Katz, David A. MPH*; Hogben, Matthew PHD†; Dooley, Samuel W. Jr MD‡; Golden, Matthew R. MD, MPH*§¶
Partner notification is a longstanding component of public health efforts to control the transmission of sexually transmitted infections, including human immunodeficiency virus (HIV). However, particularly in the early years of the HIV epidemic, the intervention was controversial, and many health departments provided partner notification services (partner services, or PS) for HIV inconsistently or not at all. We previously surveyed health departments in areas of the United States with large HIV epidemics and found that, in 2001, only 32% of persons newly diagnosed with HIV were interviewed by the health department staff for purposes of partner notification.1 Since that time, the US Centers for Disease Control and Prevention (CDC) has launched a series of initiatives emphasizing case-finding, including PS, as a means to prevent HIV infection, and in 2008, the CDC released new national PS guidelines recommending that health departments strongly consider linking PS to HIV surveillance and provide PS to all persons newly diagnosed with HIV infection.2–4 We surveyed health departments to assess the current status of HIV PS in the United States.
MATERIALS AND METHODS
The study sample included public health programs providing HIV PS in metropolitan statistical areas where the reported number of cases of acquired immunodeficiency syndrome (AIDS), gonorrhea, chlamydial infection, or primary and secondary syphilis were among the 50 highest in the United States in 2005 on the basis of the CDC Surveillance Report for AIDS and the National Electronic Telecommunications System for Surveillance for other sexually transmitted diseases (STDs).5 The survey collected information about the organization of each area's PS program, the gender and risk factors of HIV-positive persons in the area, aggregate 2006 data on the number of persons interviewed for PS and partner notification outcomes, and procedures for reporting the outcomes of partner notification investigations. Surveys were conducted by either mail or e-mail. Responding health departments were included in the analyses only if they provided HIV partner notification outcome data for their jurisdiction. Data from a previous study of PS that sampled public health programs providing HIV PS to metropolitan statistical areas reporting ≥200 AIDS cases in 20011 were available for comparison with 17 of the responding health departments and an additional 10 departments that did not participate in 2006. Questions were asked in an identical manner in both surveys except that in 2006 we (1) separated the question regarding the routine provision of PS in public HIV testing facilities into 2 questions, one regarding STD clinics and one regarding all other public testing facilities, and (2) provided additional parenthetical descriptions for some PS outcomes.
In 9 instances, the reported number of HIV cases interviewed for the purposes of PS was greater than the number of HIV cases reported in the jurisdiction in 2006 and we were unable to clarify these numbers by contacting the respondent who completed the survey; in these instances, we assumed that the number of cases interviewed equaled the number of cases. We calculated the number of HIV-infected persons that health departments needed to interview to identify 1 previously undiagnosed case of HIV (number needed to interview, or NNTI) by dividing the total number of cases interviewed in the area by the number of new cases of HIV detected among cases' sex or needle-sharing partners.
We analyzed the data by using Stata statistical software version 9.2 (College Station, TX) and compared participating jurisdictions with nonparticipating jurisdictions with respect to region and number of AIDS cases reported in 2006 using Fisher's exact and Wilcoxon rank sum tests, respectively. Correlates of the proportion of HIV cases interviewed for PS in 2006 and the NNTI were determined using Spearman's rank correlation coefficients, Wilcoxon rank sum test, and Kruskal–Wallis test. To evaluate secular trends in HIV PS, we compared PS characteristics, coverage, and outcomes in 2001 and 2006 using Pearson's chi-squared, Fisher's exact, and either Wilcoxon rank sum (all participating jurisdictions) or Wilcoxon signed-rank tests (jurisdictions participating in both years).
We contacted representatives of health departments in 71 jurisdictions where the reported number of cases of AIDS, gonorrhea, chlamydia, or primary and secondary syphilis was among the 50 highest in the United States in 2005. Three jurisdictions were unable to provide HIV partner notification outcome data, 4 were only able to provide data regarding partner notification for STDs other than HIV, and 11 refused to participate. We excluded one jurisdiction from the analysis because only 4 cases of HIV were reported in the area in 2006. Nine jurisdictions in 6 states could provide only statewide data; these were included in the analysis. Thus, data were available from 48 health departments responsible for 51 (72%) of 71 jurisdictions asked to provide data for the study. Nonparticipating jurisdictions were located in 10 states in all 4 regions of the United States (West, Northeast, South, and Midwest); the median number of AIDS cases reported in 2006 among participating and nonparticipating sites for which data were available were 271 (range, 54–4716) and 310.5 (64–1092), respectively (P = 0.92) (The number of AIDS cases reported in 2006 was not available for 3 participating sites and 0 nonparticipating sites). The 48 participating health departments were responsible for jurisdictions that reported a combined total of 22,987 AIDS cases in 2005, representing 56% of the total 41,144 cases reported nationwide and 80% of the 28,902 cases reported by the jurisdictions offered study participation. Job titles of responding individuals varied, but 37 (77%) of 48 were program directors, supervisors, managers or chiefs, with the remainder having job titles suggesting that they were either Disease Intervention Specialists (DIS; i.e., public health staff who conduct interviews with persons diagnosed with STDs, including HIV, and attempt to ensure that their partners are notified, tested, and treated) or coordinated departmental PS. Table 1 summarizes the characteristics of the 48 health agencies surveyed.
Among the 48 participating departments, 45 (94%) reported that HIV PS were routinely provided to persons testing positive in publicly funded STD clinics; in 14 (29%) health departments, HIV PS were not linked to surveillance (i.e., the PS program did not routinely obtain the names of all persons in its jurisdiction reported to the local or state health department with newly diagnosed HIV) and were therefore not provided outside public health sites unless providers or cases contacted the health department to request assistance. The majority (76%) of respondents indicated that their department placed more emphasis on HIV PS than it did in 2001. Approximately one-quarter of the health departments funded community-based organizations (CBOs) to provide PS; 9 (82%) of 11 departments that funded CBOs to provide PS indicated that CBO staff elicited names of partners and attempted to contact and test partners.
Health department procedures for when to report an outcome for a partner and what types of information public health staff used to define partner notification outcomes varied. In 27 (56%) of 48 departments, partial details of a partner's name and usable (as defined by the respondent) contact information were required to record a disposition, whereas in 19 (40%), department staff recorded partner notification dispositions for any partner described by an index patient, regardless of what information the case provided. In the remaining 2 (4%) departments, decisions were left to the DIS or supervisors (n = 1) or dispositions were recorded only if the case requested assistance in notifying partners (n = 1). Forty-two (88%) PS programs only recorded disposition codes for outcomes verified by DIS through medical records or medical provider report, whereas the remaining 6 (12%) reported both verified outcomes and unverified outcomes reported by index patients.
Forty-four departments provided data on both the number of HIV cases reported and the number of cases interviewed for PS (Table 2). Public health staff interviewed 11,270 (43%) of the 26,185 cases reported to health departments in participating jurisdictions in 2006. Departments reported interviewing a median of 66% of all reported cases. Excluding areas which reported interviewing more cases than were reported in their jurisdiction, a total of 8683 (37%) of 23,598 cases received PS, and the median percentage of cases interviewed across jurisdictions was 59%. The proportion of persons with HIV interviewed in each jurisdiction was inversely associated with the number of cases reported (r = −0.52, P < 0.001), but was not associated with the proportion of HIV cases reported in men who have sex with men (MSM; r = −0.02, P = 0.92) (Table 3). Mandatory name-based HIV reporting, routine provision of PS in publicly funded STD clinics, and linking PS to HIV surveillance were all significantly associated with higher levels of PS coverage for HIV (P < 0.02 for all). The median proportion of cases interviewed in jurisdictions in the Northeast (18%) and Western United States (56%) were lower than those in the South (83%) and Midwest (80%), though this difference was not significant. PS coverage for HIV was not correlated with PS coverage for gonorrhea, chlamydia, or primary and secondary syphilis (P > 0.05 for all; data not shown).
A total of 43 departments provided complete information on both the number of persons who received PS and partner notification outcomes. Across all jurisdictions that provided these data, public health staff interviewed 10,922 persons. These people provided information on 10,498 potentially exposed partners, of whom 2228 (21%) had been previously diagnosed with HIV, 803 (8%) were newly diagnosed with HIV, 3337 (32%) tested negative for HIV, and 4130 (39%) were not successfully notified, were notified but refused HIV testing and denied previous HIV diagnosis, or did not have a partner notification outcome recorded. The NNTI for all jurisdictions combined was 13.6; the median NNTI across departments was 13.8. However, the NNTIs were not normally distributed and partner notification success was highly variable. NNTIs did not differ between programs that restricted disposition coding to verified partner notification outcomes and those that included both verified and unverified outcomes (medians of 14.1 and 12.9, respectively; P = 0.92), nor did NNTI significantly vary on the basis of the number of HIV cases diagnosed in a jurisdiction, the proportion of HIV cases reported in MSM, the type of organization responsible for PS, whether PS were linked to surveillance, the criteria used to determine if a disposition code should be assigned, or geographical region (P > 0.05 for all; data not shown), even though NNTI was somewhat higher in the West (NNTI = 25.6) as compared with other regions of the United States (South = 13.0, Northeast = 13.9, Midwest = 14.1; P = 0.11).
Tables 4 and 5 compare the characteristics and program outcomes of PS programs participating in the 2001 and 2006 surveys. Comparing all 27 participating programs in 2001 and the 48 in 2006 (Table 4), respondents in 2006 reported fewer HIV cases and a greater proportion of their HIV cases in MSM than those in 2001 (P < 0.01 for both). In addition, the proportion of jurisdictions with separate HIV and STD programs declined from 65% in 2001 to 48% in 2006, and the proportion reporting that PS outside publicly funded HIV testing sites were restricted to instances in which the case or provider contacted the health department for assistance decreased from 44% in 2001 to 29% in 2006; these differences were not significant. In 2006, 76% of the respondents reported that more emphasis was placed on PS in 2006 than in 2001. The proportion of HIV cases that received PS increased somewhat between 2001 and 2006 (medians of 55% and 66%, respectively; P = 0.07). There were no significant differences between the 2 study periods in the proportion of index patients who requested or allowed at least one partner to be contacted by public health staff (medians of 57% in 2001 vs. 66% in 2006, P = 0.29), the proportion of investigated partners whose HIV status was defined (60% vs. 63%, P = 0.53), or the NNTI (13.5 vs. 13.8, P = 0.93).
Of the 48 health departments responding in 2006, 17 (35%) also participated in the 2001 survey and provided complete partner notification outcome data. These departments were responsible for jurisdictions reporting a combined 14,070 AIDS cases in 2006, accounting for 34% of the total AIDS cases in the United States and 61% of the AIDS cases covered by participating departments. They represented 13 states and all 4 US regions (29% West, 12% Northeast, 53% South, and 6% Midwest) and were not significantly different from the departments that only responded to the 2006 survey with respect to HIV PS program characteristics presented in Table 1 (P > 0.05 for all). Programmatic differences between the 2 years were similar to those described among all jurisdictions (Table 5), and 69% of the respondents in 2006 reported that more emphasis was placed on PS in 2006 than in 2001. As in all participating jurisdictions, the proportion of HIV cases that received PS increased somewhat between 2001 and 2006 (medians of 49% vs. 61%, respectively; P = 0.06), and there were no significant differences between the 2 study periods in the proportion of index patients who requested or allowed at least one partner to be contacted by public health staff (medians of 62% in 2001 vs. 64% in 2006, P = 0.33), the proportion of investigated partners whose HIV status was defined (55% vs. 56%, P = 0.68), or the NNTI (12.1 vs. 15.0, P = 0.78).
We conducted a national survey of health department HIV PS procedures and outcomes in 2006, and compared our results with findings from a similar survey we conducted in 2001.1 Overall, we found that health departments are providing PS to a growing proportion of persons diagnosed with HIV and that case-finding yields from these programs are stable. Comparing data from 2001 and 2006, the overall proportion of persons newly diagnosed with HIV receiving PS increased from 32% to 43%, whereas the median percentage of cases interviewed increased from 55% to 66%. Among jurisdictions providing data for both periods, the median proportion of persons newly diagnosed with HIV receiving PS increased from 49% to 61%. Overall HIV PS coverage reported by participating jurisdictions in 2006 appears to be somewhat lower than that reported in a recently completed survey of primarily state-level PS programs conducted by the National Network of STD/HIV Prevention Training Centers (PTC).6 That report found that 53% of persons newly diagnosed with HIV were offered PS, 38% accepted those services, and that the NNTI was 17. However, the STD/HIV PTC survey did not provide respondents with a definition for “offer” or “accept,” making it somewhat difficult to directly compare results from the 2 surveys. Also, the higher levels of coverage reported by STD/HIV PTC may reflect their inclusion of a larger number of areas with lower levels of HIV/STD morbidity as these areas typically provide services to a larger proportion of all reported cases.
The increase in provision of PS between 2001 and 2006 appears to reflect changes both in HIV program priorities and in policy. Most program supervisors felt that their departments were placing greater emphasis on PS than they did in 2001, and the number of jurisdictions that linked PS to HIV surveillance and routinely provided services to persons diagnosed outside of public health testing sites increased. Name-based HIV reporting, connecting PS to surveillance, and routinely providing services to persons diagnosed with HIV outside of public health settings were closely linked program characteristics strongly associated with program coverage; in areas that connected PS to surveillance, a median of 82% of persons diagnosed with HIV received PS, whereas in areas without such a connection a median of only 20% received PS. We also observed regional differences in PS coverage, which may reflect the fact that regions with lower coverage had proportionally fewer jurisdictions with name-based HIV reporting, routine provision of PS in publicly funded STD clinics, and linkage of PS to surveillance than regions with greater coverage (data not shown); however, the small numbers of jurisdictions within each region make it difficult to draw meaningful conclusions regarding these differences.
Although a growing proportion of Americans diagnosed with HIV are receiving PS, our findings suggest that most still do not receive them. In large measure, this appears to reflect the fact that there is an inverse relationship between the size of a jurisdiction's HIV epidemic and the proportion of cases in the area that receive PS. Our group has now observed this inverse relationship in 3 national studies of PS for HIV, and a similar relationship exists between the number of cases of gonorrhea in an area and the proportion of persons with that infection that receive public health partner services.7 The reasons for this relationship are unclear. It may be that health departments in areas with large epidemics are simply overwhelmed and cannot provide PS to all persons who need them. Whatever the reason, insofar as the United States now has a goal to provide PS to all persons newly diagnosed with HIV, our findings highlight the need for improvements in areas with large HIV epidemics.
Of note, we found that a sizable proportion of health departments funded CBOs to provide PS. A CDC-funded evaluation of the feasibility of incorporating rapid HIV testing into PS included 2 jurisdictions where local health departments partnered with CBOs to provide PS.8 Outcomes from these 2 jurisdictions suggested that partnerships between health departments and CBOs for the provision of PS are feasible, but that effectiveness may vary by area or program characteristics. Future studies should specifically evaluate the effectiveness of PS provided by CBOs and the procedures that foster success.
Our findings provide a measure of where the United States stands in terms of PS and demonstrate the importance of linking PS to HIV surveillance when jurisdictions make a commitment to provide this intervention. But our results should not be interpreted as evidence supporting the effectiveness of HIV PS. We collated aggregate process outcome data collected in the course of public health programs. These data are uncontrolled, and we cannot say what proportion of partners would have been notified and tested in the absence of PS, though certainly some would have been.9 Very few controlled data exist to evaluate PS program effectiveness.10–12 What data we have suggest that PS does increase the number of partners notified, but the magnitude of that effect is uncertain. Our findings, as well as previous program evaluations, suggest that the NNTI varies widely from as low as 1.3 to greater than 100,8,9,11,13–17 and cost-effectiveness likewise varies substantially.17
Unfortunately, our study provided us with little opportunity to identify factors associated with greater PS program success. Other studies have associated higher case-finding yields with the timeliness of the provision of PS and with index case youth, recent HIV testing, and acute HIV diagnosis.9,11,16,18 We cannot readily explain the relative ineffectiveness of HIV PS in the Western United States, as it was not explained by differences in the risk factor composition of the population or correlates of program coverage. A recent study comparing PS in Colorado and Louisiana similarly observed lower case-finding in Colorado.17 Future studies should attempt to better identify what factors enhance program effectiveness.
Our study has several limitations. First, we collected aggregate data from health departments and, as shown by some of our findings as well as the findings from a survey we conducted with staff providing partner services in the same jurisdictions,19 staff providing PS do not always follow identical procedures for recording partner notification outcomes. This inconsistency could be a source of error. However, our results on partner notification outcomes were remarkably consistent across 2 national surveys and are largely in agreement with previously published, single center studies.20 Second, our 2 surveys employed somewhat different sampling frames and, consequently, some of our findings related to trends could reflect differences in the areas that participated in the studies. However, our primary findings, that the proportion of persons receiving PS is increasing and that program case-finding outcomes are stable, were observed both in comparing all respondents and also when we compared only jurisdictions included in both studies. Thus, it seems unlikely that differences in the studied populations could explain the changes we observed. Third, our study collected data from a sample of US health departments, and not all departments asked to participate in the study provided us with data. As a result, our findings may not be generalizable to all areas of the United States. Sampling only those jurisdictions with the greatest number of AIDS cases may have biased the reported PS coverage downwards compared to all jurisdictions in the United States because of the inverse relationship between the number of cases reported and coverage; however, it should be most representative of the areas comprising the majority of the US HIV epidemic. Finally, we collected information on the proportion of HIV cases health departments interviewed, not the proportion they attempted to interview. Health department staff typically interview over 80% of persons diagnosed with early syphilis in the United States, suggesting that high PS coverage can be achieved. However, PS declination may be higher in persons with HIV than those with syphilis; future studies should collect data on both the number of cases health departments attempted to contact and the number interviewed.
In conclusion, we found that although the proportion of persons diagnosed with HIV in the United States who receive HIV PS has increased since 2001, most people newly diagnosed with HIV in 2006 did not receive PS. Our findings suggest that efforts to link PS to surveillance and routinely provide services to persons diagnosed with HIV outside of public health testing sites can increase PS program coverage.
1. Golden MR, Hogben M, Potterat JJ, et al. HIV partner notification in the United States: A national survey of program coverage and outcomes. Sex Transm Dis 2004; 31:709–712.
2. Janssen RS, Holtgrave DR, Valdiserri RO, et al. The serostatus approach to fighting the HIV epidemic: Prevention strategies for infected individuals. Am J Public Health 2001; 91:1019–1024.
3. CDC. Advancing HIV prevention: New strategies for a changing epidemic. Morb Mortal Wkly Rep 2003; 52:1–4.
4. CDC. Recommendations for partner services programs for HIV infection, syphilis, gonorrhea and chlamydial infection. Morb Mortal Wkly Rep 2008; 57(RR-9):1–83.
5. CDC. HIV/AIDS Surveillance Report, 2005. Atlanta, GA: Department of Health and Human Services, Centers for Disease Control and Prevention, 2007.
6. National Network of STD/HIV Prevention Training Centers. HIV partner counseling and referral services capacity needs assessment: Report highlights. Available at: http://www.nnptc.org/ncahivpcrs.pdf
. Accessed August 25, 2009.
7. Golden MR, Hogben M, Handsfield HH, et al. Partner notification for HIV and STD in the United States: Low coverage for gonorrhea, chlamydial infection, and HIV. Sex Transm Dis 2003; 30:490–496.
8. Begley EB, Oster AM, Song B, et al. Incorporating rapid HIV testing into partner counseling and referral services. Public Health Rep 2008; 123(suppl 3):126–135.
9. Golden MR, Stekler J, Kent JB, et al. An evaluation of HIV partner counseling and Referral services using new disposition codes. Sex Transm Dis 2009; 36:95–101.
10. Landis SE, Schoenbach VJ, Weber DJ, et al. Results of a randomized trial of partner notification in cases of HIV infection in North Carolina. N Engl J Med 1992; 326:101–106.
11. Spencer NE, Hoffman RE, Raevsky CA, et al. Partner notification for human immunodeficiency virus infection in Colorado: Results across index case groups and costs. Int J STD AIDS 1993; 4:26–32.
12. Golden MR, Dombrowski JC, Wood RW, et al. A controlled study of the effectiveness of public health HIV partner notification services. AIDS 2009; 23:133–135.
13. Brewer DD. Case-finding effectiveness of partner notification and cluster investigation for sexually transmitted diseases. Sex Transm Dis 2005; 32:78–83.
14. Ahrens K, Kent CK, Kohn RP, et al. HIV partner notification outcomes for HIV-infected patients by duration of infection, San Francisco, 2004 to 2006. J Acquir Immune Defic Syndr 2007; 46:479–484.
15. Malave MC, Shah D, Sackoff JE, et al. Human immunodeficiency virus partner elicitation and notification in New York City: Public health does it better. Sex Transm Dis 2008; 35:869–876.
16. Moore ZS, McCoy S, Kuruc J, et al. Number of named partners and number of partners newly diagnosed with HIV infection identified by persons with acute versus established HIV infection. J Acquir Immune Defic Syndr 2009; 52:509–513.
17. Shrestha RK, Begley EB, Hutchinson AB, et al. Costs and effectiveness of partner counseling and referral services with rapid testing for HIV in Colorado and Louisiana, United States. Sex Transm Dis 2009; 36:637–641.
18. Marcus JL, Bernstein KT, Klausner JD. Updated outcomes of partner notification for human immunodeficiency virus, San Francisco, 2004–2008 AIDS 2009; 23:1024–1026.
19. Katz DA, Hogben M, Dooley SW, et al. An evaluation of the reliability of HIV partner notification disposition coding by disease intervention specialists in the US. Sex Transm Dis 2009; 36:459–462.
20. Hogben M, McNally T, McPheeters M, et al. The effectiveness of HIV partner counseling and referral services in increasing identification of HIV-positive individuals a systematic review. Am J Prev Med 2007; 33(suppl 2):S89–S100.