A controlled study of the effectiveness of public health HIV partner notification services

Abstract

Health departments inconsistently provide partner notification services to persons with HIV. We surveyed a randomly selected sample of 370 patients at the largest HIV clinic in Washington State, USA, using a self-administered anonymous questionnaire. Patients who reported talking to health department staff about partner notification were more likely to report notifying at least one sex partner following their HIV diagnosis than patients who did not talk to health department staff (68 vs. 45%, odds ratio = 2.5, 95% confidence interval = 1.6–3.9).

Partner notification is a longstanding mainstay of public health efforts to control the transmission of sexually transmitted infections, including HIV. However, in 2001, only approximately one-third of persons diagnosed with HIV in the USA received partner notification services ^[1]. Recently released Centers for Disease Control and Prevention (CDC) guidelines recommend that health departments attempt to provide partner services to all persons with newly diagnosed HIV ^[2].

At least in part, the inconsistent provision of partner services in high-income nations reflects the paucity of data supporting the efficacy of the intervention. Only a single randomized controlled trial has evaluated HIV-partner notification ^[3]. That study found that conditional referral (i.e. the practice of having public health staff contact a patient's partners if the originally diagnosed patient does not do so him or herself) led to more partners being tested than leaving partner notification up to patients alone. However, that trial enrolled only 74 people. The only other published study presenting controlled data evaluating HIV-partner notification found that among recipients of partner services (i.e. persons interviewed by public health staff), persons who elected to notify partners themselves assured the testing of a smaller proportion of their partners than persons who allowed public health workers to contact partners for them ^[4]. We are unaware of controlled studies conducted in the era of effective antiretroviral therapy, and it may be that effectiveness of partner notification has changed as the prognosis associated with HIV has improved.

We evaluated the effectiveness of HIV-partner notification as part of a behavioral surveillance program undertaken by our health department in the largest HIV clinic in Washington State, USA. The program started in 2005, and annually surveys patients using an anonymous, self-administered written questionnaire. We randomly select the surveyed population from a list of English-speaking patients who have scheduled appointments during a 2–3-week period each spring. Previously published studies describe our survey and partner notification procedures, and surveyed persons were similar to all clinic patients with respect to age, gender and sexual orientation ^[5,6].

In 2006 and 2007, the questionnaire asked participants whether they notified any sex partners following their HIV diagnosis, how many partners they notified, and whether any of their partners tested positive for HIV ‘shortly after you contacted them.’ Following those questions, the questionnaire asked, ‘After you were originally diagnosed with HIV, did someone from the health department talk to you about notifying your partners?’ We used logistic regression to test for associations between independent variables and notifying at least one partner. Our final multivariate model included variables that were statistically significant (P < 0.05) on bivariate analysis and remained significant in a multivariate model. The University of Washington Institution Review Board approved study procedures.

We randomly selected 743 patients to receive the survey in 2006 or 2007. Of these, we excluded 223 (30%) because they did not attend their clinic appointment, 72 (10%) non-English speakers, and 14 (2%) persons for other reasons. Study staff offered the survey to 435 (58%) persons, of whom 406 completed surveys, and 370 responded to a question about partner notification. These 370 persons, 85% of those offered study participation and 50% of all randomly selected patients, comprised the study population.

One hundred and eight-three patients participated in 2006 and 187 participated in 2007. The population included 304 (82%) men, 59 (16%) women, six (2%) trangendered persons, and one person of unknown gender. Of the 304 men, 256 (84%) reported having sex with another man in the preceding year or described themselves as gay or bisexual. The median age of the population was 45 (SD = 9.4, range 21–72). A total of 245 (66%) patients were White, 63 (17%) African–American or Black, and the remaining 51 (14%) reported other or multiple races. Eleven participants did not report a race, and 34 (9%) participants were Hispanic.

A total of 199 (59%) of 340 participants reported receiving partner services. Recipients of partner services were significantly more likely to report notifying at least one sex partner than persons who did not receive partner services (Table 1). Among 269 participants who provided data on the number of partners they notified, the 151 persons who received partner services notified 243 partners (mean 1.6, median 1.0), whereas the 118 persons who did not receive partner services notified 135 partners (mean 1.1, median 0) (P = 0.0004, two-sided Wilcoxin). Persons with incomes more than $30 000/year were significantly more likely to report notifying partners than those with lower incomes, but we did not observe significant associations between notifying partners and other demographic characteristics. Persons who notified at least one partner after their HIV diagnosis were somewhat more likely to report always disclosing their HIV status to subsequent sex partners, but not more or less likely to report engaging in nonconcordant unprotected anal or vaginal intercourse. Seventeen (13%) of 134 persons who received partner services reported that at least one of their partners was diagnosed with HIV following their own HIV diagnosis compared with eight (8%) of 102 persons who did not receive partner services (P = 0.23). (One hundred and thirteen people did not answer a question about whether any of their partners tested HIV positive.)

Table 1: Association between participant's report that they notified at least one sex partner after being diagnosed with HIV and demographic factors, receipt of public health partner services and subsequent frequency of disclosing HIV positivity to sex partners.^a

Our findings support the hypothesis that public health partner services can increase the number of partners notified following a patient's HIV diagnosis, and thus are consistent with the very limited controlled data available supporting the efficacy of HIV partner notification ^[3,4]. We also observed that people who notified partners after being diagnosed with HIV were more likely to disclose being HIV positive to subsequent-sex partners, suggesting that people who notify past partners following a new HIV diagnosis may more consistently disclose their HIV status to future sex partners.

Our study included patients in a single HIV clinic in Seattle, Washington, USA, and not all patients selected for participation enrolled in the study, perhaps biasing our findings or limiting the generalizabilty of our results. Also, our data were based on participant self-report and cannot be validated. Despite these limitations, we believe our findings support new CDC recommendations to provide partner services to all persons diagnosed with HIV.

Acknowledgements

Dr M. R. Golden was responsible for the overall design of the study as well as data analysis. Mr M. Fleming collected data from study individuals and contributed to development of the research instrument. All of the authors contributed to the interpretation of study findings and writing the manuscript.

The authors thank Carol Glenn for help recruiting participants for this study and the Public Health – Seattle and King County Disease Intervention Specialist staff for providing partner services to persons in King County.

Data presented previously at the 2008 CDC STD Prevention Conference and published as an abstract.