Kerr, Jelani C. PhD, MSPH*; Stephens, Terri G. MSPH†; Gibson, James J. MD, MPH†; Duffus, Wayne A. MD, PhD†,‡
Primary care engagement (PCE) and retention after new HIV diagnosis is critical for improving survival for people living with HIV/AIDS (PLWHA). Early detection, prompt medical linkage, retention, and adherence to treatment significantly reduces mortality risk and health care expenditure.1–5 However, approximately one-fifth of PLWHA utilize hospitals and many are not linked to or retained in primary care.6–8 Inpatient hospitalization for HIV disease complications is costly and avoidable.5,9,10 Though inpatient costs have declined because highly active antiretroviral therapy (HAART) introduction, health care expenditure for people with low CD4+T- cell counts is higher than persons of less disease severity.5,11
There is little research examining relationships between HIV PCE and inpatient hospitalization. Pezzin and Fleishman12 used pre-HAART era data to examine PCE, hospitalization, and disease severity but overall did not find significant differences in ambulatory care use and inpatient treatment. Fleishman et al13 examined these relationships in urban HIV clinics and found that primary care and inpatient use may be associated with disease stage (sicker patients use more inpatient and outpatient care). However, this study was restricted to persons attending specific urban clinics and possible treatment occurrences outside of these facilities were not captured.
There is a paucity of literature utilizing complete data (capturing all medical encounters with the health care system) to examine HIV PCE and inpatient hospitalization for PLWHA though access and availability challenges may exist for individuals in rural locations. Finally, as incidence and prevalence of HIV disease increases in African-American communities, identifying characteristics that might improve patient outcomes and address disparities become critical.
This investigation overcomes limitations of previous studies by including all CD4+T cell and viral load (VL) values from mandatory reports and all statewide inpatient medical encounters including rural and urban settings; maintaining participants through study entirety; and sampling from a severely impacted population, African-Americans. The objective is to determine risk factors associated with both HIV PCE and inpatient hospitalization.
The South Carolina (SC) enhanced HIV/AIDS Reporting System (eHARS) and uniformed hospital billing data from the SC Budget and Control Board, Office of Research and Statistics (ORS) were used for this study. Institutional Review Boards for the SC Department of Health and Environmental Control (DHEC) and the ORS Data Oversight Committee approved this study.
HIV infection has been reportable by name to the SC DHEC since 1986. Data quality exceeds CDC minimum standards of reporting timeliness and completeness (T. Stephens, MSPH, SC-DHEC unpublished data, December 2009).14,15 Surveillance data quality has been improved because deceased individuals were excluded by matching to the National Death Index and the Social Security Death Master Files which minimizes miscategorization of care engagement categories and was performed 1 month before data request. Beginning January 1, 2004, state law required reporting of all CD4+T-cell and VL values to SC DHEC and recorded in eHARS. Demographics (sex, race/ethnicity, age), HIV and AIDS (when applicable) diagnosis date, behavioral risk (heterosexual, males having sex with males, injection drug use, no identified risk/no risk reported), CD4+ T cell and VL tests dates and results, and county of diagnosis (rural, urban) were extracted from eHARS.
ORS provided uniform billing data from all 62 SC inpatient facilities and all medical encounters. Health care data included International Classification of Diseases (ICD-9)16 codes, inpatient admission and discharge dates, payment source at last visit [self-payment, Medicare, Medicaid, private insurance, indigent/charity (Hill-Burton, Medically Indigent Assistance Fund)], Other government sources (eg, Tricare, Department of Corrections), and Health Maintenance Organizations.
Surveillance and ORS data were linked by name, date of birth, sex, race, social security number, and address. Data were deidentified (name, address, and social security number removed), and individuals were assigned unique identification numbers before data distribution. Linkage was conducted in a secure locale by ORS staff trained in Health Insurance Portability and Accountability Act and ORS confidentiality procedures.
Eligible individuals for this retrospective cohort study were HIV infected, age ≥18 years old, SC residents, diagnosed between 1987 and 2006, and alive at the end of the study. Individuals were excluded if they had no inpatient admissions between January 1, 2007, to June 15, 2010 post-HIV diagnosis, or diagnosed before February 1, 1986 or after December 31, 2006. Individuals with inpatient stays of less than 1 day were eliminated (n = 26). Pregnant women were excluded to eliminate confounding associated with obstetric care (n = 81). ICD-9 codes (042, V08) assigned at discharge were not ranked in frequency because individuals are known to be HIV infected at study entry. All other diagnostic codes were retained. Individuals from other race/ethnicities were excluded due to small numbers (n = 60). The final population comprised 2454 individuals.
Variable Definitions and Categorizations
The 2010 Department of Health and Human Services treatment guidelines17 recommends clinical HIV care at 3-month to 4-month intervals on initial entry to care. However, patients adherent to antiretroviral therapy, exhibiting suppressed VL, and clinically/immunologically stable for 2–3 years may receive CD4+T cell/VL testing at 6-month intervals.
CD4+T-cell/VL testing was used as a proxy for clinical visits because these tests are ordered by HIV care providers to determine antiretroviral therapy initiation and monitor treatment response. PCE was evaluated from January 1, 2007, to June 15, 2010, and categorized into 4 groups: (1) not-in-care (NIC) if individuals had no CD4+T-cell and/or VL reports during the study period; (2) low transitional care (low TC) if individuals had 1–3 of 7 intervals with at least 1 CD4+T-cell and/or VL report; (3) high transitional care (high TC) if 4–6 intervals with at least 1 CD4+T-cell and/or VL report; and (4) in-care (IC) if having a CD4+T-cell and/or VL report every 6 months. Suboptimal PCE was defined as not being IC. Tests performed during hospitalizations were excluded.
Number of inpatient visits was a count of separate hospital admission dates for each individual. Number of inpatient days was determined using time elapsed between admission and discharge dates and calculating combined lengths of stay per admission.
Predictor variables were as follows: race/ethnicity, age, sex, behavioral risk factor, urban/rural residence, reason for visit [AIDS-defining illness (ADI)] versus non-ADI (hypertension, diabetes), payment source, and disease stage [AIDS within 1 year of HIV diagnosis (AIDS ≤ 1 year); AIDS more than 1 year post-HIV diagnosis (AIDS > 1 year); or HIV-only]. Reason for visit was based on ICD-9 codes assigned at discharge.
A dual analysis strategy was implemented to describe patterns of PCE and inpatient hospitalization among PLWHA populations and to illuminate the relationship between PCE and inpatient hospitalization. Specifically, we assessed factors affecting PCE, identified factors influencing inpatient hospitalization, and determined the effect of PCE on frequency and duration of inpatient hospitalization.
Descriptive statistics were generated to observe the distribution of PCE behavior by predictor variables. The χ2 tests assessed likelihood of association. Univariate and multivariate logistic regression examined differences in PCE by predictor variables. A multivariate logistic model reflected the most appropriate model fit. Disease duration was controlled for because it is differentially associated with treatment seeking.18
Poisson regression was used to detect differences in number of visits by aforementioned independent variables, and negative binomial regression detected differences in number of inpatient days. Poisson was utilized because standard linear regression was inappropriate, and negative binomial regression was used because Poisson indicated overdispersion. Outliers reducing model fit accuracy were eliminated.
Median CD4+T cell and VL were calculated for IC, high TC, and low TC. NIC measurements were excluded because of no/outdated reports to surveillance. Frequencies and percentages of inpatient diagnoses were generated for each PCE level. Analysis was performed using SAS 9.2. (SAS Institute Inc, Cary, NC).
The study population was predominantly black (81.2%), 45–54 years old (36.4%), male (61.4%), heterosexual (37.2%), urban (67.3%), receiving Medicare (23.6%), no ADI diagnosis during hospitalizations (64.9%), and developed AIDS >1 year (46.9%) (Table 1). Approximately, 35% of the population was IC, 41% high TC, 18% low TC, and 6% NIC.
Care Pattern of Inpatients by Selected Characteristics
Blacks were more likely to be low TC [adjusted odds ratio (aOR): 1.77] than whites (Table 2). Individuals with an ADI at admission were more likely to be low TC (aOR: 1.73) and high TC (aOR: 1.34) compared with non-ADI individuals. Self-payers were more likely to be low TC (aOR: 1.80) than the privately insured. Medicare recipients had decreased likelihood of being NIC (aOR: 0.48) and high TC (aOR: 0.76). Individuals with an AIDS diagnosis were less likely to be NIC [AIDS ≤ 1 year (aOR: 0.07); AIDS > 1 year (aOR: 0.17)], and individuals with AIDS ≤1 year were less likely to be low TC (aOR: 0.43) compared with HIV-only individuals.
TABLE 2-a Multinomia...Image Tools
Frequency of Inpatient Admissions
TABLE 2-b Multinomia...Image Tools
NIC persons had the fewest median inpatient visits (1), and all other PCE groups had median visits of 2 (Table 3). Individuals NIC (incidence rate ratio [IRR]: 0.77), low TC (IRR: 0.82), and self-payers (IRR: 0.63) had fewer inpatient days than their reference groups (Table 4). Blacks (IRR: 1.08), injection drug users (IRR: 1.11), individuals presenting with ADIs (IRR: 1.74), Medicare recipients (IRR: 1.09), Medicaid recipients (IRR: 1.10), other government insurance recipients (IRR: 1.21), HMO recipients (IRR: 1.19), AIDS ≤1 year individuals (IRR: 1.11), and AIDS >1 year individuals (IRR: 1.11) had significantly more inpatient days than their referents.
Total Inpatient Days
NIC persons had the fewest median inpatient days (5), followed by IC (6), and high and low TC (8 for both) (Table 3). Self-payers (IRR: 0.57) had fewer inpatient days than the privately insured (Table 4). Low TC (IRR: 1.13), high TC (IRR: 1.17), blacks (IRR: 1.21), individuals presenting with an ADI (IRR: 2.18), Medicaid recipients (IRR: 1.31), AIDS ≤1 year individuals (IRR: 1.24), and AIDS >1 year individuals (IRR: 1.36) had significantly more inpatient days than their respective referents.
The most frequent inpatient diagnoses for the total sample were hypertension (401.9), tobacco use disorder/nondependent drug abuse (305.1), and pneumonia (486) (Table 5). The most frequent all-cause ICD-9 codes for the general SC population inpatient admissions in 2010 were hypertension (401.9), hyperlipidemia (272.4), diabetes mellitus (250.00), esophageal reflux (530.81), congestive heart failure (428.0), coronary artery disease (414.01), tobacco-use disorder (305.1), urinary tract infection (599.0), atrial fibrillation (427.31), and acute renal failure (584.9), respectively.
Median CD4+T-cell measurements were 421 cells/mm3 for individuals IC, 313 cells/mm3 for High TC individuals and 235 cells/mm3 for Low TC individuals. Median VL measurements were 74 copies/mL for individuals IC, 110 copies/mL for High TC individuals and 1723 copies/mL for Low TC. There were no measurements for NIC persons.
State-wide hospital discharge and HIV surveillance data were used to separately and then jointly characterize the association between HIV PCE and inpatient hospitalization. PCE is differentially associated with inpatient hospitalization, but this relationship is mediated by several factors.
Individuals aged 25–44 were more likely to be NIC and TC (high and low) but had fewer inpatient visits and days. Conversely, older adults typically engage in greater HIV PCE although having more inpatient admissions3,19 and longer stays.4 These inpatient episodes more likely involve chronic conditions (eg, diabetes) and side-effects of HAART than opportunistic infections or HIV disease.5 Although older adults may adopt desirable outpatient and inpatient hospitalization patterns, their health status remains at greater risk because (1) HAART is less effective in older populations,20,21 (2) late-stage diagnosis is more common in older persons, (3) opportunities for comorbid conditions increase,22 and (4) mortality risk is greater.
Consistent with prior research,7,11,12,21–23 blacks were less likely to use primary care but more likely to use inpatient hospitalization. Although blacks are disproportionately affected by HIV/AIDS, treatment access and use among this group may inhibit optimal health outcomes.
Treatment patterns of the uninsured suggest insecure health care access. As demonstrated by this study and others, persons without insurance (self-payers) are less frequently engaged in outpatient care and have fewer admissions and inpatient days.4,13 Self-payers may have less funding for primary care and may be unable to remain in inpatient treatment when subsequently requiring hospitalization. The Ryan White Care Act provides primary care treatment as a “payer of last resort24” however, its ability to meet treatment needs is concerning as several states have waiting lists for AIDS Drug Assistance Programs.25
Persons with government-sponsored health care had greater HIV PCE and inpatient hospitalization. Although public insurance is associated with greater PCE,26,27 patients of lesser disease severity may lack treatment access because Medicaid and Medicare disability benefits are not extended to the relatively immunocompetent.21,28 Thus, public insurance is associated with advanced disease stage20 and lengthy hospitalizations.22 Provisions in the Affordable Care Act (ie,broadened Medicaid eligibility standards, insurance provisions for persons with pre-existing conditions, extending parental insurance until age 26) may help expand care to groups lacking health insurance.29,30
Theoretically, consistent PCE decreases inpatient hospitalization because care retention and medication adherence improves immunological status and decreases morbidity.30,31 However, persons with an AIDS diagnosis had greater inpatient and PCE in this study. There are perhaps a few following reasons for this: (1) persons of advanced disease stage may be immunocompromised to a level that necessitates both inpatient and outpatient treatment,13 (2) HAART obtained as a result of outpatient use may cause side-effects (liver damage, cardiovascular disease) prompting inpatient hospitalization,22 or (3) a combination of both. Additional studies are needed to elucidate specific reasons for increased inpatient hospitalization for persons with AIDS though it is probable that social and structural barriers leading to late health care initiation and diagnosis increase health care cost and mortality.32,33
Diminished PCE often leads to ADI development and consequent hospitalization. It should be noted that although persons with AIDS may have acquired an ADI during the HIV disease course, this does not mean that they actually presented to the hospital with an ADI. Research suggests increases in all forms of medical care engagement as disease stage progresses.13 Care engagement patterns among patients presenting with ADI are concerning as they are less likely to engage in HIV primary care but more engaged in costly inpatient hospitalization. Risk of inpatient mortality nearly doubles if a PLWHA is diagnosed with an ADI.6 However, inpatient hospitalizations related to ADIs are often avoidable.13 Hospital practitioners have a unique opportunity to improve survival and avoid future hospitalizations if they recognize people with ADIs as a high-risk group and initiate the process to link them to HIV primary (and if necessary, ancillary) care services upon discharge.
HAART increases life spans of PLWHA, but this presents new challenges to primary care physicians treating this group. As life is prolonged, non–AIDS-related conditions are increasingly implicated causes of mortality. The most frequently assigned ICD-9 codes are similar to the general population and highlight challenges accompanying increased life expectancy. Specifically, cardiovascular disease, the most frequent reason for hospitalization, has become a leading cause of death among PLWHA.34 Further complicating matters is increased tobacco/drug use (which negatively affects cardiovascular health) within this population.35 For health care practitioners, this may introduce treatment challenges as HIV disease progression, lifestyle factors, and HAART increase cardiovascular disease risk. It is important for HIV treatment providers to maintain a comprehensive treatment focus, including traditional primary care disease prevention, for HAART benefit to be fully realized.
This study has limitations. CD4+T-cell/VL measurements were used as a proxy for PCE. This does not inform us about treatment regimen, adherence, clinical failure due to drug resistance, and how these impact inpatient hospitalization. Second, results may not be generalizable to other locales, and patients may be hospitalized out of state. Third, although eHARS and ORS data are considered comprehensive, all health care visits and diagnostic codes may not have been reported. Finally, diagnostic codes may not always accurately reflect the true reason(s) an individual was admitted for inpatient hospitalization.
The association between HIV PCE and inpatient hospitalization is contingent upon disease stage, sociodemographics, and health insurance access and type. Further investigation is needed to quantify the proportion of preventable inpatient hospitalizations attributable to these characteristics. Nevertheless, implementation of effective screening and linkage strategies should prevent disease progression and hospitalization. Developing optimal behaviors early in HIV care provides a constructive foundation to build a lifetime of care engagement and reduce preventable health conditions.
1. Kitahata MM, Gange SJ, Abraham AG, et al.. Effect of early versus deferred antiretroviral therapy for HIV on survival. N Engl J Med. 2009;360:1815–1826.
2. Walensky RP, Paltiel AD, Losina E, et al.. The survival benefit of AIDS treatment in the United States. J Infect Dis. 2006;194:11–19.
3. Shapiro MF, Morton SC, McCaffrey DF, et al.. Variations in the care of the HIV-infected adults in the United States: results from the HIV Cost and Services Utilization Study. JAMA. 1999;281:2305–2315.
4. Giordano TP, Gifford AL, White AC, et al.. Retention in care: a challenge to survival with HIV infection. Clin Infect Dis. 2007;44:1493–1499.
5. Bozzette SA, Joyce G, McCaffrey DF, et al.. Expenditures for the care of HIV-infected patients in the era of highly active antiretroviral therapy. N Engl J Med. 2001;334:817–823.
6. Gebo KA, Fleishman JA, Moore RD. Hospitilizations for metabolic conditions, opportunistic infections, and injection drug use among HIV patients; trends between 1996 and 2000 in 12 states. J Acquir Immune Defic Syndr. 2005;40:609–616.
7. Fleishman JA, Gebo KA, Reilly ED, et al.. Hospital and outpatient health services utilization among HIV-infected adults in care 2000–2002. Med Care. 2005;43(9 suppl):40–52.
8. Metsch LR, Bell C, Pereyra M, et al.. Hospitalized HIV-infected patients in the era of highly active antiretroviral therapy. Am J Public Health. 2009;99:1045–1049.
9. Crum-Cianflone NF, Grandits G, Echols S, et al.. Trends and causes of hospitalizations among HIV-infected Persons during the late HAART era: what is the impact of CD4 counts and HAART use. J Acquir Immune Defic Syndr. 2010;54:248–257.
10. Beck EJ, Harling G, Gerbase S, et al.. The cost of treatment and care for individuals living with HIV infection: implications of published studies, 1999-2008. Curr Opin HIV AIDS. 2010;5:215–224.
11. Chen RY, Accourt NA, Westfall AO, et al.. Distribution of healthcare expenditures for HIV-infected patients. Clin Infect Dis. 2006;42:1003–1010.
12. Pezzin LE, Fleishman JA. Is outpatient care associated with lower use of inpatient and emergency care? An analysis of persons with HIV disease. Acad Emerg Med. 2003;10:1228–1238.
13. Fleishman JA, Moore RD, Conviser R, et al.. Associations between outpatient and inpatient service use among persons with HIV infection: a positive or negative relationship? Health Serv Res. 2008;43:76–95.
15. Jones JL, Rion P, Hermann P, et al.. South Carolina Department of Health and Environmental Control Columbia. Improvement in AIDS case reporting, South Carolina. JAMA. 1991;265:356.
16. National Center for Health Statistics and Centers for Medicare and Medicaid Services. The International classification of diseases, 9th revision, clinical classification 2008. Available at: http://icd9cm.chrisendres.com/
. Accessed March 15, 2010.
17. Panel on Antiretroviral Guidelines for Adults and Adolescents. Guidelines for the Use of Antiretroviral Agents in HIV-1-Infected Adults and Adolescents. Washington, DC: Department of Health and Human Services; 2009:1–161. Available at: http://www.aidsinfo.nih.gov/ContentFiles/AdultandAdolescentGL.pdf
. Accessed May 19, 2010.
18. Samet JH, Freedberg KA, Stein MD, et al.. Trillion virion delay: time from testing positive for human immunodeficiency virus to presentation for primary care. Arch Intern Med. 1998;158:734–740.
19. Karon JM, Buehler JW, Byers RH, et al.. Projection of the number of persons diagnosed with AIDS and the number of immunosuppressed HIV-infected persons—United States 1992–1994. MMWR Recomm Rep. 1992;41:1–29.
20. Projection of the number of persons diagnosed with AIDS and the number of immunosuppressed HIV-infected persons—United States 1992–1994. MMWR Morb Mortal Wkly Rep. 1992;41:1–29.
21. Yehia BR, Fleishman JA, Hicks PL, et al.. Inpatient health services utilization among HIV-infected adult patients in care 2002–2007. J Acquir Immune Defic Syndr. 2010;53:397–404.
22. 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.
23. Meyerson BE, Klinkenberg WD, Perkins DR, et al.. Who's in and who's out: use of primary medical care among individuals living with HIV. Am J Public Health. 2007;97:744–749.
26. Agency for Health Care Policy and Research. Improving health care for rural populations; research in action. Available at: www.ahrq.gov/research/rural.htm
. Accessed January 8, 2011.
27. Zhang W, Mueller KJ, Chen L. Uninsured hospitalizations: rural and urban differences. J Rural Health. 2008;24:194–202.
28. Anthony MN, Gardner L, Marks G, et al.. Factors associated with use of HIV primary care among persons recently diagnosed with HIV: examination of variables from the behavioral model of health-care utilization. AIDS Care. 2007;19:195–202.
30. Ahluwalia IB, Bolen J. Lack of health insurance coverage among working-age adults, evidence from the behavioral risk factor surveillance system, 1993–2006. J Community Health. 2008;33:293–296.
31. Billings J, Zeitel L, Lukomnik J, et al.. Impact of socioeconomic status on hospital resource use in New York City. Health Aff. 1993;12:162–173.
32. Torian LV, Wiewel EW, Liu K, et al.. Risk factors for delayed initiation of medical care after diagnosis of Human Immunodeficiency Virus. Arch Intern Med. 2008;168:1181–1187.
33. Losia E, Schackman BR, Sadownik SN, et al.. Racial and sex disparities in life expectancy losses among HIV-infected persons in the United States: impact of risk behavior, late initiation, and early discontinuation on antiretroviral therapy. Clin Infect Dis. 2009;49:1570–1578.
34. Sackoff JE, Hanna DB, Pfeiffer MR, et al.. Causes of death among persons with AIDS in the era of highly active antiretroviral therapy: New York City. Ann Intern Med. 2006;145:397–406.
35. Nahvi S, Cooperman NA. Review: The need for smoking cessation among HIV-positive smokers. AIDS Educ Prev. 2009;21(suppl 3):14–27.
© 2012 Lippincott Williams & Wilkins, Inc.