It has been almost 15 years since the US Centers for Disease Control and Prevention (CDC) synthesized data from four existing randomized controlled trials (RCTs) of expedited partner therapy (EPT),1–4 and endorsed the practice for heterosexuals with Chlamydia trachomatis (chlamydia) or Neisseria gonorrhea (gonorrhea).5 When using EPT, health care providers give a patient under their care (the “index patient”) treatment to deliver to their sex partners. EPT is an important innovation in the field of sexually transmitted infection (STI) control because it allows providers to prescribe or dispense medication for their patients' partners without an interval evaluation, thereby facilitating the treatment of sex partners who might not otherwise seek care and treatment. The CDC has continued to recommend EPT in subsequent national sexually transmitted diseases treatment guidelines.6,7
By mid-year 2019, EPT had been made permissible in most US states, either through legislation, or professional board decision, and implemented to varying degrees.8,9 Accordingly, investigators studying EPT now focus on implementation and effectiveness.10,11 This continued examination is important because an intervention with demonstrated efficacy may not perform as well, or in exactly the same way, when implemented in real-world settings and in populations different from those studied in efficacy trials. Recognition of these facts has undergirded the growing and valuable discipline of implementation science.12
Initially, we set out to perform a systematic review of the growing body of literature examining the real-world effectiveness of EPT for chlamydia and gonorrhea infection, using recurrent infection in treated index patients as the outcome of interest. We selected this outcome because it was the chief outcome of interest in all three of the RCTs that formed the basis for CDC's endorsement for EPT use with heterosexuals with chlamydia or gonorrhea infection.1–3 Through a systematic search, we identified 6 studies examining the impact of EPT on recurrent infection after treatment; however, the study designs were so heterogeneous that a systematic review was not feasible.11,13–17 In addition, most of the studies had significant methodologic limitations and, perhaps as a result of these limitations, found no evidence that EPT reduced recurrent infection in treated index patients.11,13–17 In this commentary we offer our perspective on the challenges of studying EPT in real-world settings, as well as suggestions for how these challenges may be successfully navigated to obtain valid measures of EPT effectiveness.
CHALLENGE 1: KNOWING HOW EPT IS PRACTICED IN THE STUDY SETTING
Two principle modes of EPT have been described. EPT was first studied as patient-delivered partner therapy (PDPT), wherein index patients are given antibiotics in-hand to deliver to their sex partners. Using a second mode, known as “prescription-EPT,” index patients are given prescriptions to deliver directly to their sex partners, or to first fill themselves, and then deliver as medication to their partners. When implemented in the real world, a substantial proportion of EPT is provided in the form of prescription-EPT, rather than as PDPT. In New York City (NYC), a 2012 survey found that 33% of federally qualified health care centers allowed only prescription-EPT,9 and, that among a population-based sample of young women diagnosed with chlamydia who received EPT during 2014 to 2015, 52% received prescription-EPT.18 Investigators in Massachusetts found a similar proportion of EPT was provided in the form of prescription (51%) in 2012.19
Prescription-EPT differs from PDPT in that (a) a prescription necessitates additional steps to obtain partner treatment (eg, the partner or index patient must take a prescription to a pharmacy to get medication), and (b) medication must be paid for (by the person visiting the pharmacy or their insurance plan). Because of these additional steps, differences in the effectiveness of prescription-EPT versus PDPT seem likely, if not certain.18 An analysis using NYC pharmacy data suggested patients may not fill EPT prescriptions, which could result in sex partners remaining untreated.20 Electronic prescribing further complicates prescription-EPT. With electronic prescribing, a prescription is transmitted electronically from the prescriber to a pharmacist. It is an increasingly common prescribing mechanism and is mandated for all prescriptions in some states.21 The index patient must specify the pharmacy to which the EPT prescription should be sent, and the prescriber's or pharmacist's software may not be able to process “no name” EPT prescriptions, which may be common in jurisdictions where this is permissible.22
Because different methods of EPT may result in different effectiveness, it is important to specify the method of EPT to ensure study results can be contextualized and synthesized with others examining the same mode of EPT. However, in our literature review, some analyses inadequately explained the method of EPT under study.17 This is a challenge that should be easy to overcome. We recommend that investigators clearly define the method(s) of EPT assessed and consider the implications of that method for the generalizability of their findings.
CHALLENGE 2: CHOOSING THE BEST OUTCOME TO EXAMINE EPT EFFECTIVENESS
Expedited partner therapy prevents recurrent infection by facilitating treatment of sex partners. However, routine STI tests cannot distinguish between a reinfection (defined as an infection that a treated index patient ‘reacquires’ from the source of their ‘initial’ infection) and a repeat infection (defined as an infection with the same pathogen acquired from a new source). Distinguishing between these outcomes would require a combination of molecular sequencing and contact tracing which was not even done in the main EPT efficacy trials.1–4 These subsequent infections can be broadly referred to as recurrent infections.5
It was possible to establish the efficacy of EPT among heterosexuals without distinguishing between reinfection and repeat infection, but this distinction might be important when studying EPT in other populations. For example, men who have sex with men (MSM), have higher rates of STI than heterosexual men and women, and may acquire a greater proportion of recurrent infections from new sex partners compared with the same measure among heterosexuals.23 Because EPT acts by treating existing sex partners, a potentially higher proportion of repeat infections (vs. reinfections) among MSM could bias an estimate of EPT effectiveness toward the null compared with an analysis of heterosexuals only. Such a bias was likely present in a 2010 analysis of EPT effectiveness in San Francisco STI Clinics, which examined a study population that was 92% male—predominantly self-identified MSM—and found that EPT had no effect on STI recurrent infection.13 Investigators studying EPT effectiveness need to be aware of the limitations of relying on the relatively crude outcome of recurrent infection.
Perhaps recognizing the limitations of relying on recurrent infection to study EPT effectiveness, some studies have utilized alternate or additional behavioral outcomes that could be intervening factors in the causal pathway to recurrent infection, for example, self-report of the number of partners treated, or rates of disclosure to sex partners.5 In addition to the primary outcome of STI reinfection, all 4 EPT RCTs collected information on partner notification.1–4 The Schillinger and Kissinger urethritis trials found that significantly more patients in the PDPT arms talked to their partners about infection than did patients in the control arms.1,3 The Golden trial looked at the number of partners treated and found that patients in the EPT arm were significantly more likely to report that all of their sex partners were treated and significantly less likely to report having sex with an untreated sex partner.2 Two subsequent analyses found that heterosexual men in New Orleans and MSM in Peru who received PDPT were more likely to disclose their STI diagnosis.24,25 It should be noted that the data for all of these analyses came from clinical trials where resources are relatively abundant compared with real-world research. However, gathering such information is possible. An evaluation of PDPT in Seattle reported an increase in the number and percent of patients for whom all partners were treated.26
Another approach to examining EPT effectiveness is to examine the impact of EPT on rates of chlamydia and gonorrhea at regional and state levels. A 2015 step-wedge community-level randomized trial assessed how the provision of free PDPT to clinicians in Washington State affected rates of partner treatment, chlamydia positivity, and gonorrhea incidence.27 The trial found that the intervention increased the percentage of patients who received PDPT from their diagnosing medical providers.27 After adjusting for temporal trends, the intervention was associated with an approximate 10% reduction in chlamydia positivity and gonorrhea incidence at the community level, although this reduction was not statistically significant.27 The results of ecologic analyses have been conflicting. A time series analysis published in 2017 found that 84% of states that enacted laws explicitly allowing for EPT did not show a subsequent reduction in monthly chlamydia and gonorrhea morbidity as reported to the CDC.28 Another analysis that looked at state EPT legislation found that between 2000 and 2013, almost all states saw increases in chlamydia incidence, but that incidence grew significantly faster in states that did not allow EPT compared with states in which EPT was permitted.29 Ecologic analyses must always be interpreted with caution because of unmeasured and uncontrolled confounding.
Our review suggests that there are multiple possible outcomes to consider when studying EPT in the real world including: recurrent infection, number of partners notified, and regional levels of chlamydia or gonorrhea, and that none of these outcomes is perfect as each will present research challenges. There are, however, distinct advantages to choosing recurrent infection as an outcome when studying EPT in the real world. Using the same outcome that was used in the RCTs allows for a comparison of the effect of EPT in trial settings versus the real world, which can help drive hypotheses around why the intervention may work differently. Because of the advantages and relative ease of studying recurrent infection, in the next section, we examine specific challenges that arise when studying this outcome.
CHALLENGE 3: COPING WITH THE BIASES INTRODUCED WHEN USING RECURRENT INFECTION AS THE OUTCOME OF INTEREST
Measuring recurrent STI relies on patients returning for an STI retest. The challenge is that in real-world settings, rates of return are often very low. In a 2012 analysis that assessed the effectiveness of EPT as implemented in Denver Metro health clinics, only 16% (240/1478) of patients returned to clinics for STI retesting,11 and in a 2011 analysis of chlamydia partner services in California family planning clinics, only 30% (92/304) of patients were retested.16 Although there is some evidence that reminders to return to clinic for retesting can increase retesting rates (from 8% to 14% with a postcard reminder,30 and from 6% to 28% with a text message reminder),31 even with reminders, rates of return are low. Low return rates are problematic because they lead to underpowered and possibly biased estimates. Among the 6 studies we found through our systematic search, 5 were likely underpowered.11,14–17 This lack of power also means that many studies of EPT effectiveness cannot control for confounding. Of the 6 studies we identified, only 1 had sufficient power to control for multiple potential confounders of the effect of EPT on recurrent infection using a propensity score.13 To partially overcome the challenge of low retesting, future analyses that rely on recurrent infection to establish EPT effectiveness should undertake to match to the appropriate STI surveillance registry to capture infections diagnosed at locations other than the study site, as was done by Stephens et al.13
Some real-world studies of EPT have utilized long follow-up windows, possibly to decrease the impact of low retesting rates. Most studies begin measuring recurrent infection approximately 30 days after the index patients’ initial STI diagnosis to avoid misclassifying residual DNA after treatment as a new infection.32 However, there is substantial variability in the upper limit of the follow-up period for measuring recurrent infection, from as few as 90 days, to 1 year or more after initial STI diagnosis.11,13–17 The longer the follow-up period, the more likely it becomes that a recurrent infection was acquired from a new, rather than a preexisting untreated sex partner. A population with a high proportion of recurrent infections acquired from new sex partners would likely show lower EPT effectiveness, and utilizing different follow up periods will result in different magnitudes of this bias. To overcome this challenge, future analyses of EPT should decrease the follow up period to minimize the risk of misclassifying repeat infections (from a new sex partner) as reinfections (from a preexisting sex partner).
Even in studies with reasonably high retesting rates, there are often differences between patients who do, and do not return for a retest. In a 2011 study of EPT implemented in a NYC family planning clinic, 40% (189/466) of patients were retested within 3 months and 74% (343/466) were retested within 1 year; however, at both time points, patients who had received EPT were more likely to return for retesting.17 A 2013 study of EPT at Phoenix Indian Medical Center found 66% (324/492) of patients returned for retesting, but women were significantly more likely to be retested than men.14 This gender imbalance could be indicative of differences in screening guidelines. Annual STI screening is recommended by the CDC for all sexually active women younger than 25 years regardless of risk, but routine screening is only recommended for men, and for women over age 25 if sexual history reveals a higher risk of STI.6 If a study follows patients for recurrent infection beyond 1 year, it is likely to capture women returning for an annual screening; in this scenario, outcome data could be less complete for men than for women. To overcome this challenge, researchers should ensure a large enough sample size so that sex-specific differences in retesting rates can be accounted for by adjustment in regression models.
Finally, when seeking to establish effectiveness, it is important to ensure that subjects in comparison groups are at risk for recurrent infection. Index patients who bring their sex partner to the clinic (concurrent treatment) or who report that their sex partners were already treated (previous treatment) at the time of PDPT offer may have a different risk of recurrent STI. Patients with previously or concurrently treated partners were excluded from the EPT efficacy trials,1–4 but all 6 real-world studies identified in the systematic review included such patients.11,13–17 Inclusion of index patients with concurrently or previously treated partners presents a challenge because these patients are also less likely to be offered EPT and less likely to accept EPT when offered.10 Because having a concurrently or previously treated partner has been associated with both a lower risk of recurrent STI and lower EPT acceptance, including these patients in the comparison group in real-world studies can bias a measure of the effect of EPT on recurrent infection toward the null. This bias can affect a substantial proportion of potential subjects; in their 2005 trial, Golden et al. reported that 48% of patients who were interviewed reported a previously treated partner and were subsequently excluded.2 Patients with concurrently or previously treated partners were perhaps not excluded from the real-world studies because of concerns that not all partners had been treated and that the index patient remained at risk for recurrent infection, however even in the setting of a NYC sexual health clinic, patients offered EPT reported a median of only 1 sex partner in the previous 3 months.10 Excluding from effectiveness evaluations those patients who report a concurrently or previously treated partner at the time EPT is offered may result in a more valid measure of effectiveness.
THE WAY FORWARD
Studying EPT implemented in real-world settings is critical to the ongoing improvement of this important STI partner management strategy. Throughout this commentary, we have examined challenges in studying the effectiveness of EPT and have offered suggestions for mitigating various biases. In summary, future analyses examining the effect of EPT on STI reinfection should consider the following: (1) specify the method of EPT under study to ensure that real-world research can be contextualized and synthesized, (2) restrict study populations to patients with a similar risk for reinfection where possible (eg, exclude or adjust for MSM status), (3) match to a surveillance registry to increase power and lessen the potential for bias due to incomplete outcome ascertainment, (4) minimize the length of follow-up after receipt of EPT to lessen the potential for misclassification of repeat infections as reinfections and to reduce differential outcome ascertainment due to differences in screening guidelines for males and females, and (5) exclude patients who report a sex partner who has been treated or is being concurrently treated at the time EPT is offered to eliminate a source of potentially substantial confounding. By utilizing these suggestions, future analyses have the potential to overcome the research challenges that we have identified and provide a much-needed valid estimate of EPT effectiveness.
1. Schillinger JA, Kissinger P, Calvet H, et al. Patient-delivered partner treatment with azithromycin to prevent repeated Chlamydia trachomatis
infection among women: A randomized, controlled trial. Sex Transm Dis 2003; 30:49–56.
2. Golden MR, Whittington WLH, Handsfield HH, et al. Effect of expedited treatment of sex partners on recurrent or persistent gonorrhea or chlamydial infection. N Engl J Med 2005; 352:676–685.
3. Kissinger P, Mohammed H, Richardson-Alston G, et al. Patient-delivered partner treatment for male urethritis: A randomized, controlled trial. Clin Infect Dis 2005; 41:623–629.
4. Kissinger P, Schmidt N, Mohammed H, et al. Patient-delivered partner treatment for Trichomonas vaginalis
infection: A randomized controlled trial. Sex Transm Dis 2006; 33:445–450.
5. Centers for Disease Control and Prevention. Expedited partner therapy in the management of sexually transmitted diseaes: Review and guidance. Atlanta, GA: US Department of Health and Human Services, 2006.
6. Workowski KA, Bolan GA. Sexually Transmitted Diseases Treatment Guidelines, 2015. MMWR. Recommendations and reports: Morbidity and mortality weekly report. Recommendations and reports/Centers for Disease Control 2015; 64.
7. Centers for Disease Control and Prevention. Sexually Transmitted Diseases Treatment Guidelines, 2010. MMWR 2010; 59(RR-12).
9. Introcaso CE, Rogers ME, Abbott SA, et al. Expedited partner therapy in federally qualified health centers—New York City, 2012. Sex Transm Dis 2013; 40:881–885.
10. Vaidya S, Johnson K, Rogers M, et al. Predictors of index patient acceptance of expedited partner therapy for Chlamydia trachomatis
infection and reasons for refusal, sexually transmitted disease clinics, New York City, 2011 to 2012. Sex Transm Dis 2014; 41:690–694.
11. Mickiewicz T, Al-Tayyib A, Thrun M, et al. Implementation and effectiveness of an expedited partner therapy program in an urban clinic. Sex Transm Dis 2012; 39:923–929.
12. Fixsen DL, Blase KA. Implementation: the missing link between research and practice. implementation brief. Number 1. FPG Child Development Institute 2009.
13. Stephens SC, Bernstein KT, Katz MH, et al. The effectiveness of patient-delivered partner therapy and chlamydial and gonococcal reinfection in San Francisco. Sex Transm Dis 2010; 37:525–529.
14. Taylor MM, Reilley B, Yellowman M, et al. Use of expedited partner therapy among chlamydia cases diagnosed at an urban Indian health centre, Arizona. Int J STD AIDS 2013; 24:371–374.
15. Unger JA, Matemo D, Pintye J, et al. Patient-delivered partner treatment for chlamydia, gonorrhea, and trichomonas infection among pregnant and postpartum women in Kenya. Sex Transm Dis 2015; 42:637–642.
16. Yu YY, Frasure-Williams JA, Dunne EF, et al. Chlamydia partner services for females in California Family Planning Clinics. Sex Transm Dis 2011; 38:913–918.
17. Kerns JL, Jones HE, Pressman EJ, et al. Implementation of expedited partner therapy among women with chlamydia infection at an urban family planning clinic. Sex Transm Dis 2011; 38:722–726.
18. Oliver A, Rogers M, Schillinger JA. The impact of prescriptions on sex partner treatment using expedited partner therapy for Chlamydia trachomatis
infection, New York City, 2014–2015. Sex Transm Dis 2016; 43:673–678.
19. Smock L, Barker K, Hsu K. Expedited partner therapy for chlamydia infection is underutilized and underreported, Massachusetts, 2012. National STD Prevention Conference [3B 3] 2014.
20. Okah E, Arya V, Rogers M, et al. Sentinel surveillance for expedited partner therapy prescriptions using pharmacy data, in 2 New York City Neighborhoods, 2015. Sex Transm Dis 2017; 44:104–108.
22. Schillinger J, Slutsker J, Tsang L-Y, et al. Do prescriptions for expedited partner therapy get filled? Findings from a multi-jurisdictional evaluation, US, 2017–2018. Presented at the 23rd International Society for Sexually Transmitted Disease Research [P076]. Vancouver, Canada; 2019.
23. Pathela P, Braunstein SL, Schillinger JA, et al. Men who have sex with men have a 140-fold higher risk for newly diagnosed HIV and syphilis compared with heterosexual men in New York City. J Acquir Immune Defic Syndr 2011; 58:408–416.
24. Mohammed H, Leichliter JS, Schmidt N, et al. Does patient-delivered partner treatment improve disclosure for treatable sexually transmitted diseases? AIDS Patient Care STDS 2010; 24:183–188.
25. Clark JL, Segura ER, Oldenburg CE, et al. Expedited partner therapy (EPT) increases the frequency of partner notification among MSM in Lima, Peru: A pilot randomized controlled trial. BMC Med 2017; 15:94.
26. Golden MR, Hughes JP, Brewer DD, et al. Evaluation of a population-based program of expedited partner therapy for gonorrhea and chlamydial infection. Sex Transm Dis 2007; 34:598–603.
27. Golden MR, Kerani RP, Stenger M, et al. Uptake and population-level impact of expedited partner therapy (EPT) on Chlamydia trachomatis
and Neisseria gonorrhoeae
: The Washington State community-level randomized trial of EPT. PLoS Med 2015; 12:e1001777.
28. Owusu-Edusei K, Cramer R, Chesson H, et al. State-level gonorrhea rates and expedited partner therapy laws: Insights from time series analyses. Public Health 2017; 147:101–108.
29. Mmeje O, Wallett S, Kolenic G, et al. Impact of expedited partner therapy (EPT) implementation on chlamydia incidence in the USA. Sex Transm Infect 2018; 94:545–547.
30. Paneth-Pollak R, Klingler EJ, Blank S, et al. The elephant never forgets; Piloting a chlamydia and gonorrhea retesting reminder postcard in an STD clinic setting. Sex Transm Dis 2010; 37:365–368.
31. Downing SG, Cashman C, McNamee H, et al. Increasing chlamydia test of re-infection rates using SMS reminders and incentives. Sex Transm Infect 2013; 89:16–19.
32. Gaydos CA, Crotchfelt KA, Howell MR, et al. Molecular amplification assays to detect chlamydial infections in urine specimens from high school female students and to monitor the persistence of chlamydial DNA after therapy. J Infect Dis 1998; 177:417–424.