 |
|
 |
|
 |
 |
|
 |
 |
 |
 |
 |
 |
 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
a Health Economics Resource Center
b Center for Health Care Evaluation
c Cooperative Studies Program, Veterans Affairs Palo Alto Health Care System, Palo Alto, California
d Department of Health Research and Policy, Stanford University, Stanford, California
 |
ABSTRACT |
|---|
 TOP ABSTRACT METHODSRESULTSDISCUSSIONREFERENCES |
|---|
METHODS. Data for this study were obtained from the linked 2000 California birth cohort data. These data (n = 518704), provided by the California Office of Statewide Health Planning and Development (OSHPD), contain infant vital statistics data (birth and death certificate data) linked to infant and maternal hospital discharge summaries. In addition to the infant and maternal hospital discharge summaries associated with delivery, these data include discharge summaries for all infant hospital-to-hospital transfers and maternal prenatal hospitalizations. The linkage algorithm that is used by OSHPD in creating the linked cohort data file is highly accurate. More than 99% of the maternal and infant discharge abstracts were linked successfully with the birth certificates. These data were also linked successfully with the infant discharge abstracts from the receiving hospital for 99% of the infants who were transferred to another hospital. The hospital discharge records were the source of the hospital charges and length-of-stay information summarized in this study. Hospital costs were estimated by adjusting charges by hospital-specific ratios of costs to charges obtained from the OSHPD Hospital Financial Reporting data. Costs, lengths of stay, and mortality were summarized by birth weight groups, gestational age, cost categories, and types of admissions.
RESULTS. Low birth weight (LBW) and very low birth weight (VLBW) infants had significantly longer hospital stays and accounted for a significantly higher proportion of total hospital costs. The average hospital stay for LBW infants ranged from 6.2 to 68.1 days, whereas the average hospital stay for infants who weighed >2500 g at birth was 2.3 days. Overall, VLBW infants accounted for 0.9% of cases but 35.7% of costs, whereas LBW infants accounted for 5.9% of cases but 56.6% of total hospital costs. Although total maternal and infant costs were similar (
$1.6 billion), the distribution of maternal costs was much less skewed. For infants, 5% of infants accounted for 76% of total infant hospital costs. Conversely, the most expensive 3% of deliveries accounted for only 17% of total maternal costs.
CONCLUSIONS. The very smallest infants make up a hugely disproportionate share of costs; more than half of all neonatal costs are incurred by LBW or premature infants. Maternal costs are similar in magnitude to newborn costs, but they are much less skewed than for infants. Preventing premature deliveries could yield very large cost savings, in addition to saving lives.
Key Words: neonatal care • NICU • health care costs
Abbreviations: LBW—low birth weight • VLBW—very low birth weight • OSHPD—Office of Statewide Health Planning and Development • LOS—length of stay
Medical and technologic advances in the care of infants with low birth weight (LBW; <2500 g) and very low birth weight (VLBW; <1500 g) have resulted in substantial increases in survival rates1 but the NICUs and technologies that are required to achieve these improved survival rates have produced extraordinary medical costs.2, 3 Previous research has shown that the costs associated with premature births that result in LBW infants are extremely difficult to quantify.4 Although there is a small but growing literature on the economic costs associated with LBW infants, little research has quantified the maternal costs connected to pregnancy complications and hospitalizations that result in these high-risk births.5
Because of the relative frequency of delivery and birth, compared with other hospital discharges, they make up a very large share of all hospital discharges; in California in 2000, delivery/birth-related hospitalizations accounted for 32% of all hospital discharges. The renewed growth of health care costs in recent years6 and the growing reports of the adverse effects of employer health insurance costs on industrial competitiveness and labor markets7, 8 are causing increased focus on the costs of health care. Because of their sheer volume, it is almost certain that birth-related hospitalizations will be the focus of cost containment efforts. To provide information for these and other policy decisions, this article describes the current costs of newborn care by using population-based linked data for both mothers and infants and quantifies maternal hospital expenditures associated with delivery. It highlights the costs for the delivery of LBW infants and presents the maternal costs that result from prenatal and delivery hospitalizations.
|
METHODS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
The linkage algorithm that is used by OSHPD in creating the linked cohort data file is highly accurate.9 More than 99% of the maternal and infant discharge abstracts were linked successfully with the infant birth certificates (n = 518704). These data were also linked successfully to the infant's discharge abstract from the receiving hospital for 99% of the infants who were transferred to another hospital. The hospital discharge records were the source of the hospital charges and length of stay (LOS) information summarized in this study.
In addition to the information gained from linking hospitalizations, the data linkage provide complementary data. The hospital discharge abstracts are the source of information on hospital charges, LOS, and International Classification of Diseases, Ninth Revision diagnosis codes. Reabstracting studies have found that all of these data elements are coded reliably in the OSHPD discharge data.10 Using the birth certificate allows the very accurate data on birth weight, and the linkage to the death certificate yields accurate information on survival. The birth certificate is also the source of information on gestational age. Although the quality of the gestational-age data from the birth certificate is not as high as the birth weight data, birth weight can be used as a gross check on the reported gestational age. We used the criteria of Kotelchuck11 to screen out implausible gestational-age values.
Computation of LOS and Hospital Costs
LOS
The total LOS (total inpatient hospital days) was computed for both mothers and infants by summing the hospital days associated with all relevant hospitalizations. For mothers, this included any prenatal hospitalizations and the delivery hospitalization. For infants, total LOS was computed as the total number of hospital days until first discharge to home or death. This total incorporated any interhospital transfers that may have occurred.
In some instances, LOS information was missing from the hospital discharge summary. Cases that did not include complete hospital stay information were excluded from the analyses of hospital stays. For mothers, a total of 515372 delivery hospitalizations (of 518704 total linked births) with complete LOS information were identified. A total of 3332 deliveries had missing maternal LOS information. The remaining difference between maternal delivery hospitalizations and total births was attributable to multiple births. For infants, a total of 518697 births (of 518704 total linked births) with complete LOS information were identified.
Hospital Costs
Most hospital charges represent a significant markup over actual costs, and these markups vary greatly across hospitals.12 To provide a more accurate view of the actual costs, we multiplied the charges for each hospital stay by a hospital-specific cost-to-charge ratio derived from annual hospital financial data compiled by OSHPD.13 Ideally, this adjustment of charges to estimated costs would be done using department-specific cost-to-charge ratios as this yields more accurate estimates of costs.12 Unfortunately, the OSHPD discharge data report only total hospital charges, not department-specific charges. Once charges were converted to costs, costs were adjusted by the consumer price index to reflect December 2003 levels.14
Two sets of cost variables were computed for both mothers and infants who were included in this study. These included total adjusted hospital costs and cost per inpatient hospital day. For mothers, total adjusted hospital costs were computed as the sum of adjusted inpatient hospital costs for all prenatal hospitalizations and the delivery hospitalization. For infants, total adjusted hospital costs were computed as the sum of adjusted inpatient hospital costs for the delivery hospitalization and any subsequent hospitalizations (transfers) before the infant was discharged from the hospital for the first time or before death if the infant died before being discharged.
In some instances, total adjusted costs could not be computed accurately because of missing data. Some hospitals (particularly Kaiser hospitals) do not regularly report hospital charges in the OSHPD discharge summaries. For this reason, these cases and cases involving multiple hospitalizations (eg, maternal prenatal hospitalizations, interhospital transfers) that did not include complete hospital charge data for each relevant hospitalization were excluded from our analyses. For maternal and infant cost analyses, a total of 437514 deliveries (of 518704 total linked births) were identified as having complete cost data. Once total adjusted costs were computed for mothers and infants, costs per hospital day were computed by dividing total adjusted costs by total LOS.
Cost Outliers
Through an examination of the distributions of maternal and infant adjusted costs, several cases with outlying/improbable adjusted cost-per-day values were identified and excluded from our analyses. We excluded any case (infant or maternal) with an adjusted cost per day in excess of $10000 (2498 mothers and 2530 infants). We also excluded maternal cases for which adjusted cost per day was less than $300 (1070 cases) and infant cases (survivors) with an adjusted cost per day of less than $100 (2975 cases). Surviving infants with a birth weight <1500 g were excluded when adjusted costs per day were less than $400 (94 cases). Surviving infants with birth weights between 1500 and 1999 g were excluded when adjusted costs per day were less than $250 (183 cases). Infants who died and had a total hospital stay that exceeded 1 day were excluded when adjusted costs per day were less than $400 (68 cases).
Because we were interested in presenting cost analyses that were complete for both mothers and infants, if either the mother or infant was dropped as a cost outlier, then both mother and infant were excluded from the cost analyses. However, these possible outlying cases were included in the data with the LOS results. All data management and statistical analyses were conducted using the SAS Statistical Analysis System software.15
|
RESULTS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
|
Table 2 also presents infant costs by gestational-age category. Infants who were born at 
32 completed weeks (8282 cases) represented only 1.9% of cases but had mean hospital costs of $66813 and mean daily costs of $1535. Infants who were born at 33 to 36 completed weeks represented 8.0% of cases with mean hospital costs of $7081 and mean daily costs of $702. Although infants who were born at 32 completed weeks of gestation comprised only 1.9% of cases, their total combined costs ($553345754) represent more than one third (35.5%) of total infant hospital costs ($1560748560).
Table 3 summarizes the distribution of total costs and number of cases by cost category for mothers and infants and provides additional information on the skewed nature of the distribution of costs. Although 76.6% of all infants born (335269 cases) fell into the lowest cost category of $1 to $999, they accounted for only 11.1% of total costs. The cost category of $50000 or more constituted 1.3% of all infants born (5696 cases), but these infants accounted for 54.6% of the total costs. Overall, there was a broad and distorted distribution of total costs of care for infants, with the majority of infants falling into the lowest cost categories. The few who did fall into the higher categories produced staggering total costs. For mothers, the distribution across cost categories was just as broad, although the majority of mothers were concentrated into slightly higher cost categories. There were also many fewer maternal cases in the 2 highest cost categories (more than $30000). More than half (53.8%) of the maternal cases (235333 mothers) were in the $1000 to $2999 cost category, and another 27.6% of cases (120660 mothers) fell into the $3000 to $4999 cost category. In contrast to the infants, 76.6% of whom were in the very lowest cost category of $1 to $999, only 2.1% of mothers were in this low cost group. An interesting difference between mothers and infants in the distribution of costs is that 60% of mothers make up 60% of maternal costs, whereas 5% of the infants generate 75% of the costs.
|
|
In addition to the high costs associated with premature delivery, the distributional data on newborn costs and LOS reported in Tables 1 and 2 show that whereas most large infants have short hospital stays and low costs, some of these infants do require extensive care. Phibbs et al17 previously reported that congenital anomalies were a major cause of high costs for larger infants. To provide some information on the costs associated with congenital anomalies, in results not shown, we used the International Classification of Diseases, Ninth Revision codes to identify the infants who had major anomalies that required neonatal treatment. Only 2.1% of infants had a major congenital anomaly, but these infants incurred 25.5% of neonatal hospital costs.
|
DISCUSSION |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
$12000 compared with average hospital costs of nearly $119000 for infants who weighed 1000 to 1249 g at birth.
Although maternal hospital costs associated with LBW infants were substantial ($200000000), they were overwhelmed by the hospital costs for the infants, which approached $700000000 in California in 2000. Other findings that were consistent with previous research were longer hospital stays for LBW infants, with markedly longer stays for VLBW infants.
There are a variety of reasons for the very high hospital costs associated with LBW and VLBW infants as compared with infants who weighed >2500 g at birth. As previous studies have reported, a significant proportion of the infant costs result from an increasing use of advanced medical technologies and complex medical and surgical procedures. The current data support these findings, with higher costs and longer hospital stays for LBW infants and infants who are born at <32 weeks' gestation.
Increased maternal costs associated with LBW and VLBW births result from more and longer hospital stays and an increased use of obstetric medical technologies and services aimed at preventing premature and low-weight births. These obstetric interventions may include tocolytic therapy and the use of prenatal corticosteroids.18 We have also provided a population-based description of the costs of maternal prenatal hospitalizations. In round terms, there is 1 maternal hospital admission that does not result in a delivery for every 11 deliveries, and these admissions account for 8.5% of total maternal hospital costs.
As reported by Franks et al,19 most hospitalizations for complications of pregnancy are the result of preterm labor. As the authors pointed out, a variety of maternal factors are related to the frequency, length, and costs of hospital stays associated with complications of pregnancy. According to Franks et al, black women had a much higher rate of hospitalization than white women and had longer stays on average than white women. Other sociodemographic factors that influenced the number and the length of antenatal hospitalizations included maternal age, marital status, and insurance status. It is also well documented that other maternal behavioral and social factors influence complications of pregnancy and subsequent maternal and infant hospital costs. These factors may include inadequate nutrition, smoking, alcohol consumption, or inadequate weight gain.19
Although the focus of this article is on the costs associated with premature delivery, we also note that congenital anomalies are another significant cause of neonatal costs, and most of these infants are term deliveries. We note that in a simple classification of the cases with major congenital anomalies, they represent a similar disproportionate share of neonatal hospital costs. These results are consistent with findings by Phibbs et al,17 who quantified and modeled infant hospital costs by 7 risk factor groups. Additional research is needed to describe more fully the costs of infants with major anomalies. We reiterate that premature delivery is not the only cause of high hospital costs for newborns.
The current data offer a significant advance toward quantifying maternal and infant hospital costs. Documentation of these costs has historically been very difficult because of limitations with available data. The current study is unique in its use of linked maternal and infant hospital discharge records. The use of linked data allows quantification and analysis of the relationships and costs associated with maternal hospitalizations and treatment with subsequent birth outcomes and infant hospital costs. The ability to analyze data that are collected across multiple hospitalizations for mothers and their infants yields a more accurate characterization of patients, medical treatments, costs, and outcomes.
Although the current study makes a significant contribution to the documentation of hospital costs associated with pregnancy and birth, there are several limitations to this report. First, as mentioned previously, some hospitals in California (primarily Kaiser hospitals) do not report charges in the hospital discharge abstracts. As a result, all maternal and infant hospitalizations that occur at these hospitals were excluded from cost analyses. Second, it is difficult to derive accurately hospital costs from hospital charges, because hospital charges reflect different markups for different services. The method of converting charges to costs by applying a hospital-level cost-to-charge ratio used here is limited in its accuracy but is a standard and accepted method for converting hospital charges to costs.20 Although these clearly are limitations, we do not believe that these limitations have a meaningful affect on the results or conclusions. The cost results are driven by relatively small numbers of very expensive cases and large numbers of low-cost cases. Even relatively large differences in the cost estimates for individual cases will not change this dynamic, especially given that any hospital-specific corrections would affect the high- and low-cost cases equally. Given that the cases with missing cost estimates have similar distributions of LOS, it is unlikely that the cases with missing cost data have patterns of care that are dramatically different from those that remained in the sample.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
Address correspondence to Ciaran S. Phibbs, PhD, Health Economics Resource Center (152), Veterans Affairs Medical Center, 795 Willow Rd, Menlo Park, CA 94025. E-mail: cphibbs{at}stanford.edu
Ms Sneed is now a student at the University of Texas Southwestern Medical School.
|
REFERENCES |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
 |
||
|
||
Home | My Pediatrics | Journal Information | Current Issue | Past Issues | Subscriptions & Services | Contact Us Click here for faster international access © 2006 American Academy of Pediatrics. All rights reserved. |
||
|
||
