Predictors identifying those at increased risk for STDs: a theory-guided review of empirical literature and clinical guidelines

Abstract

Summary

Sexually transmitted diseases (STDs) are leading causes of substantial morbidity worldwide. Identification of risk factors for estimating STD risk provides opportunities for optimising service delivery in clinical settings, including improving case finding accuracy and increasing cost-efficiency by limiting the testing of low-risk individuals. The current study was undertaken to synthesise the evidence supporting commonly cited chlamydia and gonorrhoea risk factors. The level of empirical support for the following predictors was strong/moderate: age, race/ethnicity, multiple lifetime sexual partners, sex with symptomatic partners and concurrent STD diagnosis. The following predictors had weak evidence: socio-economic status, transactional sex, drug/alcohol use, condom use and history of STD diagnosis. The most frequently listed predictors among nine clinical guidelines were younger age and multiple sexual partners; the least consistently listed predictor was inconsistent condom use. We found reasonably good concordance between risk factors consistently listed in the recommendations and predictors found to have strong empirical support in the literature. There is a need to continue building the evidence base to explicate the mechanisms and pathways of STD acquisition. We recommend periodic reviews of the level of support of predictors included in clinical guidelines to ensure that they are in accordance with empirical evidence.

Keywords

Sexually transmitted diseases sexually transmitted infections STD STI epidemiology predictors risk factors risk assessment chlamydia gonorrhoea selective testing screening recommendations clinical guidelines

Introduction

Sexually transmitted diseases (STDs) such as chlamydia and gonorrhoea are leading causes of substantial morbidity worldwide; left untreated, they are often important exposures in the causal pathway for pelvic inflammatory disease, ectopic pregnancies and infertility. In many countries, publicly-funded sexual and reproductive health (SRH) clinics are tasked with providing free or low-cost STD testing and treatment services. However, the recent economic downturn, budgetary shortfalls and cutbacks to public health infrastructure have resulted in SRH clinics operating at maximum capacity with limited resources.^1,2 Concomitantly, there has been a shift towards the development of alternative models of service delivery in SRH clinics focused on patient-centred care and cost-effectiveness (e.g. Internet-based STD testing; self-collected specimens),^3,4 including providing ‘express visit’ options to asymptomatic patients in SRH clinics who are deemed to be at low STD risk, during which these patients do not undergo physical examination but are offered testing only.^5,6 Because these alternative service delivery options involve patients having limited interactions with clinicians, another mechanism for recommending tests and triaging patients is required. There is, however, limited guidance in the literature about how to selectively test^a

We use the phrase “selective testing” instead of “selective screening” because most studies in this review included individuals with STD signs and symptoms. By definition, only asymptomatic individuals are considered for screening purposes.

those at highest risk for STDs in the absence of clinician interaction.

Identification of risk factors predictive of STD risk provides the opportunity to deliver four key benefits in clinical settings: identifying specific patient groups with higher disease prevalence, identifying areas to target prevention services, improving case-finding accuracy, and increasing cost-efficiency by limiting the testing of low-risk individuals. Plausible candidates for selective testing criteria include known determinants of prevalent infection (e.g. age, multiple partners).⁷ However, predictors used for selective testing are often derived from clinical guidelines that are broadly perceived as evidence-based statements; because they are developed by experts, they are assumed to have the same level of certainty and authority as conclusions generated by scientific methods.^7,8 Few studies have been conducted to determine the degree to which evidence in the published literature supports the predictors included in recommendations. A better understanding of the use (or lack) of evidence in this area has the potential to inform the ways in which STD test recommendation practices unfold in the future.

The current study was undertaken to synthesise the evidence-supporting predictors of chlamydia and gonorrhoea (the two most common reportable STDs) in high-income jurisdictions in North America, Western Europe and Australia – regions with comparable STD prevalence and social determinants of sexual health. The objectives are to identify risk factors cited as being most predictive of chlamydia and gonorrhoea in clinical guidelines and in empirical studies. We also aimed to assess the strength of evidence supporting the association between these predictors and chlamydia and gonorrhoea outcomes in empirical studies. Finally, we highlight discrepancies between predictors used in STD clinical guidelines and the relevant empirical evidence.

Methods

Literature search

We identified the most recent and commonly cited clinical guidelines for screening chlamydia and gonorrhoea issued by health organisations in high-income countries in North America and Europe, and Australia. We also identified empirical studies in journals readily accessible to SRH clinicians or experts in these settings and thus limited our search to electronic searches of English-language studies published in OVID Medline, as well as manual searches of reference lists of papers identified. MeSH search terms included ‘screening’, and at least one of ‘chlamydia’, ‘gonorrhoea’, ‘STI’ or ‘STD’. Articles that measured demographical and behavioural predictors associated with STD diagnoses were included. Studies focusing on psychosocial factors (e.g., knowledge and attitudes, stress, self-esteem) were excluded, as these predictors are not consistently measured across studies. STD signs and symptoms were not examined as they are clearly indicative of infection. The search was limited to studies published 2003–2011, an era characterised by the expansion of highly-sensitive tests (e.g. Nucleic Acid Amplification Tests) and urine-based testing, in an effort to ensure comparability in specimen collection and STD outcome assessment.

Assessment of methodological quality

Articles selected for inclusion were examined to assess their methodological quality. The following nine criteria^9,10 were used to score the methodological quality of studies: study design, specimen collection, statistical methods, model selection, study sample description, variable definition, events per variable, outcome assessment and study generalisability (see Table 1). Based on the scoring system, studies with a rating of ‘unacceptable’ (≤3 points) were removed from further review. Studies rated as ‘marginal’ (4–5 points), ‘acceptable’ (6–7 points) and ‘commendable’ (8–9 points) were kept for empirical analysis stage of the review. The studies rated as ‘commendable’ or ‘acceptable’ were given greater weight when interpreting consistency or variable relationships.

Table 1.

A summary of the criteria to be used to evaluate and rate the quality of methodology for each study.^a

Element	Quality of methodology criteria
A. Study design	The study design was clearly evident (e.g. cross-sectional, cohort or case-control)
B. Specimen collection	The specimen collection methodology was clearly defined. If the same specimen collection methodology was not used consistently among study participants, methods to adjust for this difference were included in the analyses.
C. Statistical methods	Multivariate analyses were used to examine the relationship between predictors and STD outcome. Methods used to examine subgroups and interactions were stated
D. Model selection	A clear attempt at model selection was made. Step-wise procedures or an a priori specification of p-value cutoff points for variable inclusion were considered acceptable for this criterion
E. Study sample description	The characteristics of the sample and the selection criteria for the study were clearly stated such as: location, time period, sampling method, sample size, entry criteria and exclusions. The study also included a description of non-respondents. The extent and nature of missing data were assessed to estimate potential bias.
F. Variable definition	Sources of data and details of methods of assessment were given for each variable.
G. Events per variable	The final sample size that was actually used in the analysis comprised of ≥10 study outcomes per variable.
H. Outcome assessment	STD outcomes were measured by the most sensitive and specific tests available at the time of data collection.
I. Study generalisability	STD screening was evaluated in an appropriate spectrum of subjects and sampling was based upon a multicentre study or involved a single study with a relatively large sample size of consecutive visits to a clinic.

Each quality criterion was awarded one point whenever present. Overall ratings were calculated by totaling the points and designated as commendable (8–9 points), acceptable (6–7 points), marginal (4–5 points) and unacceptable (≤3 points).

Data extraction

To help clarify the relative strength or importance of each STD predictor, we categorised the predictors extracted from the empirical studies into two groups based on the proximate-determinants framework:^11,12 (1) proximate determinants, which are demographic, social or economic variables distally related to STDs, and (2) proximate determinants, which are directly associated with an individual's probability of exposure to STDs and the efficiency of STD transmission. Also, to determine whether the various factors identified in the included studies were significantly associated with STDs in multivariable analysis (if possible), a significance level of 0.05 was used. Because the studies included in this review were heterogeneous with respect to study design and predictor definition, pooling of data for meta-analyses was not possible. Therefore, a qualitative data-synthesis was performed.^13,14

Key predictors mentioned in multiple studies were retained to test the level of empirical support for their association with chlamydia or gonorrhoea infection (STDs). The strengths of the evidence for predictors were assessed as follows:¹³ (1) strong evidence: >75% of multiple ‘acceptable’ or ‘commendable’ studies found consistent and statistically significant greater risk of infection; (2) moderate evidence: 60–74% of multiple ‘acceptable’ or ‘commendable’ studies found consistent and statistically significant greater risk of infection; (3) inconclusive evidence: 45–59% of multiple ‘acceptable’ or ‘commendable’ studies found consistent and statistically significant greater risk of infection; (4) weak evidence: <45% of ‘acceptable’ or ‘commendable’ studies found consistent and statistically significant greater risk of infection. In addition, predictors were also deemed to have insufficient evidence if few studies (i.e. less than the lowest quartile of the number of studies reviewed) examined the association between the predictor and infection.

Results

Overview of clinical guidelines

We identified nine organisations issuing clinical guidelines in high-income contexts (Table 2). Six organisations listed clinical guidelines for women only, citing lack of evidence for the selective screening of men: the US Preventive Services Task Force (USPSTF),^15,16 the American Academy of Family Practice (AAFP),^15,16 the US Centers for Disease Control and Prevention (USCDC),¹⁷ the American College of Obstetrics and Gynecology (ACOG),¹⁸ the Public Health Agency of Canada (PHAC)¹⁹ and the Society of Obstetricians and Gynecologists (SOGC)²⁰ (Table 2). Three other organisations reported clinical guidelines for men and women combined.^21–23

Table 2.

Comparison of risk factors listed in chlamydia and gonorrhoea screening recommendations.^a

	Risk factors
Organisation^b	Younger age	Sexual contact of STD case	Previous STD	New sexual partners	Multiple sexual partners	Drug or alcohol use	Sex in exchange for money or gifts	Inconsistent condom use	Individuals or travellers from endemic regions	Street involvement	Sex with anyone with listed risk factors
US Preventive Services Task Force (USPSTF)	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
American Academy of Family Physicians (AAFP)	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
US Centers for Disease Control and Prevention (USCDC)	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
American Congress of Obstetricians and Gynecologists (ACOG)^c	Yes	Yes	Yes		Yes	Yes	Yes		Yes		Yes
Society of Obstetricians and Gynecologists of Canada (SOGC)	Yes	Yes	Yes	Yes	Yes	Yes	Yes		Yes
Public Health Agency of Canada (PHAC)	Yes	Yes	Yes	Yes	Yes	Yes	Yes		Yes	Yes	Yes
European STD Guidelines (IUSTI Europe)	Yes	Yes		Yes	Yes			Yes
British Association for Sexual Health and HIV (BASHH)^c	Yes	Yes	Yes		Yes	Yes			Yes
Australasian Society for HIV Medicine (ASHM)^c	Yes			Yes	Yes	Yes			Yes

The following are American organisations: USPSTF, AAFP, USCDC, ACOG; the following are Canadian organisations: SOGC, PHAC; the following are European organisations: IUSTI, BASHH; and ASHM is an Australian organisation.

The following organisations listed risk factors for women only: USPSTF, AAFP, USCDC, ACOG, PHAC, and SOGC. The IUSTI Europe ASHM and BASHH reported screening recommendations for males and females combined.

Risk factors listed are for STIs in general, not just for chlamydia and gonorrhoea.

Almost all the organisations support the universal screening of young sexually-active individuals for chlamydia and to some extent gonorrhoea (Table 2). Occasionally, clinical guidelines differ across organisations, mainly because of distinctions in mission and target audience²⁴ as well as the epidemiology of the various STDs. For example, the USCDC and PHAC focus on population health in the US and Canada, respectively, while other organisations focus on the clinical settings.²⁴ A closer examination of predictors revealed incongruities between the organisations, even among those in the same country (Table 2). For example, clinical recommendations issued by both the PHAC and the SOGC agree on screening individuals with the following risk factors for chlamydia or gonorrhoea infection: younger age, sexual contact of STD case, previous STD diagnosis, new sexual partners, multiple sexual partners, use of drugs or alcohol during sex, sex in exchange for money or gifts, and individuals or travellers from endemic regions (Table 2). However, the PHAC also identifies the following risk factors: street involvement (or homelessness) and sexual contacts of people with any of the aforementioned risk factors (Table 2).

The most consistently listed predictors among clinical recommendations were younger age and multiple sexual partners as all nine organisations recommended screening individuals with these two risk factors (Table 2). Approximately, 90% of the nine organisations recommended screening individuals who are sexual contacts of STD cases and also those who use alcohol or drugs during sexual relations. The least consistently listed predictors were street involvement (11%), inconsistent condom use (44%) and individuals or travellers from STD endemic regions (56%).

Assessment of included studies

Based on the article-selection process listed in Figure 1, 75 studies were rated, most of which included a mix of symptomatic and asymptomatic individuals (Table 3). Approximately, 68% of studies examined chlamydia infection only, 8% examined gonorrhoea outcomes only and 19% examined both outcomes (Table 3). Approximately, 32% of the studies had women-only populations and 20% included men-only populations. Nearly 50% were from the USA (Table 3). Some studies used population-based participant recruitment (21%), and most studies were cross-sectional in design (87%). A minority of the studies examined only asymptomatic populations (8%). Based on the nine criteria of methodological quality, a total of five studies (7%) received a score of ≤3/9 and thus were rated as ‘unacceptable’.^25–29 These studies were omitted from the second stage of review. Studies rated as ‘marginal’ (24%),^7,30–46 ‘acceptable’ (41%)^47–77 and ‘commendable’ (28%)^4,78–97 were retained in the review. Table 4 summarises the 70 studies that were retained for further assessment.

Figure 1.

The process of selecting articles for the inclusion in the literature review of predictors of chlamydia and gonorrhoea infection.

Table 3.

Percentage distribution of studies included in the literature review, by level of empirical support, according to selected characteristics.

Characteristic	N	Unacceptable (%)	Marginal (%)	Acceptable (%)	Commendable (%)
STD
Chlamydia only	51	9.8	25.5	39.2	25.5
Gonorrhoea only	6	0.0	33.3	66.7	0.0
Chlamydia and GC	14	0.0	14.3	35.7	50.0
Other^a	4	0.0	25.0	50.0	25.0
Gender
Females only	24	4.2	25.0	29.2	41.7
Males only	15	20.0	26.7	33.3	20.0
Both (with stratification^b)	20	0.0	15.0	55.0	30.0
Both (without stratification)	16	6.3	31.3	50.0	12.5
Country
Australia	7	14.3	28.6	28.6	28.6
Italy	3	0.0	33.3	0.0	66.7
Canada	2	0.0	50.0	50.0	0.0
Netherlands	3	0.0	0.0	0.0	100.0
Switzerland	2	0.0	0.0	100.0	0.0
UK	11	9.1	54.5	27.3	9.1
USA	36	2.8	13.9	50.0	33.3
Other	12	16.7	25.0	50.0	8.3
Setting
Population/registry	16	0.0	25.0	43.8	31.3
Emergency room	3	0.0	0.0	0.0	100.0
Sexual health clinic	11	0.0	27.3	45.5	27.3
Family planning clinic	5	0.0	40.0	40.0	20.0
GP clinic	4	0.0	75.0	0.0	25.0
HIV study	3	0.0	33.3	0.0	66.7
Military	4	25.0	0.0	50.0	25.0
Detention facility	11	9.1	9.1	63.6	18.2
Other	18	16.7	22.2	44.4	16.7
Study design
Case-control	5	0.0	20.0	60.0	20.0
Cross-sectional	65	7.7	24.6	43.1	24.6
Prospective	5	0.0	20.0	0.0	80.0
Presence of symptoms
Asymptomatic only	6	0.0	66.7	16.7	16.7
Symptomatic and asymptomatic	69	7.2	20.2	43.5	29.0

Other includes combinations of Chlamydia, gonorrhoea, syphilis and HIV outcomes.

Reported separate stratified models or an interaction term.

Table 4.

A summary of the level of empirical support for variables associated with STD outcomes according to the ratings of methodological quality.

	Proportion of studies (n) with significant findings according to ratings ^a
	Marginal		Acceptable		Commendable		Total
	N	%	N	%	N	%	N	%	Level of empirical support
Underlying determinants
Age (at study time)	20	55.0	44	59.1	20	85.0	84	64.3	Strong evidence
Age (at sexual debut)	1	0.0	9	55.6	4	25.0	14	42.9	Mixed evidence^b
Race/Ethnicity	10	70.0	37	67.6	18	72.2	65	69.2	Moderate evidence
Socio-economic status	1	0.0	11	18.2	9	55.6	21	33.3	Weak evidence
Marital status	3	100.0	10	20.0	4	25.0	17	35.3	Weak evidence
Drug/alcohol use	7	0.0	12	33.3	3	33.3	22	22.7	Weak evidence
Sex in exchange for gifts/money	2	0.0	6	16.7	--	--	8	12.5	Weak evidence^b
Proximate determinants
Exposure of susceptible to infected people
Multiple lifetime sexual partners	1	100.0	7	57.1	9	100.0	17	82.4	Strong evidence
Multiple sex partners in recent months	12	83.3	29	55.2	14	64.3	55	63.6	Moderate evidence
Sexual partners from overseas	3	0.0	4	25.0	1	100.0	8	25.0	Weak evidence^b
Sex with regular partners	4	25.0	3	33.3	3	0.0	10	20.0	Weak evidence^b
Sex with casual partners	3	33.3	6	66.7	2	0.0	11	45.5	Mixed evidence^b
Sex with new partners	5	100.0	9	22.2	10	60.0	24	54.2	Mixed evidence
Sex with infected partner	4	100.0	12	75.0	6	100.0	22	86.4	Strong evidence
Sex with symptomatic partner	1	100.0	5	80.0	5	100.0	11	90.9	Strong evidence^b
Efficiency of transmission per contact
History of STI diagnosis	7	57.1	22	45.5	12	33.3	41	43.9	Weak evidence
Concurrent STI diagnosis	5	100.0	4	75.0	2	100.0	11	90.9	Strong evidence^b
Condom use	12	25.0	19	26.3	11	36.4	42	28.6	Weak evidence

Some articles were counted more than once if they conducted separate analysis by gender or STD.

Insufficient research.

Empirical support for the following seven variables was strong or moderately strong: age (at time of study), race/ethnicity, multiple lifetime sexual partners, multiple sexual partners in recent months, sex with infected partners, sex with symptomatic partners and concurrent STD diagnosis. Three predictors had inconclusive evidence: age at sexual debut, sex with causal partners and sex with new partners. Eight predictors had weak evidence associating them with STD diagnosis: socio-economic status, marital status, sexual partners from overseas, sex in exchange for gifts/money, drug/alcohol use, condom use, sex with regular partners and history of STD diagnosis.

Predictors examined by 11 studies or less (i.e. less than the lowest quartile of the number of studies reviewed) were deemed to have insufficient research. There were insufficient empirical studies to determine the nature of the association between the following six variables and infection: sex in exchange for gifts/money, sexual partners from overseas, sex with regular partners, sex with casual partners, concurrent STD diagnosis and sex with symptomatic partners. Six variables were not included in any clinical guidelines: race/ethnicity, socio-economic status, marital status, sex with regular partners, sex with casual partners and sex with symptomatic partner.

Assessment of the predictive abilities of underlying determinants

The following underlying determinants of STD diagnosis were identified in this review: age (at study time), age at sexual debut, race/ethnicity, socio-economic status (SES), sex work, marital status and drug/alcohol use (Table 4). Younger age was consistently associated with increased risk of STD acquisition (especially chlamydia infection); 85% of ‘commendable’ studies reported a significant association between age and STD diagnosis in their multivariable analyses (Table 4). There was inconclusive evidence in the current review associating early age at sexual debut to STD diagnosis; only 25% of ‘commendable’ studies and 56% of ‘acceptable’ studies that examined this risk factor found a significant relationship between age at sexual debut and STD diagnosis (Table 4). The level of empirical evidence in the current review for the association between race/ethnicity and STD diagnosis was moderate: 72% of ‘commendable’ studies and 68% of ‘acceptable’ studies measuring this association found significant associations in their multivariable analyses (Table 4). Most of these studies found a higher risk of infection among ethnic minority individuals in multivariate analysis.^4,86 Conversely, the level of empirical evidence of the association between SES and STD diagnosis was rated as weak – only 56% of ‘commendable’ studies and 18% ‘acceptable’ studies found a significant relationship (Table 4). SES was not associated with infection in multivariable analyses using any measure, including educational level^66,87,94 and employment status.^48,57

This review found weak evidence for the association between STD diagnosis and the following predictors: drug/alcohol use and sex in exchange for drugs/money (Table 4). Only 33% of ‘commendable’ studies and 33% of ‘acceptable’ studies found associations between drug/alcohol use and STD diagnosis.

Assessment of the predictive abilities of determinants of exposure of susceptible to infected people

The following proximate determinants of exposure to infected people were identified in this review: multiple lifetime sexual partners, multiple sex partners in recent months, sexual partners from overseas, sex with regular partners, sex with casual partners, sex with new partners, sex with infected partners and sex with symptomatic partners (Table 4). The empirical evidence linking multiple lifetime sexual partners (i.e. more than one sexual partner) and STDs was strong. However, there was no consistent evidence for the association with higher cut-points (e.g. >2, >5, or >10 partners). All of the ‘commendable’ studies and 57% of ‘acceptable’ studies showed significant associations between multiple lifetime sexual partners and STD diagnosis (Table 4). We also found strong evidence for the association between having multiple sexual partners in recent months and STD diagnosis. Approximately, 64% of ‘commendable’ studies and 57% of ‘acceptable’ studies that measured this relationship found a significant association between having multiple sexual partners in recent months in multivariable analysis (Table 4).

This review also explored the association between STD diagnosis and sexual partner characteristics. The evidence for the association between STD diagnosis and the following predictors was either inconclusive, weak and/or explored in few studies: sex with regular partners, sex with casual partners and sex with partners from overseas (Table 4). We found inconclusive evidence regarding the association between having a new partner and STD diagnosis (only about 60% of ‘commendable’ studies and 33% of ‘acceptable’ studies that measured this variable found significant associations with STD diagnosis). This review found strong evidence between STD diagnosis and sex with infected partners or symptomatic partners. All of the ‘commendable’ studies and 75% of ‘acceptable’ studies measuring this association between found a significant association.

Assessment of the predictive abilities of determinants of the efficiency of transmission

The following proximate determinants of the efficiency of infection transmission were identified: history of STD diagnosis, concurrent STD diagnosis and condom use (Table 4). We found weak evidence for the association between having a history of STD diagnosis and infection. Only 33% of ‘commendable’ studies and 45% of 22 ‘acceptable’ studies that measured this association found a significant relationship between history of STD diagnosis and STD diagnosis in multivariable analyses. There was, however, strong evidence for the association between having a concurrent STD diagnosis and infection. Both of the two ‘commendable’ studies and 75% of the four ‘acceptable’ studies that explored the relationship between having a concurrent STD diagnosis and infection found a significant association, although few studies explored this relationship (n = 11).

The empirical evidence for the association between the use of condoms and infection was weak across the studies. Only 36% of ‘commendable’ studies, 26% of ‘acceptable’ studies and 25% of ‘marginal’ studies found significant associations between condom use and STD diagnosis. There were no notable trends among the studies examining the relationship between condom use and STD diagnosis, although self-reported condom use is an enduring measurement problem in this field.⁹⁸

Discussion

The increasing availability and use of non-invasive diagnostic tests for STD diagnosis has reopened discussions about the feasibility of scaling up selective testing efforts to include the general population. Thus, a better understanding of the availability (or lack) of evidence in this area may help inform the ways in which existing selective testing criteria continue to be implemented as well as the potential for novel approaches to STD testing intervention practices to unfold in the future, including Internet-based testing initiatives.⁹⁹ This review was conceptualised as a guide to help decision-makers considering undertaking selective testing programmes in alternative settings such as the Internet; specifically, the findings may help them augment selective testing by identifying variables to include in risk assessment questionnaires.

This review accomplishes two tasks. First, it summarises the literature on the most consistent predictors of chlamydia and gonorrhoea infection, which were identified after a wide range of patient characteristics was assessed. Age, number of sexual partners, sex with a symptomatic partner, sex with an infected partner and concurrent STD diagnosis are perhaps the most intuitive predictors of STD risk. The current review has reinforced the importance of these variables. Second, this review adopted a theoretical framework to contextualise the findings. The proximate-determinants framework hypothesises that that after adjustment for the proximate determinants or socio-behavioural predictors, relationships between the underlying or socio-demographic characteristics and STD diagnosis should be non-significant.^11,12 In this review, several underlying predictors (e.g. age and race/ethnicity) were found to be significantly associated with STD diagnosis even in studies that adjusted for proximate determinants. This finding suggests that some proximate determinants may potentially be measured with error.^11,12 It also highlights the complexity inherent in the self-reported nature of proximate determinants such as condom use, coupled with measurement issues in underlying predictors (e.g. race/ethnicity).¹⁰⁰

We found reasonably good concordance between risk factors consistently listed in the recommendations (e.g. younger age, multiple sexual partners and sexual contact of STD case) and predictors found to have strong empirical support in the literature (i.e. if a risk factor is included in more guidelines, then there is more evidence supporting its predictive capacity). However, when the guidelines are considered individually or in isolation, there are large inconsistencies between the STD risk factors cited by the organisations that may reflect the lack of clarity, transparency and consensus about how high-risk behaviours influence specific STD diagnosis. Another possible explanation for the differences in recommendations may be due to regional differences in prevalence. These inconsistencies could also reflect the overall priorities of organisations issuing recommendation statements (e.g. minimising harm from screening and cost-effectiveness versus concerns about missed cases).

This review has some limitations. There was a paucity of empirical studies examining the strength of predictors in asymptomatic populations. Because most STDs are asymptomatic, it is important to elucidate the predictive ability of risk factors in the absence of symptoms. However, most studies combined symptomatic and asymptomatic individuals in multivariable analyses likely masking important differences between these populations. We were unable to identify clear trends in associations between predictors and STDs among the six studies that explicitly examined asymptomatic populations. This limitation may have considerable impact on some findings. First, it is likely that the effect sizes of other predictors were underestimated in multivariable regression models that included symptoms as covariates because of the strong association between symptoms and STDs. Second, the significant associations between some predictors and STDs may be a function of selection bias in SRH clinic settings because of differential health-seeking behaviours.^101,102 It is hypothesised that when health-seeking behaviour of certain population sub-groups (e.g. older adults, racial/ethnic minorities) is low, its percentage of symptomatic clinic visits will be high because these individuals are less likely to visit SRH clinic for routine screening.^101,102 Thus, the associations found may not actually reflect a higher concentration of risk in certain sub-groups but is a reflection of the over-representation of these sub-groups in symptomatic populations presenting to clinics. Further studies assessing asymptomatic populations are, thus, warranted.

Also, we were unable to synthesise the evidence for predictors of chlamydia and gonorrhoea separately as there were few studies examining gonorrhoea outcomes only. Based on a sub-analysis, the most important predictors of gonorrhoea outcomes only were no different from those for chlamydia outcomes only, despite the differences in the epidemiology and social determinants between these two infections. In addition, we were unable to evaluate the extent of collinearity between predictors. However, a closer examination of model building in the individual articles revealed that most studies handled collinearity appropriately, e.g. including one of the collinear variables in the final model. We were unable to synthesise the evidence separately for special populations (e.g. men who have sex with men, men who have sex with women) as clinical guidelines and empirical analyses for these groups were inconsistently reported.

We would like to draw some attention to some differences in guideline recommendations for population-based screening at general practitioner settings, for example compared to SRH clinics. The identification of high-risk individuals in these two settings requires slightly differently approaches to decision-making. For example, in order to reduce STD incidence on a population level, studies of disease dynamics recommend focusing screening efforts on core groups who are more likely to have the highest concentration of risk behaviours and infection incidence (e.g. adolescents, sex workers).^103–105 However, in STD clinic settings, where individual-level screening decision-making is the priority, the development of a personalised risk-scoring tool may have more relevance.^61,106 In the future, SRH clinicians may develop risk estimation or prediction tools by combining the risk factors identified in this review. The prediction tools may have important implications for targeted risk communication and STD testing motivation among patients attending SRH clinics.¹⁰⁷ Finally, we recommend that periodic reviews of the evolving evidence base in this substantive area have good potential to contribute to the planning and evaluation of selective testing intervention practices as they unfold in the future. There is, however, a need to continue to build the evidence base, particularly for special populations, and as STD testing intervention practices expand to non-traditional settings (e.g. the Internet; the home).

Footnotes

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the Canadian Institutes of Health Research Doctoral Research Award.

References

Golden

Kerndt

. Improving clinical operations: can we and should we save our STD clinics?. Sex Transm Dis 2010; 37: 264–265.

Rietmeijer

Mettenbrink

. Why we should Save our STD Clinics. Sex Transm Dis 2010; 37: 591.

Richardus

Gotz

. Risk selection and targeted interventions in community-based control of chlamydia. Curr Opin Infect Dis 2007; 20: 60–65.

Gotz

van Bergen

Veldhuijzen

. A prediction rule for selective screening of Chlamydia trachomatis infection. Sex Transm Infect 2005; 81: 24–30.

Fairley

Vodstrcil

Read

. The importance of striving for greater efficiency. Sex Health 2011; 8: 3–4.

Shamos

Mettenbrink

Subiadur

. Evaluation of a Testing-Only “Express” Visit Option to Enhance Efficiency in a Busy STI Clinic. Sex Transm Dis 2008; 35: 336–340.

Manhart

Marrazzo

Fine

. Selective testing criteria for gonorrhea among young women screened for chlamydial infection: contribution of race and geographic prevalence. J Infect Dis 2007; 196: 731–737.

Sniderman

Furberg

. Why guideline-making requires reform. JAMA 2009; 301: 429–431.

Ownsworth

McKenna

. Investigation of factors related to employment outcome following traumatic brain injury: a critical review and conceptual model. Disabil Rehabil 2004; 26: 765–783.

10.

Turner

Fleming

Ownsworth

. The transition from hospital to home for individuals with acquired brain injury: a literature review and research recommendations. Disabil Rehabil 2008; 30: 1153–1176.

11.

Boerma

Weir

. Integrating demographic and epidemiological approaches to research on HIV/AIDS: the proximate-determinants framework. J Infect Dis 2005; 191(1): S61–S67.

12.

Lewis

Donnelly

Mare

. Evaluating the proximate determinants framework for HIV infection in rural Zimbabwe. Sex Transm Infect 2007; 83: i61–i69.

13.

de Rooij

Roorda

Otten

. Predictors of multidisciplinary treatment outcome in fibromyalgia: a systematic review. Disabil Rehabil 2013; 35: 437–449.

14.

Ariens

van Mechelen

Bongers

. Psychosocial risk factors for neck pain: a systematic review. Am J Ind Med 2001; 39: 180–193.

15.

U.S. Preventive Services Task Force. Screening for gonorrhea. http://www.uspreventiveservicestaskforce.org/uspstf/uspsgono.htm (accessed 22 July 2013).

16.

U.S. Preventive Services Task Force. Screening for chlamydial infection: U.S. preventive services task force recommendation statement. Ann Intern Med 2007; 147: 128–134.

17.

Centers for Disease Control and Prevention (CDC). Sexually Transmitted Diseases Treatment Guidelines, 2010. MMWR Recomm Rep 2010; 59: 1–110.

18.

American Congress of Obstetricians and Gynecologists. Annual women's healthcare: high-risk factors. http://www.acog.org/About_ACOG/ACOG_Departments/Annual_Womens_Health_Care/High-Risk_Factors (accessed 22 July 2013).

19.

Public Health Agency of Canada. Canadian Guidelines on Sexually Transmitted Infections – Updated January 2010. http://www.phac-aspc.gc.ca/std-mts/sti-its/cgsti-ldcits/index-eng.php (accessed 22 July 2013).

20.

The Society of Obstetricians and Gynaecologists of Canada. Screening recommendations for asymptomatic individuals, http://www.sexualityandu.ca/health-care-professionals/sexually-transmitted-infections/screening-recommendations-for-asymptomatic-individuals (accessed 22 July 2013).

21.

British Association for Sexual Health and HIV. Sexually transmitted infections: UK National Screening and Testing Guidelines. http://www.bashh.org/documents/59/59.pdf (accessed 22 July 2013).

22.

Australasian Society for HIV Medicine. Guidelines for managing sexually transmitted infections: screening of asymptomatic men and women. http://silverbook.health.wa.gov.au/Default.asp?PublicationID=1&SectionID=236 (accessed 22 July 2013).

23.

European Branch of the International Union against Sexually Transmitted Infection and the European Office of the World Health Organization. Introduction European STD Guidelines. Int J STD AIDS 2001; 12: 2–3.

24.

Meyers

Wolff

Gregory

. USPSTF Recommendations for STI screening. Am Fam Phys 2008; 77: 819–824.

25.

Kong

Hocking

Link

. Sex and sport: chlamydia screening in rural sporting clubs. BMC Infect Dis 2009; 9: 73.

26.

Greene

Stafford

. Prevalence of Chlamydia trachomatis among active duty male soldiers reporting to a troop medical clinic for routine health care. South Med J 2007; 100: 478–481.

27.

Mak

Van Renterghem

Traen

. Chlamydia trachomatis in female sex workers in Belgium: 1998–2003. Sex Transm Infect 2005; 81: 89–90.

28.

Menon-Johansson

Winston

Matthews

. The first point prevalence study of genital Chlamydia trachomatis infection in young male inmates in the UK. Int J STD AIDS 2005; 16: 799–801.

29.

Powell

O'Connor

O'hlarlaithe

. Chlamydia trachomatis prevalence in men in the mid-west of Ireland. Sex Transm Infect 2004; 80: 349–53.

30.

Annan

Sullivan

Nori

. Rectal chlamydia – a reservoir of undiagnosed infection in men who have sex with men. Sex Transm Infect 2009; 85: 176–179.

31.

Dicker

Mosure

Kay

. An ongoing burden: chlamydial infections among young American Indian women. Matern Child Health J 2008; 12: 25–29.

32.

Hince

Brett

Mak

. Opportunistic screening for chlamydia in young men. Aust Fam Phys 2009; 38: 734–737.

33.

Spasovski

Simjanovska

Taleski

. Screening of Chlamydia trachomatis urogenital infections among the male and female population of the Republic of Macedonia. J Eur Acad Dermatol Venereol 2005; 19: 427–430.

34.

Simms

Talebi

Rhia

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

35.

Hocking

Fairley

. Do the characteristics of sexual health centre clients predict chlamydia infection sufficiently strongly to allow selective screening?. Sex Health 2005; 2: 185–192.

36.

Tipple

Hill

Smith

. Is screening for pharyngeal Chlamydia trachomatis warranted in high-risk groups?. Int J STD AIDS 2010; 21: 770–771.

37.

Mossong

Muller

Majery

. Screening for Chlamydia trachomatis in secondary schools, family planning and occupational health centres in Luxembourg. Sex Transm Infect 2009; 85: 455–458.

38.

Paukku

Kilpikari

Puolakkainen

. Criteria for selective screening for Chlamydia trachomatis. Sex Transm Dis 2003; 30: 120–123.

39.

Ivens

Macdonald

Bansi

. Screening for rectal chlamydia infection in a genitourinary medicine clinic. Int J STD AIDS 2007; 18: 404–406.

40.

Landes

Thorne

Barlow

. Prevalence of sexually transmitted infections in HIV-1 infected pregnant women in Europe. Eur J Epidemiol 2007; 22: 925–936.

41.

Barry

Kent

Klausner

. Risk factors for gonorrhea among heterosexuals – San Francisco, 2006. Sex Transm Dis 2009; 36: S62–S66.

42.

Hirst

Dinsmore

. Chlamydia screening programme, HMP Hydebank Wood Young Offender Centre, Belfast Northern Ireland, UK. Int J STD AIDS 2009; 20: 360–361.

43.

Schillinger

Dunne

Chapin

. Prevalence of Chlamydia trachomatis infection among men screened in 4 U.S. cities. Sex Transm Dis 2005; 32: 74–77.

44.

Matteelli

Beltrame

Carvalho

. Chlamydia trachomatis genital infection in migrant female sex workers in Italy. Int J STD AIDS 2003; 14: 591–595.

45.

Steenbeek

Tyndall

Sheps

. An epidemiological survey of chlamydial and gonococcal infections in a Canadian arctic community. Sex Transm Dis 2009; 36: 79–83.

46.

Geisler

James

. Chlamydial and gonococcal infections in women seeking pregnancy testing at family-planning clinics. Am J Obstet Gynecol 2008; 198: 502.e1–502.e4.

47.

Bakken

Nordbo

Skjeldestad

. Chlamydia trachomatis testing patterns and prevalence of genital chlamydial infection among young men and women in Central Norway 1990–2003: a population-based registry study. Sex Transm Dis 2006; 33: 26–30.

48.

Bozicevic

Grgic

Zidovec-Lepej

. Urine-based testing for Chlamydia trachomatis among young adults in a population-based survey in Croatia: feasibility and prevalence. BMC Public Health 2011; 11: 230.

49.

Gaydos

Kent

Rietmeijer

. Prevalence of neisseria gonorrhoeae among men screened for chlamydia trachomatis in four United States cities, 1999–2003. Sex Transm Dis 2006; 33: 314–319.

50.

Jackson

Sebo

Aeby

. Prevalence and associated factors for Chlamydia trachomatis infection among undocumented immigrants in a primary care facility in Geneva, Switzerland: a cross-sectional study. J Immigr MinorHealth 2010; 12: 909–914.

51.

Joesoef

Mosure

. Prevalence of chlamydia in young men in the United States from newly implemented universal screening in a national job training program. Sex Transm Dis 2006; 33: 636–639.

52.

Richardson

Sellors

Mackinnon

. Prevalence of chlamydia trachomatis infections and specimen collection preference among women, using self-collected vaginal swabs in community settings. Sex Transm Dis 2003; 30: 880–885.

53.

Singh

Fine

Marrazzo

. Chlamydia trachomatis infection among women reporting sexual activity with women screened in family planning clinics in the Pacific Northwest, 1997 to 2005. Am J Public Health 2011; 101: 1284–1290.

54.

Steiner

Haller

Elger

. Chlamydia trachomatis infection in a Swiss prison: a cross sectional study. Swiss Med Wkly 2010; 140: w13126.

55.

Sutton

Martinko

Hale

. Prevalence and high rate of asymptomatic infection of Chlamydia trachomatis in male college reserve officer training corps cadets. Sex Transm Dis 2003; 30: 901–904.

56.

Uuskula

Kals

Denks

. The prevalence of chlamydial infection in Estonia: a population-based survey. Int J STD AIDS 2008; 19: 455–458.

57.

Stein

Kaufman

Ford

. Screening young adults for prevalent chlamydial infection in community settings. Ann Epidemiol 2008; 18: 560–571.

58.

Andersen

van Valkengoed

Olesen

. Value of self-reportable screening criteria to identify asymptomatic individuals in the general population for urogenital Chlamydia trachomatis infection screening. Clin Infect Dis 2003; 36: 837–844.

59.

Bernstein

Chow

Ruiz

. Chlamydia trachomatis and Neisseria gonorrhoeae infections among men and women entering California prisons. Am J Public Health 2006; 96: 1862–1866.

60.

Goulet

de Barbeyrac

Raherison

. Prevalence of Chlamydia trachomatis: results from the first national population-based survey in France. Sex Transm Infect 2010; 86: 263–270.

61.

Wand

Guy

Donovan

. Developing and validating a risk scoring tool for chlamydia infection among sexual health clinic attendees in Australia: a simple algorithm to identify those at high risk of chlamydia infection. BMJ Open 2011; 1: e000005.

62.

Belenko

Dembo

Weiland

. Recently arrested adolescents are at high risk for sexually transmitted diseases. Sex Transm Dis 2008; 35: 758–763.

63.

Asbel

Newbern

Salmon

. School-based screening for Chlamydia trachomatis and Neisseria gonorrhoeae among Philadelphia public high school students. Sex Transm Dis 2006; 33: 614–620.

64.

Robertson

Thomas

St Lawrence

. Predictors of infection with chlamydia or gonorrhea in incarcerated adolescents. Sex Transm Dis 2005; 32: 115–122.

65.

Kahn

Moseley

Thilges

. Community-based screening and treatment for STDs: results from a mobile clinic initiative. Sex Transm Dis 2003; 30: 654–658.

66.

Doherty

Adimora

Schoenbach

. Correlates of gonorrhoea among African Americans in North Carolina. Int J STD AIDS 2007; 18: 114–117.

67.

McDonagh

Ryder

McNulty

. Neisseria gonorrhoeae infection in urban Sydney women: prevalence and predictors. Sex Health 2009; 6: 241–244.

68.

Rao

Bacon

Evans

. Prevalence of Neisseria gonorrhoeae infection in young subjects attending community clinics in South London. Sex Transm Infect 2008; 84: 117–121.

69.

Bozicevic

Fenton

Martin

. Epidemiological correlates of asymptomatic gonorrhea. Sex Transm Dis 2006; 33: 289–295.

70.

Geisler

Chow

Schachter

. Pelvic examination findings and Chlamydia trachomatis infection in asymptomatic young women screened with a nucleic acid amplification test. Sex Transm Dis 2007; 34: 335–338.

71.

Gorgos

Fine

Marrazzo

. Chlamydia positivity in American Indian/Alaska native women screened in family planning clinics, 1997–2004. Sex Transm Dis 2008; 35: 753–757.

72.

LaMontagne

Fenton

Randall

. Establishing the national chlamydia screening programme in England: results from the first full year of screening. Sex Transm Infect 2004; 80: 335–341.

73.

Lofy

Hofmann

Mosure

. Chlamydial infections among female adolescents screened in juvenile detention centers in Washington state, 1998–2002. Sex Transm Dis 2006; 33: 63–67.

74.

Satterwhite

Tian

Braxton

. Chlamydia prevalence among women and men entering the national job training program: United States, 2003–2007. Sex Transm Dis 2010; 37: 63–67.

75.

Arcari

Gaydos

Howell

. Feasibility and short-term impact of linked education and urine screening interventions for chlamydia and gonorrhea in male army recruits. Sex Transm Dis 2004; 31: 443–447.

76.

Trick

Kee

Murphy-Swallow

. Detection of chlamydial and gonococcal urethral infection during jail intake: development of a screening algorithm. Sex Transm Dis 2006; 33: 599–603.

77.

Javanbakht

Murphy

Harawa

. Sexually transmitted infections and HIV prevalence among incarcerated men who have sex with men, 2000–2005. Sex Transm Dis 2009; 36: S17–S21.

78.

Gaydos

Howell

Quinn

. Sustained high prevalence of Chlamydia trachomatis infections in female army recruits. Sex Transm Dis 2003; 30: 539–544.

79.

Geisler

Chyu

Kusunoki

. Health insurance coverage, health care-seeking behaviors, and genital chlamydial infection prevalence in sexually active young adults. Sex Transm Dis 2006; 33: 389–396.

80.

Joffe

Rietmeijer

Chung

. Screening asymptomatic adolescent men for Chlamydia trachomatis in school-based health centers using urine-based nucleic acid amplification tests. Sex Transm Dis 2008; 35: S19–S23.

81.

McDonnell

Levy

Morton

. Risk factors for chlamydia among young women in a Northern California juvenile detention facility: implications for community intervention. Sex Transm Dis 2009; 36: S29–S33.

82.

Scott Lamontagne

Baster

Emmett

. Incidence and reinfection rates of genital chlamydial infection among women aged 16–24 years attending general practice, family planning and genitourinary medicine clinics in England: a prospective cohort study by the Chlamydia Recall Study Advisory Group. Sex Transm Infect 2007; 83: 292–303.

83.

Skjeldestad

Marsico

Sings

. Incidence and risk factors for genital Chlamydia trachomatis infection: A 4-year prospective cohort study. Sex Transm Dis 2009; 36: 273–279.

84.

van Bergen

Gotz

Richardus

. 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.

85.

Gotz

Veldhuijzen

Habbema

. Prediction of Chlamydia trachomatis infection: application of a scoring rule to other populations. Sex Transm Dis 2006; 33: 374–380.

86.

La Montagne

Patrick

Fine

. Region X infertility prevention project. Re-evaluating selective screening criteria for chlamydial infection among women in the U S Pacific Northwest. Sex Transm Dis 2004; 31: 283–289.

87.

Latino

Caneparo

Rosso

. Prevalence and risk factors for Chlamydia trachomatis infection in young women in north-west of Italy. Minerva Ginecol 2008; 60: 29–37.

88.

Salfa

Latino

Regine

. Prevalence and determinants of Chlamydia trachomatis infection among sexually active women in Turin, Italy. Italian J Public Health 2011; 8: 295–301.

89.

Berry

Ghanem

Page

. Increased gonorrhoea and chlamydia testing did not increase case detection in an HIV clinical cohort 1999–2007. Sex Transm Infect 2011; 87: 469–475.

90.

Datta

Sternberg

Johnson

. Gonorrhea and chlamydia in the United States among persons 14 to 39 years of age, 1999 to 2002. Ann Intern Med 2007; 147: 89–96.

91.

Einwalter

Ritchie

Ault

. Gonorrhea and chlamydia infection among women visiting family planning clinics: racial variation in prevalence and predictors. Perspect Sex Reprod Health 2005; 37: 135–140.

92.

Hardick

Hsieh

Tulloch

. Surveillance of Chlamydia trachomatis and Neisseria gonorrhoeae infections in women in detention in Baltimore, Maryland. Sex Transm Dis 2003; 30: 64–70.

93.

Jin

Prestage

Mao

. Incidence and risk factors for urethral and anal gonorrhoea and chlamydia in a cohort of HIV-negative homosexual men: the health in men study. Sex Transm Infect 2007; 83: 113–119.

94.

Al-Tayyib

Miller

Rogers

. Evaluation of risk score algorithms for detection of chlamydial and gonococcal infections in an emergency department setting. Acad Emerg Med 2008; 15: 126–135.

95.

Merchant

DePalo

Liu

. Developing a system to predict laboratory-confirmed chlamydial and/or gonococcal urethritis in adult male emergency department patients. Postgrad Med 2010; 122: 52–60.

96.

Mehta

Hall

Lyss

. Adult and pediatric emergency department sexually transmitted disease and hiv screening: programmatic overview and outcomes. Acad Emerg Med 2007; 14: 250–258.

97.

Chen

Rohrsheim

Donovan

. Chlamydia trachomatis infection in Sydney Women. Aust N Z J Obstet Gynaecol 2005; 45: 410–413.

98.

Vanable

Carey

Brown

. Test-retest reliability of self-reported HIV/STD-related measures among African-American adolescents in four U.S. cities. J Adolesc Health 2009; 44: 214–221.

99.

Shoveller

Knight

Davis

. Online sexual health services: examining youth's perspectives. Can J Public Health 2012; 103: 14–18.

100.

Adimora

Schoenbach

Martinson

. Social context of sexual relationships among rural African Americans. Sex Transm Dis 2001; 28: 69–76.

101.

Guan

. Self-reported sexually transmitted disease symptoms and treatment-seeking behaviors in China. AIDS Patient Care STDS 2009; 23: 443–448.

102.

Matser

Luu

Geskus

. Higher Chlamydia trachomatis prevalence in ethnic minorities does not always reflect higher sexual risk behaviour. PLoS One 2013; 8: e67287.

103.

Rockhill

. The privatization of risk. Am J Public Health 2001; 91: 365–368.

104.

Aral

. Determinants of STD epidemics: implications for phase appropriate intervention strategies. Sex Transm Infect 2002; 78: i3–i13.

105.

Rockhill

Kawachi

Colditz

. Individual risk prediction and population-wide disease prevention. Epidemiol Rev 2000; 22: 176–180.

106.

Falasinnu

Gustafson

Hottes

. A critical appraisal of risk models for predicting sexually transmitted infections. Sex Transm Dis 2014; 41: 321–330.

107.

Falasinnu

Gustafson

Gilbert

. Risk prediction in sexual health contexts: protocol. JMIR Res Protoc 2013; 2: e57.