predictors was 0.055, while the ICC in the adjusted multivariable interaction model was 0.049). Demographic and clinical characteristics of the study population overall and stratified by gestational age are presented in Table 1. Of the 81 913 late preterm infants in the study, 24 086 (29.36%) were born at 33–34 weeks of gestation and 57 827 (70.64%) were born at 35–36 weeks of gestation. Of infants 33–34 weeks, almost 90% spent more than three days in the hospital; however, the percentage
decreased by almost half (44.01%) in the older age group. Infants born at 33–34 weeks were more likely to have a weight 52500 g at birth, be delivered by C-section, have multiple gestation, develop RDS, sepsis or NEC. Table 2 presents the demographic and clinical characteristics of late preterm infants in relation to the LOS. White race, male sex, private insurance coverage, being born in Northeast and body weight 52500 g were all significant variables associated with LOS43 d. Furthermore, infants born by C-section, with multiple gestation, RDS, sepsis or NEC were more likely to stay 43 d. Table 3 examines the unadjusted bivariable model where gestational age of late preterm infants were stratified into two groups: 33–34 versus 35–36 weeks of gestation. Gestational age groups were statistically significant with each independent predictor. The effect of each specific factor on increase LOS43 d was more significant in the 35–36 weeks group more than the 33–34 weeks group. In the adjusted multivariable model with all predictors shown in Table 4, all demographic and clinical factors were significantly associated with a longer LOS in the overall late preterm infant population. In the adjusted multivariable model with interactions between gestational age and each predictor, longer LOS was significantly associated with all demographic and clinical factors, except for race and primary expected payer. High risk of a longer LOS was associated with being male, weighing 52500 g at birth, having been delivered via C-section, being born from a multiple birth and having RDS, sepsis and NEC complications after delivery in both age groups. These associations were higher in the older age group than the younger age group.
4
H. Aly et al.
J Matern Fetal Neonatal Med, Early Online: 1–7
J Matern Fetal Neonatal Med Downloaded from informahealthcare.com by University of Maastricht on 07/05/14 For personal use only.
Table 1. Demographic and clinical characteristics of the study population and comparison between the two subgroups.
Demographic Characteristics Length of stay 3 d 43 d Race White Black Hispanic Other Sex Male Female Primary expected payer Private insurance Medicaid Uninsured Other Birth region Northeast South West Midwest Clinical characteristics Birth weight 52500 g 2500 g Cesarean section Multiple gestation RDS Sepsis NEC Total
33–36 weeks N (%)
33–34 weeks N (%)
34 945 (42.7) 46 968 (57.3)
2592 (10.7) 21 494 (89.3)
35–36 weeks N (%)
p values 50.001
32 353 (56) 25 474 (44) 50.001
30 393 8972 9938 32 609
(37) (11) (11.7) (40.3)
8973 3023 2798 9292
(37.2) (12.6) (11.2) (39)
21 420 5949 7140 23 317
(37) (10.3) (11.9) (40.8) 0.043
42 902 (52.4) 38 974 (47.6)
12 775 (53) 11 300 (47)
30 127 (52.1) 27 674 (47.9)
42 669 32 796 3385 2902
(52.3) (40.1) (4.1) (3.5)
12 250 9927 967 904
(51.1) (41.2) (4) (3.7)
30 419 22 869 2418 1998
(52.8) (39.6) (4.2) (3.4)
14 076 23 417 23 452 20 968
(17.4) (28.1) (27.5) (27)
4415 7233 6591 5847
(18.5) (29.5) (26.2) (25.8)
9661 16 184 16 861 15 121
(17) (27.5) (28) (27.5)
0.01
0.047
50.001 39 768 (52.3) 36 164 (47.7) 37 660 (48.4) 17 852 (23) 8372 (10.2) 6171 (7.5) 293 (0.4) 81 913
18 686 (82.2) 4044 (17.8) 12 174 (54.1) 6428 (28.7) 4512 (18.7) 3118 (12.9) 171 (0.7) 24 086
21 082 (39.6) 32 120 (60.4) 25 486 (46.1) 11 424 (20.7) 3860 (6.6) 3053 (5.3) 122 (0.2) 57 827
50.001 50.001 50.001 50.001 50.001
Table 2. Demographic and clinical characteristics of late preterm infants in relation to the length of hospital stay. Length of stay outcome
Demographic characteristics Gestational age 33–34 weeks 35–36 weeks Race White* Black Hispanic Other Sex Male Female Primary expected payer Private insurance Medicaid* Uninsured Other Birth region Northeast* South West Midwest Clinical characteristics Birth weight 52500 g 2500 g Cesarean section Multiple gestation RDS Sepsis NEC *Reference groups.
43 LOS days N (%)
3 LOS days N (%)
21 494 (45.8) 25 474 (54.2)
2592 (7.4) 32 353 (92.6)
18 159 5290 5320 18 198
12 234 3682 4618 14 411
p values 50.001 0.002
(38.7) (11.3) (11.3) (38.7)
(35) (10.5) (13.2) (41.3) 50.001
25 129 (53.5) 21 824 (46.5)
17 773 (50.9) 17 150 (49.1)
25 308 18 090 1718 1772
(54) (38.6) (3.6) (3.8)
17 361 14 706 1667 1130
(49.8) (42.2) (4.8) (3.2)
9359 12 846 13 034 11 729
(20) (27.3) (27.7) (25)
4717 10 571 10 418 9239
(13.5) (30.3) (29.8) (26.4)
28 686 14 796 26 641 12 951 8074 5804 293
(66) (34) (60.8) (29.6) (17.2) (12.4) (0.6)
11 082 21 368 11 019 4901 298 367 0
(34) (66) (32.5) (14.4) (0.9) (1) (0)
50.001
50.001
50.001 50.001 50.001 50.001 50.001 50.001
Length of stay in late preterm infants
DOI: 10.3109/14767058.2014.927428
5
J Matern Fetal Neonatal Med Downloaded from informahealthcare.com by University of Maastricht on 07/05/14 For personal use only.
Table 3. Unadjusted bivariable model: Relative risks of having a length of stay 43 d for demographic and clinical predictors of late preterm infants stratified by gestational age.
Demographic characteristics Race White* Black Hispanic Other Sex Male versus female Primary expected payer Private insurance Medicaid* Uninsured Other Birth region Northeast* South West Midwest Clinical characteristics Birth weight 52500 g versus 2500 g Cesarean section Multiple gestation RDS Sepsis NEC
33–34 weeks gestation Relative risk (95% CI)
35–36 weeks gestation Relative risk (95% CI)
– 0.99 (0.97–1.03) 0.94 (0.91–0.98) 1.01 (0.98–1.05)
– 0.91 (0.87–0.96) 0.83 (0.78–0.88) 0.91 (0.88–0.95)
1.01 (0.99–1.02)
1.06 (1.04–1.09)
1.01 (0.99–1.03) – 0.96 (0.91–1) 1.05 (1.01–1.09)
1.08 (1.04 – 1.12) – 0.9 (0.84–0.97) 1.14 (1.03–1.25)
– 0.88 (0.81–0.94) 0.95 (0.88–1.02) 0.89 (0.83–0.96)
– 0.74 (0.66–0.82) 0.81 (0.73–0.9) 0.76 (0.68–0.84)
1.19 1.11 1.06 1.11 1.12 1.08
1.51 2.12 1.55 2.26 2.18 2.13
p value 50.001
50.001 0.002
0.004
50.001 (1.16–1.23) (1.09–1.13) (1.05–1.08) (1.09–1.19) (1.09–1.16) (1.05–1.12)
(1.47–1.56) (2–2.25) (1.49–1.61) (2.16–2.35) (2.08–2.29) (2.01–2.25)
50.001 50.001 50.001 50.001 50.001
*Reference groups.
Table 4. Adjusted multivariable models: relative risks of having a length of stay43d for demographic and clinical predictors of the entire late preterm infant population (left), and with interactions between gestational age and each predictor of late preterm infants (right). Overall 33–36 weeks Relative risk (95% CI) Demographic characteristics Gestational age 33–34 versus 35–36 weeks Race White* Black Hispanic Other Sex Male versus female Primary expected payer Private insurance Medicaid* Uninsured Other Birth region Northeast* South West Midwest Clinical characteristics Birth weight 52500 g versus 2500 g Cesarean section Multiple gestation RDS Sepsis NEC *Reference groups.
p value
33–34 weeks Relative risk (95% CI)
35–36 weeks Relative risk (95% CI)
– 0.99 (0.97–1.02) 0.97 (0.94–0.99) 1.02 (0.99–1.04)
– 0.96 (0.92–1) 0.92 (0.88–0.96) 0.98 (0.95–1.02)
p value
50.001 1.63 (1.58–1.68) 50.001 – 0.97 (0.94–0.99) 0.94 (0.91–0.96) 0.99 (0.97–1.02)
0.14
50.001 1.06 (1.05–1.07)
50.001 1.02 (1.01–1.03)
1.079 (1.06–1.1)
0.039 1.01 (0.99–1.02) – 0.95 (0.91–0.99) 1.03 (0.98–1.09)
0.21 1 (0.99–1.02) – 0.97 (0.93–1.02) 1.03 (0.99–1.07)
1.03 (0.99–1.06 – 0.92 (0.86–0.98) 1.05 (0.98–1.13)
– 0.9 (0.83–0.97) 0.99 (0.92–1.06) 0.94 (0.88–1.01)
– 0.77 (0.71–0.85) 0.91 (0.83–0.99) 0.86 (0.78–0.94)
1.19 1.09 1.02 1.09 1.1 1.06
1.42 1.88 1.16 1.85 1.79 1.68
50.001
0.01 – 0.91 (0.85–0.99) 1.01 (0.94–1.1) 0.97 (0.89–1.05) 50.001 1.36 1.46 1.08 1.33 1.37 1.19
(1.33–1.39) (1.41–1.51) (1.06–1.09) (1.29–1.36) (1.34–1.42) (1.11–1.28)
50.001 50.001 50.001 50.001 50.001
50.001 (1.16–1.22) (1.07–1.1) (1.01–1.03) (1.07–1.11) (1.07–1.13) (1.03–1.09)
(1.38–1.46) (1.77–1.99) (1.13–1.2) (1.77–1.94) (1.71–1.86) (1.45–1.94)
50.001 50.001 50.001 50.001 50.001
6
H. Aly et al.
J Matern Fetal Neonatal Med Downloaded from informahealthcare.com by University of Maastricht on 07/05/14 For personal use only.
Discussion This study showed multiple demographic and clinical factors to influence LOS in late preterm infants born both at 33–34 weeks and 35–36 weeks of gestation. Older infants were influenced more by all predictors, with increased LOS with birth weight of 52500 g, RDS, sepsis and NEC. To our knowledge, this study is the first to look at demographic and clinical factors influencing the LOS of late preterm infants across the United States and to compare the risks of having a longer LOS by stratifying the late preterm infant population into two gestational age categories. We showed that birth region has a significant influence on LOS. For example, there was about 10% increased incidence of LOS43 d in Northeast when compared to other regions. This could be related to differences in care protocols but also could be related to differences in neonatal diseases as shown in previous studies [22]. Previous data have suggested that RDS and culture-proven sepsis occur more frequently in late preterm infants [4]. We found that RDS, sepsis and NEC increased the risk of LOS among the overall late preterm infant and both stratifies groups. Studies have suggested that male sex is associated with risk of respiratory distress in late preterm infants [23]. In our study, male sex was associated with a longer LOS. Engle et al. reported that Cesarean sections were associated with longer LOS among late and early term infants. Our findings were consistent with theirs, in that we also found Cesarean sections to be associated with longer LOS both in the overall late preterm infant population and when late preterm infants were stratified [24]. The majority of studies attempted to analyze the relationship between LOS with birth weight and/or gestational age in order to predict LOS, and the analysis is usually conducted in one or a group of hospitals in one locality; however, there is little information on how demographic and clinical factors predict the LOS of the late preterm infant population across different US hospitals, while taking into consideration the presence of other perinatal, demographic and socioeconomic factors [13]. This information will be valuable for hospitals to determine expectations for LOS and to put policies for the late preterm infant population given their special needs. The large sample size of this study and weights that were used on the data make this study more representative of the late preterm infant population. One limitation of this study is the inability to breakdown infants into individual gestational weeks. Furthermore, the ‘‘Other’’ category of the race variable includes infants that had an unknown race. This is a limitation because these infants could potentially fit in any of the other race categories if their race were known, skewing our results. Finally, a few numbers of infants were transferred to other facilities and were not discharged home. One of the limitations associated with the use of a national de-identified dataset is the inability to follow these subjects. Although these infants could be sicker, the impact of eliminating them on the overall study results is minimal because of their limited numbers.
Conclusions Late preterm infants are at high risk of having a longer LOS across several demographic and clinical factors regardless if
J Matern Fetal Neonatal Med, Early Online: 1–7
they were born at 33–34 weeks or 35–36 weeks of gestation. These risk factors are more influential in the older age group compared to the younger age group. This research identifies the need for increased assessment and care of the late preterm infant population and can be used as a resource to guide this care. This in turn helps reduce the LOS of the infant while directly lowering healthcare costs. Future research is needed to stratify the late preterm infant population into separate weeks to gauge differences in LOS between each single week.
Declaration of interest The authors report no declarations of interest.
References 1. Hamilton BE, Martin JA, Ventura SJ. Births: preliminary data for 2012. Natl Vital Stat Rep 2013;62:1–20. 2. Raju TN, Higgins RD, Stark AR, Leveno KJ. Optimizing care and outcome for late-preterm (near-term) infants: a summary of the workshop sponsored by the National Institute of Child Health and Human Development. Pediatrics 2006;118:1207–14. 3. Mally PV, Bailey S, Hendricks-Munoz KD. Clinical issues in the management of late preterm infants. Curr Probl Pediatr Adolesc Health Care 2010;40:218–33. 4. Dong Y, Yu JL. An overview of morbidity, mortality and long-term outcome of late preterm birth. World J Pediatr 2011;7:199–204. 5. Pulver LS, Denney JM, Silver RM, Young PC. Morbidity and discharge timing of late preterm newborns. Clin Pediatr (Phila) 2010;49:1061–7. 6. Khashu M, Narayanan M, Bhargava S, Osiovich H. Perinatal outcomes associated with preterm birth at 33 to 36 weeks’ gestation: a population-based cohort study. Pediatrics 2009;123: 109–13. 7. Tomashek KM, Shapiro-Mendoza CK, Davidoff MJ, Petrini JR. Differences in mortality between late-preterm and term singleton infants in the United States, 1995–2002. J Pediatr 2007;151:450–6, 456 e1. 8. Young PC, Glasgow TS, Li X, et al. Mortality of late-preterm (near-term) newborns in Utah. Pediatrics 2007;119:e659–65. 9. Raju TN. Epidemiology of late preterm (near-term) births. Clin Perinatol 2006;33:751–63; abstract vii. 10. Tomashek KM, Shapiro-Mendoza CK, Weiss J, et al. Early discharge among late preterm and term newborns and risk of neonatal morbidity. Semin Perinatol 2006;30:61–8. 11. Korvenranta E, Linna M, Hakkinen U, et al. Differences in the length of initial hospital stay in very preterm infants. Acta Paediatr 2007;96:1416–20. 12. Schulman J. Studying determinants of length of hospital stay. J Perinatol 2006;26:243–5. 13. Bannwart Dde C, Rebello CM, Sadeck LS, et al. Prediction of length of hospital stay in neonatal units for very low birth weight infants. J Perinatol 1999;19:92–6. 14. Cotten CM, Oh W, McDonald S, et al. Prolonged hospital stay for extremely premature infants: risk factors, center differences, and the impact of mortality on selecting a best-performing center. J Perinatol 2005;25:650–5. 15. 2008 Introduction to the HICUP Nationwide Inpatient Sample (NIS). Agency for Healthcare Research and Quality Healthcare Cost and Utilization Project (HCUP). Available from: http://www.hcup-us.ahrq.gov/db/nation/nis/ NIS_2008_INTRODUCTION.pdf [last accessed 29 Jan 2014]. 16. 2013 Overview of the Nationwide Inpatient Sample (NIS). In HCUP Databases. Healthcare Cost and Utilization Project (HCUP), Agency for Healthcare Research and Quality. Available from: http://www.hcup-us.ahrq.gov/nisoverview.jsp [last accessed 29 Jan 2014]. 17. 2009 HCUP Sample Design: National Databases. Healthcare Cost and Utilization Project (HCUP), Agency for Healthcare Research and Quality. Available from: http://www.hcup-us.ahrq.gov/tech_assist/sampledesign/508_compliance/508course.htm [last accessed 29 Jan 2014].
DOI: 10.3109/14767058.2014.927428
J Matern Fetal Neonatal Med Downloaded from informahealthcare.com by University of Maastricht on 07/05/14 For personal use only.
18. 2009 HCUP Overview. Healthcare Cost and Utilization Project (HCUP), Agency for Healthcare Research and Quality. Available from: http://www.hcup-us.ahrq.gov/overview.jsp [last accessed 29 Jan 2014]. 19. 2010 HCUP Partners. Healthcare Cost and Utilization Project (HCUP), Agency for Healthcare Research and Quality. Available from: http://www.hcup-us.ahrq.gov/partners.jsp [last accessed 29 Jan 2014]. 20. HCUP Sample Design: National Databases. Healthcare Cost and Utilization Project (HCUP), Agency for Healthcare Research and Quality. Available from: http://www.hcup-us.ahrq.gov/tech_assist/ sampledesign/508_compliance/508course.htm [last accessed 29 Jan 2014].
Length of stay in late preterm infants
7
21. Shapiro-Mendoza CK, Tomashek KM, Kotelchuck M, et al. Risk factors for neonatal morbidity and mortality among ‘‘healthy,’’ late preterm newborns. Semin Perinatol 2006;30:54–60. 22. Mohamed MA, Aly H. Birth region, race and sex may affect the prevalence of congenital diaphragmatic hernia, abdominal wall and neural tube defects among US newborns. J Perinatol 2012;32: 861–8. 23. Anadkat JS, Kuzniewicz MW, Chaudhari BP, et al. Increased risk for respiratory distress among white, male, late preterm and term infants. J Perinatol 2012;32:780–5. 24. Engle WA, Kominiarek MA. Late preterm infants, early term infants, and timing of elective deliveries. Clin Perinatol 2008;35: 325–41, vi.
ORIGINAL ARTICLE
Factor affecting length of stay in late preterm infants: an US national database study J Matern Fetal Neonatal Med Downloaded from informahealthcare.com by University of Maastricht on 07/05/14 For personal use only.
Hany Aly1, Heather Hoffman2, Mohamed El-Dib1, Lujain Said2, and Mohamed Mohamed1 1
Department of Neonatology, The George Washington University and Children’s National Medical Center, Washington, DC, USA and 2Department of Epidemiology and Biostatistics, The George Washington University, School of Public Health and Health Services, Washington, DC, USA Abstract
Keywords
Objectives: Late preterm infants are the fastest growing segment of the premature infant population in the United States. However, it is not known if demographic and clinical factors can impact the length of hospital stay (LOS) in this population. The objectives of this study are to determine the following: (a) factors associated with a LOS43 d and (b) whether there is any difference in risks between infants born at 33–34 versus 35–36 weeks. Methods: Utilizing the Nationwide Inpatient Sample Database, a de-identified dataset produced by the Healthcare Cost and Utilization Project, analysis of 81 913 infants born at 33–36 weeks from 2007 to 2008 was conducted. LOS outcome was defined as 3 and 43 d. Bivariable and multivariable logistic regression was used to evaluate predictors of LOS among this population. Results: Only 42.7% of infants were discharged home within three days. Factors associated with a LOS43 d included gestational age of 535 weeks (RR ¼ 1.63; CI: 1.58–1.68), birth weight of 52500 g (RR ¼ 1.36; CI: 1.33–1.39), male sex (RR ¼ 1.06; CI: 1.05–1.07), delivery via C-section (RR ¼ 1.46; CI: 1.41–1.51) and multiple gestation (RR ¼ 1.08; 95% CI: 1.06–1.09). Other significant factors included race, birth region, primary insurance payer and clinical complications. In the adjusted interaction model, these variables have more impact on longer LOS in the 35–36 weeks group (p50.0001). Conclusion: Birth region in addition to gestational age, birth weight, gender, mode of delivery, multiple gestation and primary insurance payer affect LOS in late preterm infants. These variables are more critical for the 35–36 week population.
HCUP, hospitalization, neonates, NIS, premature
Introduction Premature birth is a determining factor of neonatal morbidity and mortality. Approximately 450 000 infants in the United States are born preterm, accounting for about 11.5% of live births [1]. Of these, 70% are late preterm births; approximately, 320 000 live births annually [1]. Late preterm infants, defined as births between 34 weeks and 0/7 d through 36 weeks and 6/7 d [2], are the fasting growing segment of singleton preterm births in the United States [3]. From 1990 to 2005, the birth rate of late preterm infants in the United States increased by 25% [4]. However, from 2006 to 2012, the percentage of infants born late preterm has declined by 11% [1]. Factors associated with late preterm births include spontaneous labor, preterm premature rupture of membranes and elected deliveries [4]. Late preterm delivery is associated with increased mortality and morbidities when compared to full term infants. In 2004, the infant mortality rate among late preterm infants
Address for correspondence: Hany Aly, MD, 900 23rd Street, NW, Suite G-2092, Room G-132, Washington, DC 20037. Tel: 202-715-5236. Fax: 202-715-5354. E-mail: [email protected]
History Received 28 January 2014 Revised 23 April 2014 Accepted 20 May 2014 Published online 19 June 2014
was three times higher than among full-term infants (7.3 versus 2.4 infant deaths per 1000 live births); however, studies conducted after 2004 have shown a significant decrease in mortality with each weekly increase of gestational age [5–8]. Late preterm infants are more likely to develop respiratory distress, temperature instability, hypoglycemia, kernicterus, apnea, seizures and feeding problems when compared to fullterm infants. They are more likely to be admitted to neonatal intensive care units, to be re-admitted and to have worse neurodevelopmental outcome [9,10]. Consequently, these risk factors have a significant impact on the length of hospital stay (LOS). Previous studies showed that LOS increases with decreasing gestational age and birth weight, and increases the risk of infants being exposed to adverse environmental stimuli, such as healthcare-associated infections and medical complications. For these reasons, LOS is an important indicator that reflects quality of care and efficient resource utilization and is a goal for hospitals to reduce it without compromising the quality of care [11,12]. LOS, among various outcomes and morbidities, has been studied in the very preterm infant population [11,13,14]. Yet, large studies that assess several of these factors in relation to the LOS of late preterm infants are lacking [5]. In addition, these studies
2
H. Aly et al.
do not consist of a national, population-based cohort, but rather were conducted in a single unit or a cohort of hospitals within one locality with a relatively small sample size [11]. This study aims to assess the relationship between demographic and clinical factors in relation to the LOS among a national, population-based cohort of infants born late preterm. We are also interested to identify the factors that influence the LOS of the late preterm infant at different gestation.
J Matern Fetal Neonatal Med, Early Online: 1–7
admission date from the discharge date. LOS was defined as a binary outcome of 3 d and 43 d for this study. This cutoff was used based on the LOS distribution of our dataset with a median LOS of four days (IQR 8.00), which confirms previous studies [21]. We included infants who had a LOS of zero days in our analyses. These infants only made up 0.11% of the study population, but were deemed important to since they did not fit any of the exclusion criteria. Independent variables
J Matern Fetal Neonatal Med Downloaded from informahealthcare.com by University of Maastricht on 07/05/14 For personal use only.
Patients and methods This study was approved by the George Washington University Institutional Review Board, Office of Human Research. Data were extracted through the Nationwide Inpatient Sample Database (NIS), a de-identified dataset produced by the Healthcare Cost and Utilization Project (HCUP), which is sponsored by the Agency for Healthcare Research and Quality. Data were obtained for years 2007 across 40 states and 2008 across 42 states [15]. HCUP is the largest healthcare database in the United States, reproduced from an all-payer national database collected annually from millions of inpatient hospitalization records across the United States [16]. These datasets reflect a national stratified sample of about 1000 hospitals from across the United States with various care levels (primary, tertiary), type of insurance (public, private) and academic (university, general) settings [16–19]. Datasets include more than 100 data elements for each hospital stay and represent a 20% sample of all hospital admissions during any given year for patients of all ages [16,20].
Patient selection and identification Patients included are those between 33 and 36 weeks GA. The NIS dataset groups infants into two week increments, thus we categorized late preterm infants into two groups: 33–34 and 35–36 weeks of gestation. In order to avoid duplicate inclusion, we excluded infants who were admitted from another hospital, other healthcare facility including long-term care and court/law enforcement. In addition, we excluded late preterm infants who were not discharged home, meaning they were discharged to a shortterm hospital, skilled nursing facility, intermediate care facility or another type of facility. If LOS information was missing or invalid from the HCUP NIS dataset, cases were excluded. We also excluded late preterm infants born with conditions that would have likely increased the LOS, such as infants born at a birth weight of 51000 g and central nervous system anomalies, lung anomalies, congenital heart disease, congenital diaphragmatic hernia, abdominal wall defects, multiple congenital anomalies and trisomies 13, 18 and 21. Late preterm infants who died during hospitalization were also excluded from the analysis as this could have impacted the results by decreasing the LOS.
Studied variables Outcome variable The LOS was calculated as the total number of inpatient hospital stay days until the first discharge by subtracting the
A total of 11 independent variables were analyzed in relation to LOS. The independent variables in this study fall within two categories: demographic and clinical characteristics of the infant and mother. Demographic characteristics tested in the analyses included gestational age, sex, race, primary expected payer and geographical birth region. Race was broken down into four categories: White (reference group), Black, Hispanic and Other (which includes Asian or Pacific Islanders, Native Americans, other racial groups and infants with an unknown race). Primary expected payer represents the primary payer that bears the major financial responsibility for the infants’ cost of hospitalization and was categorized into four groups: Private insurance (Blue Cross, commercial carriers, PPOs and HMOs), Medicaid (reference group), Uninsured (the infant’s hospitalization was self-paid or there was no charge) and Other (Medicare, Worker’s Compensation, CHAMPUS, CHAMPVA, Title V and other government programs). Geographical birth region represents the region in the United States in which infants were given birth to. To define birth region in our study, hospital state postal codes in the dataset were re-defined according to the four US Census Bureau regions: Northeast (reference group), South, West and Midwest. Clinical characteristics included birth weight (52500 g or 2500 g), maternal mode of delivery, maternal gestation type (singleton versus multiple), respiratory distress syndrome (RDS), sepsis and necrotizing enterocolitis (NEC). Statistical analysis methods Univariable analysis was performed to obtain descriptive statistics of the late preterm infant population. Chi-square test was used to examine the association between gestational age and each predictor for descriptive statistics. Bivariable analyses were used to examine the association between gestational age and each predictor independently against LOS. Multivariable logistic regression was performed to assess the relationship between all predictor variables on LOS among the overall late preterm infant population and to test for the interaction of gestational age with each predictor on the LOS outcome, controlling for all demographic and clinical factors. Unadjusted and adjusted relative risks, 95% confidence intervals and p values were calculated to compare differences. A log-link, generalized linear model with a Modified Poisson distribution was utilized for bivariable and multivariable logistic regression models. Compound symmetry was used in the logistic regression models as the correlation structure. The HCUP hospital identifier was applied as a cluster variable to account for differences that
Length of stay in late preterm infants
J Matern Fetal Neonatal Med Downloaded from informahealthcare.com by University of Maastricht on 07/05/14 For personal use only.
DOI: 10.3109/14767058.2014.927428
3
Figure 1. Distribution of length of stay days of late preterm infants.
may exist among and between hospital entities and among the types of patients admitted to these hospitals based on the effects of the factors in the study. Statistical analyses were performed using SAS version 9.3 for Windows (SAS Institute Inc., Cary, NC).
Results The data included 95 399 late preterm infants in years 2007 and 2008. Of these, 414 infants died during hospitalization and were excluded from the analysis. Additional records were excluded if the infants met the exclusion criteria as previously stated. Thus, a total of 81 913 late preterm infants were included in the final analysis of this study. As previously mentioned, the median LOS for the entire late preterm infant population was four days, and about 99% of infants in the study had a LOS of no more than 32 days, as shown in Figure 1. Little variation existed among infants between hospitals based on the intraclass correlations (ICC). Most of the variation occurred among infants within hospitals; thus, infants within a hospital were not more alike one another than infants between hospitals (ICC in the adjusted multivariable model with all predictors was 0.055, while the ICC in the adjusted multivariable interaction model was 0.049). Demographic and clinical characteristics of the study population overall and stratified by gestational age are presented in Table 1. Of the 81 913 late preterm infants in the study, 24 086 (29.36%) were born at 33–34 weeks of gestation and 57 827 (70.64%) were born at 35–36 weeks of gestation. Of infants 33–34 weeks, almost 90% spent more than three days in the hospital; however, the percentage
decreased by almost half (44.01%) in the older age group. Infants born at 33–34 weeks were more likely to have a weight 52500 g at birth, be delivered by C-section, have multiple gestation, develop RDS, sepsis or NEC. Table 2 presents the demographic and clinical characteristics of late preterm infants in relation to the LOS. White race, male sex, private insurance coverage, being born in Northeast and body weight 52500 g were all significant variables associated with LOS43 d. Furthermore, infants born by C-section, with multiple gestation, RDS, sepsis or NEC were more likely to stay 43 d. Table 3 examines the unadjusted bivariable model where gestational age of late preterm infants were stratified into two groups: 33–34 versus 35–36 weeks of gestation. Gestational age groups were statistically significant with each independent predictor. The effect of each specific factor on increase LOS43 d was more significant in the 35–36 weeks group more than the 33–34 weeks group. In the adjusted multivariable model with all predictors shown in Table 4, all demographic and clinical factors were significantly associated with a longer LOS in the overall late preterm infant population. In the adjusted multivariable model with interactions between gestational age and each predictor, longer LOS was significantly associated with all demographic and clinical factors, except for race and primary expected payer. High risk of a longer LOS was associated with being male, weighing 52500 g at birth, having been delivered via C-section, being born from a multiple birth and having RDS, sepsis and NEC complications after delivery in both age groups. These associations were higher in the older age group than the younger age group.
4
H. Aly et al.
J Matern Fetal Neonatal Med, Early Online: 1–7
J Matern Fetal Neonatal Med Downloaded from informahealthcare.com by University of Maastricht on 07/05/14 For personal use only.
Table 1. Demographic and clinical characteristics of the study population and comparison between the two subgroups.
Demographic Characteristics Length of stay 3 d 43 d Race White Black Hispanic Other Sex Male Female Primary expected payer Private insurance Medicaid Uninsured Other Birth region Northeast South West Midwest Clinical characteristics Birth weight 52500 g 2500 g Cesarean section Multiple gestation RDS Sepsis NEC Total
33–36 weeks N (%)
33–34 weeks N (%)
34 945 (42.7) 46 968 (57.3)
2592 (10.7) 21 494 (89.3)
35–36 weeks N (%)
p values 50.001
32 353 (56) 25 474 (44) 50.001
30 393 8972 9938 32 609
(37) (11) (11.7) (40.3)
8973 3023 2798 9292
(37.2) (12.6) (11.2) (39)
21 420 5949 7140 23 317
(37) (10.3) (11.9) (40.8) 0.043
42 902 (52.4) 38 974 (47.6)
12 775 (53) 11 300 (47)
30 127 (52.1) 27 674 (47.9)
42 669 32 796 3385 2902
(52.3) (40.1) (4.1) (3.5)
12 250 9927 967 904
(51.1) (41.2) (4) (3.7)
30 419 22 869 2418 1998
(52.8) (39.6) (4.2) (3.4)
14 076 23 417 23 452 20 968
(17.4) (28.1) (27.5) (27)
4415 7233 6591 5847
(18.5) (29.5) (26.2) (25.8)
9661 16 184 16 861 15 121
(17) (27.5) (28) (27.5)
0.01
0.047
50.001 39 768 (52.3) 36 164 (47.7) 37 660 (48.4) 17 852 (23) 8372 (10.2) 6171 (7.5) 293 (0.4) 81 913
18 686 (82.2) 4044 (17.8) 12 174 (54.1) 6428 (28.7) 4512 (18.7) 3118 (12.9) 171 (0.7) 24 086
21 082 (39.6) 32 120 (60.4) 25 486 (46.1) 11 424 (20.7) 3860 (6.6) 3053 (5.3) 122 (0.2) 57 827
50.001 50.001 50.001 50.001 50.001
Table 2. Demographic and clinical characteristics of late preterm infants in relation to the length of hospital stay. Length of stay outcome
Demographic characteristics Gestational age 33–34 weeks 35–36 weeks Race White* Black Hispanic Other Sex Male Female Primary expected payer Private insurance Medicaid* Uninsured Other Birth region Northeast* South West Midwest Clinical characteristics Birth weight 52500 g 2500 g Cesarean section Multiple gestation RDS Sepsis NEC *Reference groups.
43 LOS days N (%)
3 LOS days N (%)
21 494 (45.8) 25 474 (54.2)
2592 (7.4) 32 353 (92.6)
18 159 5290 5320 18 198
12 234 3682 4618 14 411
p values 50.001 0.002
(38.7) (11.3) (11.3) (38.7)
(35) (10.5) (13.2) (41.3) 50.001
25 129 (53.5) 21 824 (46.5)
17 773 (50.9) 17 150 (49.1)
25 308 18 090 1718 1772
(54) (38.6) (3.6) (3.8)
17 361 14 706 1667 1130
(49.8) (42.2) (4.8) (3.2)
9359 12 846 13 034 11 729
(20) (27.3) (27.7) (25)
4717 10 571 10 418 9239
(13.5) (30.3) (29.8) (26.4)
28 686 14 796 26 641 12 951 8074 5804 293
(66) (34) (60.8) (29.6) (17.2) (12.4) (0.6)
11 082 21 368 11 019 4901 298 367 0
(34) (66) (32.5) (14.4) (0.9) (1) (0)
50.001
50.001
50.001 50.001 50.001 50.001 50.001 50.001
Length of stay in late preterm infants
DOI: 10.3109/14767058.2014.927428
5
J Matern Fetal Neonatal Med Downloaded from informahealthcare.com by University of Maastricht on 07/05/14 For personal use only.
Table 3. Unadjusted bivariable model: Relative risks of having a length of stay 43 d for demographic and clinical predictors of late preterm infants stratified by gestational age.
Demographic characteristics Race White* Black Hispanic Other Sex Male versus female Primary expected payer Private insurance Medicaid* Uninsured Other Birth region Northeast* South West Midwest Clinical characteristics Birth weight 52500 g versus 2500 g Cesarean section Multiple gestation RDS Sepsis NEC
33–34 weeks gestation Relative risk (95% CI)
35–36 weeks gestation Relative risk (95% CI)
– 0.99 (0.97–1.03) 0.94 (0.91–0.98) 1.01 (0.98–1.05)
– 0.91 (0.87–0.96) 0.83 (0.78–0.88) 0.91 (0.88–0.95)
1.01 (0.99–1.02)
1.06 (1.04–1.09)
1.01 (0.99–1.03) – 0.96 (0.91–1) 1.05 (1.01–1.09)
1.08 (1.04 – 1.12) – 0.9 (0.84–0.97) 1.14 (1.03–1.25)
– 0.88 (0.81–0.94) 0.95 (0.88–1.02) 0.89 (0.83–0.96)
– 0.74 (0.66–0.82) 0.81 (0.73–0.9) 0.76 (0.68–0.84)
1.19 1.11 1.06 1.11 1.12 1.08
1.51 2.12 1.55 2.26 2.18 2.13
p value 50.001
50.001 0.002
0.004
50.001 (1.16–1.23) (1.09–1.13) (1.05–1.08) (1.09–1.19) (1.09–1.16) (1.05–1.12)
(1.47–1.56) (2–2.25) (1.49–1.61) (2.16–2.35) (2.08–2.29) (2.01–2.25)
50.001 50.001 50.001 50.001 50.001
*Reference groups.
Table 4. Adjusted multivariable models: relative risks of having a length of stay43d for demographic and clinical predictors of the entire late preterm infant population (left), and with interactions between gestational age and each predictor of late preterm infants (right). Overall 33–36 weeks Relative risk (95% CI) Demographic characteristics Gestational age 33–34 versus 35–36 weeks Race White* Black Hispanic Other Sex Male versus female Primary expected payer Private insurance Medicaid* Uninsured Other Birth region Northeast* South West Midwest Clinical characteristics Birth weight 52500 g versus 2500 g Cesarean section Multiple gestation RDS Sepsis NEC *Reference groups.
p value
33–34 weeks Relative risk (95% CI)
35–36 weeks Relative risk (95% CI)
– 0.99 (0.97–1.02) 0.97 (0.94–0.99) 1.02 (0.99–1.04)
– 0.96 (0.92–1) 0.92 (0.88–0.96) 0.98 (0.95–1.02)
p value
50.001 1.63 (1.58–1.68) 50.001 – 0.97 (0.94–0.99) 0.94 (0.91–0.96) 0.99 (0.97–1.02)
0.14
50.001 1.06 (1.05–1.07)
50.001 1.02 (1.01–1.03)
1.079 (1.06–1.1)
0.039 1.01 (0.99–1.02) – 0.95 (0.91–0.99) 1.03 (0.98–1.09)
0.21 1 (0.99–1.02) – 0.97 (0.93–1.02) 1.03 (0.99–1.07)
1.03 (0.99–1.06 – 0.92 (0.86–0.98) 1.05 (0.98–1.13)
– 0.9 (0.83–0.97) 0.99 (0.92–1.06) 0.94 (0.88–1.01)
– 0.77 (0.71–0.85) 0.91 (0.83–0.99) 0.86 (0.78–0.94)
1.19 1.09 1.02 1.09 1.1 1.06
1.42 1.88 1.16 1.85 1.79 1.68
50.001
0.01 – 0.91 (0.85–0.99) 1.01 (0.94–1.1) 0.97 (0.89–1.05) 50.001 1.36 1.46 1.08 1.33 1.37 1.19
(1.33–1.39) (1.41–1.51) (1.06–1.09) (1.29–1.36) (1.34–1.42) (1.11–1.28)
50.001 50.001 50.001 50.001 50.001
50.001 (1.16–1.22) (1.07–1.1) (1.01–1.03) (1.07–1.11) (1.07–1.13) (1.03–1.09)
(1.38–1.46) (1.77–1.99) (1.13–1.2) (1.77–1.94) (1.71–1.86) (1.45–1.94)
50.001 50.001 50.001 50.001 50.001
6
H. Aly et al.
J Matern Fetal Neonatal Med Downloaded from informahealthcare.com by University of Maastricht on 07/05/14 For personal use only.
Discussion This study showed multiple demographic and clinical factors to influence LOS in late preterm infants born both at 33–34 weeks and 35–36 weeks of gestation. Older infants were influenced more by all predictors, with increased LOS with birth weight of 52500 g, RDS, sepsis and NEC. To our knowledge, this study is the first to look at demographic and clinical factors influencing the LOS of late preterm infants across the United States and to compare the risks of having a longer LOS by stratifying the late preterm infant population into two gestational age categories. We showed that birth region has a significant influence on LOS. For example, there was about 10% increased incidence of LOS43 d in Northeast when compared to other regions. This could be related to differences in care protocols but also could be related to differences in neonatal diseases as shown in previous studies [22]. Previous data have suggested that RDS and culture-proven sepsis occur more frequently in late preterm infants [4]. We found that RDS, sepsis and NEC increased the risk of LOS among the overall late preterm infant and both stratifies groups. Studies have suggested that male sex is associated with risk of respiratory distress in late preterm infants [23]. In our study, male sex was associated with a longer LOS. Engle et al. reported that Cesarean sections were associated with longer LOS among late and early term infants. Our findings were consistent with theirs, in that we also found Cesarean sections to be associated with longer LOS both in the overall late preterm infant population and when late preterm infants were stratified [24]. The majority of studies attempted to analyze the relationship between LOS with birth weight and/or gestational age in order to predict LOS, and the analysis is usually conducted in one or a group of hospitals in one locality; however, there is little information on how demographic and clinical factors predict the LOS of the late preterm infant population across different US hospitals, while taking into consideration the presence of other perinatal, demographic and socioeconomic factors [13]. This information will be valuable for hospitals to determine expectations for LOS and to put policies for the late preterm infant population given their special needs. The large sample size of this study and weights that were used on the data make this study more representative of the late preterm infant population. One limitation of this study is the inability to breakdown infants into individual gestational weeks. Furthermore, the ‘‘Other’’ category of the race variable includes infants that had an unknown race. This is a limitation because these infants could potentially fit in any of the other race categories if their race were known, skewing our results. Finally, a few numbers of infants were transferred to other facilities and were not discharged home. One of the limitations associated with the use of a national de-identified dataset is the inability to follow these subjects. Although these infants could be sicker, the impact of eliminating them on the overall study results is minimal because of their limited numbers.
Conclusions Late preterm infants are at high risk of having a longer LOS across several demographic and clinical factors regardless if
J Matern Fetal Neonatal Med, Early Online: 1–7
they were born at 33–34 weeks or 35–36 weeks of gestation. These risk factors are more influential in the older age group compared to the younger age group. This research identifies the need for increased assessment and care of the late preterm infant population and can be used as a resource to guide this care. This in turn helps reduce the LOS of the infant while directly lowering healthcare costs. Future research is needed to stratify the late preterm infant population into separate weeks to gauge differences in LOS between each single week.
Declaration of interest The authors report no declarations of interest.
References 1. Hamilton BE, Martin JA, Ventura SJ. Births: preliminary data for 2012. Natl Vital Stat Rep 2013;62:1–20. 2. Raju TN, Higgins RD, Stark AR, Leveno KJ. Optimizing care and outcome for late-preterm (near-term) infants: a summary of the workshop sponsored by the National Institute of Child Health and Human Development. Pediatrics 2006;118:1207–14. 3. Mally PV, Bailey S, Hendricks-Munoz KD. Clinical issues in the management of late preterm infants. Curr Probl Pediatr Adolesc Health Care 2010;40:218–33. 4. Dong Y, Yu JL. An overview of morbidity, mortality and long-term outcome of late preterm birth. World J Pediatr 2011;7:199–204. 5. Pulver LS, Denney JM, Silver RM, Young PC. Morbidity and discharge timing of late preterm newborns. Clin Pediatr (Phila) 2010;49:1061–7. 6. Khashu M, Narayanan M, Bhargava S, Osiovich H. Perinatal outcomes associated with preterm birth at 33 to 36 weeks’ gestation: a population-based cohort study. Pediatrics 2009;123: 109–13. 7. Tomashek KM, Shapiro-Mendoza CK, Davidoff MJ, Petrini JR. Differences in mortality between late-preterm and term singleton infants in the United States, 1995–2002. J Pediatr 2007;151:450–6, 456 e1. 8. Young PC, Glasgow TS, Li X, et al. Mortality of late-preterm (near-term) newborns in Utah. Pediatrics 2007;119:e659–65. 9. Raju TN. Epidemiology of late preterm (near-term) births. Clin Perinatol 2006;33:751–63; abstract vii. 10. Tomashek KM, Shapiro-Mendoza CK, Weiss J, et al. Early discharge among late preterm and term newborns and risk of neonatal morbidity. Semin Perinatol 2006;30:61–8. 11. Korvenranta E, Linna M, Hakkinen U, et al. Differences in the length of initial hospital stay in very preterm infants. Acta Paediatr 2007;96:1416–20. 12. Schulman J. Studying determinants of length of hospital stay. J Perinatol 2006;26:243–5. 13. Bannwart Dde C, Rebello CM, Sadeck LS, et al. Prediction of length of hospital stay in neonatal units for very low birth weight infants. J Perinatol 1999;19:92–6. 14. Cotten CM, Oh W, McDonald S, et al. Prolonged hospital stay for extremely premature infants: risk factors, center differences, and the impact of mortality on selecting a best-performing center. J Perinatol 2005;25:650–5. 15. 2008 Introduction to the HICUP Nationwide Inpatient Sample (NIS). Agency for Healthcare Research and Quality Healthcare Cost and Utilization Project (HCUP). Available from: http://www.hcup-us.ahrq.gov/db/nation/nis/ NIS_2008_INTRODUCTION.pdf [last accessed 29 Jan 2014]. 16. 2013 Overview of the Nationwide Inpatient Sample (NIS). In HCUP Databases. Healthcare Cost and Utilization Project (HCUP), Agency for Healthcare Research and Quality. Available from: http://www.hcup-us.ahrq.gov/nisoverview.jsp [last accessed 29 Jan 2014]. 17. 2009 HCUP Sample Design: National Databases. Healthcare Cost and Utilization Project (HCUP), Agency for Healthcare Research and Quality. Available from: http://www.hcup-us.ahrq.gov/tech_assist/sampledesign/508_compliance/508course.htm [last accessed 29 Jan 2014].
DOI: 10.3109/14767058.2014.927428
J Matern Fetal Neonatal Med Downloaded from informahealthcare.com by University of Maastricht on 07/05/14 For personal use only.
18. 2009 HCUP Overview. Healthcare Cost and Utilization Project (HCUP), Agency for Healthcare Research and Quality. Available from: http://www.hcup-us.ahrq.gov/overview.jsp [last accessed 29 Jan 2014]. 19. 2010 HCUP Partners. Healthcare Cost and Utilization Project (HCUP), Agency for Healthcare Research and Quality. Available from: http://www.hcup-us.ahrq.gov/partners.jsp [last accessed 29 Jan 2014]. 20. HCUP Sample Design: National Databases. Healthcare Cost and Utilization Project (HCUP), Agency for Healthcare Research and Quality. Available from: http://www.hcup-us.ahrq.gov/tech_assist/ sampledesign/508_compliance/508course.htm [last accessed 29 Jan 2014].
Length of stay in late preterm infants
7
21. Shapiro-Mendoza CK, Tomashek KM, Kotelchuck M, et al. Risk factors for neonatal morbidity and mortality among ‘‘healthy,’’ late preterm newborns. Semin Perinatol 2006;30:54–60. 22. Mohamed MA, Aly H. Birth region, race and sex may affect the prevalence of congenital diaphragmatic hernia, abdominal wall and neural tube defects among US newborns. J Perinatol 2012;32: 861–8. 23. Anadkat JS, Kuzniewicz MW, Chaudhari BP, et al. Increased risk for respiratory distress among white, male, late preterm and term infants. J Perinatol 2012;32:780–5. 24. Engle WA, Kominiarek MA. Late preterm infants, early term infants, and timing of elective deliveries. Clin Perinatol 2008;35: 325–41, vi.
