Using E-Mail to Notify Pseudonymous E-Mail Sexual Partners

Abstract

Objective: 

Quantify the factors associated with reporting sexual partners only known by, or able to be contacted by, an e-mail address (pseudonymous partners) and evaluate the efforts of provider partner notification when only an e-mail address was available.

Study Design: 

Persons with pseudonymous e-mail sexual partners were compared with unmatched controls. Associations were quantified by odds ratios and the effectiveness of provider partner notification attempts was assessed by comparison of outcomes of pseudonymous e-mail sexual partners and sexual partners with traditional contact information.

Results: 

Factors associated with having pseudonymous e-mail sexual partners included: male sex, white non-Hispanic, history of prior sexually transmitted diseases, multiple sexual partners, and reporting male-to-male sexual contact. Efforts at provider partner notification via e-mail were successful with 49.7% of pseudonymous partners notified and 40.1% evaluated.

Conclusions: 

Internet-based forms of communications can be utilized as means to contact and identify risk partners. If this avenue is not utilized, public health agencies may be unable to intervene with a potentially growing percentage of at-risk persons.

In Brief

Examination of factors associated with having pseudonymous e-mail sexual partners and effectiveness of provider partner notification using e-mail in Austin, Texas.

PROVIDER PARTNER NOTIFICATION IS the confidential contact of sexual partners of an individual infected with a sexually transmitted disease (STD) by health department staff¹ and is a fundamental tool of public health practice at the local level.² Partner notification attempts to arrest disease transmission, identifies people requiring treatment, and provides opportunity to modify risky behaviors of individuals at risk of contracting an STD.^3–5 In provider partner notification, public health staff (commonly called Disease Intervention Specialists, DIS)¹ elicit sexual partner contact information by interviewing persons diagnosed with STDs, including HIV. Sexual partners are notified of their potential exposure status and may be screened for disease or referred for medical evaluation.¹ The success of partner notification rests on the ability of infected individuals to share sufficient identifying and contact information about their sexual partners for locating and contacting by health department staff.^6,7

The emergence of Internet venues as a locale to initiate sexual activity has exacerbated the difficulty in obtaining contact information for sexual partners.^8–10 The handles, nicknames, screen names, and aliases used in e-mail addresses, chat rooms, and instant messenger programs confer anonymity and pseudonymity.^10–13 Internet venues provide an immeasurable pool of potential sexual partners further increasing the potential number of contacts without sufficient contact or identifying information.^11,14 McFarlane and colleagues reported that utilizing the Internet to solicit partners was common among clients seeking HIV testing at local public health clinics,¹⁰ and several recent outbreaks of syphilis involved partner solicitation via the Internet.^15–17 Lack of contact and identifying information on sexual partners meeting through virtual venues hampered partner notification attempts by public health departments in these outbreaks. Even after engaging in actual sexual activity, persons with an STD may not know their partner’s name or any contact information beyond an e-mail address.^13,15–17

Of additional public health concern, individuals soliciting sexual partners through Internet venues may be engaging in behaviors, which facilitate disease transmission.^8,11,18 Previous studies suggest persons who seek sexual partners via the Internet have more sexual partners than individuals who do not use the Internet for partner finding.^9,10 Drug use during sexual activity may also be more common.^19,20 Use of the Internet as a venue to find sexual partners has also been associated with a diagnosis of early syphilis among men who have sex with men (MSM).²¹ Taylor and colleagues reported MSM diagnosed with early syphilis who used the Internet to find partners were more likely to have anonymous partners.²² While illuminating, many of the previous studies were based on surveys,^8,9,19,20 qualitative research,¹¹ or were concerned with Internet partner seeking behavior in general.²² The first objective of this article to is to quantify the association between potential risk factors and having sexual partners who can only be identified and contacted by e-mail address.

These sexual partners without sufficient contact information are not truly anonymous partners.¹³ E-mail addresses may be used to uniquely identify one person, and as such, persons can use e-mail addresses to establish an online reputation or in order to be located. Also, e-mails may be linked to other information so complete anonymity is not achieved. Sexual partners that may be contacted or identified only by e-mail address may be best considered as pseudonymous partners.²³ Pseudonymous e-mail sexual partners may be defined as sexual partners only known by an e-mail address and those known by a name but for whom the only contact information available was an e-mail address.

Given potential risks for disease transmission among pseudonymous e-mail sexual partners, intervening using the minimal contact information available to public health staff is necessary for disease prevention and control. In January 2004, the Austin/Travis County Health & Human Service Department (ATCHHSD) began providing partner notification via e-mail for sexual partners for whom only an e-mail address was known. The primary objective of e-mail notification is to elicit additional identifying and contact information so that standard provider partner notification protocols may be followed while maintaining individual confidentiality. At the time of this report, ATCHHSD staff did not include disease or exposure specific information as part of e-mail correspondence. Case reports of partner notification using e-mail messages when no other means of contact were available in other locations have been previously published.^12,16,17 These efforts by local health departments did meet with measured success in notifying contacts. However, these were descriptive reports of notification efforts, not analytic studies. Therefore, the question of the effectiveness of attempts to conduct provider partner notification via e-mail when an e-mail address is the only means of contact still remains. The second objective of this article is to answer this question by comparing notification efforts among pseudonymous e-mail sexual partners and sexual partners for whom traditional contact information was known.

Materials and Methods

Persons with pseudonymous e-mail sexual partners (where e-mail was the sole means of contact) were compared with unmatched controls. The primary data source for this analysis was the ATCHHSD’s STD surveillance database, STD Management Information System 4.0 days (STD Control Program, Centers for Disease Control & Prevention, Atlanta, GA). Surveillance data include diagnoses, demographics, and risk behaviors for all cases of reportable STDs among local residents. Data were collected through mandatory laboratory reporting, medical chart reviews, and patient interviews. Provider partner notification was attempted for persons reported with HIV (including AIDS), and early syphilis cases (duration less than 1 year) by DIS staff. Of these persons with HIV infection and early syphilis, information on sexual contact with pseudonymous e-mail partners was documented on paper logs. Cases were defined as those persons who were documented to have reported at least one sexual contact with individuals for whom only an e-mail address was known between January 1, 2004 and June 30, 2006, and were residents of the city of Austin or Travis County, TX. A total of 53 persons met these criteria during this 30-month period. Controls were randomly selected without replacement from all interviewed City of Austin or Travis County residents with diagnoses of HIV or early syphilis who did not meet the case definition during the same time period. Excluded from being potential controls were persons who were incarcerated, homeless, recently displaced by Hurricane Katrina, and one person who reported all online partners were only identifiable via chat room names. A total of 601 potential controls met these criteria. These exclusion criteria were set to ensure cases and controls came from the same overall population. Five unmatched controls were randomly selected without replacement per case (n = 265). For both cases and controls, in instances where a person had repeat infections during the study period or new diagnoses requiring subsequent interviews, only the first case of disease was considered eligible for the study. This approach was taken in order to prevent correlated observations in the data and therefore risks to statistical conclusion validity.

Each person was identified as: infected with HIV (including a diagnosis of AIDS), early syphilis, coinfected with HIV and syphilis, or finally, HIV coinfected with at least one other STD. All diagnoses of syphilis were considered in the HIV coinfected with syphilis classification. Basic demographic data included: sex, age in years at time of report (categorized as: less than 30, 30–39, 40–49, and older than 50), race, and ethnicity. The surveillance data included behavioral risks determined through interviews with cases and their partners. Risks reflected both current and past behaviors at the time of interview. Male cases with any sexual activity with other men were classified as having engaged in male-to-male sexual activity. Men were also determined to have had any sexual activity with women. Persons, regardless of gender, having sexual contact with both males and females were classified as bisexual. Injection drug use was identified from self-report or through identified needle-sharing partners. Other risk factors included: any drug use, history of prior STD infections, trading sex for drugs or money, having sexual partners whom themselves had multiple sexual partners, and number of admitted sexual partners. Number of admitted sexual partners was examined in two manners. First, persons reporting more than one sexual partner during the interview period were classified as having multiple sexual partners. If the number of admitted sexual partners was missing, having multiple sexual partners was imputed from risk interview information. Second, the number of admitted sexual partners was standardized to partners per month because the “look-back” period differs by disease.¹ The average number of sexual partners per month was categorized as less than 1 sexual partner a month, 1 partner a month, or 2 or more sexual partners a month.

Cases and controls were described according to the above variables. Differences in distributions were assessed using the χ² test of homogeneity. In instances of small cell counts, Fisher exact test was used. Bivariate associations between reporting pseudonymous sexual e-mail partners and the above factors were quantified as unadjusted odds ratios (OR) with 95% confidence intervals (CI) calculated using logistic regression.

Outcomes of provider partner notification were determined using the final disposition of elicited sexual partners. Partner dispositions, as displayed in Figure 1, were categorized to reflect the various levels of public health contact and intervention. Regarding all admitted sexual partners, not all were known or were identifiable. Individuals may not have divulged partner’s names or contact information because they have forgotten, the partners were truly anonymous, or they consciously choose not to share the name. Elicited partners were those with at least a name or contact information for which DIS staff initiated provider partner notification. Elicited partners contacted by DIS staff were considered notified, while those who were not contacted, resided out of jurisdiction, or for whom data were missing, were classified as unable to contact. Notified partners were either evaluated by DIS staff or refused evaluation. Of those evaluated, individuals could be infected or not infected. New diagnoses or newly reportable conditions constituted new infections. Outcomes for cases’ pseudonymous sexual partners and controls’ sexual partners were compared using χ² test. In addition, the number of notified partners and the number of newly detected infections per interviewed infected person were calculated for cases and controls.³

Fig. 1: Outcomes of provider partner notification.

Results

Table 1 displays the descriptive statistics for persons with pseudonymous e-mail sexual partners and controls. The most common diagnosis (62.3%) among persons with pseudonymous e-mail sexual partners was HIV (including diagnoses of AIDS). Cases differed significantly from controls by race/ethnicity. Those reporting pseudonymous e-mail sexual partners were predominately Caucasian non-Hispanic (73.6%), whereas controls were more evenly distributed. Cases were overwhelmingly male (96.2%), and in a higher proportion than controls (83.8%). Nearly all males with pseudonymous e-mail sexual partners were determined through investigation to have had male-to-male sexual contact (98.0%). A higher percentage of cases had a history of prior STDs (56.6% vs. 35.9%) and sexual partners with multiple sexual partners (83.0% vs. 62.6%). More cases had multiple sexual partners (94.3%) than controls (76.2%). Both cases (49.1%) and controls (55.1%) had large percentages of reported drug use. Cases and controls did not differ significantly by age, coinfection status, injection drug use, or bisexual activity. Male controls were more likely to report sexual activity with females (64.9%) than were cases (11.8%).

TABLE 1: Conditions, Demographics, and Risk Behaviors of Cases With and Without Pseudonymous E-Mail Sexual Partners in Travis County, Texas, 2004–2006

Half of cases (52.8%) identified only a single pseudonymous e-mail sexual partner. However, nearly a quarter (22.6%) reported more than five pseudonymous e-mail sex partners. Persons generally were not exclusive in their choice of venue for meeting sexual partners. Most persons (75.5%) with pseudonymous e-mail sexual partners identified at least one sexual partner with sufficient contact and identifying information for traditional provider partner notification.

The associations between patient characteristics and reporting sexual contact with pseudonymous e-mail sexual partners are presented in Table 2 as OR. Cases with pseudonymous e-mail partners were nearly five times as likely (OR = 4.94; 95% CI = 1.16–21.06) to be male and more likely to be white non-Hispanic (OR = 4.05; 95% CI = 2.10–7.82) Age was unassociated with reporting pseudonymous e-mail sexual partners. The strongest association was with male-to-male sexual contact. Men identified as having male-to-male sexual contact were 12 times more likely to have pseudonymous e-mail sexual partners (OR = 12.01; 95% CI = 1.61–89.40). A history of prior STD infection (OR = 2.33; 95%CI = 1.28–4.25) and having sexual partners who themselves had multiple partners (OR = 2.92; 95% CI = 1.37–6.23) were also associated with having pseudonymous e-mail sexual partners. Lastly, having pseudonymous e-mail sexual partners was significantly associated with having multiple sexual partners (OR = 5.20; 95% CI = 1.57–17.24).

TABLE 2: Association Between Reporting Pseudonymous E-Mail Sexual Partners and Selected Demographic, Behavior, and Risk Factors in Travis County, Texas, 2004–2006

The outcomes of partner notification activities for cases and controls are presented in Table 3. Controls’ sexual partners were more likely to be notified (69.7% vs. 49.7%) than cases’ pseudonymous e-mail sexual partners. Although, e-mail notification proportions were lower, the majority of e-mail addresses provided for pseudonymous e-mail sexual partners were valid; only 16.4% were “returned to sender.” Once notified, controls’ sexual partners were more likely to be evaluated (95.4% vs. 80.7%) than cases’ pseudonymous e-mail sexual partners. Among the evaluated, the proportion of infected partners did not differ significantly between cases and controls. Overall, for each infected control interviewed, 1.40 sexual partners were notified and 0.10 new infections were discovered. Efforts were not as successful among the pseudonymous e-mail sexual partners, where 0.50 persons were notified and 0.05 new infections were found per interviewed person.

TABLE 3: Comparison of Outcomes of Provider Partner Notification for Pseudonymous E-Mail Sexual Partners and Sexual Partners Identified Through Traditional Contact Information in Travis County, Texas, 2004–2006

Discussion

This research focused specifically on infected individuals who reported sexual contact with persons for whom an e-mail address was the only means of identification or contact. The factors associated with reporting pseudonymous e-mail sexual partners were generally consistent with previous research on persons seeking sexual partners online. This was true for both demographic and behavioral characteristics. Of particular interest to public health were the associations with prior STD infection, multiple sexual partners, and reporting sexual partners who themselves have multiple sexual partners. Persons with repeat infections constitute a core maintaining STD endemicity in the population,²⁴ therefore heightening the importance of public health interventions. The associations between multiple partners and having partners with multiple partners further supports the concept that the Internet may serve as an accelerator for partner finding and disease transmission.¹¹ Although, not explicitly examined in this study, these associations with partner multiplicity suggest the potential for larger sexual networks.²⁰ This assumption requires specific investigation, but it would appear that these networks potentially include few women, because male sexual contact with women was inversely associated with reporting pseudonymous e-mail sexual partners. Lastly, based on previous research an association with drug use might have been expected, but was not the case.^19,20

Results on provider partner notification via e-mail efforts may be viewed in two ways. Notifying and evaluating pseudonymous e-mail sexual partners was not as successful as notifying partners with traditional contact and identifying information. Higher percentages of sexual partners with traditional contact information were notified and evaluated. Although this difference was not surprising, efforts at provider partner notification via e-mail were not fruitless. E-mail addresses elicited by DIS staff were reliable contact information; the majority of e-mails were not automatically “returned to sender.” Provider partner notification via e-mail notified 49.7% of pseudonymous e-mail sexual partners, none of which could have been contacted unless e-mail was utilized. Additionally, those who did not respond to health department e-mails may have sought services or treatment with other providers (however, the data cannot support or disprove that assertion). In comparison, five out of five pseudonymous sexual partners of a male diagnosed with early syphilis in San Francisco responded to health department e-mails.¹⁷ In that same outbreak, 34% of all sexual partners elicited via the Internet were located and evaluated; a proportion similar to 40.1% of pseudonymous e-mail sexual partners evaluated in this study. Likewise, a case report from Los Angeles also demonstrated success in provider partner notification using e-mail. Out of 111 contacts, 26% responded to health department e-mails.¹² In general, public health agencies have success in contacting persons via e-mail about sensitive health issues.²⁵

Finally, e-mail notification resulted in the rapid delivery of health department communications, and it required no additional field work or after-hour time for staff. These benefits have the potential to increase DIS staff productivity while reducing or maintaining costs, thus resulting in a positive return on investment.

While response rates may obviously be improved, ATCHHSD notification procedures use principles designed to facilitate response. ATCHHSD conveys authenticity by only sending e-mails from a department e-mail address and by using professional language, promotes message urgency through subject line verbiage, provides exhaustive contact information for the recipient to verify the identity of the sender, utilizes read receipts, and attempts multiple contacts with message content ascending in urgency. The need for these aforementioned features, as well as proper etiquette and more comprehensive Internet-based interventions, has been previously outlined by other investigators and health promoters.^17,26–31 A traditional concern with e-mail addresses is privacy, out of the concern an e-mail address may be accessed by more than one person.²⁷ However, an e-mail address may be scenario specific,³² and the sexually explicit and descriptive pseudonymous e-mail addresses logged in this study agree with that possibility, suggesting these e-mails were not intended to be accessible by more than one person (a spouse, for example). Therefore, with this reassurance of privacy, inclusion of specific disease exposure information, such as acknowledging contact to a diagnosed case of gonorrhea or chlamydia, could improve the proportions of notified partners. Under current statutes, this would exclude specifically identifying HIV exposure. However, as public health moves toward routinization of HIV testing, e-mail notification could be an avenue for aligning HIV notification consonant with STD notification practices. Additionally, e-mail notification may be more successful with the integration of new practices, such as expedited partner therapy and through the additional use of an electronic prescription.

The results of the analysis were limited by the usual constraints imposed by a secondary analysis of surveillance data, including: missing data, and the fact that not all potential variables of interest were collected. A significant limitation was the study population, which included only those reported cases interviewed by DIS staff. This was a necessary restriction to prevent misclassification bias among cases and controls. However, those who refused to be interviewed may be different along one or more dimensions. Because the analysis only reflects reported cases of STDs in one locality, the results may not be generalizable to asymptomatic persons, persons without infection, or to other areas of the United States. Because of the relatively small number of cases, extensive multivariate modeling was not attempted; therefore, the possibility of uncontrolled confounding exists. Finally, not addressed in this article were partners identified solely by chat room names or instant messenger identifiers. Research suggests that this is another avenue of partner solicitation,^11,33 and Klausner and his colleagues effectively utilized chat room screen names for contact tracing in a syphilis outbreak.¹⁶

Unless Internet-based forms of communication are utilized as means to contact and identify at risk partners, public health agencies will be unable to intervene with a potentially growing percentage of high risk individuals. A substantial number of persons may be utilizing the Internet to locate sexual partners. A recent meta-analysis estimates 40% of MSM have looked for sexual partners using the Internet,³⁴ and an online poll reported 10% of Americans accessing the Internet looked for partners online.³⁵ Current use of Internet venues for sex partner solicitation demands a response using nontraditional contact methods such as e-mail partner notification.³⁶ The data presented here indicate e-mail partner notification is a valuable and an effective addition to standard notification practices when appropriate. If this avenue is not utilized, a subset of potentially exposed individuals, who have higher risks for disease transmission, will not be identified and evaluated by health department staff. This could lead to increases in morbidity from preventable STDs, including HIV.

Additionally, the data also illustrate the need for new skill sets for DIS staff, who traditionally are trained to elicit information from face-to-face encounters. Today, the public health workforce must have the skills to use all methods possible to respond to the changing environment of partner solicitation. DIS training programs must recognize the importance of Internet-based technologies and communication forms such as e-mail, chat rooms, instant messaging, and the use of social networking sites³⁷ and routinely adapt and use these technologies in their curriculum. Furthermore, schools of public health play a critical role in developing the workforce by integrating fundamental concepts of partner notification and contact tracing with the Internet-based communication forms familiar to, and utilized by, their students. Within government, successful and innovative state and local practices need near real time adoption, dissemination, and support nationally. Public health must have the skill sets ready to confront the potential scenario where individuals do not consent to face-to-face interviews, but only to communication with DIS staff through virtual mediums. In public health’s endeavor to locate, contact, and intervene with all potentially exposed persons in order to arrest disease transmission, electronic communication modes will have to be explored as means of partner notification.