Secondary Logo

Journal Logo

Hospitalizations for Metabolic Conditions, Opportunistic Infections, and Injection Drug Use Among HIV Patients

Trends Between 1996 and 2000 in 12 States

Gebo, Kelly A, MD, MPH*; Fleishman, John A, PhD; Moore, Richard D, MD, MHS*

JAIDS Journal of Acquired Immune Deficiency Syndromes: December 15th, 2005 - Volume 40 - Issue 5 - p 609-616
doi: 10.1097/01.qai.0000171727.55553.78
Epidemiology and Social Science

Background: Rapid changes in HIV epidemiology and highly active antiretroviral therapy (HAART) may have resulted in recent changes in patterns of inpatient utilization.

Objective: To examine trends in inpatient diagnoses and mortality in HIV patients.

Design/Setting/Patients: Serial cross-sectional analyses of HIV patients hospitalized in 1996, 1998, and 2000, using hospital discharge data from the Healthcare Costs and Utilization Project for 12 states. Each hospitalization was classified as an opportunistic illness, complication of injection drug use (IDU), liver-related complication, ischemic heart disease, cerebrovascular disease, non-Pneumocystis carinii pneumonia (PCP), diabetes, or chronic hepatitis C virus (HCV).

Main Outcome Measures: Number of hospital admissions, inpatient mortality.

Results: We evaluated 316,963 admissions that occurred between 1996 and 2000, with an overall mortality of 7%. Hospitalizations for opportunistic infections significantly decreased from 40% to 27% of all HIV-related admissions. The overall proportion of IDU complications remained relatively stable (6%) each year. Hospitalizations increased for liver-related complications from 8% to 13% and for chronic HCV from 1% to 5% in this period. The number of hospitalizations for cerebrovascular disease and for ischemic heart disease was relatively negligible in all years. Overall, inpatient mortality decreased between 1996 and 2000. Relatively higher mortality was observed among African Americans, Hispanics, those with Medicaid, those with Medicare, and the uninsured, however. Opportunistic infections and liver-related complications were associated with greater inpatient mortality.

Conclusion: Results do not show a significant recent rise in HIV-related inpatient utilization. Admissions to treat opportunistic infections have declined precipitously, consistent with the effects of HAART. Although not dramatic, liver-related disease is an increasing cause of hospitalization in HIV+ patients.

From the *Department of Medicine, Johns Hopkins University School of Medicine, Baltimore, MD; and †Agency for Healthcare Research and Quality, Rockville, MD.

Received for publication December 5, 2004; accepted May 17, 2005.

Supported by the National Institute of Drug Abuse (grants K 23 DA 00523 and K24 DA 00432) and the Johns Hopkins University Center for AIDS Research (CFAR; grant P30 AI42855).

The views expressed in this paper are those of the authors. No official endorsement by Department of Health and Human Services or the Agency for Healthcare Research and Quality is intended or should be inferred.

Reprints: Kelly Gebo, Johns Hopkins University School of Medicine, 1830 East Monument Street, Room 442, Baltimore, MD 21287 (e-mail:

Highly active antiretroviral therapy (HAART) has been shown to reduce morbidity and mortality associated with HIV-1 infection in the United States.1 Previous studies have shown that HAART reduces hospitalization rates by as much as 24% to 43%.2-11 With decreased mortality, other chronic diseases that are common to HIV-infected patients, such as viral hepatitis, and complications of HAART, such as diabetes, liver-related conditions, ischemic heart disease, and cerebrovascular disease, may now have more opportunity to become clinically apparent.12-16

A previous study, using data from 9 large urban clinic indicated that hospitalization rates decreased between 1995 and 1997 in all HIV-infected patients but then reached a plateau and may have even begun to increase between 1997 and 1998.11 Further analysis of these data showed that this increase was attributable to admissions related to hepatitis C virus (HCV) and liver disease.12 Other studies also suggest that liver failure may be related to increases in overall hospitalizations between 1998 and 2000 in HIV-infected individuals as well as to increasing mortality.13,15 A recent study in Europe indicated a decrease in hospitalization rates between 1995 and 2003 but an increased risk of hospitalization in those with HCV-HIV coinfection in 2001.16

Studies have shown increases in hospitalizations attributable to complications of cardiovascular and cerebrovascular disease, possibly linked to effects of HAART and HIV infection itself.17-27 A recent study of admissions to Veterans Affairs (VA) facilities found that rates of admissions for cardiovascular or cerebrovascular disease among persons with HIV remained relatively constant between 1993 and 2001.28 Confirmation of these results among patients outside the VA system would be valuable.

If HIV-positive patients are being hospitalized for more life-threatening conditions, including cardiac and cerebrovascular disease, in-hospital mortality may in fact be increasing. Early studies showed mortality in HIV patients to be lower in the era of HAART,1 but little is known about inpatient mortality in the first several years after the introduction of HAART.

This study uses data from several states in the United States to examine the extent to which liver-related conditions, ischemic heart disease, cerebrovascular disease, and several other conditions each account for an increasing proportion of HIV-related hospitalizations in 1996, 1998, and 2000. As points of comparison, we also examined hospitalizations for opportunistic infections, complications of intravenous drug use (which may be rising, given changes in the demographics of people with HIV infection), and diabetes (a consequence of treatment of HIV infection). In addition, we examined associations between each of these conditions and in-hospital mortality during these years.

Back to Top | Article Outline


Hospital discharge data were obtained from the Healthcare Cost and Utilization Project (HCUP) State Inpatient Database (SID).29 The SID contains hospital discharge abstract data covering inpatient stays from all short-term nonfederal community hospitals in participating states. SID data include primary and all secondary diagnoses for each inpatient stay, the patient's gender, and the primary payer for inpatient services. Some states also report the patient's race/ethnicity. SID data also include information on the length of stay (calculated as the difference between the admission and discharge dates) and whether the patient died during the hospitalization.

This study used data from 12 of the 27 states participating in the HCUP as of 2000: California, Colorado, Florida, Iowa, Illinois, Kansas, Maryland, New Jersey, New York, Pennsylvania, South Carolina, and Washington. For each state, we analyzed data from 1996, 1998, and 2000. Several of the states in the SID were excluded because they began participating in the HCUP relatively recently and did not provide data for a sufficient time period. States with relatively high HIV prevalence were selected; Iowa, Kansas, and South Carolina were included as relatively low-prevalence comparison states. Sites were also chosen to facilitate comparison with prior analyses of HIV-related hospitalizations using HCUP data.30,31

We identified HIV-related hospitalizations by examining all primary and secondary diagnoses listed in the discharge abstract. All hospitalizations with a primary or secondary International Classification of Diseases, ninth edition (ICD-9), diagnosis code that included 042.0 through 044.9 (inclusive) were selected as HIV-related hospitalizations. This algorithm has been used in other analyses of HIV-related hospitalizations using HCUP data.30,31 The V08 code has been previously used as an HIV diagnostic code. It was abandoned in 1996, however, and only 3 of the hospitalizations in the data set incorporated the V08 code. An admission with only a V08 code and no 042 code could be considered unrelated to HIV infection. After October 1994, the ICD-9 code included only 1 code for HIV disease and AIDS (042).32 We searched all codes in the 042 to 044 range in all years, however, in case those who had coded the forms persisted in using these older codes. This algorithm for identifying hospitalized patients with HIV disease is sensitive. Research suggests that the 042 code is adequate for identifying HIV-related hospitalizations. One study examined more than 6000 hospitalizations with diagnostic codes indicative of HIV disease and determined whether an individual was infected, using AIDS surveillance data from state health departments and a review of the medical charts. Nearly all (97%) of those with an 042 code in their hospital discharge abstract were infected with HIV.33 The 042 code had a positive predictive value of 89% for AIDS and 99% for HIV. Patients who were HIV infected but who were not identified as such using the ICD-9 codes specific to HIV/AIDS were not included in this analysis.

To classify the clinical conditions associated with each hospitalization, we examined the primary and the first 4 secondary discharge diagnoses. (Results changed minimally when classification was based on up to 16 secondary discharge diagnoses.) We derived indicator variables reflecting the following diagnostic categories: (1) complications of injection drug use (IDU); abscess, cellulitis, osteomyelitis, bacteremia, endocarditis, and poisoning by analgesics, using the following ICD-9 codes: 681.0-682.9, 421.0-421.9, 730.0-730.9, 965.0-965.9, and 790.7; (2) opportunistic infections, using the 1993 indicator diagnoses in the case definition of AIDS (ICD-9 codes 0031, 0072, 0078, 010.0-019.9, 031.0-031.9, 039.0-039.9, 046.3, 078.5, 112.1-112.9, 114.1-114.9, 115.0-115.9, 117.5, 130.0-130.9, 136.3, 176.0-178.9, 180.0-180.9, 202.8, 321.0, and 484.1)34; (3) liver-related complications (acute and subacute necrosis of the liver, chronic liver disease and cirrhosis, liver abscess, hepatic coma, portal hypertension, hepatorenal syndrome, hepatocellular carcinoma, gastrointestinal bleed, Mallory-Weiss tear and viral hepatitis, excluding HCV (ICD-9 codes 070.0-070.53, 070.55-070.9, 570.0-575.0, 531.0-536.9, 155.0-155.9, 530.7, and 567.2)); (4) pneumonia (non-Pneumocystis carinii pneumonia [PCP], ICD-9 code 486); (5) diabetes (ICD-9 codes 250.0-250.9); (6) ischemic heart disease (ICD-9 codes 410.0-411.9); (7) cerebrovascular disease (ICD-9 codes 434.0-436.0); and (8) chronic HCV (ICD-9 code 070.54). Diagnosis categories were not mutually exclusive. Hospitalizations for opportunistic illness were expected to decrease over this period because of the beneficial effects of HAART. Hospitalizations for complications of IDU were expected to increase, given the changes in the epidemiologic profile of persons with HIV.35 Hospitalizations for pneumonia were included in the analyses because examination of ICD-9 codes suggested that non-PCP was implicated in a notable proportion of admissions.

Because pediatric HIV-related hospitalizations may systematically differ from those for adults, we restricted analyses to hospitalizations for patients aged 18 years and older. Age was categorized as 18 to 30, 31 to 40, 41 to 50, and 51 years or older.

Payer categories were Medicare, Medicaid, private (including private health maintenance organization), self-pay, no charge, or other. The relatively small number of hospitalizations (1939 [0.6%]) with a payer code of “no charge” were combined with those for which the payer was self-pay; hospitalizations in this category were assumed to be not covered by insurance. Race/ethnicity categories were white, African American, Hispanic, Asian/Pacific Islander, Native American, other, or missing. Because the “other” category may reflect measurement error, these hospitalizations were combined with those missing race/ethnicity data. Hospitalizations for Native Americans (505 [0.16%]) were combined with those for Asian/Pacific Islanders (1641 [0.52%]) into an “other minority” category. Washington and Illinois do not report race information; thus, all admissions in these 2 states were coded as missing for this variable.

Back to Top | Article Outline


A total of 326,177 HIV-related inpatient hospitalizations were identified for the 12 states in 1996, 1998, and 2000 combined. Of these, 7915 were removed from the analysis because the patient was less than 18 years of age. Among adults, we deleted 29 observations attributable to missing data for age, 15 hospitalizations missing gender data, and 1255 hospitalizations missing payer data. The 1299 deleted observations were 0.4% of the total adult sample. The analytic sample numbered 316,963 hospitalizations. (The sample included 4276 hospitalizations reported as lasting for 0 days, 4 lasting for more than 365 days, and 111 missing information on length of stay [1.3% of the sample]. Removing these observations had virtually no effect on the results.)

For each diagnosis category, we first examined the number of HIV-related hospitalizations in each year. For multivariate analyses, data were collapsed into a 5-dimensional cross-tabulation comprising each combination of state (12 categories), year (3 categories), gender (2 categories), age (4 categories), and race (5 categories). For each combination, we calculated the total number of HIV-related admissions as well as the total number of admissions for each diagnostic category. Initial analyses using Poisson regression suggested the presence of overdispersion in the data. Consequently, we used negative binomial regression, which incorporates an additional parameter to deal with overdispersion and thereby provides better estimates of standard errors than Poisson regression.36 We compared 1998 and 2000 with the reference year of 1996. Analyses regressed the number of hospitalizations for each specific diagnostic category based on year, gender, age, state, and race. For each combination of these independent variables, the total number of HIV-related hospitalizations varied. To control for these variations, analyses also included the log of the total number of HIV-related hospitalizations as an offset (ie, its coefficient was constrained to equal 1). Because the analyses are implicitly based on a numerator representing the number of patients with a particular diagnosis and a denominator representing the total number of HIV-related admissions during the same time frame, results can be interpreted in terms of the proportion of HIV admissions of each diagnosis category per HIV-related admission. Because data on race/ethnicity were missing for Illinois and Washington, we reran the analyses excluding observations from these 2 states; the sign and significance of the coefficients were unchanged, and the magnitudes of the coefficients changed minimally. Thus, we report analyses that include data from these 2 states.

A second analysis examined in-hospital mortality, using individual hospitalization as the unit of analysis. We used logistic regression to examine in-hospital mortality for each time period as a function of state, payer, gender, race/ethnicity, age, and diagnosis group. We also included the primary payer for the hospitalization in the model. Because patients in the same hospital cannot be considered to be statistically independent, we used robust standard errors, with clustering of patients within hospitals. Statistical analysis was performed using Stata 8.0 (Stata Corp., College Station, TX).

Back to Top | Article Outline


Number of Hospitalizations

Of the 316,963 HIV-related hospitalizations for adult patients in 12 states, 128,754 occurred in 1996, 96,811 took place in 1998, and 91,398 occurred in 2000. Consistent with the introduction of HAART in 1996, hospitalizations declined 25% between 1996 and 1998 but dropped only 6% between 1998 and 2000. Overall, HIV-related hospitalizations decreased 29% from 1996 to 2000. In a negative binomial regression, estimated incidence rate ratios were 0.77 (95% confidence interval [CI]: 0.60-0.98; P = 0.03) for 1998 versus 1996 and 0.69 (95% CI: 0.54-0.88; P = 0.003) for 2000 versus 1996 (results not tabulated).

As has been shown in prior analyses of HCUP data from 7 states,31 hospitalizations decreased proportionately more for men (33%) than for women (18%) between 1996 and 2000 (Table 1). The proportionate decline in hospitalizations was greatest for the 18- to 30-year-old group (56%). In contrast, hospitalizations for patients older than 50 years of age increased 8% during this interval. Hospitalizations decreased proportionately more for whites (40%) and Hispanics (32%) than for African Americans (20%). Hospitalizations covered by private insurance decreased 45% between 1996 and 2000; those for Medicaid patients dropped by 31%. Medicare hospitalizations decreased only 3% between 1996 and 2000, however, and the number increased between 1998 and 2000.



Most HIV-related hospitalizations had none of the 8 diagnosis categories (42.3%) or exactly 1 (48.5%) (results not shown). Two of the diagnoses occurred together in 8.9% of the hospitalizations. Thus, multiple indications for hospital admission were relatively rare. In 1996 and 2000, the most common diagnosis was non-PCP. In 1996, the next most common diagnosis was Pneumocystis jiroveci pneumonia; however, in 2000, consistent with the introduction of HAART, this slipped to the eighth most common diagnosis. Other common diagnoses in both years included hypovolemia, pancreatitis, opioid dependence, and convulsions.

Table 2 reports the number of hospitalizations for each year by diagnosis category. For each diagnosis category, the proportion of hospitalizations differed significantly by year (P < 0.001 using a χ2 test). Hospitalizations for opportunistic illnesses significantly decreased from 51,050 in 1996 to 24,661 in 2000. Hospitalizations for IDU complications also decreased in number from 6390 in 1996 to 5238 in 2000, but the overall proportion remained relatively stable. In contrast, hospitalizations for liver-related complications increased from 10,481 to 11,583, or from 8% to 13% of all HIV-related hospitalizations. Hospitalizations for chronic HCV also rose from 1% to 5% in this period. Hospitalizations for diabetes rose from 3% to 5% between 1996 and 2000. The numbers of hospitalizations for cerebrovascular disease or for ischemic heart disease were relatively negligible in all years. Hospitalizations for pneumonia dropped in absolute numbers (13,230 to 12,461) but rose in proportion (10% to 14%), because the total number of hospitalizations declined even faster.



Table 3 presents negative binomial regression results for each of the 8 diagnosis categories. The proportion of hospitalizations for opportunistic illnesses decreased significantly between 1996 and 1998 and again between 1998 and 2000, controlling for state, age, gender, and race/ethnicity. The proportion of hospitalizations for cerebrovascular complications did not differ between 1996 and 1998 but was significantly higher in 2000 than in either previous year. For all other diagnostic classes, the proportion increased significantly in both periods; all differences between 1996 and 1998 were significant, as were the differences between 1998 and 2000. The strongest time period effects were for chronic HCV; between 1996 and 1998, hospitalization proportions increased by a factor of 2.25, and they increased by a factor of 3.97 between 1996 and 2000.



Women had lower hospitalization proportions for liver-related complications, opportunistic illnesses, complications of IDU, and ischemic heart disease compared with men, but women had relatively higher proportions of hospitalization for pneumonia and diabetic complications. Gender was not associated with hospitalizations for cerebrovascular complications or chronic HCV.

Compared with patients aged 18 to 30 years, the proportions of hospitalizations for patients in the 31- to 40-year-old and 41- to 50-year-old age groups were significantly higher for all conditions except opportunistic illnesses, for which they were significantly lower. Hospitalizations for patients older than 50 years of age showed a similar pattern to those of the middle-aged groups, with the exception that they did not differ from the youngest group in hospitalizations for IDU complications. The proportions of hospitalizations for the oldest group were especially high for ischemic heart disease and cerebrovascular complications.

Compared with hospitalizations for white patients, for African-American patients, the proportions of hospitalization were significantly higher for pneumonia, diabetes, and cerebrovascular complications and were significantly lower for liver-related diseases, chronic HCV, IDU complications, and ischemic heart disease. Hospitalizations for Hispanic patients showed a pattern similar to that for African Americans, with the exception of a higher proportion of hospitalizations for opportunistic illnesses than whites.

Back to Top | Article Outline

Inpatient Mortality

Death occurred in 7% of all hospitalizations. Inpatient mortality decreased across time, dropping from 8.5% of hospitalizations in 1996 to 6.3% in 1998 and 6.2% in 2000. Across all 3 years, the mortality rate was highest for cerebrovascular disease (13.8%), followed by opportunistic infections (10.1%), pneumonia (9.5%), liver-related disease (8.8%), and ischemic heart disease (8.6%). Mortality rates for other diagnoses were lower: 4.1% for IDU complications, 4.3% for diabetes, and 5.8% for chronic HCV.

In multivariate logistic regression analysis (Table 4), the adjusted odds of inpatient death were 27% lower in 1998 than in 1996 and 30% lower in 2000 than in 1996. Consistent with the bivariate results, the odds of death were more than twice as high when the hospitalization involved cerebrovascular disease, and they were nearly twice as high for opportunistic illnesses compared with hospitalizations that did not involve any of the 8 diagnosis classes. In contrast, death was less likely when the hospitalization involved complications of IDU or diabetes. Women had lower odds of death than men. The likelihood of death was higher for older patients. Medicare and Medicaid hospitalizations had lower odds of inpatient mortality than those covered by private insurance. Compared with hospitalizations for white patients, those for African-American and Hispanic patients were more likely to end in death.



Back to Top | Article Outline


Across multiple states, the total number of HIV-related inpatient hospitalizations decreased significantly between 1996 and 1998 and then plateaued between 1998 and 2000. These results extend prior analyses of HCUP data, which were based on 7 of the 12 states in the current analysis.31 Admission diagnoses changed significantly over this period, with a dramatic drop in hospitalizations for opportunistic illnesses. Hospitalizations for liver-related complications increased slightly in absolute numbers and as a proportion of all HIV-related hospitalizations. Hospitalizations for chronic HCV and for conditions related to diabetes also showed a notable rise. In contrast, hospitalizations for complications of cerebrovascular and ischemic heart disease remained relatively infrequent, consistent with prior findings from the VA.28,37

In this study, we chose to separate chronic HCV from acute liver-related conditions. Of 31,865 hospitalizations with a liver-related diagnosis, only 9% (n = 2865) also had a code for chronic HCV. Of 8792 hospitalizations with a chronic HCV code, 33% also had an acute liver-related condition. Inpatient coding of chronic HCV without a concomitant liver condition often reflects chronic HCV infection but not an acute liver diagnosis. In addition, some insurers increase reimbursement for cocoding of HIV and HCV; therefore, there is financial incentive to hospital coders to code HCV even if there is no evidence of acute liver disease. A recent analysis of hospitalizations between 1995 and 2000 demonstrated this cocoding phenomenon.12 As seen in the mortality analysis, the associations of chronic HCV and liver disease are in opposite directions, suggesting that coding of chronic HCV may in fact be a method to increase reimbursement and may not reflect an acute liver-related condition.

Because one of the most common causes of bloodstream infections and their sequelae is complications of IDU, we thought it appropriate to classify these diagnoses as complications of injection drugs. Although we did not have access to concomitant clinical information to assess whether patients with IDU diagnoses were or were not actively using drugs at the time of admission, approximately 33% of hospitalizations with an IDU diagnosis had a concomitant substance use/dependence code, indicating probable cooccurrence of these conditions. In contrast, less than 2.5% of IDU-related admissions had a concomitant diagnosis of end-stage renal disease, another condition predisposing patients to bacteremias, abscesses, and osteomyelitis.

Overall inpatient mortality decreased over time. Certain conditions increased the odds of in-hospital mortality, however, including cerebrovascular disease, opportunistic illness, liver-related complications, pneumonia, and ischemic heart disease. These results emphasize the need for continued monitoring and management of these conditions.

Female gender was strongly positively associated with hospitalizations to treat complications of diabetes and pneumonia, whereas men were more likely to be hospitalized with cardiovascular complications, liver-related complications, opportunistic illnesses, and complications of IDU. Women were significantly less likely to die in the hospital than men. This is consistent with female gender being associated with diseases with relatively lower mortality, including treatment of gynecologic conditions.

Data from several hospitals obtained during the HAART era demonstrated higher hospitalization rates for women than for men.11,12,38 National data obtained at the beginning of the HAART era also showed a similar gender differential.39 In those studies, the unit of analysis was the individual patient, whereas hospitalization was the unit of analysis in the current study. Most patients with HIV are men, and the lower proportion of hospitalizations for women in this study may reflect the relative numbers of men and women infected with HIV.

Previous studies have demonstrated higher overall hospitalization rates for HIV-positive nonwhites than for whites.30,40-42 As with gender, the current findings for racial/ethnic differences based on hospitalizations as the unit of analysis are not directly comparable to analysis based on patients as the unit of analysis. For HIV-related hospitalizations, however, racial/ethnic differences depend, in part, on specific conditions; for some conditions, such as liver-related complications and IDU-related complications, hospitalization rates for African Americans were lower than for whites.

Previous data demonstrating racial/ethnic disparities in access to HAART39,43 would lead to the expectation that the admission rate for opportunistic illnesses for African Americans would exceed that of whites. In this context, the nonsignificant difference between whites and African Americans in hospitalizations for opportunistic illnesses was unexpected. This finding could be attributable to less advanced disease in those of minority ethnicity; however, we were unable to adjust for CD4 cell count, because our data set did not provide clinical information. Future studies need to evaluate the relation between race and hospitalizations for opportunistic illnesses.

Hospitalizations for specific diagnoses were consistent with epidemiologic data on prevalence of these conditions. Hospitalizations for diabetes, pneumonia, and cerebrovascular complications were more likely to involve African Americans than whites.44 In contrast, hospitalizations for complications of IDU, ischemic heart disease, and chronic HCV were less frequent among African Americans, Hispanics, and Native Americans.45

Older age was associated with a higher proportion of hospitalizations for each diagnostic category, with the exception of opportunistic illness, as well as greater in-hospital mortality. Diseases associated with increasing age, such as diabetes and cardiovascular and cerebrovascular disease, are becoming more prevalent in the HIV population.21,22,44,46,47 The trend of a higher proportion of hospitalizations involving HIV patients older than 50 years of age in more recent years (10% in 1996 vs. 15% in 2000) may presage increases in the number of hospitalizations attributable to diseases associated with increasing age.

Hospitalizations covered by private insurance decreased proportionately more than those covered by Medicaid (45% vs. 31%), but hospitalizations for persons covered by Medicare increased between 1998 and 2000. These results could have significant cost implications for publicly funded programs if the number of HIV patients qualifying for Medicare continues to increase.32

Hospitalizations for patients with Medicaid and Medicare coverage had lower in-hospital mortality than hospitalizations for those with private insurance. It is possible that patients with private coverage were more integrated into a system of outpatient care and that they were hospitalized only for more serious conditions compared with those with public coverage. The difference between self-pay (ie, uninsured) and private insurance was not significant for in-hospital mortality.

This study has some limitations. First, the same patient could have been admitted multiple times and would therefore be counted more than once in the analysis. Unfortunately, the database did not include unique patient identifiers that would enable linking multiple hospitalizations for the same patient. Ignoring such clustering could result in standard errors that were too low. In most cases, however, the z tests of the significance of the coefficients reported in Tables 3 and 4 were 3.0 or higher, suggesting that increasing the standard error by as much as 50% would not alter most conclusions (4 coefficients with significant z values below 3.0 are marked with asterisks in Table 3). Moreover, from the standpoint of examining aggregate trends in inpatient resource utilization, the aggregate temporal, demographic, and clinical differences do provide important utilization information, even if they include multiple hospitalizations for the same individual.

Another limitation concerns the accuracy of coding. We identified HIV-related hospitalizations by the presence of specific ICD-9 codes as primary or secondary diagnoses. It is possible that some HIV-related hospitalizations were not assigned ICD-9 codes reflecting HIV infection. In addition, we were unable to assess 30-day postdischarge mortality. Also, because we did not have clinical data, we were unable to adjust for severity of illness, which may have resulted in an underestimate of inpatient mortality rates. Finally, the results do not provide direct information on changes in the prevalence of the diagnostic categories over time. There may be an increase in the prevalence of certain conditions that are being treated successfully on an outpatient basis; such conditions would not appear in the HCUP data.

In conclusion, admission diagnoses for HIV-related hospitalizations changed significantly between 1996 and 2000. Consistent with the introduction of HAART, there was a concomitant decrease in opportunistic illnesses, suggesting that HIV patients are benefiting from HAART. Treatment complications of HAART, including diabetes and cardiovascular, and cerebrovascular complications, have increased during this period, but they still comprise a relatively small proportion of the total number of HIV hospitalizations in these states.

Back to Top | Article Outline


The authors acknowledge the state data organizations that participate in the HCUP: California Office of Statewide Health Planning and Development, Colorado Health and Hospital Association, Florida Agency for Health Care Administration, Iowa Hospital Association, Illinois Health Care Cost Containment Council, Kansas Hospital Association, Maryland Health Services Cost Review Commission, New Jersey Department of Health and Senior Services, New York State Department of Health, Pennsylvania Health Care Cost Containment Council, South Carolina Budget and Control Board, and Washington State Department of Health.

Back to Top | Article Outline


1. Palella FJ Jr, Delaney KM, Moorman AC, et al. Declining morbidity and mortality among patients with advanced human immunodeficiency virus infection. HIV Outpatient Study Investigators. N Engl J Med. 1998;338:853-860.
2. Paul S, Gilbert HM, Ziecheck W, et al. The impact of potent antiretroviral therapy on the characteristics of hospitalized patients with HIV infection. AIDS. 1999;13:415-418.
3. Mocroft A, Barry S, Sabin CA, et al. The changing pattern of admissions to a London hospital of patients with HIV: 1988-1997. Royal Free Centre for HIV Medicine. AIDS. 1999;13:1255-1261.
4. Baum SE, Morris JT, Gibbons RV, et al. Reduction in human immunodeficiency virus patient hospitalizations and nontraumatic mortality after adoption of highly active antiretroviral therapy. Mil Med. 1999;164:609-612.
5. Mouton Y, Alfandari S, Valette M, et al. Impact of protease inhibitors on AIDS-defining events and hospitalizations in 10 French AIDS reference centres. Federation National des Centres de Lutte contre le SIDA. AIDS. 1997;11(Suppl):F101-F105.
6. Torres RA, Barr M. Impact of combination therapy for HIV infection on inpatient census. N Engl J Med. 1997;336:1531-1532.
7. Beck EJ, Mandalia S, Williams I, et al. Decreased morbidity and use of hospital services in English HIV-infected individuals with increased uptake of anti-retroviral therapy 1996-1997. National Prospective Monitoring System Steering Group. AIDS. 1999;13:2157-2164.
8. Altes J, Guadarrama M, Force L, et al. The impact of highly active antiretroviral therapy on HIV-related hospitalizations in 17 county hospitals in Catalonia, Spain. Catalonian County Hospitals HIV Infection Study Group. AIDS. 1999;13:1418-1419.
9. Paul S, Gilbert HM, Lande L, et al. Impact of antiretroviral therapy on decreasing hospitalization rates of HIV-infected patients in 2001. AIDS Res Hum Retroviruses. 2002;18:501-506.
10. Spira R, Marimoutou C, Binquet C, et al. Rapid change in the use of antiretroviral agents and improvement in a population of HIV-infected patients: France, 1995-1997. Groupe d'Epidemiologie Clinique du SIDA en Aquitaine (GECSA). J Acquir Immune Defic Syndr Hum Retrovirol. 1998;18:358-364.
11. 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.
12. Gebo KA, Diener-West M, Moore RD. Hospitalization rates differ by hepatitis C status in an urban HIV cohort. J Acquir Immune Defic Syndr. 2003;34:165-173.
13. Martin-Carbonero L, Soriano V, Valencia E, et al. Increasing impact of chronic viral hepatitis on hospital admissions and mortality among HIV-infected patients. AIDS Res Hum Retroviruses. 2001;17:1467-1471.
14. Gardner LI, Klein RS, Szczech LA, et al. Rates and risk factors for condition-specific hospitalizations in HIV-infected and uninfected women. J Acquir Immune Defic Syndr. 2003;34:320-330.
15. Soriano V, Martin-Carbonero L, Garcia-Samaniego J, et al. Mortality due to chronic viral liver disease among patients infected with human immunodeficiency virus. Clin Infect Dis. 2001;33:1793-1795.
16. Mocroft A, Monforte A, Kirk O, et al. Changes in hospital admissions across Europe: 1995-2003. Results from the EuroSIDA study. HIV Med. 2004;5:437-447.
17. Heath KV, Hogg RS, Chan KJ, et al. Lipodystrophy-associated morphological, cholesterol and triglyceride abnormalities in a population-based HIV/AIDS treatment database. AIDS. 2001;15:231-239.
18. Behrens G, Schmidt H, Meyer D, et al. Vascular complications associated with use of HIV protease inhibitors. Lancet. 1998;351(9120):1958.
19. Gallet B, Pulik M, Genet P, et al. Vascular complications associated with use of HIV protease inhibitors. Lancet. 1998;351:1958-1959.
20. Vittecoq D, Escaut L, Monsuez JJ. Vascular complications associated with use of HIV protease inhibitors. Lancet. 1998;351(9120):1959.
21. Moore RD, Keruly JC, Lucas GM. Increasing incidence of cardiovascular disease in HIV-infected persons in care [abstract]. Presented at: 10th Conference on Retroviruses and Opportunistic Infections; 2003; Boston.
22. The Data Collection on Adverse Events of Anti-HIV Drugs (DAD) Study Group. Combination antiretroviral therapy and the risk of myocardial infarction. N Engl J Med. 2003;349(21):1993-2003.
23. Klein D, Hurley LB, Quesenberry CP, et al. Do protease inhibitors increase the risk for coronary heart disease in patients with HIV-1 infection? J Acquir Immune Defic Syndr. 2002;30:471-477.
24. Currier JS, Taylor A, Boyd F, et al. Coronary heart disease in HIV-infected individuals. J Acquir Immune Defic Syndr. 2003;33:506-512.
25. Holmberg SD, Moorman AC, Greenberg AE, et al. The DAD Steering Committee. Trends in rates of myocardial infarction among patients with HIV. N Engl J Med. 2004;350:730-732.
26. Mary-Krause M, Cotte L, Simon A, et al. Increased risk of myocardial infarction with duration of protease inhibitor therapy in HIV-infected men. AIDS. 2003;17:2479-2486.
27. Holmberg SD, Moorman AC, Williamson JM, et al. Protease inhibitors and cardiovascular outcomes in patients with HIV-1. Lancet. 2002;360:1747-1748.
28. Bozzette SA, Ake CF, Tam HK, et al. Cardiovascular and cerebrovascular events in patients treated for human immunodeficiency virus infection. N Engl J Med. 2003;348:702-710.
29. Steiner C, Elixhauser A, Schnaier J. The healthcare cost and utilization project: an overview. Eff Clin Pract. 2002;5:143-151.
30. Fleishman JA, Hellinger FJ. Trends in HIV-related inpatient admissions from 1993 to 1997: a seven-state study. J Acquir Immune Defic Syndr. 2001;28:73-80.
31. Fleishman JA, Hellinger FH. Recent trends in HIV-related inpatient admissions 1996-2000: a 7-state study. J Acquir Immune Defic Syndr. 2003;34:102-110.
32. Fasciano NJ, Cherlow AL, Turner BJ, et al. Profile of Medicare beneficiaries with AIDS: application of an AIDS casefinding algorithm. Health Care Financ Rev. 1998;19:19-38.
33. Rosenblum L, Buehler JW, Morgan MW, et al. HIV infection in hospitalized patients and Medicaid enrollees: the accuracy of medical record coding. Am J Public Health. 1993;83:1457-1459.
34. Centers for Disease Control and Prevention. 1993 Revised classification system for HIV infection and expanded surveillance case definition for AIDS among adolescents and adults. JAMA. 1993;269:729-730.
35. Centers for Disease Control and Prevention. HIV/AIDS Surveillance Supplemental Report. Atlanta: Centers for Disease Control and Prevention; 2001:7.
36. Long JS. Regression Models for Categorical and Limited Dependent Variables. Thousand Oaks, CA: Sage Publications; 1997.
37. Betz ME, Gebo KA, Barber E, et al. Patterns of diagnosis in hospital admissions in a multi-state cohort of HIV+ adults in 2001. Med Care. (in press).
38. Floris-Moore M, Lo Y, Klein RS, et al. Gender and hospitalization patterns among HIV-infected drug users before and after the availability of highly active antiretroviral therapy. J Acquir Immune Defic Syndr. 2003;34:331-337.
39. Shapiro MF, Morton SC, McCaffrey DF, et al. Variations in the care of HIV-infected adults in the United States: results from the HIV Cost and Services Utilization Study. JAMA. 1999;281:2305-2315.
40. Crystal S, Lo Sasso AT, Sambamoorthi U. Incidence and duration of hospitalizations among persons with AIDS: an event history approach. Health Serv Res. 1999;33:1611-1638.
41. Fleishman JA, Mor V, Laliberte LL. Longitudinal patterns of medical service use and costs among people with AIDS. Health Serv Res. 1995;30:403-424.
42. The HIV Research Network. Hospital and outpatient health services utilization among HIV-infected patients in care in 1999. J Acquir Immune Defic Syndr. 2002;30:21-26.
43. Gebo K, Fleishman J, Conviser R, et al. Racial and gender disparities in receipt of HAART persist in a multistate sample of HIV patients in 2001. J Acquir Immune Defic Syndr. 2005;38:96-103.
44. Centers for Disease Control and Prevention. National Diabetes Fact Sheet: General Information and National Estimates on Diabetes in the United States, 2003. Atlanta: US Department of Health and Human Services; 2003.
45. Substance Abuse and Mental Health Services Administration. Results from the 2002 National Survey on Drug Use and Health: National Findings (Office of Applied Studies, NHSDA Series H-22, DHHS Publication No. SMA 03-3836). Rockville, MD: Department of Health and Human Services; 2003.
46. Carr A, Samaras K, Burton S, et al. A syndrome of peripheral lipodystrophy, hyperlipidaemia and insulin resistance in patients receiving HIV protease inhibitors. AIDS. 1998;12(Suppl):F51-F58.
47. Hommes MJ, Romijn JA, Endert E, et al. Insulin sensitivity and insulin clearance in human immunodeficiency virus-infected men. Metabolism. 1991;40:651-656.

HIV; hospitalization; highly active antiretroviral therapy; liver disease; injection drug use

© 2005 Lippincott Williams & Wilkins, Inc.