Exploration of gender differences of Chlamydia trachomatis infection amongst young people reveals limitations of using sexual histories to assess risk in high-prevalence areas

Abstract

Summary

In England, chlamydia positivity in young men occurs at a slightly older age group (20–24 years) than positivity among young women (16–20 years) but total rates of infection among the population aged under 25 years are similar. Where there is variation, explanations often focus on individual sexual risk behaviours. We aimed to explore the extent to which variations in chlamydia positivity could be explained by reasons for attendance and sexual behaviour at a sexual health clinic in a high-prevalence area of England. Data routinely collected during clinic appointments were extracted retrospectively from the medical records of 952 clinic users (634 women) aged 16 to 24. We tested for associations with chlamydia positivity using the Chi square statistic and multiple logistic regression for men and women separately; 19.5% of men tested positive (95% CI: 15.1–23.8) compared to 11.5% of women (95% CI: 9.0–14.0). Reporting a partner with symptoms or a sexually transmitted infection diagnosis was significantly associated with a positive diagnosis (Men OR: 3.14, 95% CI: 1.5–6.25; Women OR: 3.78, 95% CI: 1.83–7.83). All other reasons for attendance and all sexual behaviour variables were not significantly associated with a positive diagnosis. Differences in chlamydia positivity between men and women attending this service cannot be explained by individual sexual behaviours found to be associated with higher risk of infection in national studies. Our findings question the utility of individual behavioural data routinely collected during clinic appointments for predicting risk of sexually transmitted infections in high-prevalence areas.

Keywords

sexually transmitted infections young people gender differences sexual histories high prevalence cross-sectional study

Background

Genital Chlamydia trachomatis infection is the most commonly diagnosed sexually transmitted infection (STI) in the UK.¹ Infection is usually asymptomatic and is associated with increased risk of pelvic inflammatory disease, ectopic pregnancy and tubal factor sub-fertility.² Positivity is highest in specialist sexual health services with 11% of 16-to-24-year-olds testing positive within genito-urinary medicine clinics in England in 2011/12.³ There is no evidence to support a gender difference in infection rates although positivity peaks among women at a slightly younger age (15−19) than men (20−24). We were therefore surprised when a multi-site study of chlamydia testing and positivity identified significantly higher results in men across the 15−24 age group at a sexual health clinic in South London,⁴ where prevalence of chlamydia positivity is among the highest in the country.⁵ This was not replicated in any of the other sites, suggesting something unusual about this population.

We considered two possible causes for the observed gender differences based on the current literature on risk factors for chlamydial infection. Firstly, differences between men and women’s reasons for attending the clinic could explain higher chlamydia positivity for men. The most common reason for attendance at a sexual health service is having symptoms, or a partner with symptoms/who has been diagnosed with an STI, which is associated with a higher likelihood of STI diagnosis.^6,7 As an integrated sexual health clinic, Camberwell Sexual Health Centre also provides a contraceptive service to women with no symptoms or known STI contacts. This means that some female clients attend for contraception only and are offered opportunistic chlamydia screening. This group of women may be at lower risk for chlamydia than women and men attending for STI screening.

The second possible cause for the gender differences in chlamydia prevalence is that the sexual behaviour of the men using this clinic is different from that of the women. In other populations, an increased risk of a chlamydia diagnosis in men and women has been associated with two or more partners in the year prior to testing and there is strong evidence that condom use is associated with reduced risk of a diagnosis.^8,9 It is not known whether men using our clinic have a higher prevalence of high-risk behaviour or whether the association between sexual behaviours and chlamydia observed in other populations is also observed in this population.

We conducted a cross-sectional study to examine differences in chlamydia positivity using routinely collected data on sexual behaviour and reasons for attendance from clinical notes of chlamydia-positive and -negative clients who participated in the larger multisite study.⁴ Our aim was to examine the extent to which differences in chlamydial infection could be explained by reported sexual behaviour and reasons for attendance among 16−24-year-olds.

Methods

The study took place in Camberwell Sexual Health Centre, King’s College Hospital NHS Foundation Trust, South-East London. The clinic is located in an area with high levels of socioeconomic deprivation and an ethnically diverse population where just under 50% of the population identify as white British.¹⁰ All clients attending from November 2010 to April 2011, between 16 and 25, were eligible to take part in the study.

The study used a cross-sectional design examining factors associated with chlamydia positivity. When individuals aged 16–24 registered at the clinic, an alert appeared for receptionists to attach an information sheet to eligible clients’ notes and clinicians invited and consented patients into the study. Clients were asked a series of well-established questions relating to socio-demographic characteristics, using pre-defined response categories, and chlamydia tests results were extracted from clinic systems entered into a study database.

During routine consultations, clinicians record clients’ reasons for attendance, symptoms and sexual behaviour as free text on clients’ medical notes. A pro-forma was later used to extract this information from medical notes. A standardised coding framework was developed to standardise across the free text entries reasons for attendance, symptoms and sexual behaviour.

Symptomatic clients were defined as those reporting any genito-urinary symptom and not only those associated with chlamydial infection. Women attending for ‘Contraceptive Service’ were defined as those who attended for contraception, emergency contraception or a pregnancy test and did not report having symptoms, a partner with symptoms or a diagnosis of an STI. Concurrent relationships were not consistently recorded within sexual histories and were deemed to have occurred where medical notes indicated one or more partners within the same time period of a regular relationship. Ethnicity, recorded using census categories, were combined for statistical purposes in four groups – White (British, Irish, Other), Black (mixed, African, Caribbean, British, other), Other and Missing.

Data analysis

Data on demographic characteristics and chlamydia test results from the original study were merged with data extracted from client notes using a non-identifiable number in place of clinic client ID. Data analysis was performed using SPSS 20.0 (IBM Corporation, Somers, NY, USA). χ² tests were used to compare proportions and multivariate logistic regression to examine factors associated with chlamydia positivity. Multivariate models were built using forced entry method including all confounders significant at the 5% level in univariate analysis.

Results

Sample characteristics

In all, 965 16–24-year-old clients taking a chlamydia test during the study period responded. Of this sample, 955 clinic records were available. Three clients (two women, one man) were excluded from the analysis as notes stated no prior sexual experience. Of the remaining 952 clients, 634 (66.6%) were women. Chlamydia positivity among men was significantly higher, 19.5% (95% confidence intervals [CI] 15.1%, 23.8%) vs 11.5% in women (95% CI 9.0%, 14.0%), which is representative of the clinic population of the same age (15.2% in men vs 9.1% in women) who attended in 2011. The median age for men was 22 (inter-quartile range [IQR]: 19–23) with 27% under 20 and 21 (IQR 19–23) for women with 32.8% under 20.

Chlamydia positivity between gender, age and ethnic groups

Younger women had a significantly higher proportion of positive chlamydia results; 17.3% in the <20 compared to 8.69% in the >20 age groups (p = 0.001). No significant association was found for differences in chlamydia positivity between younger and older men. Black ethnicity was significantly associated with higher positivity in men compared to the other ethnic groups (p = 0.001). A moderate to weak association was seen between ethnic groups for women (p = 0.056) and was included into the multivariate model. Table 1 displays chlamydia positivity for all independent variables analysed.

Table 1.

Tests of association between independent variables and Chlamydia trachomatis among clients attending Camberwell Sexual Health Centre.

	Men			Women
	n	CT+ (%)	p Value	n	CT+ (%)	p Value
All	318	62 (19.5)		634	73 (11.5)
Age group (years)			0.095			0.001^a
16–19	86	22 (25.6)		208	36 (17.3)
20–24	232	40 (17.2)		426	37 (8.7)
Ethnic groups			0.001^a			0.056
White	81	5 (6.2)		209	14 (6.7)
Black	179	49 (27.4)		273	40 (14.7)
Missing	19	3 (15.8)		46	6 (13.0)
Other	39	5 (12.8)		106	13 (12.3)
Reason for attendance ^b
Symptom reported			0.588			0.274
Yes	179	33 (18.4)		292	38 (13.0)
No	139	29 (20.9)		342	35 (10.2)
Partner with symptoms / diagnosis.			0.000^a			0.000^a
Yes	52	21 (40.4)		41	13 (31.7)
No	266	41 (15.4)		593	60(10.1)
Contraceptive service						0.124
Yes				188	16 (8.5)
No				446	57 (12.8)
Behavioural characteristics
Same sex male partner in last 3 months
Yes	18	0 (0.0)	NA
No	271	56 (20.7)
Not recorded	29	6 (20.7)
LSI with regular partner			0.257			0.871
No	121	18 (14.9)		100	11 (11.0)
Yes	168	38 (22.6)		491	58 (11.8)
Not recorded	29	6 (20.7)		43	4 (9.3)
# of partners in last 3 months			0.649			0.827
<2	145	25 (17.2)		488	55 (11.3)
≥2	136	29 (21.3)		98	13 (13.3)
Not recorded	37	8 (21.6)		48	5 (10.4)
Used condom at last sexual intercourse			0.733			0.132
No	210	43 (20.5)		470	61 (13.0)
Yes	79	13 (16.5)		121	8 (6.6)
Not recorded	29	6 (20.7)		43	4 (9.3)
Evidence of concurrency recorded			0.044^a			0.110
No	235	39 (16.6)		555	61 (11.0)
Yes	54	17 (31.5)		36	8 (22.2)
Not recorded	29	6 (20.7)		43	4 (9.3)

CT+: Chlamydia trachomatis positive; LSI: last sexual intercourse.

Significant at p < 0.05.

Clients may have multiple reasons for attendance.

Reasons for attendance

The presence of a symptom as a reason for attendance and therefore the trigger for use of the service did not show any association with chlamydia positivity (men, p = 0.588; women, p = 0.274). There was a significant association between men and women reporting a partner with symptoms or an STI diagnosis (p < 0.000). These clients accounted for 34% of all male and 17% of female clients testing positive for chlamydia. Just under a third of women attended for a contraceptive service. There was no significant difference in chlamydia positivity between this group and women who had attended for an STI-related reason (p = 0.124).

Sexual behaviour

Men who have sex with men have higher rates of diagnosis of STI in this clinic but in our sample only 18 men reported a same sex partner in the last three months and none of these 18 clients tested positive for chlamydia. The lack of any positive cases amongst this high-risk group is likely due to chance. Unexpectedly we found no difference in chlamydia positivity among those reporting last sex with a casual partner among either men (p = 0.257) or women (p = 0.871). Similarly, we found no association with condom use at last sexual intercourse (men p = 0.73 and women p = 0.132) or for number of partners in the last three months (men p = 0.649 and women p = 0.827). Fifty-four men (18.7%) reported one or more concurrent partnership within the last three months and this was associated with chlamydia infection with almost one-third (31%) testing positive for chlamydia, p = 0.044. In contrast, there was no association between chlamydia and concurrent partners among women (p = 0.11).

Factors associated with Chlamydia positivity

Table 2 shows crude and adjusted odds ratios for all variables significantly associated with chlamydia positivity. In the multivariate analyses, older age remained significantly associated with chlamydia infection among women (adjusted odds ratio [AOR] 0.54, 95% CI: 0.32–0.9). Black ethnicity remained independently associated with chlamydia positivity for men (AOR 4.86, 95% CI: 1.82–12.94) and for women (AOR 2.05, 95% CI: 1.06–3.97). Clients reporting partners with STI symptoms or diagnosis had significantly higher odds of chlamydia positivity (men AOR 3.14 95% CI: 1.57–6.25; women AOR 3.78 95% CI: 1.83–7.83). After adjusting for other variables, concurrency was not associated with an increased risk in men.

Table 2.

Crude and adjusted odds ratios for demographic, attendance reasons and behavioural risk factors among clients attending Camberwell Sexual Health Centre.

	OR	(95% CI)	AOR^a	(95% CI)
Men
Combined ethnic groups
White	1		1
Black	5.73	(2.19−15.03)	4.86^b	(1.82−12.94)
Missing	2.85	(0.62−13.16)	3.24	(0.69−15.32)
Other	2.24	(0.61−8.23)	2.4	(0.64−8.98)
Partner with symptom/diagnosis
No	1		1
Yes	3.72	(1.95−7.09)	3.14^b	(1.57−6.25)
Concurrency
No	1		1
Yes	2.31	(1.18−4.51)	1.74	(0.86−3.54)
Not recorded	1.31	(0.50−3.43)	1.43	(0.52−3.91)
Women
Age
<20	1		1
≥20	0.45	(0.28−0.74)	0.54^b	(0.32−0.9)
Combined ethnic groups
White	1		1
Black	2.39	(1.26−4.52)	2.05^b	(1.06−3.97)
Missing	2.09	(0.76−5.77)	1.94	(0.69−5.48)
Other	1.95	(0.88−4.31)	1.75	(0.78−3.93)
Partner with symptom/diagnosis
No	1		1
Yes	4.12	(2.03−8.39)	3.78^b	(1.83−7.83)

CI: confidence interval; OR: odds ratios; AOR: adjusted odds ratios.

Adjusted for all other variables in model. Criteria for inclusion in the multivariate model is p <0.05 (see methods).

Significant at p < 0.05.

Discussion

Main findings

Differences in chlamydia positivity between young men and women attending this service cannot be explained by most of the routinely collected sexual behavioural risk factors associated with higher risk of infection in national studies. However, clients reporting a partner with symptoms or a diagnosis of an STI were significantly more likely to test positive. Our findings question the use of the current sexual risk behaviour questions at identifying young people with chlamydia in this setting.

Comparisons with other studies/possible explanation for our findings

Our sample size was large enough to find statistical associations between other variables and chlamydia consistent with national cross-sectional studies such as higher chlamydia positivity in male Black ethnic groups and women between 16 and 20.^8,9 However, these larger studies found a statistically significant association between condom use and partner numbers and chlamydia whereas we did not. There is evidence that high prevalence of STIs within a community is associated with higher STI risk independent of demographic and behavioural risk^11–13 possibly because the probability of having an infected partner is higher, even among those with low partner numbers. In addition, the magnitude of association between these sexual behaviours and chlamydia was much higher in The National Survey of Sexual Attitudes and Lifestyles 2, a general population, than in the National Chlamydia Screening Programme, which is characterised as a higher risk population. This may suggest that the excess risk of chlamydia associated with individual sexual behaviours in higher risk populations is not as great as in lower risk populations.

The introduction of the diagnostic rate within the National Chlamydia Screening Programme (positive tests per 100,000 population under 25 years) and its inclusion within the English Public Health Outcomes Framework has focused attention on the ability to predict chlamydia positivity among young people and to target testing at high-risk groups.¹⁴ Our data suggest that in high-risk populations, using broad categories of risk is not a predictor of chlamydia positivity.

Methodological considerations

The detailed process of extracting data from hand-written clinical notes resulted in recorded sexual history information collected for 99% of clients within the sample, thus limiting selection bias in our study. Many studies report associations between sexual behaviour and chlamydia positivity using self-completed questionnaires. Whilst there may be advantages for the validity of data collection, it is problematic to extrapolate from these studies into clinic settings, where behavioural data are most commonly collected by clinic staff as part of a consultation.¹⁵ In contrast, our study uses clinic notes as the source of sexual behaviour data, and therefore it directly illustrates limitations in the value of sexual history taking for risk-assessment within these settings.

Clearly this study only reports on the epidemiology of infection at one sexual health centre so its findings may well not be generalisable to other areas. The clinic serves a socioeconomically and ethnically diverse urban population in London that is often underrepresented in population-based surveys. This population is characterised by a high prevalence of infection and other markers of poor sexual health (e.g. unplanned pregnancies). Therefore, it may well provide insights into different risk factors from STIs facing a high-risk population that are not routinely captured by population-based surveys or that are masked in national or even regional analyses. There are also likely to be distinct differences between ethnic groups that determine cultural norms around sex and in turn influence individual behaviours and sexual networks in ways that could affect the risk of Chlamydia.^16,17 In this analysis, we used aggregated ethnicity categories because of the small sample sizes within each group. As a result, this study is unable to explore these in any depth.

Clients’ notes contained incomplete information for sexual behaviour in around 12% of men and 10% of women. The chlamydia positivity profile of these groups with missing data was similar to the men with higher risk behaviour and women with lower risk behaviour. The triage system used in this clinic may contribute to missing behaviour data. If clients report no risk factors or symptoms, partner contact or pregnancy when they present, they may see a client support worker for a urine sample or self-taken swab sample (and hence be seen quicker). Other studies have found that men may adopt a strategy of non-disclosure when they attend clinics due to reluctance to share information, have an examination or a belief they will be seen quicker.^18,19 If men visiting our clinic adopted this strategy, then it is possible that those with missing data are more similar in their behaviours to higher risk men.

Conclusion

Current strategies for risk assessment in a sexual health service in South East London did not accurately predict risk of chlamydia infection or explain significant differences in chlamydia positivity between male and female clients. Our findings provide evidence of the limitations of using individual sexual behaviour as a determinant of risk for chlamydia within a high-prevalence population.

Footnotes

Acknowledgments

We thank Rachel Scott & Emma Thorman for their work in supporting this study. We also thank clinical staff at CHSC for their work in recruiting clients.

Conflict of interest

The authors declare no conflict of interest.

Funding

This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

Ethics

Ethical approval was obtained from the Outer West London Research Ethics Committee ref: 10/H0709/054.

References

HPA. Testing times: HIV and other sexually transmitted infections in the United Kingdom, London: Health Protection Agency, 2007.

Baraitser

Alexander

Sheringham

. Chlamydia trachomatis screening in young women. Curr Opin Obstet Gynecol 2011; 23: 315–320.

NSCP England. Chlamydia testing Data 2011/12, http://wwwchlamydiascreeningnhsuk/ps/resources/data-tables/CTD-Q1–4-2011_2012pdf (2012, accessed Jan 2013).

Sheringham JR. Investigating socio-economic variations in access to chlamydia testing in young people in England. Doctoral Thesis, UCL (University College London), wwwuclacuk (2012, accessed Nov 2012).

NSCP England. Chlamydia testing Data 2011/12, http://wwwchlamydiascreeningnhsuk/ps/resources/data-tables/CTD-Q1–4-2011_2012pdf (2012, accessed 20 Jan 2013).

Dave

French

Jungmann

. The need for innovative sexually transmitted infection screening initiatives for young men: evidence from genitourinary medicine clinics across England. Int J STD AIDS 2011; 22: 600–603.

Mercer

Aicken

Estcourt

. Building the bypass–implications of improved access to sexual healthcare: evidence from surveys of patients attending contrasting genitourinary medicine clinics across England in 2004/2005 and 2009. Sex Transm Infect 2012; 88: 9–15.

Riha

Mercer

Soldan

. Who is being tested by the English National Chlamydia Screening Programme? A comparison with national probability survey data. Sex Transm Infect 2011; 87: 306–311.

Simms

Talebi

Rhia

. The English National Chlamydia Screening Programme: variations in positivity in 2007/2008. Sex Transm Dis 2009; 36: 522–527.

10.

ONS. 2007–2011 London Data, http://wwwonsgovuk/ons/about-ons/what-we-do/FOI/foi-requests/population/london-population-figures-by-ethnic-group/london-population-figures-by-ethnic-groupxls (2012, accessed 20 January 2013).

11.

Bunnell

Dahlberg

Rolfs

. High prevalence and incidence of sexually transmitted diseases in urban adolescent females despite moderate risk behaviors. J Infect Dis 1999; 180: 1624–1631.

12.

Jennings

Taylor

Iannacchione

. The available pool of sex partners and risk for a current bacterial sexually transmitted infection. Ann Epidemiol 2010; 20: 532–538.

13.

Sucato

Celum

Dithmer

. Demographic rather than behavioral risk factors predict herpes simplex virus type 2 infection in sexually active adolescents. Pediatr Infect Dis J 2001; 20: 422–426.

14.

DoH. Healthy lives, healthy people: Improving outcomes and supporting transparency, http://wwwdhgovuk/en/Publicationsandstatistics/Publications/PublicationsPolicyAndGuidance/DH_132358 (2012, accessed 23 June 2012).

15.

Copas

Wellings

Erens

. The accuracy of reported sensitive sexual behaviour in Britain: exploring the extent of change 1990–2000. Sex Transm Infect 2002; 78: 26–30.

16.

Low

Sterne

Barlow

. Inequalities in rates of gonorrhoea and chlamydia between black ethnic groups in south east London: cross sectional study. Sex Transm Infect 2001; 77: 15–20.

17.

Gerressu

Elam

Shain

. Sexually transmitted infection risk exposure among black and minority ethnic youth in northwest London: findings from a study translating a sexually transmitted infection risk-reduction intervention to the UK setting. Sex Transm Infect 2009; 85: 283–289.

18.

Heseltine

Foley

Alborough

. Patient views on examinations as part of asymptomatic screening in sexual health clinics. Int J STD AIDS 2012; 23: 305–307.

19.

David

. Are they really asymptomatic? Int J STD AIDS 2008; 19: 575–576.