Ninth Annual Maternal and Child Health Epidemiology Conference / December 10-12, 2003
REBECCA RUSSELL: Good morning. Going to switch focus just for a little minute. I think we heard a very nice presentation from Melissa on what is the picture of pre-term birth locally in a state, and I’m going to focus on a national picture, focusing on the economic consequences of pre-term birth. And I’d like to begin by acknowledging my co-authors from the Perinatal Data Center at the March of Dimes, also from the office of the medical director, and our office of government affairs. Here’s a little background. Since 1981, the rate of pre-term birth in the United States, which is defined as less than 37 completed weeks of gestation, has risen 28% to 12% of live births in 2002 according to preliminary data from the National Center for Health Statistics. This is the highest level reported since NCHS started collecting data on gestational age.
Pre-maturity combined with low birth weight is the leading cause of death in the first month of life, and the second leading cause of death during the first year of life, second only to birth defects. It accounts for nearly 16% of all infant deaths and is the leading cause of infant death among African/American infants. In addition to having a higher risk of mortality, premature infants face a much higher risk of health problems compared to other newborns, and often require care in a neonatal intensive care unit. In a response to the increasing rates of pre-term birth, the March of Dimes launched a five year national pre-maturity campaign in January of this year and hopefully, most of you were at the Plenary session yesterday where our medical director, Nancy Green, spoke specifically on the campaign, so I’m not going to talk in any detail about that right now. But as part of this campaign, it has been important to try to understand the costs that are related to pre-maturity. Previous studies have estimated the cost associated with pre-maturity to be substantial. One study estimated medical costs of pre-term birth to be nearly five billion dollars each year, while others have shown that delivery costs for pre-term births increase with decreasing gestational age.
Others have estimated that health care costs for low birth weight infants can be ten times as high as for a baby born at a normal weight. Hospital charges for infants specifically related to pre-maturity are an important indicator of the costs associated with pre-term birth in the U.S. So, the purpose of this study was to determine the magnitude of hospital charges for infant stays related to pre-maturity. We conducted a retrospective analysis of the 2001 nation wide inpatient sample data. The NIS is part of the Health Care Cost and Utilization Project or HCCUP, which is sponsored by the Agency for Health Care Research and Quality. It approximates a 20% sample of community hospitals, which included 986 hospitals from 33 states in 2001.
A community hospital is defined by the American Hospital Association as a non-federal short term general and other specialty hospital, so this can include teaching, academic and children’s hospitals and it would exclude long term hospitals and hospital units of institutions. The data are weighted to provide national estimates for approximately 35 million hospital stays in 2001. This makes it a very robust data set and the largest all-payer data set publicly available. Each hospital stay in the file lists a principal diagnosis identified on the medical record as the primary reason for that stay, and also has diagnosis for up to 14 secondary conditions. These diagnosis are coded using the *(inaudible) vision of the International Classification of Disease. In the NIS, each ICD-9 code is then categorized using the clinical classification software, or CCS codes. This classification system was created by ARC and classifies the ICD-9 codes into mutually exclusive clinically homogenous categories.
We identified infant stays by an age of admission less than one year, so this means that our estimates of charges and discharges can include readmissions into the hospital after the initial stay for the delivery. We identified stays due to pre-maturity and low birth weight as those with a CCS code of 219, which is defined as pre-maturity, low birth weight, and fetal growth restriction. We included those stays with both a principal diagnosis of pre-maturity and those with any secondary diagnosis of pre-maturity, and it was important to include both of these categories as many of the infant stays with secondary diagnosis of pre-maturity has, as their principal diagnosis, live birth so their delivery was the reason they were in the hospital to begin with. So, excluding those with a secondary diagnosis would have resulted in a large underestimate of the discharges and charges due to pre-maturity. Using the ICD-9 codes, we were able to then further break down the broader category of prematurity/low birth weight in a more specific category, so we dichotomized those as stays with codes for slow growth and fetal malnutrition, which had the ICD-9 codes listed on the slide, and those with codes for short gestation and unspecified low birth weight.
We were also able to further breakdown the short gestation and unspecified low birth weight, but I’ll get into that a little bit later. In order to account for stays that were missing data on charges, which was a very small amount, less than 3% of stays with any diagnosis of pre-maturity. In order to account for those, we assigned those stays the mean charge for stays in that group. This first slide shows an over view of all infant hospital stays in the United States. In 2001, there were a total for 4.6 million hospital stays for infants in the United States, which is shown by the bar on the left. This also includes healthy newborns, so the 4.6 million seems to be a pretty accurate number. Of those 4.6 million stays, 384,000 or 8%, which is represented by the pink section of the bar had either a principal or a secondary diagnosis of pre-maturity. These 384,000 stays, however, had a very disproportionate influence on the charges.
The bar on the right shows the distribution of charges for all infants, with the section in pink representing the charges for those 384,000 infant stays with any diagnosis of pre-maturity. In 2001, charges for stays due to pre-maturity totaled 13.6 billion dollars, nearly half of the 25 billion for all stays including healthy newborns-- I’m sorry, half of the 29 billion. The first bar on the left of this slide represents the 384,000 infant stays with any diagnosis of pre-maturity. Pre-maturity was identified as being the primary reason for the hospitalization, and 260,000 of those 384,000 stays. When we looked at discharges by the two categories of ICD-9 codes that I mentioned earlier-- and those are the peach colored bars to the right of the slide-- 59,000 stays had one of the ICD-9 codes for slow growth or fetal malnutrition. A large majority of the stays, more than 88%, had one of the ICD-9 codes for short gestation or unspecified low birth weight, which accounted for 335,000 stays.
It is possible for a stay to have one ICD-9 code that falls into the category of slow growth and another in short gestation, which is why the two groups total more than the 384,000 total stays due to pre-maturity. So this slide shows total adjusted charges for infant stays due to pre-maturity/low birth weight, and just to remind you, included in those stays are the missing charges data that were assigned the mean charge. In 2001, inpatient hospital charges for the 384,000 infant stays as I showed earlier totaled 13.6 billion dollars. Infant stays with a principal diagnosis of pre-maturity accounted for 1.9 billion of that 13.6 billion. The majority of charges could be attributed to stays with one of the ICD-9 codes for short gestation or unspecified low birth weight, which totaled 13.1 billion dollars compared to 1 billion first stays with any diagnosis of slow growth or fetal malnutrition. This slide further breaks down the category of short gestation and unspecified low birth weight. This is a large group of stays encompassing a wide range of severity and conditions, so we broke down the ICD-9 codes into those for extreme immaturity, which are the ICD-9 codes 765.00-765.09, which are the ICD-9 codes that usually imply a birth weight less than 1,000 grams or gestational age less than 28 completed weeks.
And then our other group was other pre-term infants, which usually implies a birth weight of 1,000 to 2,499 grams, or a gestational age of 28 to 36 completed weeks. When we do this, we see that 29,000 infant stays had any diagnosis of extreme immaturity with a mean charge of $155,000 and charges totaling 4.5 billion, approximately 1/3 of the 13.1 billion total charges for infant stays in the larger category of short gestation. Charges for the 305,000 infants stays with any ICD-9 code and the other pre-term infants category averaged $28,000, and totaled $8.7 billion dollars. This slide shows a comparison of stays due to pre-maturity, both those with any diagnosis, and those with a principal diagnosis, to those for an uncomplicated newborn. We defined uncomplicated newborn as an infant stay with a principal diagnosis of live born, and no secondary diagnosis of disease. The CCS code for live born, surprisingly, also includes stillbirths and we removed those from the analysis. In 2001, there were approximately 1.9 million stays matching this definition of an uncomplicated newborn.
When we examined mean charges infant stays for any diagnosis of pre-maturity averaged $35,000 and $75,000 for the subset of stays with a principal diagnosis of pre-maturity. This is compared to an average charge of $1,300 for an uncomplicated newborn. When we compared length of stay in 2001, the mean length of stay for infants with any diagnosis of pre-maturity was 12.9 days, and 24.7 days for those with a principal diagnosis of pre-maturity, and this is compared to 1.9 days for an uncomplicated newborn. These analyses represent only charges for infant stays and do not include the maternal charges. Had we included charges for the mothers, our estimates of the charges due to pre-maturity and the cost for pre-maturity hospitalizations could increase substantially. Unfortunately, using the NIS we were unable to identify or link those stays for the mother with those for the infants.
Another limitation is that these charges are not reflective of actual payments or what the actual cost for the stay was, only what was billed by the hospital for the stay. Also, charges in the NIS are only for acute care hospital stay and do not include other charges associated with other costly aspects of pre-maturity which I’ll mention in a minute. In conclusion, charges for infant stays due to pre-maturity accounted for nearly half of the charges for all infant stay in 2001, while accounting for less than 10% of the total discharges. Second, the majority of hospital charges for stays due to pre-maturity can be attributed to disorders relating to short gestation and low birth weight, while charges for slow growth and fetal malnutrition contributed a relatively small amount. Finally, mean charges for the most severe stays defined as having pre-maturity as the principal diagnosis were nearly 60 times higher than the mean charges for uncomplicated newborns, and their length of stay was 13 times longer. While these data are helpful they’re just the first step in understanding the magnitude and scope of the cost associated with pre-maturity in the United States. Analysis of the contribution of other conditions, such as respiratory distress syndrome or congenital anomalies along with the contributions of certain procedures, such as mechanical ventilation to the overall charges would help provide a more complete picture of charges associated with the hospitalization of a premature infant.
Additional data are needed to represent the full range of costs associated with pre-maturity and low birth weight. These would include costs for the physician and other medical professionals, rehabilitation expenses, and the cost associated with follow-up care or homecare for these infants. And finally, while these data provide a nice overall picture of the hospital bill for pre-maturity in the nation as a whole, we have heard that state and local organizations are very involved in this topic. Having state and local level estimates of pre-maturity related costs could help create support for these local activities aimed at reducing pre-maturity. For more information on pre-maturity and the campaign, you can visit the March of Dimes website or for pre-maturity related statistics including state and county level pre-term birth rates. I’d just like to mention the *(inaudible) stats website if you all aren’t familiar with it.