A Comparison of Partner Notification Effectiveness in African-,Caribbean-,and United States-Born HIV-Infected Blacks in New York City

Introduction

N ationwide and in New York City (NYC), blacks are disproportionately impacted by human immunodeficiency virus (HIV) infection. ^1,2 While blacks comprised only about 23% of New Yorkers, ³ they represented 47.7% (N=1661) of total HIV diagnoses in 2010. ⁴ To help address this health disparity, the NYC Department of Health and Mental Hygiene (DOHMH) established the Field Services Unit (FSU) within the Bureau of HIV Prevention and Control in the summer of 2006. ⁵ FSU collaborates with large hospitals and other community clinics located in areas with high HIV morbidity and mortality and frequent late diagnosis of HIV infection, where more than 60% of reported cases are among blacks. ⁵

FSU has a team of experienced disease intervention specialists (DIS) stationed at participating clinical sites to work with providers and HIV-infected individuals to offer HIV Partner Services (PS). PS is the process of identifying HIV-infected individuals, eliciting from them the names of their sex or needle-sharing partners for exposure notification, and testing notified partners for HIV infection. FSU's delivery of PS in clinical and community settings has had a high acceptance rate both among providers, as well as patients. ⁶ In addition, FSU's experience supports the growing evidence that HIV partner notification is an effective method to identify new cases of HIV. ^{5,7 –9} Since implementing HIV field testing of notified partners in 2008, the FSU has observed a significant increase in the number of notified partners who tested for HIV, as well as more than a twofold increase in the number of individuals newly diagnosed with HIV at participating sites. ¹⁰

Despite FSU's success, PS has been shown to be less effective among black New Yorkers than among other racial or ethnic groups, particularly Hispanics, the second largest case group in NYC. Between July 2006 and November 2007, the FSU interviewed 894 blacks and 491 Hispanics, and observed a significantly lower new HIV case finding among blacks, mainly due to fewer elicited partners per interview among blacks. ¹¹ Lower partner elicitation rates for blacks may be related to distrust of FSU staff as some patients may believe in governmental AIDS conspiracy theories. ^12,13 Evidence exists that cultural and other differences may affect access to care and adherence to HIV treatment among blacks and may be playing a role here as well. ^{14 –16}

Because NYC blacks are a culturally heterogeneous group that includes African Americans, and Caribbean and African immigrants, an examination of PS outcomes by country of origin may deepen our understanding of cultural factors that may be related to the differences in PS outcomes observed among blacks, and also help to identify opportunities for improved PS effectiveness by subgroup.

Methods

Data collected by the DOHMH HIV FSU was used for this analysis. Laboratory-related information on patients' CD4 T-cell counts and HIV plasma viral load was primarily obtained from the HIV/AIDS Reporting System (HARS). The NYC HARS is a population-based registry of all persons diagnosed with AIDS in NYC since 1981 and HIV since 2000. Only HIV-infected index patients interviewed for PS from July 2006 to December 2008 whose self-reported race/ethnicity is non-Hispanic black were included. Index patients were categorized into three mutually exclusive groups based on birthplace: the United States (US) and its territories, sub-Saharan Africa, or the Caribbean. Data were collected for routine PS program purposes through interviews and medical record abstractions conducted by Disease Intervention Specialists (DIS). Information included demographics, including birth country and number of years living in the US, HIV transmission risk (e.g., sexually transmitted infection history, drug use, and sexual practices), HIV-testing history, HIV-related laboratory data, and names and contact information of exposed sex and needle-sharing partners. HIV diagnosis date was defined as the earliest-known diagnosis date obtained from patient interviews, chart review, and/or the HARS.

All PS outcomes were recorded using a standardized form, which includes information on whether a partner agrees to HIV testing at time of notification, and these outcomes are entered into the FSU database. The PS outcomes compared among the three groups were: proportions of named, notified, and tested partners among all partners reported; proportions of partners with a new diagnosis of HIV infection as a result of PS among all notified partners; and the number of index cases that needed to be interviewed to identify one new HIV case. Data were analyzed using SAS software, version 9.1 (SAS Institute, Inc., Cary, NC), and the χ² test was used to compare the groups' laboratory values at initial diagnosis (i.e., CD4 and viral load) and PS outcomes. In African and Caribbean blacks, the effect of length of stay in the US on PS outcomes was assessed using binary logistic regression.

Results

PS interviews were conducted from July 2006 to December 2008 with 1560 non-Hispanic black individuals. Of these, 1502 were born in the US and its territories, sub-Saharan Africa, or the Caribbean; 45 were born elsewhere; and country of birth was not available for 13. Among the 45 non-Hispanic blacks born outside of the US, sub-Saharan Africa, and the Caribbean, 32 (71%) were born in Guyana and none were from areas of Africa outside sub-Saharan Africa. Of the 1502 individuals from the three regions of interest, 1049 (69.8%) were African Americans, 285 (19.0%) were Caribbean-born, and 168 (11.2%) were born in sub-Saharan Africa (Table 1).

Compared with African Americans, a higher proportion of Africans were female (53.6% vs. 43.1%, p=0.01), and compared to African-American females (n=452), a higher proportion of African females reported current pregnancy at PS interview (27.8% vs. 13.7%, p=0.0009). Africans were more likely to report a married or co-habiting status than African Americans (38.7% vs. 11.1%, p<0.0001) and were more likely to be widowed (p=0.04). Compared with Caribbeans, Africans reported living in the US for a shorter time: 6.3 years versus 13.9 years.

With respect to HIV transmission risk, Africans and Caribbeans were less likely to report an injection drug use history than African Americans (0.6%, 0.4%, vs. 9.0%, respectively) or being men who have sex with men (1.8%, 13.3%, vs. 25.6%, respectively). Africans had the highest proportion of individuals with no identified HIV transmission risk (83.9% vs. 39.2% for African Americans and 62.8% for Caribbeans, Table 1).

All groups had a high proportion of cases with an AIDS-defining CD4 count at initial HIV diagnosis (30.8%, 34.5%, 47.9% for African American, Africans, and Caribbeans, respectively, vs. 23.4% among all newly-diagnosed New Yorkers citywide in 2008). ² Caribbeans were more likely to have an AIDS-defining CD4 count at initial diagnosis than African Americans (47.9% vs. 30.8%, p<0.0001) or Africans (34.5%, p=0.03) (Table 2).

Although African Americans were less likely to accept a PS interview, they were more likely to name sex and/or needle-sharing partners for notification than Africans (56.6% vs. 48.2%, p=0.04) and were more likely to name partners who were previously diagnosed with HIV than Africans (20.4% vs. 11.7%, p=0.03) and Caribbeans (12.2%, p=0.0009). African Americans had a lower overall partner notification rate (defined as proportion of partners notified among all partners named with negative or unknown HIV status) than both Africans (33.1% vs. 43.9%, p=0.03) and Caribbeans (38.6%, p=0.09). Notified partners of African American index patients were more likely to take an HIV test than partners of Caribbeans (70.7% vs. 58.3%, p=0.02). Partner index (i.e., the number of named partners for each person interviewed) was lowest among Africans.

Binary logistic regression was used to assess the effect of US residence length on the likelihood of naming partners for African and Caribbeans, and no statistically significant effect was observed (OR=1.07, 95% CI: 0.77–1.49).

Discussion

To our knowledge, this is the first study to compare PS outcomes among HIV-infected blacks by region of birth. Overall, the number needed to interview to find one newly diagnosed case was high and similar among US-, African-, and Caribbean-born index cases. While we observed some differences in the rates of partner elicitation, notification, and HIV testing among the groups, we did not observe a significant difference in the rate of new HIV case identification. High frequencies of AIDS-defining CD4 levels at initial HIV diagnosis consistent with late diagnosis were observed in all three groups, but the frequency was highest among Caribbeans. New HIV cases were found among all groups, indicating NYC blacks benefit from current Partner Services outreach. However, further work is needed to improve HIV case finding from Partner Services among blacks, and additional measures to promote early diagnosis are needed.

Africans were less likely to name HIV-exposed partners, but because of higher partner notification and testing rates among their named partners, their overall PS outcomes remained similar to that of the other two groups. While it is possible Africans may in fact have fewer partners than the other groups, here we find them naming on average much fewer than one partner per index patient (0.66 per index), indicating that partners are going unnamed. Only one African reported an intravenous drug use history, suggesting that for the rest, the most likely route of infection was sexual contact that would require a sexual partner. One potential explanation for recent African immigrants naming fewer partners is that many of their partners lived in their home countries and thus were considered not notifiable by the index patients or worth naming on that basis. This explanation does not entirely account for the findings, since those in the US longer were not more likely to name partners, although they would presumably have more US-based partners to name. Married index patients were not less likely to name partners, suggesting that the higher rate of marriage in this group did not explain this finding either.

Another possibility is that cultural influences and perceptions of HIV/AIDS may also be perpetuating stigma around the disease and reducing African-born blacks' PS participation. We do not have data to investigate this hypothesis; however, other studies have identified these issues as significant deterrents to accessing care and/or open discussion of HIV/AIDS with medical providers and health workers. ^14,15 This suggests that cultural context and perceptions are important components in understanding the HIV/AIDS experience of immigrants. It is highly probable that cultural issues that inhibit linkage to HIV care also influence the likelihood of open discussions around other HIV/AIDS issues, including PS. Specifically, if fear of isolation and a fatalistic view about HIV/AIDS keeps patients away from seeking care, it is reasonable to assume that they will be less likely to name partners and risk further isolation.

We found Caribbean-born patients to be at highest risk for late HIV diagnosis, with almost half diagnosed with AIDS at the same time as initial HIV diagnosis. While it is not clear why Caribbeans were more likely to have an AIDS-defining CD4 count, FSU data has shown that among individuals concurrently diagnosed with AIDS, lack of perception of risk was the most common reason for not testing for HIV sooner. ¹⁷ FSU does not collect information on immigration status, so we were not able to evaluate the possible effects of immigration-related issues in care-seeking behavior. The higher frequency of late diagnosis among the foreign-born also may be caused by reduced access to HIV testing in their home countries, as suggested by our finding that the proportion of Caribbeans, whose AIDS-defining CD4 levels at initial HIV diagnosis was lower among those in the US longer. These high rates of late diagnosis indicate additional measures to promote early diagnosis among Caribbeans and others blacks are needed.

Limitations of this study include having groups of unequal size (i.e., fewer Africans and Caribbeans than African-Americans), which may have restricted the ability to detect statistically significant differences among groups. Also, country of birth and other behavioral data were self-reported and not verified through an independent source.

NYC blacks and their partners are benefiting from current Partner Services outreach. PS did identify new cases of HIV infection among patients' partners from all 3 groups. Blacks are also likely accruing other benefits from PS services not measured by our study, such as the impact of exposure notification, linkage to care, and HIV prevention counseling. However, the efficiency of PS was low in all groups. Additional interventions to improve HIV case finding from Partner Services and additional measures to promote early diagnosis among blacks are needed.