Kahn, Richard H. MS*; Peterman, Thomas A. MD, MSc*; Arno, Janet MD†; Coursey, Emmett John BS‡; Berman, Stuart M. MD, ScM*
SEXUALLY TRANSMITTED DISEASE (STD) CONTROL programs engage in many activities in a broad effort to prevent syphilis transmission. This effort has taken on increased importance given the commitment, announced by the Surgeon General in 1999, to undertake the National Plan to Eliminate Syphilis from the United States.1 As part of this effort, public health officials in Indianapolis implemented screening programs in the jail and one hospital emergency department, increased partner notification efforts, including an emphasis on cluster interviewing (interviewing cases or their partners to elicit persons who are symptomatic, have been exposed to a known STD case, or are high risk and would therefore benefit from testing) and enhanced community partnerships working with community-based organizations. Health officials in Nashville implemented screening in their local jail, maintained traditional partner notification activities, and enhanced community partnerships. Although rates have decreased in these cities and elsewhere, it is not clear which of the activities were most effective.
Previous studies have suggested that syphilis persistence may be maintained in communities by relatively small groups (4.0–15.0%) of high-risk persons (i.e., core transmitters) centrally placed among a larger group with low to moderately risky behavior.2–4 Core transmission theory maintains that a small group of persons with high rates of partner change contribute disproportionately to the spread of disease.5 Prompt identification and treatment of these individuals should contribute disproportionately to decreasing disease rates. However, developing interventions to identify these individuals has been challenging. Public health officials have traditionally relied on partner notification, targeted screening, and private provider testing as primary case-detection strategies.6 Few studies have addressed the extent to which these strategies succeed or vary in the identification of high-risk persons (i.e., probable core transmitters) who disproportionately contribute to syphilis persistence. However, identification of high-risk individuals when they are no longer infectious is not of great value; an assessment of prevention potential must also consider stage at the time of case identification.
We evaluated data from 2 syphilis-elimination demonstration sites to determine which case-detection methods identified the most cases of early syphilis and the most cases among persons with high transmission risk.
We selected 2 cities where recent outbreaks of syphilis occurred among heterosexuals (Indianapolis, Indiana, and Nashville, Tennessee). Data from syphilis interview records, laboratory records, and morbidity reports were abstracted for all reported syphilis cases from 1997 through 2002. Syphilis stage was based on the stage that was reported to Centers for Disease Control and Prevention (CDC) and both programs used standard CDC case definitions.7 Major programmatic case-detection categories were developed using 2 routinely collected variables: the method of case detection and the information source. The “method of case detection” is how the case was discovered (provider referral, cluster, patient referral, prenatal, delivery, institutional screening, community screening, reactor, provider report, volunteer, other). The “information source” is the type of provider that first identified the case (STD clinic, HIV counseling and testing site, drug treatment, family planning, prenatal/obstetrics, tuberculosis, other clinic, private physician/HMO, hospital [inpatient], emergency room, correctional facility, laboratory, blood bank, delivery, prenatal, job corps, Indian health service, military, other). Using both variables, we recoded the data to create a single variable that identified both process and source for major programmatic approaches to case identification: partner notification (by the health department or patient), cluster interviews, walk-ins to STD clinics (i.e., volunteers), private medical doctors and HMOs, jails, hospitals (inpatient or emergency department), and other. The “other” category is comprised of the “information source” categories for which 5 or fewer cases were coded (i.e., HIV counseling and testing site, drug treatment, family planning, prenatal, tuberculosis clinic, blood bank, job corps, military, and other).
We used the disease stage and total number of reported sex partners to estimate the relative likelihood and magnitude of future transmission had the case not been identified and treated (Fig. 1). For each stage of disease, we estimated the number of infectious days that would have occurred had the case not been treated. We assumed that all transmission occurs during primary or secondary syphilis, the risk of transmission (per day) was the same for every day while infectious, and that there was a 0.24 rate of relapse (recurrence) to secondary for early-latent cases.8 We also assumed that cases were detected in the middle of the interval for the stage and the intervals were: primary 12 days, early-latent (before secondary) 34 days, secondary 110 days, and early-latent (after secondary) 184 days.8 Using these parameters, we estimated that primary cases have almost twice the transmission potential (and prevention potential) as secondary cases and almost 5 times the transmission potential as early-latent cases (primary = 4.3, secondary = 2.5, and early-latent = 1.0) (Fig. 1). These scores were weighted (multiplied) by the number of reported sex partners to determine the prevention value score. The interview period for partners was 3 months for primary, 6 months for secondary, and 12 months for early latent. Based on these weighted scores, cases were assigned to either low- or high-prevention-value categories. Cases scoring >10 were considered high value; this translates to primary cases that reported 3 or more partners, secondary cases that reported 5 or more partners, and early-latent cases that reported 11 or more partners.
Demographic, risk, and case-detection information were compared between groups. Log transformation was used to compare the mean number of sex partners between groups.9 We computed the logarithm of each value, computed means, and transformed the results back to the original scale for comparison. Statistical comparisons were made using the chi-squared or Fisher exact test (P < 0.01). SAS (version 8.2; Cary, NC) was used to analyze the data.
In Indianapolis, the rate of infectious syphilis began increasing in 1997, peaked in 1999, and decreased dramatically by 2002 (Fig. 2). During this time, 1559 early syphilis cases were reported, including 1126 (72%) primary or secondary cases and 433 (28%) early-latent cases. There were 825 (53%) women; 69% of these women had primary or secondary syphilis. There were 734 men; 76% of these men had primary or secondary syphilis. Among female cases, the mean age was 28.7 years (median, 28 years; range, 14–66 years). Among male cases, the mean age was 35.2 years (median, 34 years; range, 17–82 years). Overall, 90% of cases were black, 7% were non-Hispanic white, and 3% were classified as “other.” Eighty-nine percent of men reported sex exclusively with women and 6% reported sex with other men. Ninety percent of women reported sex with men and 5% reported sex with other women.
Nashville experienced an outbreak in the early 1990s followed by a marked decline in 1995 and a reemergence from 1996 to 2002 (Fig. 2). From 1997 through 2002, 2011 cases of early syphilis were reported, including 998 (50%) primary or secondary cases and 1013 (50%) early-latent cases. There were 882 (44%) women; 49% of these women had primary or secondary syphilis. There were 1129 men, 50% of whom had primary or secondary syphilis. Among male cases, the mean age was 36.5 years (median, 36 years; range, 14–83 years) compared with a mean of 31.1 years (median, 31 years; range, 14–68 years) in females. Overall, 82% of cases were black and 18% were non-Hispanic whites. A higher proportion of female cases than male cases were non-Hispanic whites (23% vs. 14%). Eighty-six percent of men reported sex with women and 9% reported sex with other men. Eighty-nine percent of women reported sex with men and 7% reported sex with other women.
Number of Sex Partners
The interview period for partner elicitation varies by stage; primary is 3 months, secondary is 6 months, and early-latent is 12 months. However, the reported number of partners was similar regardless of the syphilis stage. In Indianapolis, the mean number of reported partners by stage and 95% confidence intervals were: primary, 1.8 (1.5–2.1); secondary, 2.3 (2.1–2.5); and early-latent, 2.4 (2.1–2.7). In Nashville, the mean number of reported partners by stage and 95% confidence intervals were as follows: primary, 1.6 (1.4–1.9); secondary, 1.8 (1.7–2.0); and early-latent, 2.0 (1.8–2.2).
Similarly, the reported number of partners among males was similar regardless of stage of syphilis. In Indianapolis, the mean number of reported partners by stage and 95% confidence intervals were: primary, 2.1 (1.9–2.2); secondary, 2.5 (2.2–2.6); and early-latent, 2.6 (2.3–2.7). In Nashville, we found the following: primary, 1.5 (1.4–1.9); secondary, 1.7 (1.6–1.8); and early-latent, 1.7 (1.6–1.8).
Transmission Risk and Case Detection
In Indianapolis, 13% of cases in women were high prevention value. High-value cases were more likely than low-value cases to be in primary or secondary stage (78% vs. 69%) and report more sex partners (14.8 vs. 1.7) (Table 1). High-value cases were also more likely to report exchanging sex for drugs or money (73% vs. 13%) and report crack cocaine use (52% vs. 16%). Early syphilis cases in general were most frequently identified by private physicians, including health maintenance organizations (HMOs) (28%), followed by partner notification (19%), hospitals (15%), and jails (12%). However, high-value cases were most frequently identified by the local jail (34%) followed by private physicians (18%) and the STD clinic (13%).
In Nashville, 7% of cases in women were high value. High-value cases were more likely than low-value cases to be in primary or secondary stage (61% vs. 49%) and report more sex partners (17.8 vs. 1.6). High-value cases were also more likely to be white (40% vs. 21%) and report exchanging sex for money or drugs (84% vs. 28%). Early syphilis cases were most commonly identified by private physicians, including HMOs (29%), followed by the jail (25%), partner notification (13%), the STD clinic (9%), and hospitals (4%). However, high-value cases were most frequently identified by the local jail (49%) followed by the STD clinic (14%) and private physicians (13%).
In Indianapolis, 20% of cases in men were high value. High-value cases in men were more likely than low-value cases to be in primary or secondary stage (95% vs. 76%) and report more sex partners (5.5 vs. 1.9) (Table 2). High-value cases were also more likely to report exchanging drugs or money for sex (40% vs. 21%). The STD clinic (self-referred) identified the largest proportion of all male cases (32%) followed by private physicians (21%), partner notification (21%), jails (9%), and hospitals (9%). The largest proportion of high-value cases was identified by the STD clinic (44%) followed by private physicians (19%) and partner notification (12%).
In Nashville, 7% of cases in men were high value. High-value cases were more likely than low-value cases to be in primary or secondary stage (85% vs. 50%) and report more sex partners (6.7 vs. 1.5). High-value cases were also more likely to report exchanging drugs or money for sex (48% vs. 29%). The Davidson County Jail identified the largest proportion of all male cases (39%) followed by private physicians (14%), the STD clinic (14%), partner notification (9%), and hospitals (4%). The largest proportion of high-value cases was identified by the STD clinic (28%) followed by the jail (24%) and private physi-cians (14%).
The best methods for finding high-value cases were different from the best methods for finding cases in general. An implicit assumption of syphilis control programs is that finding and treating every syphilis case contributes equally to syphilis control.10 However, if the goal is to reduce disease transmission, health officials should evaluate and prioritize strategies that identify high-risk individuals, early in the course of syphilis, rather than promoting those that are directed at lower-risk individuals or individuals with late-stage syphilis. Previous studies have focused on evaluating specific interventions in terms of the number of cases identified.11–17 Other studies have developed methods to characterize core transmitters and define their role in sustaining syphilis persistence.18–20 However, no studies that we know of have combined these concepts to determine if interventions varied in the identification of high-frequency transmitters.
In both Indianapolis and Nashville, roughly 40% of high-value female cases were identified in jails and 16% were identified by private providers. Jail screening programs are an important means of identifying high-risk women who may not be reached by partner notification or other traditional methods. Female high-value cases (high-frequency transmitters) were more likely than low-value cases to use cocaine, exchange sex for money or drugs, and reported 15 to 20 times as many sex partners as low-risk cases.
Approximately 40% of high-value cases in males were identified at the STD clinic. The majority of these cases were self-referred as a result of symptoms. Overall, 21% of male cases were identified in STD clinics compared with only 9% of cases in females. Thus, in these cities, public STD clinics are important venues for syphilis prevention in males who may otherwise have limited access to healthcare systems. Private providers identified 17% of all male early cases and high-value cases, suggesting that programs targeting providers to increase syphilis testing and reporting are important. Male high-value cases were more likely than low-value cases to exchange money or drugs for sex and they reported roughly 4 times as many sex partners.
The high-prevention-value women usually had 10 or more sex partners. High-prevention-value men were more likely to have primary syphilis than other cases. In both cities, the jail was the most productive site for identifying women with many sex partners, but few of them had primary or secondary syphilis. The STD clinic was the most important site for identifying men with lesions; however, the number of reported sex partners was no higher among these men than those identified in other settings.
Partner notification and cluster interviewing identified relatively few high-value cases. Moreover, the resources required for these activities are substantial. Partner notification interviews of primary and secondary cases identify significantly more infectious partners compared with interviews of early-latent cases; focusing partner notification activities on infectious cases has been proposed previously but has not been widely adopted and deserves consideration in areas with limited resources.21 The marginal success of partner notification as a method for detecting high-value cases should be interpreted cautiously. An equally important purpose of partner notification is the identification and preventive treatment of exposed sex partners who may be incubating infection. In theory, more syphilis is prevented by treating incubating syphilis than by treating primary syphilis. However, quantifying the impact on transmission is difficult. Additionally, our model does not account for transmission from sex partners of index cases.
Our study is subject to limitations. We relied on disease stage to estimate the number of infectious days prevented over a 1-year period. These estimates were calculated using parameters from old studies that would be difficult to replicate today; therefore, the current validity of these parameters is unknown. We used the number of reported sex partners from the partner notification interviews to help estimate the number of future sex partners that would have been exposed if the case had not been treated. The syphilis interview period varied by stage; primary is 3 months, secondary is 6 months, and early-latent is 12 months. However, the number of reported partners in the interview period did not differ significantly by stage, suggesting that this was a fairly stable estimate to use in weighting each case. The basic assumption that the number of past sex partners is a reasonable estimate of the number of near-future sex partners may not be accurate in all circumstances. Misclassification of the stage of disease may be more likely to occur in jails than other settings because of lack of time, space, and expertise for conducting examinations and may have resulted in an underestimation of the number of primary and secondary cases identified in jails. Despite these limitations, we believe that the method described here offers a practical evaluation tool that is translatable to many state and local syphilis control programs.
Our results highlight potentially important findings and opportunities for evaluating syphilis control methods. The case-detection methods that best identified high-value cases in these 2 cities may not be generalizable to other settings; however, it is important that similar evaluations of syphilis prevention activities occur. We estimated that primary cases have almost twice the transmission potential (days of infectiousness) as secondary cases and almost 5 times the transmission potential as early-latent cases. These ratios are very similar to those that were used by disease intervention specialists for prioritizing disease prevention activities in the 1970s and 1980s (primary = 0.9, secondary = 0.45, early-latent = 0.2) and may be useful for other evaluations.22 We found that case-detection methods vary in their ability to identify high-risk transmission cases. However, our analysis was limited to outbreaks among 2 heterosexual populations; additional work is needed to define high-frequency transmitters in outbreaks among men who have sex with men and identify interventions with the greatest potential for reaching them.
1. Division of STD Prevention. The National Plan to Eliminate Syphilis From the United States. National Center for HIV, STD, and TB Prevention, Centers for Disease Control and Prevention, 1999.
2. Rosenberg D, Moseley K, Kahn RH, et al. Networks of persons with syphilis and at risk for syphilis in Louisiana: Evidence of core transmitters. Sex Transm Dis 1999; 26:108–114.
3. Koumans EH, Farley TA, Gibson JJ, et al. Characteristics of persons with syphilis in areas of persisting syphilis in the United States. Sex Transm Dis 2001; 28:497–503.
4. Bernstein KT, Curriero FC, Jennings JM, et al. Defining core gonorrhea transmission utilizing spatial data. Am J Epidemiol 2004; 160:51–58.
5. Brunham RC. The concept of core and its relevance to the epidemiology and control of sexually transmitted diseases. Sex Transm Dis 1991; 18:67–68.
6. Parran T. Shadow on the Land. New York: Reynal and Hitchcock, 1937.
7. Centers for Disease Control and Prevention. Case definitions for infectious conditions under public health surveillance. MMWR Morb Mortal Wkly Rep 1997; 46:34–38.
8. Garnett GP, Aral SO, Hoyle DV, et al. The natural history of syphilis: Implications for the transmission dynamics and control of infection. Sex Transm Dis 1997; 24:185–200.
9. Keene ON. The log transformation is special. Stat in Med 1995; 14:811–819.
10. Oxman GL, Doyle L. A comparison of the case-finding effectiveness and average costs of screening and partner notification. Sex Tranms Dis 1996; 23:51–57.
11. Brewer DD. Case-finding effectiveness of partner notification and cluster investigation for sexually transmitted disease and HIV. Sex Transm Dis 2005; 32:78–83.
12. Kohl KS, Farley TA, Ewell J, et al. Usefulness of partner notification for syphilis control. Sex Transm Dis 1999; 26:201–207.
13. 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.
14. Kahn RH, Scholl DT, Shane SM, et al. Screening for syphilis in arrestees: Usefulness for community-wide syphilis surveillance and control. Sex Transm Dis 2002; 29:150–156.
15. Gardella C, Marfin AA, Kahn RH, et al. Persons with early syphilis identified through blood or plasma donation screening in the United States. J Infect Dis 2002; 185:545–549.
16. Patrick DM, Rekart ML, Jolly A, et al. Heterosexual outbreak of infectious syphilis: Epidemiological and ethnographic analysis and implications for control. Sex Trans Infect 2002; 78(suppl 1):164–169.
17. Kahn RH, Moseley K, Thilges JN, et al. Community-based screening and treatment for STDs: Results form a mobile clinic initiative. Sex Transm Dis 2003; 30:654–658.
18. Rothenberg R, Kimbrouh L, Lewis-Hardy R, et al. Social network methods for endemic foci of syphilis: A pilot project. Sex Transm Dis 2000; 27:12–18.
19. Cates W, Rothenberg RB, Blount JH. Syphilis control: The historic context and epidemiologic basis for interrupting sexual transmission of Treponema pallidum. Sex Transm Dis 1996; 23:68–75.
20. Oxman GL, Smolkowski K, Noell J. Mathematical modeling of epidemic syphilis transmission: Implications for syphilis control programs. Sex Transm Dis 1996; 23:30–39.
21. Gunn RA, Eldred SE, Mathews C. Letter to the editor. Sex Transm Dis 2001; 28:455–456.
22. Guidelines for STD Control Programs Operation. Atlanta: Department Health and Human Services, Public Health Service, Centers for Disease Control, Center for Prevention Services, Division of Sexually Transmitted Diseases, Field Operations Section, p 15.