SEXUALLY TRANSMITTED INFECTIONS (STIs) are spread through contact networks with a much lower average numbers of contacts than most other types of contagious diseases.1 The risk of an individual becoming infected is therefore not only determined by the behavior of the individual, but also by the structure of the sexual network and the individual's position in this structure.2 Since the emergence of HIV, the study of sexual networks in the transmission dynamics of STIs has become quite extensive.3,4 We now know that the structural properties of contact networks that facilitate the spread of contagious diseases such as a large variation in number of contacts,5 assortative interaction,6,7 and concurrent relationships8 can be found in sexual networks.
The speed at which a disease is spread in a network is closely related to the average path distance, that is, the average number of links needed to “walk” the shortest path between all individuals in the network. One recent important discovery that is likely to affect the transmission dynamics is the so-called small-world effect.9 The implication of this effect is that only a small fraction of the links (connections) between pairs of random nodes (individuals in this case) is needed to drastically lower the average path distance in any structured interconnected network to the short average path distance that we would observe in a random network with a similar number of nodes and links. The fact that individuals usually have sexual contact with people living close by with similar sociodemographics (ethnicity and class) may therefore have only a minor limiting effect on the speed of the spread of STIs if a small fraction of contacts transcends the geographic and social space. The effect of local intervention such as contact tracing may also be reduced if the disease is continually imported into the region. The study of such contacts (bridges) is therefore of great importance because intervention targeted at individuals (bridgers) with sexual contacts that transcend geographic and social space may drastically increase the average path distance between the individuals, thereby reducing the speed at which STIs are spread.
When talking about sexual bridges, one typically distinguishes between 2 types: bridges between different subgroups in a population and spatial (geographic) bridges between populations.
Mixing between high-prevalence and low-prevalence groups in the population has been shown to be important for the spread of STIs and thus for sustaining an epidemic and/or endemic state,10 and these phenomena have been widely studied (see, for example,8,11–14). The indirect mixing pattern between geographically separate locations, also called spatial bridging, has also been studied, but not as thoroughly.
When outbreaks of STI occur in different geographic areas, it may appear that they are isolated phenomena; however, individuals who live far away from each other can still be part of the same network. In an analysis of a gonorrhea outbreak in Alberta, Canada, sexual contact was reported between individuals as far as 220 km apart; in this case, a motel bar was the link.15
This phenomenon is also clearly seen in the current outbreak of lymphogranuloma venereum (LGV) in Europe. Here, infected individuals mainly reside in metropolitan areas such as Amsterdam, Paris, and London, and LGV transmission takes place at international meeting places in these cities. Links are then found to sporadic cases of the infection in countries as distant as Sweden and Spain.16 In this case, individuals are linked in a network by high-risk behaviors rather than by geographic proximity.
In a study conducted in Manitoba, Canada, Wylie and Jolly7 used network analysis to examine the spread of chlamydia and gonorrhea. Among other things, they found that almost 80% of the network components defined contained individuals from at least 2 different geographic locations, many of whom had sexual contacts both inside and outside high-prevalence areas and thus acted as bridgers between them. The bridgers found in the material linked together 20 communities inside Manitoba and Ontario as well as 4 other provinces.7
Individuals can, of course, be defined both as bridgers between core groups and the general population and as spatial bridgers. Sex workers clearly constitute a core group that can bridge populations. In a recent study, Williams et al17 examined a group of male sex workers in Houston, Texas. They found that slightly less than half of the male sex workers examined acted as bridgers between sexual networks in different cities and that these connected Houston to almost all parts of the country. They also showed that spatial bridging behavior was significantly associated with a number of other high-risk behaviors for contracting STIs such as a large number of partners, intravenous drug use, and homosexuality. Geographic bridging behavior was also correlated with being HIV-positive and having spent time in jail.17
Another study that examined the characteristics of spatial bridgers was conducted by Kerani et al.18 They used randomized data derived from partner notification to study 2912 heterosexuals who had been diagnosed with chlamydia, gonorrhea, or both. In this population, compared with the male sex workers mentioned here, a much smaller fraction (5.2%) was labeled as spatial bridgers. They, too, found that spatial bridging behavior was associated with some types of high-risk behavior, in this case having a large number of partners and more frequent concurrent sexual relationships.18 However, their results also show a more complex picture, with spatial bridgers of both sexes being older than nonbridgers, male bridgers having a significantly higher educational level than did male nonbridgers, and the female bridgers generally residing in areas with a higher socioeconomic status than women not labeled as spatial bridgers. These are sociodemographic characteristics usually associated with subpopulations having a lower prevalence of STIs. Kerani et al showed that many of these individuals may also be defined as population bridgers. They are therefore probably responsible for both importing STIs and for transmitting them to subpopulations that are more vulnerable for contracting and spreading STIs.18
A more general study of sexual behavior that also raised the question of geographically distant sexual partners was the British National Survey of Sexual Attitudes and Lifestyles (Natsal 2000). Fourteen of the men and 7.1% of the women reported that they had had one or more new sexual partners from outside the United Kingdom in the past 5 years. This proportion was largest in the 16-year to 24-year age group for both women and men and declined with age.19
To better understand the phenomenon of spatial bridging behavior both in subgroups and in the general population, we present a study covering all detected chlamydia cases and their traced partners in a Swedish county over the course of 1 year. The main objective of the study is to analyze the sociodemographic characteristics of spatial bridgers and to compare these with the rest of the study population.
Materials and Methods
Study Population/Data Material
Värmland County is located in midwestern Sweden and has 274,000 inhabitants. All Chlamydia trachomatis tests in Värmland County were analyzed at the Department of Clinical Microbiology, Central Hospital in Karlstad (the major city in the county) using a commercial polymerase chain reaction-based method20; in the study year 2001, a total of 9538 examined individuals generated 11,019 tests.* Of the eligible population (all individuals in the age group 15–40 living in the county of Värmland in 2001), approximately 18% of the women and 3.5% of the men were tested for chlamydia during the study year (Table 1).
Most of the individuals tested had only been tested once during the year (Table 2), but one single individual had been tested as many as 6 times.† The major reasons for testing (approximately 80% of the tests) were contact tracing (see subsequently), clinical symptoms, and opportunistic screening.
Although 81% (n = 7752) of the individuals tested were women (vs. n = 1621 for men), they only accounted for 57% (n = 368) of the positive test results. The chlamydia rate was 4.8% in tested women and 17% in tested men (Table 1).
A significantly higher proportion of those who lived in Karlstad (12.1%, n = 3504) were tested for chlamydia than of those in the same age groups who resided in the rural areas (9.5%, n = 5869) (P = 0.000). However, of those tested, individuals in the rural areas were as likely as those living in the city to test positive (n = 405, 6.9% vs. n = 225, 6.4%, P = 0.4).
The Contact Tracing Procedure/Data Collection
Since 1988, cases of chlamydia must by law be reported in Sweden. Contact tracing (partner notification) is mandatory for each new case.21 The contact tracing procedure asks about partners at least 6 months back in time, and it is partly performed by physicians who are not specialists in the procedure. Referring to this lack of specialization and to the fact that physicians are not able to spend the required time on the task, some researchers 22–24 have raised questions about the efficacy of this procedure. Therefore, specially trained midwifes were assigned this task as part of this study. The result of this reorganization was an increase from 1.6 contacts found per index case to 2.1 compared with a pilot study in the same county 2 years earlier.25
Contact tracing was performed in approximately 98% of all cases in Värmland that tested positive, and the dataset derived from this procedure consists of information on 868 individuals. In addition to the information collected by contact tracing, we were also able to gather data on 851 of the individuals from Statistics Sweden, data that consists of information on the sociodemographic variables shown in Table 3. These 851 individuals were included in our analysis.
The bivariate analysis was carried out through a comparison of odds ratios. The variables that were shown to be significant at a 5% confidence level in the bivariate analysis were then included in a multivariate logistic regression. SPSS was used for all statistical analysis.26
In this study, we have defined a spatial bridger as an individual who in the contact tracing procedure reported 2 or more partners, at least one of whom resided outside the county of Värmland and at least one of whom resided in the county. These individuals have the potential to import chlamydia into Värmland, that is, to contract the infection outside and then infect someone who resides in the county. We also identified potential spatial bridgers, defined as individuals who had 2 or more distant partners (outside the county of Värmland) but no partners who could be labeled as local. The definitions used in this study thus follow the ones used by Kerani et al,18 which means that the results obtained here can be compared with those from that study. In contrast to the Kerani study, gonococcal infections were not included in our study, but because the prevalence is very low in Sweden (3–6 annual cases per 105 inhabitants in 1997–200427), it would not alter the number of bridgers. Unfortunately, the timespan of the contact tracing is not the same in this study as in the previously mentioned one and therefore the number of partners cannot be compared directly.
Characteristics of the Study Population
Of the 851 individuals who were included in this study, 68% were index cases and, consequently, 32% partners. Four hundred sixty-nine individuals (55%) had been tested as a result of the contact tracing procedure and 45% (382 individuals) for other reasons. Fifty percent of the 851 individuals were women. The age distribution ranged from 13 to 37, the mean age for women being 22 and for men 24 (P = 0.00).
Most of the individuals in the study population, 68% (Table 2), had only been tested once during the year; the highest number of tests was 5. The number of partners reported ranged from zero to 13 (Fig. 1). Men reported more partners (2.8 vs. 2.3 for women), but the median number of partners reported was 2 for both sexes.
The Spatial Bridgers
A total of 68 (8.0%) of the 851 individuals who were included in this study were labeled as spatial bridgers. There was a significant difference in the gender distribution—68% of the spatial bridgers were women (P = 0.00). None of the spatial bridgers were married.
When looking at place of residence, we found that most of the spatial bridgers (56%) lived in the rural areas, and this was also the case with the nonbridgers (59% lived outside the city of Karlstad). None of these differences are statistically significant, however.
The spatial distance between individuals infected with chlamydia and their sexual partners is shown (using the center of the individuals' county of residence as the point of origin) in Figure 2.
The postal code areas of the cases and partners showed that 59% of all relationships in the dataset were within a distance of 10 km, 20% between 10 and 100 km, 16% between 100 and 1,000 km, and only 4% over 1,000 km (Fig. 3). This is consistent with previous findings that indicate that most sexual partners are found within the closest geographic area.28,29 Using our definition of spatial bridgers, 16 of Sweden's 21 counties were linked together. A total of 46 individuals (26 spatial bridgers and 20 potential bridgers) reported having one or more sexual partners abroad. Together, these bridgers and potential bridgers linked Sweden to 17 different countries, including Thailand, Cyprus, and South Africa, among others.
Eleven percent of the women in the dataset (n = 46 of 429) were identified as spatial bridgers. The proportion of women under the age of 25 was significantly higher among female bridgers than among females not labeled as bridgers (89% vs. 73%, P = 0.02), and the female bridgers were also significantly more likely to be registered as students at the university in the county than were nonbridgers (odds ratio [OR], 2.7, P = 0.00). Bridging behavior among females was also associated with being employed/working (OR, 2.0, P = 0.03) (Table 4).
The women in the dataset who were labeled as bridgers were more likely to be both first- (OR, 2.8) or second-generation (OR, 2.2) immigrants than were the nonbridgers. They also had a significantly lower yearly income,‡ and 7% of these women received social welfare compared with 11% of the women not labeled as spatial bridgers (Table 4).
Furthermore, 11% of the bridgers were tested as a result of contact tracing compared with 44% of the nonbridgers (P = 0.00), and over half of the bridgers (54%) had undergone more than one test for chlamydia during the year (vs. 42% for nonbridgers, not significant) (Table 4).
When adjusting for all the variables significant at a 5% risk level in the bivariate analysis, female bridgers were still more likely to be registered students at Karlstad University and to have been tested for reasons other than contact tracing (Table 5).
Of all the men in the sample (n = 422), 5.2% (n = 22) were identified as spatial bridgers. In contrast to the females, a larger proportion of males was older than 25 (54% vs. 41% in nonbridgers) (Table 4). None of the male bridgers was registered at Karlstad University compared with 7.5% of the nonbridgers. Bridging behavior among males was not associated with being employed/working (OR, 0.70, not significant) or with receiving social welfare (134% of the bridgers vs. 17% of the nonbridgers) (Table 4).
All of the men labeled as spatial bridgers were born in Sweden, but they were more likely than nonbridgers to be second-generation immigrants (OR, 1.19). The male bridgers had a higher mean yearly income than the nonbridgers, although this difference was not significant (Table 4). As was the case with females, male bridgers were more likely to have been tested for reasons other than contact tracing (OR, 5.38, P = 0.000), and to a greater extent than nonbridgers, they had, like the females, undergone more than one test during the year (27% vs. 19%, not significant) (Table 4).
Because only one variable was significant at a 5% confidence level (“reason for testing”), there was no need for multivariate analysis.
The Potential Bridgers
We also identified 30 potential bridgers, that is, individuals who reported having 2 or more partners outside the county but no partners in Värmland. To determine whether there were any differences between these individuals and the ones labeled as spatial bridgers, we tried to include them in this analysis. This did not result in any significant changes, and we decided not to take these individuals into account.
For individuals, spatial bridging behavior may not in itself be a high-risk activity, but sexual mixing (bridging) between geographic areas may fuel the transmission of STIs if they represent low- and high-prevalence populations.30 Such geospatial distribution is well established for both gonococcal31 and chlamydial infections 32–34 in larger cities, but has also been noted in moderately urbanized areas.35–37 In our study, we found that the majority of sexual partners resided within 10 km of each other (like also in 28,29) and that no difference in the incidence of chlamydia could be found when comparing Karlstad with other municipalities in the county. This indicates that chlamydia is endemic throughout the county and is widespread in the population, a finding that is consistent with official figures on the prevalence of chlamydia in different regions of Sweden.27
As mentioned at the beginning of this article, the prevalence of spatial bridgers in a sexual network can have serious effects on the spread of STIs. In our study, 8% of the individuals in the study population were identified as spatial bridgers, a fraction that (as was shown by Watts and Strogartz in 19989) is more than enough to characterize the network as a small-world network. This finding indicates that the removal of these links that is targeted information to individuals labeled as spatial bridgers is one interventional approach that would be effective in reducing the speed at which STIs are spread between populations with different levels of incidence.
To carry out an intervention, one must identify the individuals to whom information should be targeted. Previous studies have indicated that spatial bridging behavior may be associated with different types of high-risk behavior,17 being older, having a higher educational level, and living in areas with a higher socioeconomic status.18 We did not find any support for this in our population or much evidence that individuals who engage in spatial bridging behavior are easy to distinguish from others. When using multivariate analysis separately for the sexes, the only significant associations with being a bridge was (for females) being a student and (for both sexes) the reason why they chose to be tested in the first place.§
The conclusion of the results is that spatial bridgers can fuel the transmission of chlamydia between geographically separated subpopulations with different rates of prevalence. Thus, there is a significant risk that spatial bridges will reduce the effect of contact tracing in local populations as a result of their ability to reintroduce the infection.30 Therefore, efficient contact tracing intervention requires going beyond the local clinic level and should be coordinated at the county level and, ideally, at higher levels as well. Spatial bridgers should therefore be considered as potentially important in the design of interventions against STIs.
1. Hethcote H, Yorke JA. Gonorrhea: Transmission Dynamics and Control. New York: Springer, 1984.
2. Klovdahl AS. Social networks and the spread of infectious diseases: the AIDS example. Soc Sci Med 1985; 21:1203–1216.
3. Liljeros F, Edling CR, Amaral LAN. Sexual networks: implications for the transmission of sexually transmitted infections. Microbes Infect 2003; 5:189–196.
4. Doherty IA, Padian NS, Marlov C, et al. Determinants and consequences of sexual networks as they affect the spread of sexually transmitted infections. J Infect Dis 2005; 191:S42–S54.
5. Klovdahl AS, Potterat JJ, Woodhouse DE, et al. Social networks and infectious disease: The Colorado Springs Study. Soc Sci Med 1994; 38:79–88.
6. Rothenberg, RB, Woodhouse, DE, Potterat JJ, et al. Social networks in disease transmission: The Colorado Springs Study. In: Needle RH, Genser SG, Trotter II RT, eds. Social Networks, Drug Abuse and HIV Transmission. National Institute of Drug Abuse Research Monograph No 151 (NIH Publication No. 95-3889), 1995:3–19.
7. Wylie JL, Jolly A. Patterns of chlamydia and gonorrhea infection in sexual networks in Manitoba, Canada. Sex Transm Dis 2001; 28:14–24.
8. Morris M, Zavisca J, Dean L. Social and sexual networks: Their role in the spread of HIV/AIDS among young gay men. AIDS Educ Prev 1995; 7(suppl):24–35.
9. Watts DJ, Strogatz SH. Collective dynamics of ‘small-world' networks. Nature 1998; 393:440–442.
10. Service S, Blower SM. HIV transmission in sexual networks: An empirical analysis. Proc Biol Sci 1995; 260:237–244.
11. Aral SO, Hughes JP, Stoner B, et al. Sexual mixing patterns in the spread of gonococcal and chlamydial infections. Am J Public Health 1999; 89:825–833.
12. Gorbach P, Sopheab MH, Phalla T, et al. Sexual bridging by Cambodian men: Potential importance for general population spread of STD and HIV epidemics. Sex Transm Dis 2000; 27:320–326.
13. Ford K, Sohn W, Lepkowski J. American adolescents: Sexual mixing patterns, bridge partners, and concurrency. Sex Transm Dis 2002; 29:13–19.
14. Kraut-Becher JR, Aral SO. Gap length: An important factor in sexually transmitted disease transmission. Sex Transm Dis 2003; 30:221–225.
15. De P, Singh AE, Wong T, et al. Sexual network analysis of a gonorrhoea outbreak. Sex Transm Infect 2004; 80:280–285.
16. Fenton KA, Imrie J. Increasing rates of sexually transmitted diseases in homosexual men in Western Europe and the United States: Why? Infect Dis Clin North Am 2005; 19:311–331.
17. Williams ML, Atkinson J, Klovdahl A, et al. Spatial bridging in a network of drug-using male sex workers. J Urban Health 2005; 82(suppl 1):i35–42.
18. Kerani RP, Golden MR, Whittington WL, et al. Spatial bridges for the importation of gonorrhea and chlamydial infection. Sex Transm Dis 2003; 30:742–749.
19. Johnson AM, Mercer CH, Erens B, et al. Sexual behaviour in Britain: Partnerships, practices, and HIV risk behaviours. Lancet 2001; 358:1835–1842.
20. Cobas Amplicor C. trachomatis
Test. Branchburg, NJ: Roche Diagnostics.
21. Ministry of Health and Social Affairs. The Communicable Diseases Act and Other Legislation on Control of Communicable Disease. Stockholm: International Secretariat, October 1989.
22. Eitrem R, Erenius M, Meeuwisse A. Contact tracing for genital Chlamydia trachomatis
in a Swedish county. Sex Transm Dis 1998; 25:433–436.
23. Gustafsson B, Parment PA, Ramstedt K, et al. [A questionnaire study in the county of Stockholm on transmission control of chlamydia infections. Too many physicians neglect the contact tracing.] Lakartidningen 2000; 97:3269–3272.
24. Gotz H, Lindback J, Ripa T, et al. Is the increase in notifications of Chlamydia trachomatis
infections in Sweden the result of changes in prevalence, sampling frequency or diagnostic methods? Scand J Infect Dis 2002; 34:28–34.
25. Osterlund A, Persson T, Persson I, et al. Improved contact tracing of Chlamydia trachomatis
in a Swedish county—Is genotyping worthwhile? Int J STD AIDS 2005; 16:9–13.
28. Zenilman JM, Ellish N, Fresia A, et al. The geography of sexual partnerships in Baltimore: Applications of core theory dynamics using a geographic information system. Sex Transm Dis 1999; 26:75–81.
29. Rothenberg R, Muth SQ, Malone S, et al. Social and geographical distance in HIV risk. Sex Transm Dis 2005; 32:506–512.
30. Woodhouse DE, Potterat JJ, Muth JB, et al. A civilian–military partnership to reduce the incidence of gonorrhea. Public Health Rep 1985; 100:61–65.
31. Rothenberg RB. The geography of gonorrhea. Empirical demonstration of core group transmission. Am J Epidemiol 1983; 117:688–694.
32. Shahmanesh M, Gayed S, Ashcroft M, et al. Geomapping of chlamydia and gonorrhoea in Birmingham. Sex Transm Infect 2000; 76:268–272.
33. Elliott LJ, Blanchard JF, Beaudoin CM, et al. Geographical variations in the epidemiology of bacterial sexually transmitted infections in Manitoba, Canada. Sex Transm Infect 2002; 78(suppl 1):i139–44.
34. Monteiro EF, Lacey CJ, Merrik D. The interrelation of demographic and geospatial risk factors between four common sexually transmitted diseases. Sex Transm Infect 2005; 81:41–46.
35. Herrmann B, Egger M. Genital Chlamydia trachomatis
infections in Uppsala County, Sweden, 1985–1993: Declining rates for how much longer? Sex Transm Dis 1995; 22:253–260.
36. Law DC, Serre ML, Christakos G, et al. Spatial analysis and mapping of sexually transmitted diseases to optimise intervention and prevention strategies. Sex Transm Infect 2004; 80:294–299.
37. van Bergen J, Gotz HM, Richardus JH, et al. Prevalence of urogenital Chlamydia trachomatis
increases significantly with level of urbanisation and suggests targeted screening approaches: Results from the first national population based study in the Netherlands. Sex Transm Infect 2005; 81:17–23.
*In Osterlund et al,24 the number of tests is said to be 12,761. The difference in this number in this article is the result of the fact that we have removed the tests that were part the contact tracing but were performed in the year before and after.
†Like in most counties in Sweden, follow-up testing is not common in Värmland. As a result, most retesting was due to having symptoms or because the individual wanted to be tested after changing partners.24
‡This variable is, however, not tested in the multivariate analysis as a result of correlation problems with the variable “working” (r = 0.71).
§Individuals who were tested as a result of a contact tracing were less likely to be labeled as bridgers. This association is probably the result of the fact that a larger actual number of individuals who are tested and found to be positive on screening than through contact tracing.