You are here

  • Home
  • Archive
  • Volume 94, Issue 2
  • Can enhanced screening of men with a history of prior syphilis infection stem the epidemic in men who have sex with men? A mathematical modelling study

Article Text

Article menu

Download PDFPDF

Epidemiology
Original article
Can enhanced screening of men with a history of prior syphilis infection stem the epidemic in men who have sex with men? A mathematical modelling study
  1. Ashleigh R Tuite1,
  2. Souradet Shaw2,3,
  3. Joss N Reimer3,
  4. Craig P Ross3,
  5. David N Fisman4,5,6,
  6. Sharmistha Mishra4,5,6,7
  1. 1 T H Chan School of Public Health, Harvard University, Boston, Massachusetts, USA
  2. 2 Department of Community Health Sciences, University of Manitoba, Winnipeg, Canada
  3. 3 Winnipeg Regional Health Authority, Winnipeg, Manitoba, Canada
  4. 4 Division of Infectious Diseases, Department of Medicine, University of Toronto, Toronto, Canada
  5. 5 Dalla Lana School of Public Health, University of Toronto, Toronto, Canada
  6. 6 Institute of Medical Sciences, University of Toronto, Toronto, Canada
  7. 7 Centre for Urban Health Solutions, Li Ka Shing Knowledge Institute, University of Toronto, Toronto, Canada
  1. Correspondence to Dr Sharmistha Mishra, University of Toronto, Room 363, Li Ka Shing Knowledge Institute, 30 Bond Street, Toronto, ON M5B 1W8, Canada; sharmistha.mishra{at}utoronto.ca

Abstract

Objectives The aim of this study is to determine the transmission impact of using prior syphilis infection to guide a focused syphilis screening intervention among men who have sex with men (MSM).

Methods We parameterised a deterministic model of syphilis transmission in MSM to reflect the 2011–2015 syphilis outbreak in Winnipeg, Canada. Enhanced screening of 75% of men with prior syphilis every 3 months (A) was compared with distributing equivalent number tests to all MSM (B) or those with the highest partner number (C). We compared early syphilis incidence, diagnoses and prevalence after 10 years, relative to a base case of 30% of MSM screened annually.

Results Strategy A was expected to avert 52% of incident infections, 44% of diagnosed cases and reduce early syphilis prevalence by 89%. Strategy B had the least impact. Strategy C was most effective, averting 59% of incident cases. When screening frequency was semiannual or annual, strategy A was the most effective.

Conclusions Enhanced screening of MSM with prior syphilis may efficiently reduce transmission, especially when identification of high-risk men via self-reported partner numbers or high-frequency screening is difficult to achieve.

  • SYPHILIS
  • SCREENING
  • MATHEMATICAL MODEL
  • TRANSMISSION DYNAMICS

https://doi.org/10.1136/sextrans-2017-053201

Statistics from Altmetric.com

    Request Permissions

    If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

    Introduction

    Frequent syphilis screening remains the best available biomedical tool for syphilis control. Mathematical modelling studies suggest that frequent (every 3–4 months) screening may be an effective and cost-effective approach to syphilis control among populations with high syphilis incidence.1–4 However, considerable resources are often required to achieve high-frequency screening of all at-risk individuals,5 and increased screening may lead to a paradoxical increase in cases6 or resurgence following apparent control if screening cannot be sustained.7 Implementable approaches to focus screening to those most at risk are needed to maximise impact in an efficient and sustained manner.

    Focusing on individuals who are disproportionately vulnerable to acquisition and transmission of infection (core groups) remains a strategic focus for STI control. However, identifying core group members may be challenging, with STI screening guidelines often using self-reported behaviours to identify individuals at greatest risk for STI acquisition. The Canadian city of Winnipeg recently initiated a novel approach for increasing screening in at-risk individuals. Under this programme, all individuals diagnosed with syphilis are offered the option to receive testing reminders every 3 months. As of May 2017, greater than 80% of individuals diagnosed with syphilis had agreed to participate. The rationale for focusing on those with a prior infection relates to the observation that individuals with a prior syphilis infection experience an increased future risk of syphilis and other STIs.8 Importantly, a prior infection may signal ongoing vulnerability to STI acquisition and transmission due to sexual network factors, independent of self-reported sexual behaviours such as number of sexual partners.9–14 The extent to which a prior syphilis infection could help identify core group members without requiring elicitation of sexual behaviours has not been examined.

    In parallel with the implementation of this new strategy in Winnipeg, we used mathematical modelling to examine the potential effectiveness of using previous infection as a marker of risk in a focused screening intervention and compared its impact on reducing syphilis burden with that of other screening strategies—universal screening or focused screening based on number of sexual partners. Our model specifically focused on men who have sex with men (MSM), who experience elevated rates of infection and was informed by data from the ongoing syphilis epidemic in Winnipeg.

    Methods

    Model overview

    We modified a previously described deterministic compartmental model of syphilis transmission in MSM6 to capture the demographic and epidemiological features of the 2011–2015 outbreak in Winnipeg, Manitoba. Table 1 presents the model parameters, which were drawn from the biomedical literature, where possible, or by assumption or model calibration otherwise. Model equations and additional details are provided in online supplementary file 1.

    Supplementary file

    View this table:
    Table 1

    Model parameter values

    The natural history of syphilis infection was modelled by including the following states: susceptible, incubating, primary, secondary, early latent and late latent infection (see online supplementary figure 1). Transmission occurred through sexual contact between susceptible and infectious individuals. The primary and secondary stages were assumed to be infectious. In the absence of treatment, infected individuals progressed through the various stages of infection and remained in the late latent stage until they exited the modelled system. We assumed that untreated individuals with syphilis could not be infected with another strain (ie, no superinfection). Men treated for early (primary, secondary or early latent) infection were assumed to be protected from reinfection for 1 week following treatment, reflecting the persistence of penicillin in the body following administration.15 Treatment of late syphilis was followed by an average of 5 years of immunity from reinfection.16 17

    Supplementary figure 1

    The model population was separated by previous infection status: men with a reported syphilis infection entered a ‘susceptible, previously treated’ compartment (SR) following treatment. The natural history was identical for first and repeat infections. We included a ‘susceptible, treated but unreported’ compartment (SABX) to distinguish men with treated syphilis whose initial infection was not reported in the local public health system. This might occur due to inadvertent treatment of syphilis from background antibiotic use for non-syphilis indications or treatment in another public health jurisdiction. Such cases are not flagged as individuals with a history of infection but must be considered when calibrating the model to seroprevalence estimates based on serology which remains positive after treatment.18 If infected again, men with ‘treated but unreported’ syphilis transitioned through the same compartments as infection-naive individuals. Following a reported syphilis infection, men entered and remained in ‘previously treated’ part of the model.

    The model included three levels of sexual activity (high, moderate and low), with the proportions of sexually active men in each category based on self-reported partner numbers in the most recent (2006) biobehavioural survey of a convenience sample of 121 Winnipeg MSM.19 The high-activity group comprised 20% of MSM with rates of partner change 20-fold higher than rates in the low-activity group. Mixing between sexual activity groups could range from proportionate (no preferential partnerships between groups, ε=0) to assortative (ε=1, where individuals partner exclusively with individuals of the same risk group). ε was estimated by model calibration. Additional details on the calculation of force of infection by sexual activity are provided in the technical appendix (see online Supplementary file 1). Men remained in the model for 20 years on average. The modelled population size remained constant and individuals did not move between activity groups.

    Screening and treatment

    Syphilis was diagnosed and treated if individuals with infection actively sought medical care for symptoms, through partner notification or by participating in opportunistic screening for asymptomatic syphilis. For individuals actively seeking care, treatment was assumed to occur at the midpoint of the infection stage.

    We assumed a constant hazard of screening (α), with the probability of screening converted to a rate, assuming an exponential distribution20:

    Embedded Image

    where the coverage reflected the proportion of the population screened in each screening interval (base case, 1 year).21 For simplicity, we assumed 100% sensitivity and specificity of the diagnostic tests and 100% treatment efficacy.

    Model calibration

    We used Winnipeg early syphilis case data from 2011 to 2015 for model fitting. Over the 5-year period, there were 171 early syphilis cases in men reporting sex with male partners (bisexual and MSM), accounting for 80% of early syphilis cases among all male cases reported in Winnipeg. Rates in MSM were estimated to range from a minimum of 140 per 1 00 000 in 2011 to a maximum of 996 per 100 000 in 2014. Of early syphilis cases, 7.7% were repeat infections. Additional details on model fitting and estimation of the size of the Winnipeg MSM population are provided in online supplementary file 1.

    Interventions

    After the calibration period, we projected the potential impact of different screening interventions applied to the modelled population over a 10-year time period (see online supplementary table 1). Men with a syphilis diagnosis up to 5 years prior to the start of the intervention were included in the previously infected stratum of the model to correspond with the cut-off used in the actual Winnipeg intervention. Our base case scenario was 30% screening of the entire modelled population of MSM annually, and we compared this to the following enhanced screening interventions, with an equivalent volume of additional tests used for each strategy. For each of these strategies, we assumed that 30% screening of the entire population continued, with additional screening tests distributed as follows:

    • Focused screening every 3 months of men with prior reported syphilis infections

    • Uniform screening every 3 months (ie, uniform distribution of additional tests)

    • Focused screening every 3 months of men in the high sexual activity group

    Supplementary Table 1

    In strategy A, 75% of men with a prior reported infection were screened every 3 months. The additional number of tests under strategy A varied over time because the number of men with prior syphilis infection changed over the course of the epidemic. Therefore, we calculated the number of additional tests performed each year under strategy A compared with the base case (constant 30% annual screening across the intervention period) and redistributed these additional tests across the entire MSM population (strategy B) or to men with the highest number of partners (strategy C), with this extra screening occurring every 3 months. In every scenario, 30% of the MSM population not reached by the additional volume of tests continued to receive annual screening.

    Outcomes included the following relative changes over a 10-year time horizon compared with the base-case: (1) cumulative diagnosed cases of early syphilis; (2) cumulative incidence of syphilis and (3) early syphilis prevalence at the end of the intervention period.

    Sensitivity analyses: natural history and intervention assumptions

    We conducted three sensitivity analyses. First, to explore the influence of screening intensity on outcomes, we repeated the comparative analysis assuming annual or semiannual screening frequencies. Second, we explored the influence of scaling down the enhanced screening programme in strategy A after 2 years of implementation. The scale down involved three possibilities: screening in men with prior infection was reduced in frequency from every 3 months to (1) semiannually or (2) annually, with coverage maintained at 75%, or (3) all screening returned to baseline (30% MSM receiving annual screening).

    Third, due to uncertainty surrounding the existence and duration of protective immunity following treatment of late syphilis,16 17 we recalibrated the model and repeated our analyses under the assumption that men treated for late latent infection became susceptible after 1 week or 1 year.

    Ethics approval

    Research ethics board approval was not required because syphilis surveillance data from the Winnipeg Regional Health Authority were provided to the research team as aggregate data, with no personal identifying information.

    Results

    Model calibration

    The model reproduced the overall trend of increasing early syphilis diagnoses in MSM in Winnipeg (see online supplementary figure 2). Estimated syphilis seroprevalence at the end of the calibration period matched survey data19 at 3.3%. At the end of the calibration period, 5.6% of annual diagnosed early syphilis cases were repeat infections (ie, occurred in men with a recognised previous infection).

    Supplementary figure 2

    Base case scenario

    Under base case assumptions, prevalence and incidence were projected to decline over the intervention period (figure 1). Reinfections accounted for 16% of diagnosed early syphilis infections and 14% of all incident infections by the end of the 10-year intervention period, when early syphilis prevalence was projected to be 4 per 1000 MSM.

    Figure 1
    Figure 1

    Model projections of syphilis prevalence and diagnosed cases over a 10-year intervention period. The dashed lined indicates the start of the intervention period. The base case assumed 30% population screening annually (applied uniformly across the population). The ‘prior infection’ scenario (A) involved 75% of the population with a recognised prior syphilis infection being screened every 3 months. The ‘entire population’ (B) and ‘high rate of partner change’ (C) scenarios involved distributing an equivalent number of tests to those used in scenario A to either the general population or men with the highest rates of partner change every 3 months. MSM, men who have sex with men.

    Enhanced screening strategies

    In all intervention scenarios, we projected an initial transient increase in diagnosed early syphilis cases (figure 1) relative to the base case. The increase was largest under strategy C, which was focused on increasing testing among men with the highest number of partners.

    Compared with the base case, enhanced screening of men with prior diagnosed syphilis (strategy A) was projected to reduce the prevalence of early syphilis by an additional 89% at the end of the intervention period. This strategy was expected to avert 52% of incident infections and 44% of diagnosed cases relative to the base case.

    Uniform enhanced screening (strategy B) was expected to have less of an impact than focused screening, reducing prevalence by 39% and averting 19% of incident and 14% of diagnosed cases. Focused screening of men with a high number of partners (strategy C) was projected to be the most effective strategy, averting 59% of incident cases and reducing prevalence by 95% relative to the base case.

    Of the interventions evaluated, focused screening in men with a prior infection had the greatest impact on reducing syphilis prevalence syphilis in this group (see online supplementary figure 3). Screening men with high rates of partner change also reduced prevalence among men with a prior infection, but to a lesser degree. At the end of the intervention period, the majority of prevalent early syphilis cases were in men with the highest rate of partner change, regardless of the screening strategy used. Strategy C was the most effective for reducing the relative infection burden in this group (see online supplementary figure 4).

    Supplementary figure 3

    Supplementary figure 4

    Sensitivity analyses

    When frequency of enhanced screening was reduced to semiannually or annually, focusing on men with prior infection was equally or more effective than allocating tests to men with high rates of partner change (figure 2).

    Figure 2
    Figure 2

    Impact of different screening frequencies on intervention effectiveness. The proportionate reduction in incident cases, diagnosed cases and prevalence at the end of the intervention period was determined relative to the base case scenario (30% annual population screening). Results are shown for different frequencies of screening. For the prior infection strategy (A), 75% of the eligible population was screened at the indicated frequency, with the required test numbers for the entire population (B) and high rate of partner change (C) strategies calculated based on the number of men with previous infection at the start of each intervention year.

    Scaling down the intensity of strategy A after 2 years reduced its overall impact, compared with what was observed when enhanced screening was sustained, but the prevalence of early syphilis still remained lower than the base case (figure 3). Notably, if screening returned to baseline levels after 2 years of enhanced screening, syphilis prevalence was projected to rebound during the 10-year evaluation period.

    Figure 3
    Figure 3

    Impact of reducing intervention intensity over time. We evaluated the impact of reducing the frequency or coverage of screening in MSM with a prior syphilis infection on prevalence of early syphilis. The first vertical line indicates the start of the intervention period, while the second dashed line indicates the time at which intervention intensity could be reduced. At the start of the intervention, 75% of men with a prior infection were screened every 3 months. After 2 years, this could be maintained (blue), reduced to 75% of men with prior infection screened every 6 months (red), reduced to 75% of men with prior infection screened annually (green) or reduced to 30% of men (regardless of prior infection history) screened annually (purple). The base case of 30% of MSM screened annually (grey) is shown for comparison.

    Reducing the duration of protective immunity following treatment for late latent syphilis produced fewer oscillations in prevalence in base case projections (see online supplementary figure 5). Nonetheless, the relative differences in outcomes across screening strategies remained consistent across assumptions about duration of protective immunity.

    Supplementary figure 5

    Discussion

    Using a model of syphilis transmission that explicitly captures reinfection, we compared three strategies to enhance screening and interrupt transmission among MSM. We found that for an equivalent number of additional tests, focused screening was more effective than universal screening. Prioritised screening of MSM with a history of syphilis was potentially the most effective strategy when high-frequency (every 3 months) screening could not be achieved, and provides an alternate approach to self-reported sexual behaviour for identifying those most at risk of infection. Short-term, high-frequency screening of MSM with prior syphilis was projected to produce large and sustained reductions in syphilis cases even if screening intensity is scaled back after 2 years, but screening must remain higher than baseline to avoid a resurgence of cases in the medium term.

    After treatment for early syphilis, standard of care includes repeat serology at 3, 6 and 12 months and 24 months. The enhanced screening approach would involve one additional test in the first year post treatment (ie, at 9 months after treatment) and 3–4 additional tests per year thereafter. Few settings in practice achieve screening intervals of 3 months with high coverage. Barriers to increasing the frequency of screening to all or even a subset of MSM include: requiring that individuals return to clinics for syphilis (and other STI) screening in the absence of any other healthcare needs and resource limits on clinic and outreach capacity. Several settings have attempted routinised syphilis testing to increase frequency, employing measures such as computerised reminders or ‘opt-out’ syphilis tests with routine HIV viral load tests among persons living with HIV.5 22–24 Our model projected that if enhanced screening frequency was semiannual or annual, then prioritising those with a prior syphilis history was the most effective strategy.

    Enhanced screening may lead to an initial rise in diagnosed (as opposed to incident) cases. For all of the intervention strategies we evaluated, there was an approximately 2-year period when diagnosed cases would be higher than baseline while ‘true’ underlying prevalence fell. The observation is an artefact of disease surveillance: as screening is intensified among those at risk, we find and treat more cases. Thus, an initial increase in diagnosed cases may be a sign that screening programmes are having the desired effect and should not be interpreted by surveillance as an increase in incidence or prevalence. Misinterpretation of such trends can lead to concerns that enhanced screening is ineffective and further stigmatise communities vulnerable to STIs. While this issue could be addressed by adjusting incidence for testing denominators, in practice such denominator data are seldom available for communicable diseases of public health importance.25

    To our knowledge, this is the first modelling study to explore the impact of enhanced screening in individuals with a history of prior infection. Strengths of the study include the inclusion of a history of prior syphilis infection and the fact that it was informed by the best available local data on the epidemiological characteristics of the Winnipeg syphilis epidemic. There are important limitations to the study. The regional, sexual behavioural parameters were collected over 10 years and from a small sample size; the limitations of these inputs were offset by calibrating the input parameters to observed rates of syphilis infections over time. Our results are conditional on the epidemiological context under study, and the relative impact of focused screening of MSM with prior syphilis may vary by the rates of reinfection in the region of interest. Relatedly, the syphilis epidemic in Winnipeg has evolved, making significant inroads into heterosexual populations26; further research is needed to understand the impact of focused screening of populations susceptible to reinfection on mixed or parallel syphilis epidemics.

    Nonetheless, this study provides insights that can lead directly to pragmatic programme change, as public health authorities often do have case history data that would identify individuals as having prior syphilis infection, whereas the validity of self-reported sexual behaviour data appears to vary widely across populations.27

    Simpler, less subjective, risk markers could allow for a more efficient use of screening resources than behavioural screening. Identifying those at highest risk of syphilis acquisition based on self-reported behaviour requires that providers ask the pertinent questions and risk-stratify patients accordingly that patients disclose details of sexual partnerships and their sexual orientation to providers and that both sides engage in a discourse surrounding sexual health. In one study, less than half of MSM report talking to their primary care provider about sexual health,28 while another study found those who had not disclosed to their providers were less likely to receive HIV/STI testing.29 Behavioural screening for syphilis screening also places the burden of decision making and the delivery of a public health strategy on individual patient–provider interactions. Furthermore, individuals classified as ‘low risk’ by behavioural surveillance may have a small number of sexual partners within a sexual network with high-syphilis prevalence that places them at high risk of acquisition and repeat infection. Using a clinical history of syphilis infection as the risk marker means that public health teams can deliver the screening strategy and support providers accordingly. In this study, we found that even in the face of perfect information on sexual behaviour, screening strategies based on prior syphilis infection history either approached or outperformed behaviour-based strategies. The insights generated by this model provide helpful and practical guidance to public health practitioners and decision makers working to control syphilis resurgence.

    Key messages

    • Modelling studies suggest that frequent screening can reduce syphilis burden, but high-frequency screening has been challenging to achieve in practice.

    • We investigated if prior infection history could be used to focus limited public health resources to those most likely to benefit from regular syphilis screening.

    • We used a mathematical model to compare enhanced screening in MSM with a prior infection to standard screening approaches.

    • Even in the face of perfect information on sexual behaviour, screening strategies based on prior syphilis infection history either approached or outperformed behaviour-based strategies.

    References

    1. ↵
      2. Tuite A ,
      3. Fisman D
      . Go big or go home: impact of screening coverage on syphilis infection dynamics. Sex Transm Infect 2016;92:49–54.doi:10.1136/sextrans-2014-052001
      OpenUrlAbstract/FREE Full Text
    2. ↵
      2. Tuite AR ,
      3. Burchell AN ,
      4. Fisman DN
      . Cost-effectiveness of enhanced syphilis screening among HIV-positive men who have sex with men: a microsimulation model. PLoS One 2014;9:e101240.doi:10.1371/journal.pone.0101240
    3. ↵
      2. Tuite AR ,
      3. Fisman DN ,
      4. Mishra S
      . Screen more or screen more often? Using mathematical models to inform syphilis control strategies. BMC Public Health 2013;13:606.doi:10.1186/1471-2458-13-606
      OpenUrlCrossRefPubMed
    4. ↵
      2. Gray RT ,
      3. Hoare A ,
      4. Prestage GP , et al
      . Frequent testing of highly sexually active gay men is required to control syphilis. Sex Transm Dis 2010;37:1–305.doi:10.1097/OLQ.0b013e3181ca3c0a
      OpenUrlCrossRefPubMed
    5. ↵
      2. Zou H ,
      3. Fairley CK ,
      4. Guy R , et al
      . The efficacy of clinic-based interventions aimed at increasing screening for bacterial sexually transmitted infections among men who have sex with men: a systematic review. Sex Transm Dis 2012;39:382–7.doi:10.1097/OLQ.0b013e318248e3ff
      OpenUrlCrossRefPubMed
    6. ↵
      2. Tuite A ,
      3. Fisman D
      . Go big or go home: impact of screening coverage on syphilis infection dynamics. Sex Transm Infect 2016;92.doi:10.1136/sextrans-2014-052001
    7. ↵
      2. Rekart ML ,
      3. Patrick DM ,
      4. Chakraborty B , et al
      . Targeted mass treatment for syphilis with oral azithromycin. Lancet 2003;361:313–4.doi:10.1016/S0140-6736(03)12335-5
      OpenUrlCrossRefPubMedWeb of Science
    8. ↵
      2. Burchell AN ,
      3. Allen VG ,
      4. Moravan V , et al
      . Patterns of syphilis testing in a large cohort of HIV patients in Ontario, Canada, 2000-2009. BMC Infect Dis 2013;13:246.doi:10.1186/1471-2334-13-246
      OpenUrlCrossRefPubMed
    9. ↵
      2. Phipps W ,
      3. Kent CK ,
      4. Kohn R , et al
      . Risk factors for repeat syphilis in men who have sex with men, San Francisco. Sex Transm Dis 2009;36:331–5.doi:10.1097/OLQ.0b013e3181990c85
      OpenUrlCrossRefPubMedWeb of Science
    10. ↵
      2. Katz KA ,
      3. Lee MA ,
      4. Gray T , et al
      . Repeat syphilis among men who have sex with men--San Diego County, 2004-2009. Sex Transm Dis 2011;38:349–52.doi:10.1097/OLQ.0b013e3181fe650b
      OpenUrlPubMed
    11. ↵
      2. Cohen SE ,
      3. Chew Ng RA ,
      4. Katz KA , et al
      . Repeat syphilis among men who have sex with men in California, 2002-2006: implications for syphilis elimination efforts. Am J Public Health 2012;102:e1–8.doi:10.2105/AJPH.2011.300383
      OpenUrlCrossRefPubMed
    12. ↵
      2. Brewer TH ,
      3. Peterman TA ,
      4. Newman DR , et al
      . Reinfections during the Florida syphilis epidemic, 2000-2008. Sex Transm Dis 2011;38:12–17.doi:10.1097/OLQ.0b013e3181e9afc7
      OpenUrlCrossRefPubMed
    13. ↵
      2. Ogilvie GS ,
      3. Taylor DL ,
      4. Moniruzzaman A , et al
      . A population-based study of infectious syphilis rediagnosis in British Columbia, 1995-2005. Clin Infect Dis 2009;48:1554–8.doi:10.1086/598997
      OpenUrlCrossRefPubMedWeb of Science
    14. ↵
      Centers for Disease Control and Prevention (CDC). Notes from the field: repeat syphilis infection and HIV coinfection among men who have sex with men--Baltimore, Maryland, 2010-2011. MMWR Morb Mortal Wkly Rep 2013;62:649–50.
      OpenUrlPubMed
    15. ↵
      Pfizer Canada Inc. Product monograph: bicillin L-A (penicillin G benzathine). http://www.pfizer.ca/sites/g/files/g10028126/f/201511/Bicillin-LA_PM_E.pdf (accessed 17 Feb 2017).
    16. ↵
      2. Garnett GP ,
      3. Aral SO ,
      4. Hoyle DV , et al
      . The natural history of syphilis. implications for the transmission dynamics and control of infection. Sex Transm Dis 1997;24:185–200.
      OpenUrlCrossRefPubMedWeb of Science
    17. ↵
      2. Magnuson HJ ,
      3. Thomas EW ,
      4. Olansky S , et al
      . Inoculation syphilis in human volunteers. Medicine 1956;35:33–82.doi:10.1097/00005792-195602000-00002
      OpenUrlCrossRefPubMed
    18. ↵
      2. Seña AC ,
      3. White BL ,
      4. Sparling PF
      . Novel Treponema pallidum serologic tests: a paradigm shift in syphilis screening for the 21st century. Clin Infect Dis 2010;51:700–8.doi:10.1086/655832
      OpenUrlCrossRefPubMedWeb of Science
    19. ↵
      Public Health Agency of Canada. M-Track: enhanced surveillance of HIV, sexually transmitted and blood-borne infections, and associated risk behaviours among men who have sex with men in Canada. Phase 1 Report 2011.
    20. ↵
      2. Miller DK ,
      3. Homan SM
      . Determining transition probabilities: confusion and suggestions. Med Decis Making 1994;14:52–8.doi:10.1177/0272989X9401400107
      OpenUrlCrossRefPubMedWeb of Science
    21. ↵
      2. Conway JM ,
      3. Tuite AR ,
      4. Fisman DN , et al
      . Vaccination against 2009 pandemic H1N1 in a population dynamical model of Vancouver, Canada: timing is everything. BMC Public Health 2011;11:932.doi:10.1186/1471-2458-11-932
      OpenUrlCrossRefPubMed
    22. ↵
      2. Trubiano JA ,
      3. Hoy JF
      . Taming the great: enhanced syphilis screening in HIV-positive men who have sex with men in a hospital clinic setting. Sex Health 2015;12:176–8.doi:10.1071/SH14164
      OpenUrl
    23. ↵
      2. Bissessor M ,
      3. Fairley CK ,
      4. Leslie D , et al
      . Frequent screening for syphilis as part of HIV monitoring increases the detection of early asymptomatic syphilis among HIV-positive homosexual men. J Acquir Immune Defic Syndr 2010;55:211–6.doi:10.1097/QAI.0b013e3181e583bf
      OpenUrlCrossRefPubMedWeb of Science
    24. ↵
      2. Cohen CE ,
      3. Winston A ,
      4. Asboe D , et al
      . Increasing detection of asymptomatic syphilis in HIV patients. Sex Transm Infect 2005;81:217–9.doi:10.1136/sti.2004.012187
      OpenUrlAbstract/FREE Full Text
    25. ↵
      2. Fisman DN ,
      3. Tang P ,
      4. Hauck T , et al
      . Pertussis resurgence in Toronto, Canada: a population-based study including test-incidence feedback modeling. BMC Public Health 2011;11:694.doi:10.1186/1471-2458-11-694
      OpenUrlCrossRefPubMed
    26. ↵
      2. Shaw SY ,
      3. Ross C ,
      4. Nowicki DL , et al
      . Infectious syphilis in women: what’s old is new again? Int J STD AIDS 2017;28:77–87.doi:10.1177/0956462415627397
      OpenUrl
    27. ↵
      2. Weinhardt LS ,
      3. Forsyth AD ,
      4. Carey MP , et al
      . Reliability and validity of self-report measures of HIV-related sexual behavior: progress since 1990 and recommendations for research and practice. Arch Sex Behav 1998;27:155–80.doi:10.1023/A:1018682530519
      OpenUrlCrossRefPubMedWeb of Science
    28. ↵
      2. Coleman TA ,
      3. Bauer GR ,
      4. Pugh D , et al
      . Sexual orientation disclosure in primary care settings by gay, bisexual, and other men who have sex with men in a Canadian city. LGBT Health 2017;4:42–54.doi:10.1089/lgbt.2016.0004
      OpenUrl
    29. ↵
      2. Ng BE ,
      3. Moore D ,
      4. Michelow W , et al
      . Relationship between disclosure of same-sex sexual activity to providers, HIV diagnosis and sexual health services for men who have sex with men in Vancouver, Canada. Can J Public Health 2014;105:e186–91.
      OpenUrl

    Footnotes

    • Handling editor Jackie A Cassell

    • Contributors All authors participated in study conception and design and interpretation of results. AT built the model and performed the analyses. SS, JR and CR assisted with data acquisition. AT and SM drafted the manuscript, and all authors assisted with critical revision of the manuscript.

    • Competing interests None declared.

    • Provenance and peer review Not commissioned; externally peer reviewed.

    • Data sharing statement Model code is available upon request by contacting the corresponding author.