Partner notification, a principal means of controlling sexually transmitted diseases, has traditionally been performed by public health professionals. They interview infected persons and contact the sex partners of these persons to notify them and convince them of the need to seek evaluation and treatment (known as “provider referral”). This notification method is labor intensive; the typical alternative to provider referral is to leave notification to the infected person (known as “patient referral”). However, innovations in partner notification, often created by public health professionals responsible for the practice, have yielded adjuncts and complements to both provider and patient referral. The present review article covers 4 areas of innovation: (1) enhancements to patient referral instructions and provider interview techniques, (2) use of the Internet in partner notification, (3) the emerging influence of network methods, and (4) expedited partner therapy, principally through patient-delivered medications or prescriptions. Partner notification remains necessary, and flexibility, openness to the use of multiple methods, and collaboration are likely to be helpful.
An integral part of the treatment of persons with most sexually transmitted diseases (STDs) is the treatment of their sex partners. This assertion has been operationalized in the United States through notification of sex partners of infected persons of their exposure to an STD, followed by some effort to convince the sex partners to seek evaluation and, typically, treatment. This is known as “partner notification.” The specific primary aims of partner notification are finding and treating new cases of STDs and providing prophylactic treatment to individuals exposed to STDs. For syphilis and, to a lesser extent, for HIV infection and gonorrhea, the task of notification has historically been the province of trained public health professionals, who are most commonly known as “disease intervention specialists” (DISs). The most common alternative notification method has been for providers to encourage patients to notify their own sex partners and refer them for treatment: this is known as “patient referral” and “self-referral.”
For syphilis and, sometimes, for other STDs, DISs interview infected persons at least once to obtain the names of their sex partners and to locate as much other identifying information as possible [1,2–3]. They then trace the sex partners, notify them that they have been exposed to the STD in question, and try to convince the partners to present for evaluation and treatment. Many DISs draw blood in the field for testing, and, in some jurisdictions, DISs are permitted to carry oral treatment to dispense to partners who they notify (sometimes an accompanying nurse is the dispensing agent). DISs are prohibited from revealing the name of the patient who provided the partner information, so notification is confidential for both patient and partner. DISs sometimes trace selected high-risk social contacts of infected persons by use of the same procedures. If an interview is conducted, DISs and infected persons can arrange for patient referral within a defined time window. If named partners do not present for evaluation within that window, DISs begin provider referral.
Extent and effectiveness of traditional partner notification strategies.In spite of a lack of consistent national-level data on the magnitude of case reporting, provider referral is widely considered to be the reference standard for partner notification in terms of its primary aims and its effect on syphilis control. Systematic reviews of partner notification strategies endorse provider referral as being more effective than patient referral [4,5–6]. However, partner notification is not perfect, as is shown by the number of infected persons who DISs need to interview to find a newly infected sex partner [7]. Furthermore, current labor resources simply preclude the use of the DIS model for prevalent gonorrhea and chlamydial infections, not to mention trichomoniasis. A 2002 survey of jurisdictions with high morbidity associated with ⩾1 of the following STDs—syphilis, AIDS, gonorrhea, or chlamydial infection—found that, although 89% of persons with primary and secondary cases of syphilis were interviewed, only 52% of persons with HIV infection, 17% of persons with gonorrhea, and 12% of persons with chlamydial infection were interviewed [8]. In a subsequent survey, the addition of other jurisdictions lowered the estimate for HIV infection to 32% [9]. In the absence of DIS-based provider referral, patients are typically left with the responsibility of notifying their partners themselves. The effectiveness of this approach is not known, although data from sources such as program evaluations [10, 11] and the control conditions of partner notification trials [12] suggest that up to one-half of infected persons notify ⩾1 partner. The efficacy of partner notification through any method based on diagnosis in private settings is also unknown. What is known is that physicians are ill disposed to taking charge of notification, compared with case reporting and encouraging patient referral [13], and that few physicians engage in provider referral [14].
Focus of the current review.In summary, DIS-based provider referral is likely to be confined mainly to syphilis and to a minority of cases of gonorrhea and, currently, HIV infection. Consequently, this article, in serving as background to a review of partner notification strategies for the 2006 STD treatment guidelines, focuses on 4 areas. The first area pertains to standard approaches—provider and patient referral. The second area presents data on Internet-based versions of partner notification (that follow the logic model of provider and patient referral). The third area examines research centered around network-enhanced approaches to partner notification. This topic was a main feature of the background report [15] to the previous guidelines [16]. In this article, I examine some empirical investigations using network methods. The fourth area pertains to expedited partner therapy (EPT), which is the treatment of sex partners of infected persons without an intervening examination or diagnosis. The 2002 guidelines [16] made brief note of EPT under the guise of patient-delivered partner therapy (PDPT), which is the most likely mechanism to be employed by users of the current guidelines. Since 2002, more trial-based data concerning EPT have been published, and the Centers for Disease Control and Prevention currently suggests that clinicians consider this approach when they are not likely to secure a personal evaluation of sex partners [17, 18]. The current guidelines maintain “PDPT” as the dominant acronym [19]. One point of note is that DISs will plausibly remain integral to partner notification in all 4 areas but especially in the first 3 areas.
Review strategy.Because the Centers for Disease Control and Prevention STD treatment guidelines rely on published material, I confined my search to databases of publications, principally Medline, but also to reference sections of reviews and commentaries about partner notification. Specific criteria were that the data be (1) from 1999 to 2005, (2) published in a peer-reviewed setting, and (3) published in the English language. Trial and program evaluation data were both eligible (to rely on trial data alone would raise serious questions about domain-sampling adequacy and, therefore, generalizability). I searched databases, using generic terms for STDs (e.g., “sexually transmitted,” “STI” [i.e., sexually transmitted infection], and “STD”), as well as specific names of STDs combined with either (1) the term “partner notification” or the words “contact” and any of the permutations of “tracing” or (2) the word “partner” and any of the words used to indicate partner notification (e.g., “manage,” “management,” “notification,” and “services”). After 2 readings, 41 articles were retained for review, including several surveys that illustrated important points, such as current practices. I did not consider surveys in recommendations for the guidelines.
Quality rankings were not comments on the inherent qualities of the studies but, rather, denoted specific definitions. “Good” studies were those that contained statistically and substantively relevant causal evidence for partner notification outcomes (i.e., notification, evaluation, and treatment) or biological outcomes (i.e., number of cases found and reduced reinfection rates). “Fair” studies contained evidence for partner notification or biological outcomes but had study designs limiting the scope or generalizablity of conclusions. “Weak” studies lacked substantively relevant end points (such as reinfection rates, number of cases found, and number of partners notified, evaluated, or treated) or had major methodological limitations. Several retrospective evaluations and observations without demonstrated potential for prospective use fell into this category.
Standard approaches to partner notification in studies from 1999 to 2005 are shown in table 1. There are few completed efforts to actually improve patient referral services without introducing some novel feature (e.g., medications for partners and referral over the Internet, both of which are covered below). Some recent studies from Uganda have examined the determinants of self-referral in some detail [31, 40], although their applicability to US partner notification via patient referral is unknown. In qualitative research, the authors found that respondents acknowledged the importance of self-referral and their responsibility to notify partners, but they also cited negative consequences, such as relationship dissolution. Among infected persons, past successful referral, current intentions, and self-efficacy (the latter for women only) were predictors of subsequent patient referral.
Partner notification (PN) studies published between 1999 and 2005: standard approaches to PN (n = 26).
Kissinger et al. [43] tested patient referral with counseling versus patient referral with counseling and a booklet of referral cards; a third study arm, which evaluated patient referral with provision of medications to sex partners, is discussed below. The overall follow-up rate was 79%, but only 38% of patients who completed follow-up (i.e., 30% of the full sample) agreed to provide urine samples for biological testing. There was no clear-cut reason for this low rate, but the investigators surmised that general reluctance was a reason and that it was alleviated only by the suspicions of the participants that they were (re)infected. The notification rates per partner, among those given booklets, were similar to those noted for partners receiving self-referral (53% vs. 48%). In contrast to relatively similar notification behaviors (53% in the booklet arm vs. 48% in the control arm), the risk of infections at follow-up among participants in the booklet arm was substantially lower than that among those in the control arm (OR, 0.22; 95% CI, 0.11–0.44). According to results from South Africa, clinical training with simple written instructions can improve case management, including referral instructions [22]. Many of the evaluations in table 1 lead to the same conclusion. Because patient referral instructions are more common across STD and practice settings than are any other instructions from US physicians [14, 37], improved instruction may be useful.
For provider referral, interview training helps produce more named sex partners [44]. On an anecdotal basis, continued training (or periodic quality assurance) and practice improve the performance of DISs. As noted above, DISs may be closely involved in the adjuncts and additions to provider referral discussed below. There has also been some recent research into optimal interviewing strategies. Brewer et al. [45] found that people do underreport the number of sex partners that they have had, for reasons that include forgetting the exact number. They found that reporting could be increased by nonspecific prompts (e.g., by asking “Who else have you had sex with in the past 12 months?”). Prompting and reading back a list of partners elicits, on average, 5%–10% additional partners beyond those already mentioned in the interview [45,46–47]. Sets of cues referring to locations where people have sexual contact with their partners, locations where they first meet their partners, and first names of partners each elicit reporting of up to 20% more sex partners, with similar results found for social contacts [45, 46]. In a randomized trial of various recall cues administered in routine partner notification for STDs, these supplementary techniques increased by 12% the number of new cases of STDs found (i.e., brought to treatment) [47].
Various reports (usually convenience samples) cite the availability of sex partners over the Internet [48, 49]. An emerging response is the use of the Internet as part of infection-control efforts, even as part of useful new communication technology for partner notification [50]. Privacy concerns during partner notification imply a need for secure Web sites and a general assurance that electronic notification will go to and be read by the partner exposed to the STD (this also complies with state laws about confidential notification and Centers for Disease Control and Prevention endorsement of that policy) [2, 3].
Two studies from California report the outcomes of Internet-based referral (table 2). During a 1999 outbreak of syphilis among men using an Internet chat room, San Francisco Health Department personnel sent e-mail messages to the individuals behind the screen names of sex partners provided to DISs by 2 infected persons; the screen names were the only contact information that the infected individuals had [51]. Respondents whose screen names matched a screen name on the list provided by the index case patients were considered to have been notified. By this process and on the basis of subsequent presentations for evaluation, 5 cases (4 new STD cases and 1 STD case from earlier in 1999) were identified. Investigators reported that 42% of “named” partners were notified (the mean number of partners per case across the 7 total cases was 12.4 partners). Data also revealed that case patients who had larger numbers of sex partners were more likely to have a higher proportion of untested partners (as well as larger absolute numbers of untested partners).
Partner notification studies published between 1999 and 2005: Internet-based partner notification (n = 2).
On a smaller scale, in Los Angeles, 2 unrelated cases of syphilis were diagnosed among men with Internet contacts [52]. One man reported 134 partners and 111 e-mail addresses. The health department sent notices to all e-mail addresses, and 29 individuals responded to the health department's notification attempt. The second man reported 16 sex partners and contacted 13 of those partners himself via e-mail messages, with copies of the e-mails sent to the health department. Seven of the 13 partners eventually underwent testing. Internet contact information did not preclude gathering other information, where available: the first man also provided 23 phone numbers and 8 complete names of sex partners. The main deficiency of Internet-based partner notification, compared with in-person provider referral, is the reduced ability to persuade a partner to seek evaluation. However, Internet-based partner notification clearly has promise, given that the alternative to Internet partner notification by DISs, in this case, was simply self-referral (not in-person provider referral).
Rothenberg [15] wrote extensively about the theory and emerging use of network methodology to inform partner notification prospectively (e.g., an outbreak investigation) as well as retrospectively (e.g., how network information explains the spread of disease in a recent outbreak). Eight network-based evaluations of partner notification are listed in table 3. Network structure has been used to explain changes in endemicity [57] and bridging from one population to another [58]. To some extent, a network-influenced approach has always existed in partner notification through the cluster approach (i.e., interviews of persons other than sex partners who are socially related to the infected case patient, with follow-up testing prescribed).
The 2002 guidelines contain a cogent summary of the theoretical usage: “...networks have an influence on disease transmission that is independent of personal behaviors [and]...network structure is related directly to prevalence and underlying disease transmission dynamics” [16, p. 4]. Rothenberg et al. [60] reported the results of a prospective evaluation of a syphilis outbreak by use of social and sexual network methods for data gathering and analysis. DISs collected the usual information on sex partners in clinic settings, but they also spent up to 80% of their time in street settings, interviewing index case patients and others to determine drug-use partners and important social contacts, as well as sex partners.
Drug-use partners and social contacts of infected persons were informed of the high local prevalence of syphilis and were encouraged to seek testing. Sex contacts of 48 infected case patients yielded 30 additional syphilis infections from 130 sex partners and 9 infections from 153 other contacts. Contacts of uninfected persons yielded 2 infections from 37 sex partners and 4 infections from 76 other contacts. Thus, interviewing the social contacts of infected persons yielded a 30% increase in the number of infections found (9 of 39 infections), whereas interviewing uninfected persons yielded a prevalence of 5.3% (6 infected contacts among 113 contacts identified by uninfected persons). Network diagrams revealed several heavily embedded members of the overall sociosexual network, who could be useful to interview in subsequent STD outbreaks.
The point of interviewing these individuals would be to (1) produce a more efficient series of investigations over time, and (2) co-opt community members into the case-finding process—as a part of the solution rather than the problem. Even retrospective construction of a network from prospectively gathered venue and other social information [53] could help set the stage for prospective use in a future outbreak. Ideally, network analysis incorporating economic, cultural, political, and other types of data [62] would become part of forming and maintaining an in-depth and dynamic portrait of a community with a given level of a given STD.
EPT is a broad term covering methods whereby partners of index case patients receive treatment before undergoing a personal evaluation [18]. Five studies of EPT are listed in table 4. The underlying principle is prophylactic treatment, and EPT is not designed to be used in lieu of personal evaluation. Retrospective evaluations suggested that EPT might reduce the number of STDs (measured by the proxy of reinfection of the index case patient) [68, 69], and the practice of giving infected persons extra medications (or prescriptions) to give to their sex partners is widespread nationally [67], although it is virtually unregulated outside of a few states. Survey reports suggest that ∼50% of physicians surveyed have ever given medications to patients with chlamydial infections or gonorrhea to give to their partners and that 10%–15% do so “usually” or “always” [67, 70, 71]. Two later surveys from California and New York City provided higher estimates of “frequent” practice [72, 73].
Partner notification studies published between 1999 and 2005: expedited partner therapy (n = 5).
Three trials comparing EPT with patient referral [43, 63, 64] largely used index case patients as the sole means of bringing therapy to partners via medications or prescriptions (one trial incorporated DIS assistance when requested). Across trials, the risk of reductions in the reinfection of participants in the EPT arm versus the risk of such reductions in participants in the self-referral control arm were as follows: OR, 0.80 (95% CI, 0.62–1.05) [63]; relative risk [RR], 0.76 (95% CI, 0.59–0.98) [64]; and OR, 0.38 (95% CI, 0.19–0.74) [43]. Effects across studies were larger for gonorrhea than for chlamydial infection (Golden et al. [64] had direct comparison via subanalyses, Schillinger et al. [63] enrolled women with chlamydial infection, and Kissinger et al. [43] enrolled men with largely gonorrhea-driven urethritis).
Behaviorally, EPT, as it has been most commonly tested, comprises patient referral plus provision of medication. That is, the index case patient is enjoined to notify his or her partners and, in fact, refer them for evaluation, just as one would for patient referral. Notification by participants of their sex partners appears to be equivalent or better than patient referral, depending on the study. Both Golden et al. [64] and Kissinger et al. [43] reported substantial increases in the proportions of partners receiving treatment via EPT versus self-referral. In the trial by Golden et al. [64], 61% of participants in the EPT arm reported that all partners were “very likely” treated, compared with 49% of participants in the patient referral arm (P < .01). Kissinger et al. [43] found that 56% of men who were randomized to give medications to their partners reported that their partners told them that they took the medication, and 48% also saw the partner take the medications. Corresponding percentages for the self-referral arm were 45% and 32%, both of which were significantly lower (P < .001).
Potential unintended consequences of EPT include the possibility of partners avoiding evaluation because they have been treated—a point that is particularly salient with respect to missed comorbidity and to women incubating pelvic inflammatory disease. Results from a multiple-city survey suggest that the rate of HIV infection comorbidity among men who have sex with men may be sufficiently high in some locations that DIS-based partner notification is preferable to EPT [74]. Adolescents with chlamydial infections or gonorrhea at STD clinics may also have high rates of trichomoniasis comorbidity [75]. To date, there is no evidence to suggest that providing medication either facilitates or retards evaluation-seeking by partners; regardless, medication should be accompanied by instructions to seek evaluation and to avoid sex during the course of medication, among other health education instructions. Of course, the same ought to be true of instructions accompanying self-referral, because these issues apply equally to both strategies.
As of the time of writing, there are several complementary approaches to partner notification that involve either provider or patient referral as a base. One lesson is that the use of one approach need not preclude the use of another; this lesson is already widely practiced across STDs (provider referral is used for syphilis, and patient referral is used for chlamydial infections). However, different methods might be applied to different infected persons who have the same STD, depending on case-by-person limitations. A patient who is willing to notify his partner by phone and bring the partner to the clinic is a good candidate for patient referral; a patient who can notify his partner but is unlikely to bring the partner to the clinic may need provider referral (and, if provider referral is unavailable, may perhaps need EPT). These examples show that flexibility and openness to innovation remain important, and, therefore, so do policy conditions that enable both. Finally, the role of the Internet in particular emphasizes the need for broad collaboration. Treatment guideline planning is certainly not the only activity during which collaboration is discussed; however, collaboration among jurisdictions conducting partner notification and facilitation by cross-cutting organizations, such as the Centers for Disease Control and Prevention, should be encouraged.
Supplement sponsorship.This article was published as part of a supplement entitled “Sexually Transmitted Diseases Treatment Guidelines,” sponsored by the Centers for Disease Control and Prevention.
Potential conflicts of interest.M.H.: no conflicts.
IDSA Members: For your free access to this journal, log in via the IDSA members area.
Open access options for authors visit Oxford Open
This journal enables compliance with the NIH Public Access Policy