HARDICK, JUSTIN BS*; HSIEH, YU-HSIANG PhD*†; TULLOCH, SCOTT BS‡; KUS, JAMES PhD*; TAWES, JENNIFER BS*; GAYDOS, CHARLOTTE A. DrPH*
ONE OF THE CORNERSTONES of HIV infection prevention is management of sexually transmitted infections (STIs). 1 These infections have been considered to be hidden epidemics in the population of the United States, which result in costs in excess of $10 billion per year. 2 An effective means of controlling and preventing STIs, particularly Chlamydia trachomatis (CT) and Neisseria gonorrhoeae (GC) infections, is the implementation of targeted screening and treatment programs. CT and GC infections are readily curable STIs, and screening and treatment programs are a highly effective means for decreasing the prevalence of current infections and for preventing future infections.
Most detention facilities throughout the United States test for STIs on the basis of symptoms. 3,4 Testing of asymptomatic women is rarely performed because of a lack of medical staff at many of these facilities, as well as the large number of intakes performed on a daily basis. When testing for STIs is performed, the results are often obtained after the arrestee is released. 3,4 Despite the problems, this segment of the population presents a high-risk group for STIs and should be considered a reasonable target for public health programs. 5,6 The Centers for Disease Control and Prevention performed a feasibility study that examined the effectiveness of screening women for CT and GC upon intake at several correctional facilities. 3,4 That study demonstrated that the prevalence rates for CT and GC were high in the populations of surveyed women. 3,4
Recently, the availability of urine testing for CT and GC with use of nucleic acid amplification tests has greatly expanded screening venues. 7–10 In addition, urine testing for CT and GC has been previously shown to be cost-effective. 11,12 Therefore, screening incarcerated women at intake may now be feasible.
As part of a larger study further examining the effectiveness of screening programs in detention settings, we determined the prevalence of screening for CT and GC in women arrestees in a high prevalence area. 13 Specific demographic risk factors associated with infections were determined.
A previous study has indicated that a geographic information system (GIS) is useful in determining geographic areas of high GC prevalence, identifying people with high risk and targeting areas where resources may be most needed. 14 We attempted to expand on these previous findings, by utilizing GIS software to study the possible association among infection, race, and area of residence in a high-prevalence area. This information could be used in future interventions, allowing specific targeting of areas and individuals who are at highest risk for CT and GC infection.
Design and Participants
As part of a broader study protocol, women entering the Baltimore Women's Detention Center between July 28, 1999, and June 27, 2000, were offered screening for CT, GC, and pregnancy, in addition to standard intake medical procedures. 13 The study was approved by the Centers for Disease Control and Prevention, the state of Maryland, and the institutional review board of The Johns Hopkins University. Women were informed that the study was voluntary and confidential and were also informed of both the risks and benefits involved. Informed consent was obtained from women volunteering to participate in the study, and demographic information including age, race, pregnancy status, and the residential zip code was collected by questionnaire for statistical analysis. Symptoms of infection were defined as the presence of discharge or dysuria and were self-reported. Demographic information was also collected for women who did not wish to participate in the study.
Detention center staff performed pregnancy testing on urine specimens, and an additional aliquot of urine was sent to The Johns Hopkins University STD laboratory for CT and GC testing. Upon receipt, specimens were accessioned, processed, and tested by ligase chain reaction (LCR; Abbott Laboratories, Abbott Park, IL), which has a sensitivity of 93.1% and a specificity of 97.1%, in accordance with the manufacturer's instructions. Specimens that had been collected more than 4 days before receipt were rejected. Records of the rejected specimens were kept, and the detention center was notified. The study participants were reapproached, if still available. Rejected specimens were not included in the statistical analysis portion of the study.
The detention center was notified of infected participants, and their treatment was administered by the Maryland State Department of Health and Hygiene. Disease intervention specialists (DISs) attempted to locate and counsel women testing positive who had been released before test results were available. Records were also kept indicating the time from testing to treatment for infected participants.
Contingency table analysis compared characteristics of the participants. Age-specific rates in the age groups of <20 years, 20 to 24 years, 25 to 29 years, 30 to 34 years, and >34 years, by race, were calculated and compared in order to evaluate the racial differences in prevalence of CT and GC among female detainees. Age was further categorized into three age groups (<25 years, 25 to 34 years, and >34 years) on the basis of age-specific rates and age distribution. In addition, age was standardized and compared between both racial groups. Participants who were not white or African American were excluded from data analysis because of the small sample size (n = 25). Coinfection was defined as the presence of both CT and GC in the same individual.
Because the numbers of detainees (from 0 to 260) from each zip code varied dramatically by race during the 11-month study period, the zip code–specific rates by each race were not stable and reliable. Instead, race-specific CT and GC prevalences of pooled zip codes were analyzed on the basis of geographic vicinity and the total number of infections in each zip code. Core areas were defined for statistical and geographic analysis as zip codes that were geographically adjacent and contained at least six total infections.
Multivariate logistic analysis was used to identify the risk factors for CT and GC infections. Multivariate analysis was not used to identify risk factors for coinfection because of the small sample size (n = 21). Age, race, and core area were included in the multivariate model. Pregnancy and symptoms of discharge and dysuria were not evaluated in the multivariate analysis because of the small sample size. Interaction between race and core area was further explored in the final multivariate model. In order to explore the interaction between race and core area on prevalent CT/GC infections, the race and core area product term (race*core area) was included in the multivariate logistic regression model. If the interaction term was significant in the multivariate model, race*core group was categorized into four groups—white core area, white noncore area, African American core area, and African American noncore area—in order to present effect modification in these four groups on CT/GC infections. Statistical analysis was performed with SAS version 6.12 software (SAS Institute, Cary, NC).
Women who tested positive for either CT or GC were mapped according to the residential zip code they provided. Participants whose residential zip code was outside the Baltimore area were excluded from geographic analysis, as were participants who provided no residential zip code. Mapping was performed with use of the GIS software Mapinfo (Mapinfo Corporation, Troy, NY). Baltimore city zip codes were used with the permission of the Baltimore City Health Department.
Between July 28, 1999, and June 27, 2000, women were asked to participate. There were 1931 total participants. From February 15, 2000, to June 2000, the participation rate was 88.9% (634/713), with an 11.1% (79/713) refusal rate. Refusals were not monitored before February 2000.
Of the 1931 participants, 1858 (96.1%) were included in the study, while specimens from 73 participants (3.8%) were not processed. This was due to extended length of transit (49.3%), mismatched identifiers on the container and questionnaire (32.8%), and container leakage (17.9%).
Of the 1858 persons tested, 17.4% (323/1858) were white, 81.3% (1510/1858) were African American, and 1.3% (25/1858) indicated that they were of a different racial background. Ages of the participants ranged from 16 to 63 years, with a mean age of 32.8 and median of 33 years. The population of participants who agreed to participate in the study and met the appropriate criteria (1858) was compared to the population of participants whose specimens were rejected from the study (73) and to the population of participants who refused to participate (79) (Table 1). There was no significant difference in age and race between these populations (P > 0.05).
The prevalence of CT was 5.9% (109/1858), and that of GC was 3.4% (63/1858). The African American population had a prevalence of 5.1% (77/1510) for CT and a prevalence of 2.3% (34/1510) for GC. In the white population the CT prevalence was 8.9% (29/323) and the GC prevalence was 8.6% (28/323). The prevalences among women who were nonwhite or non–African American were 12.0% (3/25) and 4.0% (1/25) for CT and GC, respectively. Detainees aged 20 to 24 years had the highest prevalence of CT and GC infections (CT: 18.8%; GC: 8.8%), followed by those aged <20 years (CT: 9.5%; GC: 6.4%), those aged 25 to 29 years (CT: 7.6%; GC: 3.1%), and those aged 30 to 34 years (CT: 4.2%; GC: 3.8%). Arrestees older than 34 years of age had the lowest prevalences (CT: 2.7%; GC: 1.8%). Racial differences in prevalence of CT and GC among female detainees were observed in age-specific prevalences of CT and GC infections. Whites had higher prevalences of GC infection than African Americans in every age stratum: <20 years, 20 to 24 years, 25 to 29 years, 30 to 34 years, and >34 years (data not shown). Whites also had higher prevalences of CT infection in the age strata of 20 to 24 years and 25 to 29 years (data not shown). Age was further categorized into three age groups (<25, 25 to 34, and >34 years) on the basis of age-specific rates and age distribution. The difference in prevalence by age group between white and African American participants was statistically significant (P < 0.05) (Table 2).
The coinfection rate was 1.1% (21/1858). Of the women positive for CT, 19.2% (21/109) were positive for GC, and 33.3% (21/63) of the GC-positive women were infected with CT. Of the women who were pregnant, 10.9% (7/64) were positive for CT and 7.8% (5/64) were positive for GC. Of the women who reported symptoms, 9.5% (2/21) were positive for CT and 4.8% (1/21) were positive for GC. The majority of infections were found in women who did not report symptoms. The CT prevalence in the asymptomatic group was 5.9% (107/1822), and the GC prevalence was 3.4% (62/1822).
There were a total of 151 women who tested positive for CT, GC, or both. These women were tested for CT and GC between 0 and 27 days of arrest, with a mean testing time of 11 days and a median testing time of 11 days.
Time to treatment for infected participants was a mean of 6 days, with a median of 2 days. Of the women who received treatment, 93.2% were treated at the Women's Detention Center, while the remaining women were treated through the Baltimore City Health Department (3.8%), a noncity department of correction (1.5%), or another provider (1.5%).
Of the CT-positive women, 81.5% (88/108) were treated at the Women's Detention Center; 13 (12.0%) could not be located for treatment and 7 (6.5%) were out of Baltimore City jurisdiction. Of the GC-positive women, 69.9% (44/63) received treatment, while 14 (22.2%) could not be located for treatment. Five (7.9%) women were outside city jurisdiction. There was no statistical difference between the percentages of African American women and white women who received treatment.
There were no statistical differences in prevalence of CT or GC infection among participants living in Baltimore City and those living outside of the city. However, enrollees who were living in the city were more likely to be African American than were participants living outside of the city, consistent with the Baltimore City population. In addition, participants living in the city were older than those living outside the city (P = 0.034).
The geographic clustering of these infected individuals living in Baltimore City is shown in Figures 1 and 2. Of the African American women who were CT-positive, 81.8% (63/77) could be mapped, and 72.4% (21/29) of white women who were CT-positive could be mapped. For CT, there was no statistical difference between the percentages of African Americans and whites who were mapped (P = 0.29). Of the GC-positive African American women, 94.1% (32/34) were mapped, and of the GC-positive white women, 85.7% (24/28) were mapped. For GC, there was also no statistical difference between the percentages of African Americans and whites who were mapped (P = 0.27).
The core area for infected white women was in the southern part of the city, including the adjacent zip codes 21223, 21230, and 21225 and neighboring zip codes (just across the harbor) 21224 and 21222. The African American core area was distributed in the central east (adjacent zip codes 21202, 21205, 21213, and 21218) and west parts of the city (neighboring zip codes 21215, 21216, and 21217). In addition, for each zip code area that contained both white and African American participants, the CT prevalence was higher among whites; the GC prevalence was higher among whites except for within one zip code. Pooled zip code–specific rates by race were calculated on the basis of identified core area. Whites in the white core area had the highest prevalences of CT and GC infections (CT: 10.7%; GC: 11.8%), while those in noncore areas had prevalences of 1.5% for CT and 3.0% for GC. African Americans in the African American core area had the highest prevalence of CT (4.8%), in comparison with those in other areas (3.8% and 2.5%). There was no difference in prevalence of GC infection for African Americans in relation to core area (2.1–3.0%).
Factors associated with CT or GC infection were identified by multivariate logistic analysis. Young age, a white racial background, and residing within the core area were associated with infection. The strength of the association between age, racial background, and core area decreased with increased age. Interaction between ethnic background and core area was observed in the multivariate logistic regression model. The effects of race and core area in each stratum (adjusted by age group) on CT and GC infection are presented in Table 3. White women with a CT or GC infection were more likely to be clustered within the core area (CT OR = 2.49; GC OR = 5.60), while African Americans with a CT or GC infection were more likely to live throughout the city (Table 3).
As part of a larger study protocol, the feasibility of implementing a routine screening program for CT and GC at the Women's Detention Center in Baltimore, Maryland, was examined. 13 Women were enrolled during intake between July 28, 1999, and June 27, 2000, and demographic data were collected for risk factor analysis. Residential zip code was collected to determine if there was an association among infection, race, and area of residence. Significant findings included high CT and GC prevalences, at 5.9% and 3.4%, respectively, as well as unusually high prevalences among the white population, at 8.9% for CT and 8.6% for GC, that were significant (P < 0.05) in comparison with CT and GC prevalences among African Americans. Geographic analysis did not completely explain this discrepancy, although geographic analysis did demonstrate clustering of infected individuals within core areas. This information could be useful in future interventions. If infected individuals are more likely to cluster within a core area, these areas could be targeted, thus enabling identification and treatment of more infected individuals while maximizing resources.
From February 2000 to June 2000, the volunteer rate was 88.8%. Most important, 82.4% of the women testing positive for CT received treatment, and 69.9% of the women testing positive for GC received treatment. Treatment occurred in a timely fashion, within a median of 6 days. Although the overall coinfection rate was low, 1.1%, the individual prevalences for coinfection were high, at 9.2% for CT and 33.3% for GC. The ability to detect and successfully treat a majority of these infected women suggests that similar screening programs could be effective in identifying and treating CT and GC infections, especially in high-prevalence areas.
Several risk factors for CT and GC infection were demonstrated. Multivariate analysis indicated that age, race, and residential location were significantly associated with CT and GC infection. It is interesting that women of white race were more likely to be infected than women of African American race. In many studies, prevalence of STIs in the African American population is higher than in the white population. 7,8,15–17 There have been previous studies of STIs and incarceration. 18–24 The majority of these studies have been of HIV and its effects, with a limited number on CT and GC infection. 18,21,22 The previous CT and GC studies of women in detention did not involve collection of the same demographic data as in this study, which makes comparisons difficult. However, a recent study of CT and GC prevalences in jails and juvenile detention centers showed a high prevalence of both CT and GC for the entering population, which is consistent with findings in our study. 17
One large study of CT infection in military women demonstrated a prevalence of 14.9% and 5.5% for African Americans and whites, respectively. 7 Among high school students, Marrazzo et al, 8 found an 8.7% CT prevalence among African Americans and 5.0% prevalence among whites in Seattle, Washington. Between 1981 and 1996 in the United States, reported cases of GC were 35 times higher for African Americans than for whites. 25
In our study, the opposite was true; the prevalence rates were higher for white women, at 8.9% for CT and 8.6% for GC, compared with 5.1% and 2.3% for African American women. We believe this finding to be the product of several factors.
First, when racial background, infectivity status, and residential zip code were compiled and mapped, as in Figures 1 and 2, some geographic differences between the white and African American populations became apparent. The majority of the CT- and GC-positive whites were clustered in the southern portion of Baltimore City, while the majority of the African American positives clustered in the east–central and western areas of the city. There may be some behaviors that are endemic to one geographic region and not to others that account for the discrepancy in infection by race. For example, certain regions of the city may contain a higher proportion of white commercial sex workers, accounting for both geographic clustering of CT and GC infections among whites as well as the discrepancy in infection by race. Other high-risk behaviors, such as drug use, could be localized in certain regions of the city and could also contribute to the discrepancy in infection by race. Therefore, inhabitants of a particular region or core area may be at a higher risk for being infected by CT- or GC-infected partners or infected core groups.
However, it is important to note that geographic analysis was performed on a small number of participants (African Americans = 95, whites = 45), over a zip code as opposed to a smaller geographic area. In addition, geographic analysis was performed on the basis of zip codes, which are large geographic areas, and no information about the background populations of city zip codes was obtained. In future studies, it would be beneficial to obtain more detailed geographic information from participants to determine the full usefulness of geographic analysis in detention studies. Geographic analysis did demonstrate clustering of infected individuals, and that information could be a useful tool in targeting intervention programs to specific areas. Ultimately, the geographic differences between African American and white participants do not account for the discrepancy between infection and race.
Second, CT and GC infections are associated with those younger than the age of 25 years. If the majority of the younger participants were white, this could account for the discrepancy between prevalences among African Americans and whites. 7,15,25–27 However, this is not the case. Age standardization of the two racial groups by the direct method revealed that whites had the higher CT and GC prevalence. 28
We hypothesize that the discrepancy may be explained by behavioral factors associated with arrest. The risk behavior responsible for the participant's arrest may be associated with the behavioral risk factors for CT and GC infection, such as prostitution. Other behaviors, such as drug use, could also account for this discrepancy. Unfortunately, the reason for participant arrest was not recorded during the study. In addition, arrest records from the majority of Baltimore City districts could not be obtained. However, arrest records of the Baltimore City Police Department Southern District (zip codes 21223, 21225, 21226, and 21230) were obtained and indicated that during the study period, of 143 women arrested for prostitution, 88% were white and 12% were African American (personal communication, B. Baker). The fact that these zip codes were associated with higher CT and GC prevalences among whites than African Americans may support but does not prove our hypothesis. As previously mentioned, other high-risk behaviors such as drug use could also account for the discrepancy in the data. In future studies, it would be useful to obtain arrest records and determine if there was a correlation between infection and reason for arrest.
Some limitations to this study exist. These include the number of samples that were not processed due to extended length of transit time, mismatched identifiers, and container leakage. In total, 3.8% of the women who volunteered could not be tested due to these issues. Of the women who were excluded due to these logistical issues, only 4.1% (3/73) resubmitted samples and were tested.
Of the women who were positive for CT, 20/108 (17.6%) could not be treated, and 19/63 (30.1%) of the GC-positive women could not be treated. Of the women released without treatment, only nine were later located and treated. Reasons for this are unknown, but those untreated women represent a potential reservoir for future transmission, contributing to the difficulty of limiting these readily curable STIs. In future programs, it would be beneficial to expend extra effort to ensure follow-up of this population of women. Testing time also had an impact on the number of women effectively treated.
The interval for testing of samples was a mean of 11 days. Testing samples in a timelier manner would have allowed for identification of more positive women while they were still in detention. Since a large percentage of the women were treated while in detention (93.2%), timelier testing could have resulted in more efficient treatment of the infected women.
In conclusion, this study demonstrated the feasibility of STI screening programs based on noninvasive specimen obtainment for women in detention centers in high-prevalence areas. Although geographic analysis of the data did not completely describe racial and geographic differences in prevalence, arrest records in addition to geographic information could become a useful tool in future studies exploring these differences. Geographic analysis could also be a useful tool in targeting future interventions to specific regions of high-prevalence areas. The implementation of screening programs for CT and GC in detention centers will be a valuable tool in reducing infections and controlling future transmission of STIs, particularly in areas where CT and GC are highly prevalent.
1. Wasserheit JN. Epidemiological synergy: interrelationships between human immunodeficiency virus infection and other sexually transmitted diseases. Sex Transm Dis 1992; 19: 61–77.
2. Institute of Medicine. The Hidden Epidemic: Confronting Sexually Transmitted Diseases. Washington, DC: National Academy Press, 1997.
3. Hammet TM, Harmon P, Maruschak LM. 1997 Update: HIV/AIDS, STDs, and TB in Correctional Facilities. Washington, DC: US Department of Justice, National Institute of Justice/Centers for Disease Control and Prevention, July 1999.
4. Parece MS, Herrera GA, Voigt RF, et al. STD testing policies and practices in U.S. city and county jails. Sex Transm Dis 1999; 26: 431–437.
5. Glaser JB. Sexually transmitted diseases in the incarcerated: an underexploited public health opportunity. Sex Transm Dis 1998; 25: 308–309.
6. Skolnick AA. Look behind bars for key to control of STDs. JAMA 1998; 279: 97–98.
7. Gaydos CA, Howell RM, Pare B, et al. Chlamydia trachomatis infections in female military recruits. N Engl J Med 1998; 339: 739–744.
8. Marrazzo JM, White CL, Krekeler B, et al. Community-based urine screening for Chlamydia trachomatis with a ligase chain reaction assay. Ann Intern Med 1997; 127: 796–803.
9. Smith KR, Ching S, Lee H, et al. Evaluation of ligase chain reaction for use with urine for identification of Neisseria gonorrhoeae in females attending a sexually transmitted disease clinic. J Clin Microbiol 1995; 33: 455–457.
10. Chernesky MA, Jang D, Lee H, et al. Diagnosis of Chlamydia trachomatis infections of men and women by testing first-void urine by ligase chain reaction. J Clin Microbiol 1994; 32: 2682–2685.
11. Gaydos, CA, Crotchfelt KA, Howell, MR, et al. Molecular amplification assays to detect Chlamydia trachomatis infections in urine specimens from high school students and to monitor the persistence of chlamydial DNA in urine after therapy. J Infect Dis 1998; 177: 417–424.
12. Howell, MR, Quinn TC, Gaydos CA. Screening for Chlamydia trachomatis in asymptomatic women attending family planning clinics: a cost-effectiveness analysis of three strategies. Ann Intern Med 1998; 128: 277–284.
13. Mertz KJ, Schwebke JR, Gaydos CA, et al. Screening women for chlamydial and gonococcal infection using urine tests: feasibility, acceptability, prevalence and treatment rates. Sex Transm Dis 2002; 29: 271–276.
14. Becker KM, Glass GE, Brathwaite W, et al. Geographic epidemiology of gonorrhea in Baltimore, Maryland, using a geographic information system. Am J Epidemiol 1998; 147: 709–716.
15. Klausner JD, 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.
16. Laumann EO, Youm Y. Racial/ethnic group differences in the prevalence of sexually transmitted diseases in the United States: a network explanation. Sex Transm Dis 1999; 26: 250–261.
17. Zenilman JM. Ethnicity and sexually transmitted infections. Curr Opin Infect Dis 1998; 11: 47–52.
18. Centers for Disease Control and Prevention. High prevalence of chlamydial and gonococcal infection in women entering jails and juvenile detention centers–Chicago, Birmingham, and San Francisco. MMWR Morb Mortal Wkly Rep 1999; 48: 793–796.
19. Centers for Disease Control and Prevention. Assessment of sexually transmitted diseases services in city and county jails–United States, 1997. Morb Mortal Wkly Rep 1998; 47: 429–431.
20. New approaches to syphilis control: finding opportunities for syphilis treatment and congenital syphilis prevention in a women's correctional setting. Sex Transm Dis 1997; 24: 218–226.
21. Holmes MD, Safyer SM, Bicknell NA, et al. Chlamydial cer-vical infection in jailed women. Am J Public Health 1993; 83: 551–555.
22. Bicknell NA, Vermund SH, Holmes MD, et al. Human papillomavirus, gonorrhea, syphilis, and cervical dysplasia in jailed women. Am J Public Health 1991; 81: 1318–1320.
23. Gaitch HD, Fleming P. Epidemiology of AIDS in incarcerated persons in the United States. AIDS 1999; 13: 2429–2435.
24. Rich JD, Dickinson BP, Macolino C, et al. Prevalence and incidence of HIV among incarcerated and reincarcerated women in Rhode Island. J AIDS 1999; 22: 161–166.
25. Fox KK, Whittington WL, Levine WC. Gonorrhea in the United States, 1981–1996: demographic and geographic trends. Sex Transm Dis 1998; 25: 386–393.
26. Hook EW 3rd, Reichart CA, Upchurch DM. Comparative behavioral epidemiology of gonococcal and chlamydial infections among patients attending a Baltimore, Maryland, sexually transmitted disease clinic. Am J Epidemiol 1992; 136: 662–672.
27. Cook RL, St. George K, Lassak M, et al. Screening for Chlamydia trachomatis infections in college women with a polymerase chain reaction. Clin Infect Dis 1999; 28: 1002–1007.
28. Kahn HA, Sempos CT, Statistical Methods in Epidemiology. Oxford: Oxford University Press, 1989: 87–95.