HISTORICALLY, GONORRHEA HAS been a nationally notifiable sexually transmitted infection (STI) in Canada since 1924. However, genital chlamydia infection is a more recently recognized disease: The Public Health Agency of Canada began national surveillance in 1990.1 Since treatment for gonorrhea is not effective against chlamydia, many people treated for gonorrhea had a concomitant, silent chlamydia infection that remained untreated.2 Canadian guidelines now recommend additional treatment for chlamydia when the presence of gonorrhea is suspected; however, chlamydia remains a ubiquitous disease.3
In the Calgary Health Region (CHR), gonorrhea and chlamydia diagnoses increased after years of decline. Gonorrhea infections decreased from 51.2 to 15.5 per 100,000 between 1990 and 1995, and then increased to 19.75 per 100,000 in 2001. Chlamydia infections decreased from 309.7 to 177.4 per 100,000 between 1990 and 1997 and subsequently increased to 200.84 per 100,000 in 2001.1,4,5
Transmission of an STI is not homogeneous across the population; for example, a few individuals are rapidly, frequently, or heavily infected whereas the majority of the population either evade infection or suffer infrequent or light infections.6–8 In STI clinics in Alberta, the highest rates of reportable STIs are found in the 15- to 24-year-old age group and particularly in women.9 Infections with chlamydia and gonorrhea are often difficult to diagnose in women since they are frequently asymptomatic.10,11 This results in a higher risk of continued transmission due to such factors as asymptomatic infections, failure to recognize symptoms of infection, and failure to seek diagnosis and treatment in a timely fashion.12 These factors may be influenced by high prevalence or exposure to prostitution, illicit drug use, low economic and academic achievement, low utilization or limited access to health care resources, and related socioeconomic and environmental factors.13
There are 2 main approaches to identifying groups who are at high risk of acquiring an STI: Identify individuals or identify high prevalence areas.14 Individuals with repeat infections (within a 12-month period) or who are coinfected with gonorrhea and chlamydia have been found to be useful markers for identifying individuals at high risk of acquiring an STI.15 In general, it is difficult to identify individuals at high risk since most other characteristics [such as race, socioeconomic status (SES), or age] are risk markers for behaviors of subgroups and are stigmatizing to individuals.3,5,13,15,16 In addition, the individuals comprising high-prevalence groups often do not present to the health care system for testing or for treatment; hence, they continue to transmit the disease. The other strategy is to identify key geographic areas where STI transmission rates are highest through mapping of STI prevalence rates. This geographic clustering of STI infections probably results from the existence of localized sexual networks at the neighborhood level. STI control programs targeting these areas will then theoretically impact high-prevalence group members.6,10,15
Geographic information systems (GIS) are software tools that allow for the processing of spatial data for visualization and exploratory data analysis.14,17 Spatial analysis is the ability to manipulate spatial data (such as Postal Codes, PCs) into different forms and extract additional meaning as a result. Layers of data, such as geographic location, type of STI, and SES variables, can be entered into a GIS and a map generated to identify statistically significant clusters of disease.14
The purpose of this article is to use an ecological study design to describe the geographic distribution of people with reported gonorrhea and chlamydia infections, and to identify “high-risk” areas for targeting resources for STI prevention and control programs.
The STI study sample includes 3 subgroups of the local general population residing in the CHR: those patients seen through the Family Planning Clinic, the STI clinic, and the remainder of service providers of the CHR. Patients attending the STI clinic are tested because they may be symptomatic and seeking medical attention or may be asymptomatic contacts of an index case of STI. Other patients seen through the Family Planning Clinic or in the remainder of the health care system (private physicians, walk-in clinics, and hospital emergency rooms) are primarily tested for STI because they present with symptoms, are tested for routine STI screening, or are tested at the patient's request. The STI Services surveillance database also contains data from any positive diagnostic test performed by Calgary Laboratory Services or by the Provincial Laboratory of Public Health (Microbiology) for southern Alberta.
The study population was extracted from the Alberta Health and Wellness STI surveillance database and consisted of all individuals reported as having a diagnosis of gonorrhea between January 1, 1998, and June 21, 2002, or having a diagnosis of chlamydia between December 23, 1997, and July 8, 2002. Individuals were included who had a valid PC within the Calgary city limits. Only those individuals with a diagnosis of chlamydia or gonorrhea were included: individuals with other STIs such as nongonococcal urethritis were excluded from the study population.
The STI data were imported into the GIS, ArcView 8.1 (ArcGIS, ESRI, Redlands CA). The Statistics Canada Postal Code Conversion File was used to link the cleaned STI data to geographic coordinates in order to map the data in the GIS software. The Postal Code Conversion File is a digital file that provides a link between the 6 character PC and Statistics Canada's standard geographical areas for which Census data are produced. It also provides the x, y (latitude/longitude) coordinates for a point representing the approximate location of the PC to support mapping.18 The numbers of cases were then aggregated first at the PC level and then at the Census Tract level. Rates of chlamydia and gonorrhea per 100,000 population were calculated at the Census Tract level. The STI rates were then converted to core (top 25% of rates), adjacent (next 25% of the rates), and peripheral (bottom 50% of the rates) categories and plotted on the map of the CHR, replicating previous authors' methodology.6,14,15
The core model break points were analyzed spatially to determine the pattern of core, adjacent, and peripheral areas for chlamydia and gonorrhea in the CHR. Spatial analysis was performed using CrimeStat, a program provided on the Internet by the United States Department of Justice for “hotspot” analysis of GIS data (www.icpsr.umich.edu/NACJD/crimestat.html). The program analyzed the distribution of gonorrhea or chlamydia cases by PC using nearest neighbor hierarchical spatial clustering.
Data from the Canadian 2001 Census were used to estimate mobility, marital status, presence of detached houses in the Census Tract, employment rate, university education, lone female parent, median household income, proportion of population 15 to 40 years, and owning a dwelling for Calgary Census Tracts (Statistics Canada, Microdata Release File). Pearson's correlation coefficients were calculated at the Census Tract level to look for ecological associations between the calculated STI rates and the socioeconomic rates derived from the 2001 Census data. The data were aggregated such that at least 5 cases of chlamydia or gonorrhea occurred for each geographic aggregation to provide stable rates for comparison and to mask individuals to preserve anonymity.10 The aggregation was performed visually by clustering Census Tracts that were spatially close together and considering their similarity in SES variables from the 2001 Census. The corresponding Census data for the clustered Census Tracts were then likewise aggregated.
Alberta is a province in Canada with a population of approximately 3.2 million people. Calgary is the larger of the 2 main centers in the province, with a population of nearly 1 million. The 2001 Canadian Census shows that Calgary is a young, wealthy, and well-educated city. The city has experienced the largest population increase of all the large Canadian centers when compared with the last Census done in 1996 with nearly 16% growth. The median age of the population is the youngest at 34.9 years, while median family income at $58,861 is amongst the highest in Canada.
There were a total of 7456 records after the data cleaning was completed. Mapped cases (5890 of 7456; 78.9%) were linked through a valid residential PC to geographic coordinates and were able to be mapped in the GIS. Unmapped cases (1566 of 7456; 21%) could not be linked because a valid PC was not included with the STI data. These were unmapped because the address given was for the Calgary STI Clinic or a Post Office box (19.7%) or because the PC given was incorrect (1.3%).
Overall, the mapped and unmapped cases were similar, indicating no bias in the mapped populations for chlamydia or gonorrhea. There were more females in the mapped populations than in the unmapped populations for both chlamydia and gonorrhea with 69.6% versus 53.9% (P <0.0001) females with chlamydia and 37.1% versus 24.7% (P =0.0016) females with gonorrhea.
Within the mapped population, major differences were seen in the demographic characteristics of individuals diagnosed with chlamydia and gonorrhea (Table 1). The chlamydia population was comprised of younger people (23.6 years vs. 27.9 years) and more females (69.6%) than the gonorrhea population (37.1%). Differences in ethnic origin reported were also significant but since a high proportion of the population did not provide this information, the data should be viewed with caution.
To describe the sociodemographic characteristics of the Calgary population, Pearson's correlation coefficients were calculated. A pattern of correlations emerged for the 11 independent variables chosen from the 2001 Canadian Census data. Areas reporting a larger population between 15 and 40 years of age also were positively correlated with a reported move within the past 5 years, indicating Calgary has a young, mobile population. These same areas were negatively correlated with having a high median income, being married and owning a home, indicating that this young population is more likely to be single and rent. A positive correlation existed between a high median income, being married, owning a dwelling, and the presence of a high proportion of detached houses in the geographic area. Areas with a high lone female parent population also reported lower university attendance.
In the analysis of correlations of SES variables from the 2001 Canadian Census, variables indicating high SES (high median household income, owning a dwelling, lower housing density, married) were found to be highly correlated with each other. These high SES variables occurred in suburban Census Tracts in Calgary. Similarly, variables indicating lower SES (single mothers, lower education, high housing density, renters, single marital status) were also found to be highly correlated with each other. These low SES variables tended to be located in the downtown and northeast (NE) quadrant of the city.
Correlations between sociodemographic characteristics and the chlamydia and gonorrhea rates were found indicating some ecological associations (Table 2). For each a strong positive correlation exists between STI and a young population (population between 15 and 40 years of age) and the presence of lone single female parent in the Census Tract. High rates of gonorrhea correlated with areas that had a higher proportion of males in the Census Tract. Strong negative correlations were found between the areas with high chlamydia or gonorrhea rates and high income, being married, the presence of detached houses, and owning a dwelling.
The core (top 25% of chlamydia and gonorrhea rates), adjacent (next 25% of the chlamydia and gonorrhea rates), and peripheral (bottom 50% of the chlamydia and gonorrhea rates) points were plotted on the map of the CHR. The maps are presented in Figures 1 and 2. Hotspot analysis using the CrimeStat program showed a significant clustering of PCs with high rates of gonorrhea near the downtown, whereas there was a significant clustering of PCs with high rates of chlamydia near the downtown and in the NE quadrant. Outlier points with high rates were seen in the suburban areas for each map.
The Census Tracts with high rates of chlamydia and gonorrhea infection positively correlate with the low SES variables and negatively correlate with the high SES variables. Those areas with top 25% of STI rates (core areas) had the lowest SES while those areas with bottom 50% of STI rates (peripheral areas) have the highest SES.
There were no significant differences in age between the mapped and unmapped populations with gonorrhea or chlamydia, and women were more likely to report a mappable PC than men. However, the mapped population with chlamydia had more females and was younger than the mapped gonorrhea population. One possible explanation is that gonorrhea is often symptomatic in men and so men may present for gonorrhea testing more frequently than do women. Additionally, women may be screened for chlamydia more frequently than men are since there are often opportunities for screening (such as pregnancy and PAP smear testing) that do not occur for men. Men who have sex with men is an important population for gonorrhea,13 but was not a 2001 Census variable and was not investigated in this study. Other studies have shown no significant differences in the ages of their study populations with chlamydia and with gonorrhea,19,20 but this may be an artefact of the chosen study populations in those centers, such as only specialty STI clinic patients, rather than the population-based sample for this study.
The significant SES variables were not fundamentally different for the chlamydia and gonorrhea positive populations (Table 2). Jolly et al. have shown differences between these populations, perhaps indicating multiple reservoirs of infection.21 Further investigations using a qualitative study design may be needed to determine if Calgary's core members have stronger network connections than was seen in other centers and whether the similarities were the result of exposure to a single reservoir of infection present in the CHR.
Census Tracts with a high prevalence of chlamydia and gonorrhea (core areas) were associated with low SES. Individuals in the populations from these core areas were more likely to be male, single, have low income, and be less educated than those from the peripheral area populations. Frohlich found a similar association between low SES variables and other socially unacceptable behaviors in Quebec.22 In addition to being a marker of low SES, living in single-parent families has also been associated with earlier onset of sexual activity, which is another risk factor for STIs.23 Although these associations are ecological, the core area pattern of high density housing and low SES has been widely reported in the literature.1,13,15,24,25 Low SES may be associated with an ecological risk of exposure to STIs that is independent of individual risk behavior.21
Spatial analysis of the maps of chlamydia and gonorrhea prevalence rates showed that the core (high prevalence) areas clustered in the downtown core and NE section of the city (Figs. 1 and 2). The adjacent (intermediate prevalence) rates were geographically closer to the core areas whereas the peripheral (low prevalence) rates primarily appeared in the suburban Census Tracts. There were some outlier core (high prevalence) areas seen in suburban Census Tracts in the south and Northwest sections of the city. This may be a result of mapping the residential PCs of people with reported gonorrhea and chlamydia infection and may not reflect the areas in the CHR where infection was actually acquired.
The pattern of core (high-prevalence) areas for STIs follows the pattern of low SES areas within the CHR. The finding of the predominant core (high-prevalence) areas occurring in areas of the CHR with lower SES Census Tracts seems intuitive since it is well known in the literature.13,15,25–27 The core model showing patterns of high-prevalence STI areas is supported by the Calgary data where several core (high prevalence STI) areas were detected. Blanchard et al.15 detected core (high-prevalence), adjacent (intermediate prevalence), and peripheral (low prevalence) areas in Winnipeg based on average annual incidence rates for gonorrhea and chlamydia. They found that the 4 PCs in the downtown that comprised the core areas for gonorrhea and for chlamydia overlapped. The Winnipeg core (high-prevalence) area had high population density with greater mobility, higher unemployment rates, lower mean household incomes, and a higher proportion of male residents than the peripheral areas. Becker et al.6 repeated this study in Baltimore and found the core area covered a larger proportion of the city than in Winnipeg and clustered in 2 primary areas. In these studies and the CHR study, the adjacent (lower-prevalence) areas were found at the edge of the core areas and the peripheral (low-prevalence) areas were outside of the adjacent areas.
STI prevention and control activities typically target decreasing the number of new sexual partners and exposed individuals (e.g., through increased condom usage), and decreasing the duration of disease through effective diagnosis and treatment of STIs.27,28 To date, provision of diagnostic services and educating at-risk populations have been key control strategies for both chlamydia and gonorrhea.28 The core model of STI transmission suggests that directing efforts towards members of a core (high-prevalence STI) group yields the largest decrease in STI rates. However, community-based interventions can help to change community social norms and are more effective than trying to change individual behavior since it is difficult for an individual to step out of line with their peers.29,30 This study identified those high prevalence communities (downtown and NE) where focused interventions must be directed if the prevalence of chlamydia and gonorrhea in the CHR is to decrease.
STIs are spread through an interconnected networks of sexual partnerships that are strongly influenced by social and geographic factors.31,32 The geographic clustering of STI prevalence rates found in this study probably results from the existence of localized sexual networks at the neighborhood level.15 Geographic approaches to STI prevention in areas where core groups have been found to cluster have been advocated as a viable alternative to case-by-case contact tracing and have shown success in lowering disease incidence.33 However, this geographic approach of targeting the high-prevalence areas could miss certain individuals in a sexual network who then can reintroduce infection into the core group.31 In any sexual network, the risk of STI depends both on individual behaviors and the environmental risk, such as the infection status and behaviors of people in the sexual network with whom the individual interacts.34 The magnitude of risk for infection with an STI depends on the prevalence of STI in that population: there is no risk of STI transmission between 2 individuals with risky sexual behavior if they are both disease free.23,31,34 A possible intervention to address both the individual and environmental risk factors for transmission of STIs is an enhanced education program in the school system that addresses both individual characteristics (enhanced self-esteem and better decision-making skills for younger age groups) and group characteristics such as knowledge about STIs and prevention activities (e.g., condom usage).
This study relied on an objective data source for STI information and did not rely on self-reporting, which is notorious for errors in reporting of socially sensitive diseases. The cases were reported as part of Alberta's legislated public health reporting requirements from the entire CHR by physicians and laboratories. Completeness of the data are critical to ensure that case detection is representative of the population at large.35 Quartiles of the rates were used to provide an objective definition for the core (high prevalence), adjacent (intermediate), and peripheral (low) areas. Other researchers have used subjective, visual aggregations of the STI prevalence rates but this may not be reproducible in all settings.15 This study was an important first step to establish a baseline of the spatial pattern of gonorrhea and chlamydia in the CHR. STI surveillance activities can be supplemented by repeating the study to determine if there has been a change in the geographic pattern of chlamydia and gonorrhea over time. This study is a valuable addition to what was known by the CHR and by Alberta Health and Wellness about the status of STI infections in Alberta's largest urban centre. This study also added valuable information to help guide program delivery and the allocation of resources in the core geographic areas of high chlamydia and gonorrhea prevalence for prevention and control of STIs in the CHR.
The rates found in this study, although high in some areas, were underestimates of the true rates. The STI services database contains only detected cases; therefore, asymptomatic and undetected cases are not represented. Unmapped cases were not included in the numerator, and so there was an underestimate of the number of STI cases. Underreporting by physicians in areas of high SES has been estimated at 30% to 50%, suggesting that the true rates were higher than reported11,34 and rendered the core (high prevalence) area rates and low SES association weaker. The denominator used for rate calculations was the whole population of each CT. It can be argued that a smaller population of those most likely to be sexually active with multiple partners (i.e., the 15–39 years of age population) should be used to more closely approximate the true STI prevalence rate.6
This study design involved mapping the individuals' residential PC. This location was not necessarily where sexual activity had taken place. This may account for some of the high-prevalence (core) rates that were visible outside the main clusters. Additionally, if all the unmapped cases resided in the same areas, this would have skewed the locations of the core (high-prevalence) rate points. Geocoding was performed using the Postal Code Conversion File and then rates were aggregated to the Census Tract level, which reduced potential error in the true geospatial location of individual PCs.36 Although rate stabilization in small geographic areas in public health research is an issue,37 the choice of aggregation to 5 cases was primarily to preserve individual's anonymity in Census Tracts where there were fewer than 5 cases.
Medical geography serves to increase consciousness about the community and environment as determinants of health.38 In this study, the community may be a marker of other characteristics that place people at high risk of acquiring and transmitting STIs, such as low SES, low education, close proximity to geographic core (high prevalence) areas, and an increased likelihood of selecting a core group member as a sexual partner.
2. Schachter J. Chlamydia trachomatis
: The more you look, the more you find—how much is there? Sex Transm Dis 1998; 25:229–231.
3. Health Canada. LCDC Expert Working Group on Canadian Guidelines for STDs. Highlights: 1998 Edition of the Canadian STD Guidelines. Cat. Number H49-119/1998–1E.
5. AB Health and Wellness. STD Services. 2001. Table B-8. Incident cases and incident rates per 100 000 population of STD by type of disease and region of examination. Available at www.health.gov.ab.ca/rhas/require/b8–00.htm
. Accessed July 14, 2002.
6. Becker K, Glass G, Brathwaite W, et al. Geographic epidemiology of gonorrhea in Baltimore, Maryland, using a geographic information system. Am J Epidemiol 1998; 147:709–716.
7. Hart G. Risk profiles and epidemiologic interrelationships of sexually transmitted disease. Sex Transm Dis 1993; 20:126–136.
8. Woolhouse M, Dye C, Etard J-F, et al. Heterogeneities in the transmission of infectious agents: Implications for the design of control programs. Proc Nat Acad Sci USA 1997; 94:338–342.
9. Romanowski B, Campbell P, Preiksaitis J, et al. Human immunodeficiency virus seroprevalence and risk behaviors in patients attending sexually transmitted disease clinics in Alberta. Sex Transm Dis 1997; 24:487–494.
10. Armstrong M, Rushton G, Zimmerman D. Geographically masking health data to preserve confidentiality. Stat Med 1999; 18:497–525.
11. Belongia E, Moore S, Steece R, et al. Factors associated with the geographic variation of reported chlamydia infection in Minnesota. Sex Transm Dis 1994; 21:70–75.
12. Rice R, Aral S, Blount J, et al. Gonorrhea in the United States 1975–1984: Is the giant only sleeping? Sex Transm Dis 1986; 14:83–87.
13. Rice R, Roberts P, Handsfield H, et al. Sociodemographic distribution of Gonorrhea incidence: Implications for prevention and behavioral research. Am J Public Health 1991; 81:1252–1258.
14. Clarke K, McLafferty S, Tempalski B. On epidemiology and geographic information systems: A review and discussion of future directions. Emerg Infect Dis 2001; 2:85–92.
15. Blanchard J, Moses S, Greenaway C, et al. The evolving epidemiology of chlamydial and gonococcal infections in response to control programs in Winnipeg, Canada. Am J Public Health 1998; 88:1496–1502.
16. Kilmarx P, Zaidi A, Thomas J, et al. Sociodemographic factors and the variation in Syphilis rates among US counties, 1984 through 1993: An ecological analysis. Am J Public Health 1997; 87:1937–1943.
17. Fotheringham A, Brunsdon C, Charlton M. Quantitative geography: Perspectives on spatial data analysis. London: Sage Publications, 2000:15–29.
18. Statistics Canada. Postal Code Conversion File September 2002 Postal Codes Reference Guide. Cat. No. 92F-0153-GIE. 2002.
19. Gunn R, Fitzgerald S, Aral S. Sexually transmitted disease clinic clients at risk for subsequent gonorrhea and chlamydial infections: Possible “core” transmitters. Sex Transm Dis 2000; 27:343–349.
20. Stoner B, Whittington W, Hughes J, et al. Comparative epidemiology of heterosexual gonococcal and chlamydial networks: Implications for transmission patterns. Sex Transm Dis 2000; 27:215–223.
21. Jolly A, Orr P, Hammond G, et al. Risk factors for infection in women undergoing testing for Chlamydia trachomatis
and Neisseria gonorrhoeae
in Manitoba, Canada. Sex Transm Dis 1995; 22:289–295.
22. Frohlich K, Potvin L, Chabot P, et al. A theoretical and empirical analysis of context: Neighbourhoods, smoking and youth. Soc Sci Med 2002; 54:1401–1417.
23. Santelli J, Lowry R, Brener N, et al. The association of sexual behaviors with socioeconomic status, family structure, and race/ethnicity among US adolescents. Am J Public Health 2000; 90:1582–1588.
24. Klausner J, McFarland W, Bolan G, et al. Knock-knock: A population-based survey of risk behavior, health care access, and Chlamydia trachomatis
infection among low-income women in the San Francisco Bay area. J Infect Dis 2001; 183:1087–1092.
25. Bunnell R, Dahlberg L, Rolfs R, et al. High prevalence and incidence of Sexually Transmitted Disease in urban adolescent females despite moderate risk behaviors. J Infect Dis 1999; 180:1624–1631.
26. Martens P, Mayer T, Derksen S. Factors affecting adolescent reproductive health in Manitoba. Can J Public Health 2002; 93:S39–S43.
27. Brunham R. The concept of core and its relevance to the epidemiology and control of Sexually Transmission Disease. Sex Transm Dis 1991; 18:67–68.
28. Brunham R, Plummer F. A general model of sexually transmitted disease epidemiology and its implications for control. Med Clin North Am 1990; 74:1339–1353.
29. Rose G. Sick individuals and sick populations. Int J Epidemiol 1985; 14:32–38.
30. Dunn J, Frohlich K, Ross N, et al. Geography matters: The role of place in explaining social inequalities in health and human development. In: Hertzman C, Heyman J, Barer M, Evans R, eds. Why are Some Populations Healthy and Others Not? 2003.
31. Wylie J, Jolly A. Patterns of Chlamydia and Gonorrhea infection in sexual networks in Manitoba, Canada. Sex Transm Dis 2001; 28:14–24.
32. Shiboski S, Padian N. Population and individual-based approaches to the design and analysis of epidemiologic studies of sexually transmitted disease transmission. J Infect Dis 1996; 174:S188–S200.
33. Zenilman J, 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.
34. Thomas J, Tucker M. The development and use of the concept of a sexually transmitted disease core. J Infect Dis 1996; 174:S134–S143.
35. Jolly A, Wylie J. Gonorrhoea and chlamydia core groups and sexual networks in Manitoba. Sex Transm Infect 2002; 78:i145–i151.
36. Schootman M, Sterling DA, Struthers J, et al. Positional accuracy and geographic bias of four methods of geocoding in epidemiologic research. Ann Epidemiol 2007; 17:464–469.
37. Devine OJ, Louis TA. A constrained empirical bayes estimator for incidence rates in areas with small populations. Stat Med 1994; 13:1119–1133.
38. Bithell J. A classification of disease mapping methods. Stat Med 2000; 19:2203–2215.
© Copyright 2008 American Sexually Transmitted Diseases Association
This article has been cited