THE SOUTHEASTERN REGION OF THE United States has consistently had higher rates of sexually transmitted infections (STIs) than other regions of the country, perhaps as a result of the racial/ethnic composition and sexual mixing patterns of persons living in this region, poverty, and reduced access to health care.1–3 Although increases in early syphilis rates have been reported as a result of risk behaviors among men who have sex with men (MSM) from several U.S. cities,4–7 syphilis outbreaks in rural communities have not been well characterized. An understanding of the factors associated with syphilis infection in rural and urban areas with high morbidity is important to direct national syphilis elimination efforts.
In 2001, the rate of primary and secondary (P&S) syphilis in North Carolina was 5.5 cases per 100,000 population, the third highest among the 50 states.2 Columbus County had the highest reported P&S syphilis rate (74.1 cases per 100,000 population) in the state in 2001, which represented a 20-fold increase from the rate of 3.7 cases per 100,000 in 2000. During 2000 to 2001, the rate of early latent syphilis (ELS) in Columbus County increased 3-fold (from 7.3–23.5 cases per 100,000).8
Columbus County is a rural county with a land area of 937 square miles and a population of 54,749 residents, 22.7% of whom live below the poverty line. The county population is predominantly white with approximately one third comprised of racial/ethnic minority groups (30.9% black, 2.3% Hispanic) (Table 1).9 Columbus County is situated in the southeastern part of North Carolina bordering South Carolina and Robeson County, NC, a county which had a syphilis outbreak during 2000–2001.2 The syphilis outbreak in Robeson County, which peaked with a rate of 71.7 cases per 100,000 population in 2001, was thought to be related to prostitution and/or drug use similar to prior syphilis epidemics in the South (defined as the U.S. region consisting of the District of Columbia and 16 states located between the Atlantic Coast and Texas).10–12 We conducted a retrospective chart review of early syphilis (P&S, ELS) cases reported in Columbus County to evaluate factors contributing to an outbreak of syphilis and the associated sociosexual network among case patients and contacts in this rural community.
Materials and Methods
We identified cases of early syphilis reported in Columbus County from January 2001 to February 2002 (14-month period) to the NC HIV/STD Prevention and Care Branch. Syphilis is reportable in the state from both laboratories and healthcare providers. Probable or confirmed P&S syphilis or ELS case reports are investigated by disease intervention specialists (DIS), who are responsible for interviewing persons with syphilis and identifying their contacts to assure that they are evaluated and treated appropriately. DIS interviews with infected persons were conducted using a case report form developed by the Centers for Disease Control and Prevention (CDC), which records demographic information, stage of syphilis, risk factors, sexual partners, and social acquaintances of persons with syphilis. The case report form also allows a narrative in which DIS can record comments that are not captured elsewhere on the form (e.g., activities reported by interviewees and places where they socialize, meet partners, and have sex). Cluster interviewing was applied as a strategy in this syphilis outbreak to gain information regarding the index patients’ social circles, which may include additional persons exposed through unobserved sexual networks.13 Cluster interviews were conducted to identify individuals (in addition to sexual partners) who are in the case patients’ community or social networks of friends or acquaintances who may have symptoms suggestive of syphilis, be partners of other persons known to be infected, or may otherwise benefit from an STI examination. Efforts to communicate with sexual partners and social acquaintances were carried out using standard procedures to preserve the confidentiality and privacy of all individuals involved. Contacts were notified by DIS of their exposure to early syphilis, but were not provided the names of the case patients who identified them in the interviews.
Contact tracing was performed by the identification and comparison of contacts, which we defined to include both sexual partners and social acquaintances identified from a cluster interview with a case patient, with contacts reported by other infected persons. Because case patients may have multiple partners or common acquaintances, contact tracing in this syphilis investigation included contacts that were identified by more than one case patient and did not represent unique individuals.
Retrospective chart reviews were conducted on case reports of persons with early syphilis using a chart abstraction form for collecting demographic information and behavioral and sexual risk factors during the incubation period for each case patient. Incubation periods were defined as 90 days for persons with primary syphilis (PS), 6 months for persons with secondary syphilis (SS), and 1 year for persons with ELS. “Period partners” were defined as individuals reported to be sexual partners of case patients during the incubation period for syphilis. The data were abstracted from both the standardized fields and from the narrative sections of the case reports. A single investigator conducted all the chart reviews to maintain consistency in the interpretation and recording of case report data on the chart abstraction forms. The data were double-entered and analyzed using Epi-Info 2002 (CDC, Atlanta, GA) for the descriptive analyses.
The network analysis was conducted independent of the chart reviews using the available contact tracing data for syphilis from Columbus and Robeson counties during the study period that were systematically collected using the STD Management Information System (STD-MIS). STD-MIS is an electronic data system that allows health departments to manage information from healthcare providers, laboratories, clinics, and DIS regarding infected patients as well as data collected in the course of contact investigations. SAS (SAS Institute, Cary, NC) was used to determine connected groups of individuals, and Pajek14 software was used to produce network visualizations or diagrams. The network diagrams were constructed to illustrate infected persons and their contacts (sexual partners and social acquaintances) who were linked from the contact tracing data.
During the 14-month study period, Columbus County reported 61 cases of early syphilis: 20 PS, 25 SS, and 16 ELS cases. The male-to-female case ratio among persons with early syphilis was 1.8 and the mean age was 35 years (range, 17–62 years). Fifty (82%) case patients were black, although only one third of Columbus County residents are black (Table 1). Forty-one (67%) case patients lived in one zip code area, the largest town in Columbus County, with a population of 5,148 residents.9 Of 30 (49%) persons with early syphilis who had HIV test results available, 2 (7%) had a prior diagnosis of HIV infection and none were newly diagnosed with HIV at the time of their syphilis diagnosis.
Over 90% of case patients reported only having sexual partners of the opposite sex (Table 2). The mean number of sexual partners reported by case patients during their incubation period was 2.4 (range, 1–5 partners), and 38% of infected persons reported at least 2 period partners (Table 2). Twelve percent of the case patients reported meeting sexual partners in clubs or bars in the area, and 8% reported meeting sexual partners on one block in Columbus County that was associated with general drug-using activity according to local law enforcement personnel. Although no case patients reported travel to Robeson County to meet sexual partners during their incubation period, 13% reported travel across the border to South Carolina to meet sexual partners.
Twenty-seven (44%) case patients reported using illicit drugs defined as crack cocaine, cocaine other than crack, marijuana, and/or heroin (Table 2). Eighteen case patients (30%) reported crack cocaine use, 28 (46%) indicated having sex with a period partner who used crack cocaine, and 31 (51%) reported use of crack and/or having sex with a period partner who was a crack user. Nineteen (31%) persons acknowledged exchanging sex for drugs or money, of whom 12 (63%) reported using crack (although the rest of the results focus on exchange of sex for crack, “exchange of sex for drugs or money” was a variable directly taken from the case reports and the CDC interview form).
The case patients reported a total of 143 period partners during the incubation period but provided individual names for only 117 partners during the interviews of whom 101 (86%) were located by DIS (Table 3). In addition to the sexual partners, the cluster interviews yielded 189 named social acquaintances from the case patients, of whom 168 (89%) were located by DIS. Overall, 59 case patients named 306 contacts (sexual partners  and acquaintances  who were not necessarily unique), providing an average of 5.0 contacts (range, 0–22) per infected person. Forty-four (72%) case patients were found to have at least one period partner with documented early syphilis, and 6 (10%) of these had ≥2 infected partners. Among these 44 case patients, 3 were sexual partners to a Robeson County resident with early syphilis, and one case patient was a sexual partner to a person with early syphilis residing in another state. In addition to the case patients with an infected sexual partner, 6 case patients were found to have at least one acquaintance with documented early syphilis. Overall, 82% of the case patients identified in Columbus County had at least one contact (sexual partner or acquaintance) with documented early syphilis.
STD-MIS data for network analysis were available from the 61 Columbus County case patients, the infected person from Robeson County, and the contacts reported by case patients during the outbreak. The network analysis revealed complex interactions between infected and noninfected individuals identified as sexual partners or social acquaintances. Figure 1 depicts the sexual partners of case patients with men represented as squares and women as circles. The size of the nodes (case patients and their contacts) is proportional to the number of concurrent sexual partners, defined as sexual relationships occurring during an overlapping time frame.15 Details of the calculation of the concurrency numbers, which appear as labels in Figure 1, are provided in the Appendix. Nodes without number labels represent persons for whom concurrency was not demonstrated. The node labeled “R” represents the case patient from neighboring Robeson County. The sexual network demonstrated multiple concurrent sexual partnerships, especially among male members of the largest connected group of individuals. Same-sex partnerships were not evident from the data. The sexual network was largely dendritic (treelike). However, there were 2 linked “4 cycles,” a term used to indicate cyclic connections among 4 individuals illustrated at the top center of Figure 1 that involved a total of 6 persons with multiple concurrent sexual partners. One hundred twenty-seven persons were included in Figure 1 (including case patients who were not linked to other infected persons in the sexual network and those who did not report any contacts). Seventy-five (59%) of the case patients and contacts in Columbus County were connected in one large sexual network.
Figure 2 shows the network of sexual partners and acquaintances of case patients with sexual relationships represented by solid lines and nonsexual relationships represented by dotted lines. The sociosexual network illustrates a densely interconnected pattern of interactions among case patients, sexual partners, and acquaintances, underscoring the close connections between case patients, even those who may not be directly linked to one another as sexual partners. The overall sociosexual network included 262 individuals, of whom 232 (89%) were linked in one highly interconnected group. The remaining 30 persons comprised 10 smaller, disconnected groups of ≤6 individuals.
A risk representation (Fig. 3) illustrates the sociosexual network of case patients, sexual partners, and acquaintances who reported crack cocaine use or exchange of sex for crack. Exchange of sex for crack cocaine appears to be a unifying feature among the individuals with multiple concurrent sexual partners who comprise the core of the sexual network.
The investigation of early syphilis cases in Columbus County demonstrated that the outbreak in this rural community primarily involved heterosexual transmission among blacks. Possible factors among infected persons that contributed to the magnitude of this outbreak include having multiple period partners, illicit drug use (particularly crack cocaine), and exchange of sex for drugs. These factors have been associated with syphilis transmission in prior epidemics.16–19 The sociosexual network analysis illustrated a densely interconnected, highly cyclic network of individuals, which suggests that additional unobserved sexual connections may have facilitated disease transmission. A high percentage of case patients were found to have at least one contact (sexual partner or acquaintance) with documented early syphilis. The links between infected persons in Columbus County to an individual with early syphilis from Robeson County, a county with high P&S syphilis rates since 1997, suggest syphilis transmission among residents of 2 adjacent counties that have had fairly distinct outbreaks. Although the Robeson County outbreak was also associated with prostitution and/or drug use, only 40% of early syphilis cases reported in that outbreak occurred among blacks and over one third occurred among Native Americans.12
The rate of acquisition of Treponema pallidum per sexual encounter with an infectious partner has been estimated to be 30%.20 Approximately 30% of period partners reported by case patients in Columbus County were identified with syphilis. However, this percentage may be an over- or underestimate, because DIS located only 71% of reported partners. Seventy-two percent of case patients in Columbus County had at least one sexual partner with documented early syphilis and 82% had at least one infected contact, which is higher than the proportion of contacts infected with syphilis reported in the literature. In a review of over 12,000 patients with early syphilis identified in Louisiana, only 55% had at least one infected contact.21 The high percentage of infected contacts among case patients in Columbus County may have been the result of the tightly associated sociosexual network of at-risk individuals as well as the intensive efforts by DIS to identify potentially infected persons in this outbreak through cluster interviews.
Cluster interviewing has been recognized for decades as a strategy to enhance the disease intervention process for syphilis and has been recommended in outbreak settings to expand the base of information about high-risk groups associated with infected persons.13 Cluster interviews conducted during this outbreak investigation assisted in the identification of infected individuals who might have been missed by partner notification (Table 3) and provided valuable information about where contacts of case patients live and socialize. Data gathered from the cluster interviews were important to the determination of the sociosexual network pattern.
Network analysis has been used to study disease transmission in gonorrhea,22,23 chlamydia,24 syphilis,25 HIV,26 and tuberculosis,27 as well as bloodborne disease transmission among high-risk cohorts of drug users.28,29 A network approach also facilitates identification of persons in an infected individual’s social group who may benefit from diagnostic screening and appropriate treatment.30 The sexual network analysis in our study revealed a predominantly dendritic chain, indicating a high degree of interchange among sexual partners. Typically, a dense sexual network pattern signifies high contact rates among individuals, which increase the risk of disease transmission.31 Adding social acquaintances to the network illustrated numerous interconnections and dense, cyclic patterns among contacts, suggesting close associations between infected case patients and contacts (perhaps through unidentified sexual relationships) than would have been determined from a sexual network alone.
This sociosexual network pattern underscores the importance of syphilis control strategies directed toward individuals as well as social circles comprised of persons with multiple concurrent partners or linked through drug use. A “network-informed approach” may potentially guide prevention efforts by directing DIS assignments to specific neighborhoods or venues to meet sexual partners in addition to traditional case finding. Contact can then be initiated with persons at risk as well as persons active in the network who can influence other individuals’ risk and health care-seeking behaviors.30
The sexual network in the Columbus County syphilis outbreak did include 2 linked cycles illustrating sexual connections among 6 individuals at the top of Figure 1. Linked 4 cycles indicating cyclic relationships between 4 individuals with multiple concurrent sexual partners (in this case, with alternating male and female partners) occur infrequently in romantic relationships and represent a hallmark of core heterosexual transmission.15 Concurrency has been shown in simulation models to increase the intensity, variability, and size of an epidemic.32 These 4 cycle structures indicate a high concurrency of sexual partnerships among 4 individuals who reported use of crack and/or sex in exchange for crack cocaine (Fig. 3), which are factors that likely fueled this syphilis outbreak.
During the 1990s, rates of early syphilis declined in all racial/ethnic groups and all regions. However, there have been increases in the rates of infectious syphilis since 2001 and in the male-to-female ratio since 2000, which may be attributed to the rise in P&S syphilis cases among MSM nationally.33 Although high rates of P&S syphilis among MSM have been reported mainly by large urban areas such as New York City, San Francisco, and Los Angeles,6,34 66% of the counties in the United States with 2001 P&S syphilis rates above the Healthy People 2010 objective (<0.2 cases per 100,000 population) were located in the South.1
A comparison of the factors associated with the recent syphilis outbreaks among MSM in urban areas and this rural outbreak reveal some similarities. Multiple sexual partners and drug use were also high-risk behaviors identified among MSM with syphilis.5,6 However, whereas alcohol, marijuana, poppers, and crystal methamphetamines were the most common drugs used by MSM with syphilis in New York City,7 crack cocaine was most frequently reported in Columbus County (Fig. 3). The outbreaks of syphilis among MSM were associated with high rates of HIV coinfection. Unfortunately, information about HIV testing was not available for roughly 50% of Columbus County case patients, so the actual extent of HIV coinfection in this outbreak is unknown.
The Columbus County outbreak was not unique from other syphilis epidemics reported among heterosexual populations in the last 20 years. Crack cocaine use and exchange of sex for drugs or money have been linked to syphilis outbreaks among heterosexual populations in both urban and rural areas.16,18,35–37 An outbreak of syphilis in San Diego County that began in the mid-1980s involved illegal drug use, especially crack cocaine, that was related to prostitution in the inner-city areas.18 In the early 1990s, outbreaks of syphilis in rural Texas towns were concentrated in neighborhoods where crack cocaine dealers conducted business and where exchange of sex for drugs or money was common.16 During the same time period, syphilis rates rose substantially among rural nonwhite women of all ages in North Carolina, which may have been associated with drug activity along an interstate highway located in the eastern part of the state.11,38 Columbus County is not situated along this interstate highway, but the outbreak likely occurred from the introduction of infectious syphilis from surrounding areas with higher syphilis morbidity to a core group of residents practicing illicit drug use and exchange of sex for drugs.
Similar to other STI outbreaks reported in the southeastern United States, the Columbus County outbreak primarily affected economically disadvantaged minority groups and exhibited geographic clustering of infected persons. Previous investigators hypothesized that the risk of syphilis is determined not only by individual behavior, but also by the neighborhood socioeconomic dynamics that could affect sexual behaviors, including unemployment, incarceration rates, low levels of education, alcohol and drug marketing.39 In the southeastern United States, the combined effects of poverty, racial and ethnic minority populations, and geographic clustering (of racial/ethnic minorities) are thought to contribute to the relatively high incidence of syphilis infection.40
In October 1999, the CDC launched its National Plan to Eliminate Syphilis with a goal to reduce P&S syphilis cases to <0.4 cases per 100,000 nationwide by 2005.41 Key strategies of the plan include enhanced surveillance, strengthened community involvement and partnerships, enhanced clinical and laboratory services, enhanced health promotion and rapid response to outbreaks. A successful outbreak response is facilitated by an interdisciplinary approach by DIS, outreach workers, clinicians, laboratory workers and epidemiologists, and involved communities to develop targeted interventions.42 Outbreak responses involving intense case finding and enhanced screening at crack houses have been previously conducted in the South, which may be in part responsible for the decline in syphilis cases in the 1990s.43,44
The increase in early syphilis cases in Columbus County prompted the NC HIV/STD Prevention and Care Branch to enhance case detection and surveillance activities through intensive contact investigation, cluster interviews, and outreach screening in the area. The state worked in collaboration with the local health department and several community-based organizations to form a rapid intervention outbreak team (RIOT), which provided syphilis and HIV education and testing in the field. The RIOT identified 15 new cases of syphilis and one case of HIV from 313 and 212 individuals screened for syphilis and HIV, respectively. The Columbus County P&S syphilis rate declined from 74.1 cases per 100, 000 population in 2001 to 27.1 cases per 100,000 in 2002,5 perhaps as a result of the intensive RIOT activities involving door-to-door screening of persons living in high-risk areas of the county or from the natural course of the outbreak, which had already affected the majority of individuals in the core group of transmitters. The rates of early syphilis statewide have also fallen since 2001, which may have resulted from continued statewide syphilis elimination efforts. In 2004, Columbus County reported no cases of early syphilis.45
Our study has several limitations, primarily as a result of the small number of syphilis cases that were reviewed and that represented only one rural community in eastern North Carolina. Our assessment did not include a control group of individuals from Columbus County for comparison; therefore, the strength of the association between factors such as crack use with acquisition of syphilis among the case patients in this outbreak could not be assessed. Furthermore, variations may have occurred in interview style and documentation because the information in the narratives may not have been gathered and/or recorded in a systematic manner. Factors that may have contributed to syphilis risk among case patients (such as locations visited to meet sexual partners) may have been underestimated because the information is not routinely required on the case report forms. The measurement of concurrency of partnerships is subject to recall error, although this error is not likely to be substantial when evaluating data that involve recent sexual partnerships.46 Underreporting of sexual partners from case patients may have also occurred, although the effect of this error is mitigated by cluster interviewing of social contacts, which better highlights the affected groups.
Many of the persons involved in the syphilis outbreak in Columbus County had multiple concurrent partners, used illicit drugs, exchanged sex for drugs or money, and were closely associated through a dense sociosexual network. Crack cocaine use and exchange of sex for crack likely fueled the outbreak by contributing to the high degree of interchange of sexual partners among persons at high risk for syphilis. Although the rates of HIV coinfection were not fully assessed in this outbreak, the spread of syphilis in this small community could have had a substantial impact on HIV transmission in the local population. Cluster interviewing methods and a rapid outbreak response with extensive community involvement assisted in decreasing further syphilis transmission by enhancing the identification and treatment of infected persons and contacts. The application of rapid outbreak response activities and other interdisciplinary syphilis control strategies in both urban and rural areas are imperative if syphilis elimination is to be achieved at local and national levels.
Calculation of Concurrency Score
Concurrency (as opposed to serial monogamy) occurs when one person has sex with another, switches to yet another person, and then returns to the original partner. Although detailed partner exposure histories are not routinely collected in the contact-tracing process, it is sometimes possible to deduce concurrent partnerships from routinely collected dates of first and most recent exposure. The following hypothetical example illustrates the calculation of the number of concurrent partnerships an index person has with 5 sexual partners (A, B, C, D, and E), each with associated first and last exposure dates. T0 represents the earliest exposure date with any of the 5 partners. Each line segment represents the time period between first and last exposure.
In this example, the index person begins a sexual partnership (with A) at T0, which terminates at T1. A second partnership (with B) begins later at T2. Such disjoint partnerships would be an example of serial monogamy; however, a third partnership with C occurs in one single instance at time T3 before the partnership with B terminates at T5. One can therefore deduce that the index person must have had sex first with B, then had sex with C, and then had sex (at least one more time) with partner B. Similarly, there is demonstrated temporal overlap (from T4 to T5) between partners B and D, which indicates that the index person had sex with B, then D, then again with B, and finally again with D. There may have been more instances in which the index had gone back and forth between these 2 persons, but with only 2 time points per partner, it is sufficient to show that temporal overlap implies partnerships B and D to be concurrent. Finally, we have the partnership with E, which is shown to begin coincidentally when the partnership with D ended. For the purposes of measuring concurrency with contact-tracing data (in which dates are sometimes approximated to the first of the month), such partnerships are considered to be serially monogamous.
The concurrency number calculated for persons in this network can be restated as the number of partnerships that are shown to be concurrent with at least one other partnership. In this example, partnerships B, C, and D are shown to be concurrent (B with C and B with D). Thus, the calculated concurrency number for this index person is 3.
One other consideration arises in network data—people sometimes name each other, and their dates of exposure do not exactly agree. Sometimes this is the result of recall error or other misreporting, and other times this can occur simply because of the difference in interview dates of the 2 case patients. For the purpose of estimating concurrent partnerships, the earliest and latest reported dates were used—a process that has been shown to work reasonably well with real-world data.46 Cited Here...
1. Laumann E, Youm Y. Race/ethnic group differences in the prevalence of sexually transmitted diseases in the United States: A network explanation. Sex Transm Dis 1999; 26:250–261.
2. Centers for Disease Control and Prevention. Sexually Transmitted Disease Surveillance, 2001. Atlanta: US Department of Health and Human Services, 2002.
3. Centers for Disease Control and Prevention. Primary and secondary syphilis—United States, 2000–2001. MMWR Morb Mortal Wkly Rep 2002; 51:971–973.
4. Centers for Disease Control and Prevention. Resurgent bacterial sexually transmitted disease among men who have sex with men—King County, Washington, 1997–1999. MMWR Morb Mortal Wkly Rep 1999; 48:773–777.
5. Centers for Disease Control and Prevention. Outbreak of syphilis among men who have sex with men—Southern California, 2000. MMWR Morb Mortal Wkly Rep 2001; 50:117–120.
6. Centers for Disease Control and Prevention. Primary and secondary syphilis among men who have sex with men—New York City, 2001. MMWR Morb Mortal Wkly Rep 2002; 51:853–856.
7. Paz-Bailey G, Meyers A, Blank S, et al. A case–control study of syphilis among men who have sex with men in New York City. Sex Transm Dis 2004; 31:581–587.
8. North Carolina HIV/STD Surveillance Report. North Carolina Department of Health and Human Services HIV/STD Prevention and Care Branch, 2002.
9. State and County QuickFacts. US Census Bureau, 2000.
10. Thomas JC, Kulik AL, Schoenbach VJ. Syphilis in the South: Rural rates surpass urban rates in North Carolina. Am J Public Health 1995; 85:1119–1122.
11. Cook RL, Royce RR, Thomas JC, et al. What’s driving an epidemic? The spread of syphilis along an interstate highway in rural North Carolina. Am J Public Health 1999; 89:369–373.
12. Foust EM, Leone PA, Ashby RM, et al. The impact of prostitution and drugs on a rural syphilis outbreak. Program and Abstracts of the National STD Conference, California, 2002:C2A.
13. Division of STD Prevention. Program Operation Guidelines for STD Prevention. Atlanta: Centers for Disease Control and Prevention, 2000.
14. Batagelj V. PAJEK: Package for Large Network Analysis. Slovenia: University of Ljubljana, 2002.
15. Beaman PS, Moody J, Stovel K. Chains of affection: The structure of adolescent romantic and sexual networks. Am J Sociol 2004; 110: 44–91.
16. Schulte JM, Ramsey HA, Paffel JM, et al. Outbreaks of syphilis in rural Texas towns, 1991–1992. South Med J 1994; 87:493–496.
17. Centers for Disease Control and Prevention. Outbreak of primary and secondary syphilis—Guilford County, North Carolina, 1996–1997. MMWR Morb Mortal Wkly Rep 1998; 47:1070–1072.
18. Gunn RA, Montes JM, Toomey KE, et al. Syphilis in San Diego County 1983–1992: Crack cocaine, prostitution, and the limitations of partner notification. Sex Transm Dis 1995; 22:60–66.
19. Koumans EH, Farley TA, Gibson JJ, et al. Characteristics of persons with syphilis in areas of persisting syphilis in the United States: Sustained transmission associated with concurrent partnerships. Sex Transm Dis 2001; 28:497–503.
20. Schroeter AL, Turner RH, Lucas JB, et al. Therapy for incubating syphilis: Effectiveness for gonorrhea treatment. JAMA 1971; 218:711–713.
21. Kohl KS, Farley TA, Ewell J, et al. Usefulness of partner notification for syphilis control. Sex Transm Dis 1999; 26:201–207.
22. Potterat JJ, Rothenberg RB, Woodhouse DE, et al. Gonorrhea as a social disease. Sex Transm Dis 1985; 12:25–32.
23. Centers for Disease Control and Prevention. Gang-related outbreak of penicillinase-producing Neisseria gonorrhoeae
and other sexually transmitted diseases—Colorado Springs, Colorado, 1989–1991. MMWR Morb Mortal Wkly Rep 1993; 42:25–28.
24. Potterat JJ, Muth SQ, Rothenberg RB, et al. Sexual network structure as an indicator of epidemic phase. Sex Transm Infect 2002; 78(Suppl 1):i152–i158.
25. Rothenberg RB, Sterk C, Toomey KE, et al. Using social network and ethnographic tools to evaluate syphilis transmission. Sex Transm Dis 1998; 25:154–160.
26. Potterat JJ, Phillips-Plummer L, Muth SQ, et al. Risk network structure in the early epidemic phase of HIV transmission in Colorado Springs. Sex Transm Infect 2002; 78(Suppl 1):i159–i163.
27. McElroy PD, Rothenberg RB, Varghese R, et al. A network-informed approach to investigating a tuberculosis outbreak: Implications for enhancing contact investigations. Inter J Tuberc Lung Dis 2003; 7:486–493.
28. Potterat JJ, Woodhouse DE, Muth SQ, et al. Network dynamism: history and lessons of the Colorado Springs study. In: Morris M, ed. Network Epidemiology: A Handbook for Network Survey Design. Oxford University Press, 2004:87–114.
29. Rothenberg R, Long D, Sterk C, et al. The urban and rural networks project: Atlanta and Flagstaff. In: Morris M, ed. Network Epidemiology: A Handbook for Network Survey Design. Oxford University Press, 2004:115–138.
30. Rothenberg R, Narramore J. Commentary: The relevance of social network concepts to sexually transmitted disease control. Sex Transm Dis 1996; 23:24–29.
31. Doherty IA, Padian NS, Marlow C, et al. Determinants and consequences of sexual networks as they affect the spread of sexually transmitted infections. Clin Infect Dis 2005; 191(suppl 1):S42–S54.
32. Morris M, Kretzschmar M. Concurrent partnerships and the spread of HIV. AIDS 1997; 11:641–648.
33. Centers for Disease Control and Prevention. Primary and secondary syphilis—United States, 2002. MMWR Morb Mortal Wkly Rep 2003; 52:1117–1120.
34. Centers for Disease Control and Prevention. Trends in primary and secondary syphilis and HIV infections in men who have sex with men—San Francisco and Los Angeles, California, 1998–2002. MMWR Morb Mortal Wkly Rep 2004; 53:575–578.
35. Edlin BR, Irwin KL, Faruque S, et al. Intersecting epidemics—crack cocaine use and HIV infection among inner-city young adults. N Engl J Med 1994; 331:1422–1427.
36. Nakashima AK, Rolfs RT, Flock ML, et al. Epidemiology of syphilis in the United States, 1941–1993. Sex Transm Dis 1996; 23:16–23.
37. Centers for Disease Control and Prevention. Outbreak of primary and secondary syphilis—Guilford County, North Carolina, 1996–1997. MMWR Morb Mortal Wkly Rep 1998; 47:1070–1071.
38. Thomas JC, Kulik AL, Schoenbach VJ. Syphilis in the South: Rural rates surpass urban rates in North Carolina. Am J Public Health 1995; 85:1119–1121.
39. Farley TA, Kahn RH, Johnson G, et al. Strategies for syphilis prevention: Findings from surveys in a high-incidence area. Sex Transm Dis 2000; 27:305–310.
40. Aral SO. The social context of syphilis persistence in the southeastern United States. Sex Transm Dis 1996; 23:9–15.
41. Division of STD Prevention. The National Plan to Eliminate Syphilis From the United States. Atlanta: Centers for Disease Control and Prevention, October 1999.
42. Finelli L, Levine WC, Valentine J, et al. Syphilis outbreak assessment. Sex Transm Dis 2001; 28:131–135.
43. Centers for Disease Control and Prevention. Epidemic early syphilis—Montgomery County, Alabama, 1990–1991. MMWR Morb Mortal Wkly Rep 1992; 23:790–794.
44. Centers for Disease Control and Prevention. Selective screening to augment syphilis case-finding—Dallas, 1991. MMWR Morb Mortal Wkly Rep 1993; 42:424–427.
45. North Carolina HIV/STD Surveillance Report. North Carolina Department of Health and Human Services HIV/STD Prevention and Care Branch, 2004.
46. Brewer DD, Rothenberg RB, Muth SQ, et al. Agreement in reported sexual partnership dates and implications for measuring concurrency. Sex Transm Dis 2006; 33:277–283.