National Estimates of Hospital Use by Children With HIV Infection in the United States: Analysis of Data From the 2000 KIDS Inpatient Databa
http://www.100md.com
《中华首席医学网》
ABSTRACT
OBJECTIVES. The purpose of this research was to describe hospital use patterns of HIV-infected children in the United States.
STUDY DESIGN. We analyzed a nationwide, stratified probability sample of 2.5 million hospital discharges of children and adolescents during the year 2000, weighted to 7.3 million discharges nationally. We excluded discharges after hospitalizations related to pregnancy/childbirth and their complications and discharges of neonates <1 month of age and of patients >18 years of age. Diagnoses were identified through the use of the Clinical Classification Software with grouping of related diagnoses.
RESULTS. We estimated that there were 4107 hospitalizations of HIV-infected children in 2000 and that these hospitalizations accounted for $100 million in hospital charges and >30000 hospital days. Infections, including sepsis and pneumonia, were among the most frequent diagnoses in such hospitalizations, followed by diagnoses related to gastrointestinal conditions, nutritional deficiencies and anemia, fluid/electrolyte disorders, central nervous system disorders, cardiovascular disorders, and respiratory illnesses. Compared with hospitalizations of non-HIV-infected children, hospitalizations of HIV-infected ones were more likely to be in urban areas, in pediatric/teaching hospitals, and in the Northeast, and the expected payer was more likely to be Medicaid (77.6% vs 37.2%). Compared with children without HIV, those with HIV tended to be older (median age: 9.5 years vs 5.2 years), to have been hospitalized longer (mean: 7.8 days vs 3.9 days), and to have incurred higher hospital costs (mean: $23221 vs $11215); HIV-associated hospitalizations ended in the patient's death more frequently than non-HIV ones (1.8% vs 0.4%), and complications of medical care were also more common (10.8% vs 6.2%).
CONCLUSIONS. Infections account for the majority of hospitalizations of HIV-infected children in the United States, although nutritional deficiencies, anemia and other hematologic disorders, gastrointestinal and renal disorders, and complications of medical care are also more common among hospitalized children with HIV than among those without HIV.
Key Words: HIV ? children ? hospitalizations ? United States ? HCUP ? diagnosis
Abbreviations: HAART—highly active antiretroviral therapy ? KID—KIDS Inpatient Database ? HCUP—Healthcare Cost and Utilization Project ? DRG—diagnosis-related group ? CCS—Clinical Classification Software
There is little published information on the impact of pediatric HIV/AIDS on hospitalizations and costs or on the effects of highly active antiretroviral therapy (HAART) on these rates and costs. Data from adult studies have indicated that the number of HIV-related hospitalizations and the average length of hospital stay among HIV patients initially declined after the advent of HAART1–5 but that these trends have since leveled off.6,7 The results of 1 adult study also suggested that racial/ethnic disparities in HIV-related hospitalizations have persisted since the advent of HAART.8 However, few national estimates of the effect of pediatric HIV on the use of inpatient hospital services have been published, although such estimates are vital in this era of rapidly changing health care practices.
Our purpose in this study was to describe the use of inpatient hospital services by children with HIV infection in the United States. Specifically, we sought to describe the burden of pediatric HIV hospitalizations; the most common diagnoses among child HIV patients; the distribution of pediatric HIV hospitalizations by age group; and the distribution of such hospitalizations by expected payer/insurance provider, hospital type (teaching versus nonteaching and pediatric versus nonpediatric), and geographic location. We also estimated the average charges associated with these hospitalizations, estimated the average length of hospital stay, and compared the diagnostic codes most frequently assigned to HIV-infected children and adolescents (<19 years of age) with those most frequently assigned to HIV-uninfected ones.
METHODS
Description of the Databases
We used the 2000 KIDS Inpatient Database (KID) of the Healthcare Cost and Utilization Project (HCUP).8 HCUP is a partnership between state medical data collecting organizations and the Agency for Healthcare Research and Quality in which the state partners contribute their statewide hospital discharge data to HCUP. Data elements that states have in common are recorded into a uniform coding scheme, and data elements unique to individual states are retained if they are useful for research purposes. KID, part of the HCUP databases, is a nationwide sample of pediatric discharges from nonrehabilitation community hospitals weighted to reflect all pediatric discharges from US community hospitals open during any part of a particular calendar year. An in-depth description of the HCUP KID 2000 is available.9
Briefly, the 2000 KID contains information on 2.5 million discharges from 2784 hospitals. These discharges were stratified by reason for hospital stay (uncomplicated birth, complicated birth, and not birth related) and were also sorted by state, hospital, diagnosis-related group (DRG), and a random number within each DRG. Systematic random sampling was used to select from each sampling frame 10% of discharges after hospitalization related to an uncomplicated birth and 80% of discharges after hospitalization for other reasons.
To obtain national estimates, HCUP created discharge weights using American Hospital Association survey data. Six hospital characteristics were used to poststratify the hospitals: geographic region, hospital control, urban or rural location, teaching status, bed size, and hospital type. The KID database includes demographic information on discharged patients; diagnosis codes; procedure codes; and information about hospital type and geographic region, hospital charges, and length of hospital stay. In 2000, 27 states contributed data to KID, and the pediatric patients in these states accounted for 72.6% of all US hospital discharges of children in 2000.
The hospital universe is defined as all US hospitals open during any part of a particular calendar year and designated as community hospitals in the American Hospital Association Annual Survey. Charge data are based on the charges for the hospitalization only and not on physician fees. Charges do not necessarily reflect costs or reimbursements actually received by hospitals. Other studies have demonstrated that HCUP is a powerful tool for estimating the charges of hospital resource use.8,10
Study Design and Statistical Analysis
The diagnoses recorded on hospital discharge abstracts are coded in accordance with the International Classification of Diseases, Ninth Revision, Clinical Modification.11 Because >12000 diagnosis codes are in the International Classification of Diseases, Ninth Revision, Clinical Modification, we combined discharge diagnoses into a relatively small number of DRGs or "major diagnostic categories" using Clinical Classification Software (CCS)12,13 to aggregate illnesses and conditions into 259 mutually exclusive categories, most of which are clinically homogeneous. For this report, we further aggregated CCS categories into the following clinically relevant groupings: malignancies: CCS 11–45; cardiovascular disorders: 96–121; accidents/injuries: all E codes except E2616 and E2617 plus 225–236 and 239–244; complications of medical care: 237, 238, E2616, and E2617; mental/psychiatric disorders: 65 and 68–75; alcohol and drug abuse: 66 and 67; central nervous system/eye disorders: 79–91 and 93–95; gastrointestinal disorders: 137–155; renal disorders: 156–158, 160–163, and 165–166; sexually transmitted diseases: 9 and 168; congenital anomalies: 213–217; shock: 249; nutritional deficiencies and anemia: 52 and 59; fluid/electrolyte disorders: 55; sickle cell disease: 61; social admissions: 255; asthma: 128; other respiratory disorders (except infections and asthma): 56, 127, and 130–134; endocrine/metabolic disorders: 48–51, 53, and 58; other hematologic disorders: 57, 60, and 62–64; tuberculosis: 1; bacterial infections/sepsis: 2, 3, 76, 92, 124, 125, 135, 159, 197, 201, 247, and 248; mycoses: 4; viral/other infections: 6–8, 77, 78, 123, 126, and 134; pneumonia: 122 and 129; fever of unknown origin: 246; and allergic reactions: 253.
We analyzed data by age group only for those states that disclose the age of the child. Texas does not release age information for sensitive groups (including HIV-infected patients). Pennsylvania recorded age for sensitive groups (including HIV-infected patients) to the midpoint of 5-year age range.
Costs were calculated using the total charges that were reported for each hospitalization. We excluded neonate delivery hospitalizations (variable neomat: 2,3) and hospitalizations for conditions related to pregnancy and delivery (CCS codes 177-196). HIV-infection was identified by the CCS code of 5.
Standard methods for analyzing weighted survey data were employed using 2 statistical software packages: SAS (SAS Institute, Inc, Cary, NC) and SUDAAN (Research Triangle Institute, Research Triangle Park, NC) statistical. A 2 test was used to compare categorical variables and t tests to analyze means. Programming and data results were confirmed by 2 independent researchers.
RESULTS
We estimated that there were 4107 pediatric discharges associated with an HIV diagnosis in the United States in 2000, that the discharged patients spent a total of 31940 days in the hospital at a cost of $99.8 million (Table 1), and that children with HIV accounted for 0.2% of all hospital discharges, 0.4% of all hospital charges, and 0.7% of all hospital days among children hospitalized in the United States for reasons other than normal birth or conditions associated with pregnancy. Compared with non-HIV-infected children, those with HIV had longer mean hospital stays (7.8 days vs 3.9 days; P < .001), higher mean hospital charges per hospitalization, and a higher mean number of diagnoses per hospitalization (5.10 vs 3.03; P < .01).
The median age of discharged children with HIV was 9.5 years, compared with 5.2 years for those without an HIV code. Infants (<1 year of age) accounted for only 5% of all HIV-related pediatric hospitalizations but for >20% of all non-HIV related pediatric hospitalizations, whereas children aged 5 to 14 years accounted for more than half of all HIV-related pediatric hospitalizations but for only about one third of all non-HIV-related hospitalizations (Table 1). Medicaid was the expected payer for 77.6% of HIV-infected pediatric patients, and private insurance was the expected payer for 15.7% (vs 37.2% Medicaid and 54.1% private insurance for children without HIV). We also found that hospitalizations of children with HIV were more likely than those without HIV to be in an urban teaching hospital (84.2% vs 59.5%), less likely to be in a hospital or hospital unit not specifically for children (37.1% vs 62.3%), more likely to be from the Northeast or South (87.8% vs 58.2%), and more likely to lead to death during the hospitalization (1.7% vs 0.4%; P < .001).
The most common discharge diagnoses for hospitalizations of HIV-infected children were, in order of frequency, bacterial infections and sepsis, viral/and other (including parasitic) infections, gastrointestinal diagnoses, pneumonia, nutritional deficiencies and anemia, fluid/electrolyte disorders, fungal infections, cardiovascular conditions, and respiratory illnesses (Table 2). Categories of conditions more common among pediatric patients with HIV than among those without HIV included all infections (tuberculosis and bacterial, viral, and fungal infections), cardiovascular disorders (mostly cardiomyopathies or hypertension), central nervous system diseases (mostly seizures or encephalopathy), gastrointestinal conditions (mainly gastrostomy or diarrhea), renal conditions (mainly chronic renal failure or nephritis), endocrine and metabolic disorders (mostly lack of normal development), sexually transmitted diseases, nutritional deficiencies, anemia and other hematologic disorders, mental/psychiatric disorders, and complications of medical care (mainly infections and inflammation associated with the use of vascular devices; Fig 1). Categories of conditions more common among hospitalizations of uninfected children included congenital anomalies, accidents/injuries, asthma, malignancies, alcohol and drug abuse, and sickle cell disease.
In 53% of hospitalizations of HIV-infected children in 2000, there was 1 procedure performed compared with 42.9% of hospitalizations of uninfected children; hospitalizations of children with HIV averaged 1.33 procedures, whereas those without HIV averaged 0.89 (P = .03). Antibiotic injections, immune globulin administration, esophagogastro-duodenoscopy, gastrostomy, and enteral and parenteral feeding were more common among children with HIV-related hospitalizations, whereas appendectomies, cancer chemotherapy, mechanical ventilation, and computerized tomography of the head were more common among uninfected children (Table 3).
DISCUSSION
The Centers for Disease Control and Prevention reported that 5400 US children <19 years of age were living with AIDS in 2000.14 Although the annual number of infants infected with HIV perinatally has been decreasing, the annual number of adolescents infected with HIV has been increasing.14 This increase in the rate of HIV infections among adolescents is particularly pronounced among blacks, Hispanics, and girls of all racial/ethnic groups.14,15 Despite the increase of HIV infections among adolescents, the number of AIDS cases and HIV-related deaths have been declining, largely because of the use of combination drug regimens, such as HAART.1,3,14
Although comprehensive data on HIV-related morbidity among US pediatric patients since the advent of HAART are currently not available, the results of some limited studies suggest that HAART has been effective in reducing morbidity among children with HIV. For example, a retrospective analysis of 525 charts from HIV-infected children participating in the Ryan White program indicated that as use of combination therapy increased, the rate of opportunistic infections and hospitalizations decreased.16 In addition, results of a recent study of 129 children followed at 3 HIV clinics in California showed that the rate of HIV-related deaths, as well as the hospitalization rate and average length of hospital stay among perinatally infected children, all decreased significantly between 1994 and 2001,16 a period during which the use of HAART was increasing. They also showed that pneumonia and sepsis were the main causes of hospitalization among these children from 1994 through 1999 but that the rate of hospitalizations for these conditions decreased in 2000 to 200117 after the introduction of complex pharmacologic regimens in the care of these children. The problem of psychiatric disease among HIV-infected children was also highlighted recently when results from the Pediatric AIDS Collaborative Trials Study cohort of 1808 HIV-infected children under 15 years of age followed over a period of 3 years showed them to have a greater risk for psychiatric hospitalization than noninfected children of the same age.18
To our knowledge, our study is the first to provide national estimates of hospital use by US children with HIV, as well as the first to describe the patterns of pediatric HIV-related morbidity in the era of HAART. We showed that the distribution of conditions for which HIV-infected children were hospitalized and the type of hospital to which they were admitted (mainly urban, pediatric teaching hospitals in the Northeast and South) were different from those of children without HIV infection. Hospitalizations of HIV-infected children were, on average, longer, more costly, and more complex, as reflected in a higher number of concomitant diagnoses and higher rates of medical complication and death. Because Medicaid is the expected payer for the majority of pediatric HIV patients, their hospital care poses an incremental burden on the national health care budget. Infections; hematologic disorders (mainly anemia); nutritional disorders; gastrointestinal, renal, and cardiovascular conditions; and medical complications/iatrogenic conditions are more common in hospitalizations of HIV-infected children compared with non-HIV-infected ones. In comparison, hospitalizations resulting from injuries, appendicitis, malignancies, congenital anomalies, and alcohol and substance abuse assume a lesser proportion in HIV-infected compared with HIV-uninfected children. Using Centers for Disease Control and Prevention estimates that 5400 US children <19 years of age and 330000 US adults were living with AIDS 2000,14 we estimated that the hospitalization burden would translate to 0.76 hospitalizations and $18480 in hospitalization costs per US child with AIDS versus 0.67 hospitalizations and $14290 in hospitalization costs per US adult with AIDS during the year 2000.
A limitation of this study is that we were unable to analyze hospitalization data by primary discharge diagnosis, which would be expected to more accurately reflect the primary reason for the hospitalization, because HIV infection itself was the primary diagnosis in about one third of the hospitalizations involving HIV-infected children. This probably reflects nonuniform coding practices, because usually there is a more specific HIV disease manifestation leading to a particular hospitalization of an HIV-infected child. Instead, we analyzed data by all of the discharge diagnoses, which reflect comorbidities that may not always have led to the hospitalization. Our analysis, thus, provides a "snapshot" of the complex morbidities afflicting children with HIV infection who need to be hospitalized. Another limitation to this study is that because the data we used was for discharges rather than for patients, some children were represented more than once in the database. In addition, given nonuniversal HIV testing, some HIV-infected children might have been missed if they were hospitalized for an unrelated cause. Underreporting of procedures, particularly nonoperative ones, such as administration of antibiotics, is likely to be occurring, given the rates we observed, which would be expected to be higher when the prevalence of infections was so high. These limitations notwithstanding, this nationwide look at patterns of hospital use and morbid conditions among HIV-infected children in the United States in the era of HAART provides data that can be used as a starting point for more in-depth examinations of specific disease processes and current medical practices affecting children with HIV and as a baseline for analyses of trends as further improvements in HIV medical care are realized.
FOOTNOTES
The authors have indicated they have no financial relationships relevant to this article to disclose.
The views expressed in this article are those of the authors and do not necessarily reflect those of the Agency for Healthcare Research and Quality or the Centers for Disease Control and Prevention.
REFERENCES
De Martino M, Tovo PA, Balducci M, et al. Reduction in mortality with availability of antiretroviral therapy for children with perinatal HIV-1 infection. JAMA. 2000;284 :190 –197
Gortmaker SL, Hughes M, Ceriva J, et al. Effect of combination therapy including protease inhibitors on mortality among children and adolescents infected with HIV-1. N Engl J Med. 2001;345 :1522 –1528
Mocroft A, Ledergerber B, Katlama C, et al. Decline in the AIDS and death rates in the EuroSIDA study: an observational study. Lancet. 2003;362 :22 –29
Canani RB, Spagnuolo MI, Cirillo P, Guarino A. Decreased needs for hospital care and antibiotics in children with advanced HIV-1 disease after protease-inhibititor-containing therapy. AIDS. 1999;13 :1005 –1006
Sanchez JM, Ramos Amador JT, Fernandez de Miguel S, et al. Impact of highly active antiretroviral therapy on the morbidity and mortality in Spanish HIV-infected children. Pediatr Infect Dis J. 2003;22 :863 –867
Fleishman JA, Hellinger FJ. Recent trends in HIV-related inpatient admission, 1996–2000. Epidemiol Soc Sci. 2003;34 :102 –110
Gebo KA, Diener-West M, Moore RD. Hospitalization rates in an urban cohort after the introduction of highly active antiretroviral therapy. J Acquir Immune Defic Syndr. 2001;27 :143 –152
McCormick M, Kass B, Elixhauser A, et al. Annual report on access to and utilization of health care for children and youth in the United States—1999. Pediatrics. 2000;105 :219 –230
Healthcare Cost and Utilization Project. Design of the HCUP Kids' Inpatient Database (KID), 2000. Jan 29, 2003. Available at: www.hcup-us.ahrq.gov/db/nation/kid/kidrelatedreports.jsp. Accessed January 20, 2006
Bentham WD, Cai L, Schulman KA. Characteristics of hospitalizations of HIV-infected patients: an analysis of data from the 1994 Healthcare Cost and Utilization Project. J Acquir Immune Defic Syndr. 1999;22 :503 –508
Public Health Service and Health Care Financing Administration. International Classification of Diseases, Ninth Revision, Clinical Modification, Volumes I, II, and III. 5th ed. Washington, DC: Public Health Service; 1994
Elixhauser A, Steiner CA, Whittington CA, McCarthy E. Clinical Classifications for Health Policy Research: Hospital Inpatient Statistics, 1995. Healthcare Cost and Utilization Project, HCUP-3 Research Note. Rockville, MD: Agency for Health Care Policy and Research; 1998
Cowen ME, Dusscau DJ, Toth BG, et al. Casemix adjustment of managed care claims data using the clinical classification for health policy research method. Med Care. 1998;36 :1108 –1113
Centers for Disease Control and Prevention. HIV/AIDS Surveillance Report, 2003 (Vol. 15). Atlanta, GA: US Department of Health and Human Services; 2004:1–46. Available at
www.cdc.gov/hiv/stats/2003surveillancereport.pdf. Accessed February 5, 2006
Centers for Disease Control and Prevention. Sexually Transmitted Disease Surveillance 2003, STDs in Racial and Ethnic Minorities. Atlanta, GA: Centers for Disease Control and Prevention; 2003
Caudill S, Goldman T, Marconi K. Evaluation of pediatric HIV care provided in Ryan White CARE Title IV Women, Infants, Children and Youth Clinics. AIDS Patient Care STDS. 2003;17 :65 –73
Viani RM, Araneta MR, Deville JG, Spector SA. Decrease in hospitalization and mortality rates among children with perinatally-acquired HIV type 1 infection receiving highly active antiretroviral therapy. Clin Infect Dis. 2004;39 :725 –731
Gaughan DM, Hughes MD, Oleske JM, et al. Psychiatric hospitalizations among children and youths with HIV infection. Pediatr. 2004;113 :e544 –e551
a Division of Reproductive Health, National Center for Chronic Disease Prevention and Health Promotion, Centers for Disease Control and Prevention, Atlanta, Georgia
b Eastern Virginia Medical School, Norfolk, Virginia
c CONRAD Program, Arlington, Virginia
d Agency for Healthcare Research and Quality, Rockville, Maryland(Athena P. Kourtis, MD, Ph)
OBJECTIVES. The purpose of this research was to describe hospital use patterns of HIV-infected children in the United States.
STUDY DESIGN. We analyzed a nationwide, stratified probability sample of 2.5 million hospital discharges of children and adolescents during the year 2000, weighted to 7.3 million discharges nationally. We excluded discharges after hospitalizations related to pregnancy/childbirth and their complications and discharges of neonates <1 month of age and of patients >18 years of age. Diagnoses were identified through the use of the Clinical Classification Software with grouping of related diagnoses.
RESULTS. We estimated that there were 4107 hospitalizations of HIV-infected children in 2000 and that these hospitalizations accounted for $100 million in hospital charges and >30000 hospital days. Infections, including sepsis and pneumonia, were among the most frequent diagnoses in such hospitalizations, followed by diagnoses related to gastrointestinal conditions, nutritional deficiencies and anemia, fluid/electrolyte disorders, central nervous system disorders, cardiovascular disorders, and respiratory illnesses. Compared with hospitalizations of non-HIV-infected children, hospitalizations of HIV-infected ones were more likely to be in urban areas, in pediatric/teaching hospitals, and in the Northeast, and the expected payer was more likely to be Medicaid (77.6% vs 37.2%). Compared with children without HIV, those with HIV tended to be older (median age: 9.5 years vs 5.2 years), to have been hospitalized longer (mean: 7.8 days vs 3.9 days), and to have incurred higher hospital costs (mean: $23221 vs $11215); HIV-associated hospitalizations ended in the patient's death more frequently than non-HIV ones (1.8% vs 0.4%), and complications of medical care were also more common (10.8% vs 6.2%).
CONCLUSIONS. Infections account for the majority of hospitalizations of HIV-infected children in the United States, although nutritional deficiencies, anemia and other hematologic disorders, gastrointestinal and renal disorders, and complications of medical care are also more common among hospitalized children with HIV than among those without HIV.
Key Words: HIV ? children ? hospitalizations ? United States ? HCUP ? diagnosis
Abbreviations: HAART—highly active antiretroviral therapy ? KID—KIDS Inpatient Database ? HCUP—Healthcare Cost and Utilization Project ? DRG—diagnosis-related group ? CCS—Clinical Classification Software
There is little published information on the impact of pediatric HIV/AIDS on hospitalizations and costs or on the effects of highly active antiretroviral therapy (HAART) on these rates and costs. Data from adult studies have indicated that the number of HIV-related hospitalizations and the average length of hospital stay among HIV patients initially declined after the advent of HAART1–5 but that these trends have since leveled off.6,7 The results of 1 adult study also suggested that racial/ethnic disparities in HIV-related hospitalizations have persisted since the advent of HAART.8 However, few national estimates of the effect of pediatric HIV on the use of inpatient hospital services have been published, although such estimates are vital in this era of rapidly changing health care practices.
Our purpose in this study was to describe the use of inpatient hospital services by children with HIV infection in the United States. Specifically, we sought to describe the burden of pediatric HIV hospitalizations; the most common diagnoses among child HIV patients; the distribution of pediatric HIV hospitalizations by age group; and the distribution of such hospitalizations by expected payer/insurance provider, hospital type (teaching versus nonteaching and pediatric versus nonpediatric), and geographic location. We also estimated the average charges associated with these hospitalizations, estimated the average length of hospital stay, and compared the diagnostic codes most frequently assigned to HIV-infected children and adolescents (<19 years of age) with those most frequently assigned to HIV-uninfected ones.
METHODS
Description of the Databases
We used the 2000 KIDS Inpatient Database (KID) of the Healthcare Cost and Utilization Project (HCUP).8 HCUP is a partnership between state medical data collecting organizations and the Agency for Healthcare Research and Quality in which the state partners contribute their statewide hospital discharge data to HCUP. Data elements that states have in common are recorded into a uniform coding scheme, and data elements unique to individual states are retained if they are useful for research purposes. KID, part of the HCUP databases, is a nationwide sample of pediatric discharges from nonrehabilitation community hospitals weighted to reflect all pediatric discharges from US community hospitals open during any part of a particular calendar year. An in-depth description of the HCUP KID 2000 is available.9
Briefly, the 2000 KID contains information on 2.5 million discharges from 2784 hospitals. These discharges were stratified by reason for hospital stay (uncomplicated birth, complicated birth, and not birth related) and were also sorted by state, hospital, diagnosis-related group (DRG), and a random number within each DRG. Systematic random sampling was used to select from each sampling frame 10% of discharges after hospitalization related to an uncomplicated birth and 80% of discharges after hospitalization for other reasons.
To obtain national estimates, HCUP created discharge weights using American Hospital Association survey data. Six hospital characteristics were used to poststratify the hospitals: geographic region, hospital control, urban or rural location, teaching status, bed size, and hospital type. The KID database includes demographic information on discharged patients; diagnosis codes; procedure codes; and information about hospital type and geographic region, hospital charges, and length of hospital stay. In 2000, 27 states contributed data to KID, and the pediatric patients in these states accounted for 72.6% of all US hospital discharges of children in 2000.
The hospital universe is defined as all US hospitals open during any part of a particular calendar year and designated as community hospitals in the American Hospital Association Annual Survey. Charge data are based on the charges for the hospitalization only and not on physician fees. Charges do not necessarily reflect costs or reimbursements actually received by hospitals. Other studies have demonstrated that HCUP is a powerful tool for estimating the charges of hospital resource use.8,10
Study Design and Statistical Analysis
The diagnoses recorded on hospital discharge abstracts are coded in accordance with the International Classification of Diseases, Ninth Revision, Clinical Modification.11 Because >12000 diagnosis codes are in the International Classification of Diseases, Ninth Revision, Clinical Modification, we combined discharge diagnoses into a relatively small number of DRGs or "major diagnostic categories" using Clinical Classification Software (CCS)12,13 to aggregate illnesses and conditions into 259 mutually exclusive categories, most of which are clinically homogeneous. For this report, we further aggregated CCS categories into the following clinically relevant groupings: malignancies: CCS 11–45; cardiovascular disorders: 96–121; accidents/injuries: all E codes except E2616 and E2617 plus 225–236 and 239–244; complications of medical care: 237, 238, E2616, and E2617; mental/psychiatric disorders: 65 and 68–75; alcohol and drug abuse: 66 and 67; central nervous system/eye disorders: 79–91 and 93–95; gastrointestinal disorders: 137–155; renal disorders: 156–158, 160–163, and 165–166; sexually transmitted diseases: 9 and 168; congenital anomalies: 213–217; shock: 249; nutritional deficiencies and anemia: 52 and 59; fluid/electrolyte disorders: 55; sickle cell disease: 61; social admissions: 255; asthma: 128; other respiratory disorders (except infections and asthma): 56, 127, and 130–134; endocrine/metabolic disorders: 48–51, 53, and 58; other hematologic disorders: 57, 60, and 62–64; tuberculosis: 1; bacterial infections/sepsis: 2, 3, 76, 92, 124, 125, 135, 159, 197, 201, 247, and 248; mycoses: 4; viral/other infections: 6–8, 77, 78, 123, 126, and 134; pneumonia: 122 and 129; fever of unknown origin: 246; and allergic reactions: 253.
We analyzed data by age group only for those states that disclose the age of the child. Texas does not release age information for sensitive groups (including HIV-infected patients). Pennsylvania recorded age for sensitive groups (including HIV-infected patients) to the midpoint of 5-year age range.
Costs were calculated using the total charges that were reported for each hospitalization. We excluded neonate delivery hospitalizations (variable neomat: 2,3) and hospitalizations for conditions related to pregnancy and delivery (CCS codes 177-196). HIV-infection was identified by the CCS code of 5.
Standard methods for analyzing weighted survey data were employed using 2 statistical software packages: SAS (SAS Institute, Inc, Cary, NC) and SUDAAN (Research Triangle Institute, Research Triangle Park, NC) statistical. A 2 test was used to compare categorical variables and t tests to analyze means. Programming and data results were confirmed by 2 independent researchers.
RESULTS
We estimated that there were 4107 pediatric discharges associated with an HIV diagnosis in the United States in 2000, that the discharged patients spent a total of 31940 days in the hospital at a cost of $99.8 million (Table 1), and that children with HIV accounted for 0.2% of all hospital discharges, 0.4% of all hospital charges, and 0.7% of all hospital days among children hospitalized in the United States for reasons other than normal birth or conditions associated with pregnancy. Compared with non-HIV-infected children, those with HIV had longer mean hospital stays (7.8 days vs 3.9 days; P < .001), higher mean hospital charges per hospitalization, and a higher mean number of diagnoses per hospitalization (5.10 vs 3.03; P < .01).
The median age of discharged children with HIV was 9.5 years, compared with 5.2 years for those without an HIV code. Infants (<1 year of age) accounted for only 5% of all HIV-related pediatric hospitalizations but for >20% of all non-HIV related pediatric hospitalizations, whereas children aged 5 to 14 years accounted for more than half of all HIV-related pediatric hospitalizations but for only about one third of all non-HIV-related hospitalizations (Table 1). Medicaid was the expected payer for 77.6% of HIV-infected pediatric patients, and private insurance was the expected payer for 15.7% (vs 37.2% Medicaid and 54.1% private insurance for children without HIV). We also found that hospitalizations of children with HIV were more likely than those without HIV to be in an urban teaching hospital (84.2% vs 59.5%), less likely to be in a hospital or hospital unit not specifically for children (37.1% vs 62.3%), more likely to be from the Northeast or South (87.8% vs 58.2%), and more likely to lead to death during the hospitalization (1.7% vs 0.4%; P < .001).
The most common discharge diagnoses for hospitalizations of HIV-infected children were, in order of frequency, bacterial infections and sepsis, viral/and other (including parasitic) infections, gastrointestinal diagnoses, pneumonia, nutritional deficiencies and anemia, fluid/electrolyte disorders, fungal infections, cardiovascular conditions, and respiratory illnesses (Table 2). Categories of conditions more common among pediatric patients with HIV than among those without HIV included all infections (tuberculosis and bacterial, viral, and fungal infections), cardiovascular disorders (mostly cardiomyopathies or hypertension), central nervous system diseases (mostly seizures or encephalopathy), gastrointestinal conditions (mainly gastrostomy or diarrhea), renal conditions (mainly chronic renal failure or nephritis), endocrine and metabolic disorders (mostly lack of normal development), sexually transmitted diseases, nutritional deficiencies, anemia and other hematologic disorders, mental/psychiatric disorders, and complications of medical care (mainly infections and inflammation associated with the use of vascular devices; Fig 1). Categories of conditions more common among hospitalizations of uninfected children included congenital anomalies, accidents/injuries, asthma, malignancies, alcohol and drug abuse, and sickle cell disease.
In 53% of hospitalizations of HIV-infected children in 2000, there was 1 procedure performed compared with 42.9% of hospitalizations of uninfected children; hospitalizations of children with HIV averaged 1.33 procedures, whereas those without HIV averaged 0.89 (P = .03). Antibiotic injections, immune globulin administration, esophagogastro-duodenoscopy, gastrostomy, and enteral and parenteral feeding were more common among children with HIV-related hospitalizations, whereas appendectomies, cancer chemotherapy, mechanical ventilation, and computerized tomography of the head were more common among uninfected children (Table 3).
DISCUSSION
The Centers for Disease Control and Prevention reported that 5400 US children <19 years of age were living with AIDS in 2000.14 Although the annual number of infants infected with HIV perinatally has been decreasing, the annual number of adolescents infected with HIV has been increasing.14 This increase in the rate of HIV infections among adolescents is particularly pronounced among blacks, Hispanics, and girls of all racial/ethnic groups.14,15 Despite the increase of HIV infections among adolescents, the number of AIDS cases and HIV-related deaths have been declining, largely because of the use of combination drug regimens, such as HAART.1,3,14
Although comprehensive data on HIV-related morbidity among US pediatric patients since the advent of HAART are currently not available, the results of some limited studies suggest that HAART has been effective in reducing morbidity among children with HIV. For example, a retrospective analysis of 525 charts from HIV-infected children participating in the Ryan White program indicated that as use of combination therapy increased, the rate of opportunistic infections and hospitalizations decreased.16 In addition, results of a recent study of 129 children followed at 3 HIV clinics in California showed that the rate of HIV-related deaths, as well as the hospitalization rate and average length of hospital stay among perinatally infected children, all decreased significantly between 1994 and 2001,16 a period during which the use of HAART was increasing. They also showed that pneumonia and sepsis were the main causes of hospitalization among these children from 1994 through 1999 but that the rate of hospitalizations for these conditions decreased in 2000 to 200117 after the introduction of complex pharmacologic regimens in the care of these children. The problem of psychiatric disease among HIV-infected children was also highlighted recently when results from the Pediatric AIDS Collaborative Trials Study cohort of 1808 HIV-infected children under 15 years of age followed over a period of 3 years showed them to have a greater risk for psychiatric hospitalization than noninfected children of the same age.18
To our knowledge, our study is the first to provide national estimates of hospital use by US children with HIV, as well as the first to describe the patterns of pediatric HIV-related morbidity in the era of HAART. We showed that the distribution of conditions for which HIV-infected children were hospitalized and the type of hospital to which they were admitted (mainly urban, pediatric teaching hospitals in the Northeast and South) were different from those of children without HIV infection. Hospitalizations of HIV-infected children were, on average, longer, more costly, and more complex, as reflected in a higher number of concomitant diagnoses and higher rates of medical complication and death. Because Medicaid is the expected payer for the majority of pediatric HIV patients, their hospital care poses an incremental burden on the national health care budget. Infections; hematologic disorders (mainly anemia); nutritional disorders; gastrointestinal, renal, and cardiovascular conditions; and medical complications/iatrogenic conditions are more common in hospitalizations of HIV-infected children compared with non-HIV-infected ones. In comparison, hospitalizations resulting from injuries, appendicitis, malignancies, congenital anomalies, and alcohol and substance abuse assume a lesser proportion in HIV-infected compared with HIV-uninfected children. Using Centers for Disease Control and Prevention estimates that 5400 US children <19 years of age and 330000 US adults were living with AIDS 2000,14 we estimated that the hospitalization burden would translate to 0.76 hospitalizations and $18480 in hospitalization costs per US child with AIDS versus 0.67 hospitalizations and $14290 in hospitalization costs per US adult with AIDS during the year 2000.
A limitation of this study is that we were unable to analyze hospitalization data by primary discharge diagnosis, which would be expected to more accurately reflect the primary reason for the hospitalization, because HIV infection itself was the primary diagnosis in about one third of the hospitalizations involving HIV-infected children. This probably reflects nonuniform coding practices, because usually there is a more specific HIV disease manifestation leading to a particular hospitalization of an HIV-infected child. Instead, we analyzed data by all of the discharge diagnoses, which reflect comorbidities that may not always have led to the hospitalization. Our analysis, thus, provides a "snapshot" of the complex morbidities afflicting children with HIV infection who need to be hospitalized. Another limitation to this study is that because the data we used was for discharges rather than for patients, some children were represented more than once in the database. In addition, given nonuniversal HIV testing, some HIV-infected children might have been missed if they were hospitalized for an unrelated cause. Underreporting of procedures, particularly nonoperative ones, such as administration of antibiotics, is likely to be occurring, given the rates we observed, which would be expected to be higher when the prevalence of infections was so high. These limitations notwithstanding, this nationwide look at patterns of hospital use and morbid conditions among HIV-infected children in the United States in the era of HAART provides data that can be used as a starting point for more in-depth examinations of specific disease processes and current medical practices affecting children with HIV and as a baseline for analyses of trends as further improvements in HIV medical care are realized.
FOOTNOTES
The authors have indicated they have no financial relationships relevant to this article to disclose.
The views expressed in this article are those of the authors and do not necessarily reflect those of the Agency for Healthcare Research and Quality or the Centers for Disease Control and Prevention.
REFERENCES
De Martino M, Tovo PA, Balducci M, et al. Reduction in mortality with availability of antiretroviral therapy for children with perinatal HIV-1 infection. JAMA. 2000;284 :190 –197
Gortmaker SL, Hughes M, Ceriva J, et al. Effect of combination therapy including protease inhibitors on mortality among children and adolescents infected with HIV-1. N Engl J Med. 2001;345 :1522 –1528
Mocroft A, Ledergerber B, Katlama C, et al. Decline in the AIDS and death rates in the EuroSIDA study: an observational study. Lancet. 2003;362 :22 –29
Canani RB, Spagnuolo MI, Cirillo P, Guarino A. Decreased needs for hospital care and antibiotics in children with advanced HIV-1 disease after protease-inhibititor-containing therapy. AIDS. 1999;13 :1005 –1006
Sanchez JM, Ramos Amador JT, Fernandez de Miguel S, et al. Impact of highly active antiretroviral therapy on the morbidity and mortality in Spanish HIV-infected children. Pediatr Infect Dis J. 2003;22 :863 –867
Fleishman JA, Hellinger FJ. Recent trends in HIV-related inpatient admission, 1996–2000. Epidemiol Soc Sci. 2003;34 :102 –110
Gebo KA, Diener-West M, Moore RD. Hospitalization rates in an urban cohort after the introduction of highly active antiretroviral therapy. J Acquir Immune Defic Syndr. 2001;27 :143 –152
McCormick M, Kass B, Elixhauser A, et al. Annual report on access to and utilization of health care for children and youth in the United States—1999. Pediatrics. 2000;105 :219 –230
Healthcare Cost and Utilization Project. Design of the HCUP Kids' Inpatient Database (KID), 2000. Jan 29, 2003. Available at: www.hcup-us.ahrq.gov/db/nation/kid/kidrelatedreports.jsp. Accessed January 20, 2006
Bentham WD, Cai L, Schulman KA. Characteristics of hospitalizations of HIV-infected patients: an analysis of data from the 1994 Healthcare Cost and Utilization Project. J Acquir Immune Defic Syndr. 1999;22 :503 –508
Public Health Service and Health Care Financing Administration. International Classification of Diseases, Ninth Revision, Clinical Modification, Volumes I, II, and III. 5th ed. Washington, DC: Public Health Service; 1994
Elixhauser A, Steiner CA, Whittington CA, McCarthy E. Clinical Classifications for Health Policy Research: Hospital Inpatient Statistics, 1995. Healthcare Cost and Utilization Project, HCUP-3 Research Note. Rockville, MD: Agency for Health Care Policy and Research; 1998
Cowen ME, Dusscau DJ, Toth BG, et al. Casemix adjustment of managed care claims data using the clinical classification for health policy research method. Med Care. 1998;36 :1108 –1113
Centers for Disease Control and Prevention. HIV/AIDS Surveillance Report, 2003 (Vol. 15). Atlanta, GA: US Department of Health and Human Services; 2004:1–46. Available at
www.cdc.gov/hiv/stats/2003surveillancereport.pdf. Accessed February 5, 2006
Centers for Disease Control and Prevention. Sexually Transmitted Disease Surveillance 2003, STDs in Racial and Ethnic Minorities. Atlanta, GA: Centers for Disease Control and Prevention; 2003
Caudill S, Goldman T, Marconi K. Evaluation of pediatric HIV care provided in Ryan White CARE Title IV Women, Infants, Children and Youth Clinics. AIDS Patient Care STDS. 2003;17 :65 –73
Viani RM, Araneta MR, Deville JG, Spector SA. Decrease in hospitalization and mortality rates among children with perinatally-acquired HIV type 1 infection receiving highly active antiretroviral therapy. Clin Infect Dis. 2004;39 :725 –731
Gaughan DM, Hughes MD, Oleske JM, et al. Psychiatric hospitalizations among children and youths with HIV infection. Pediatr. 2004;113 :e544 –e551
a Division of Reproductive Health, National Center for Chronic Disease Prevention and Health Promotion, Centers for Disease Control and Prevention, Atlanta, Georgia
b Eastern Virginia Medical School, Norfolk, Virginia
c CONRAD Program, Arlington, Virginia
d Agency for Healthcare Research and Quality, Rockville, Maryland(Athena P. Kourtis, MD, Ph)