At-home specimen self-collection as an additional testing strategy for chlamydia and gonorrhoea: a systematic literature review and meta-analysis

Description of the infections

Chlamydia and gonorrhoea are the leading reported bacterial sexually transmitted infections (STIs) worldwide with an estimated 211 million combined cases in 2020, according to the WHO.1 Many of these infections are asymptomatic, making screening crucial for detection, treatment and prevention of sequelae in both individuals and their partners. Over the past decade, reported cases of chlamydia and gonorrhoea have risen in many countries with established screening programmes.2–6 However, these reports likely underestimate the true number of infected individuals.

Several countries including the USA,7 Canada8 and England9 have annual screening recommendations for chlamydia and/or gonorrhoea for sexually active populations at higher-risk for acquiring STIs. These populations include, but are not limited to, young women and adolescents (under 25), pregnant individuals and men who have sex with men (MSM). Despite these recommendations, less than 62% of women under 25 in the USA10 and less than 21% in England11 participate in screening. This is particularly concerning as untreated Chlamydia trachomatis (Ct) and Neisseria gonorrhoeae (Ng) infections can lead to serious health complications such as pelvic inflammatory disease (PID), infertility and ectopic pregnancy.12 13 Annual Ct and Ng screening at sites of exposure is recommended for MSM by the CDC and other agencies, with more frequent testing recommended for MSM on HIV pre-exposure prophylaxis (PrEP).14 In 2017, only 42% of MSM reported being screened for Ct and Ng and only 16% reported extragenital testing in the past 12 months.15 Recent studies suggest that STI screening in MSM reduces chlamydia incidence by 15%,16 and increased screening for MSM taking PrEP would result in a 17% decrease in asymptomatic STIs. Therefore, strategies to boost screening, especially among young women and MSM, should be fully explored.

Description of the intervention

Nucleic acid amplification tests (NAATs) are recommended for the screening and diagnosis of both Ct and Ng due to their high sensitivity and specificity.14 17 18 Variations of NAATs include ligase chain reaction, transcription-mediated amplification, PCR and strand displacement amplification. NAATs can be used with a variety of specimen types. However, self-collected vaginal swabs from females and urine from males are the most sensitive for detecting urogenital Ct and Ng infections.17 19 If extragenital testing is needed, NAATs approved for use with extragenital specimens are also highly sensitive and specific.11 To increase screening, self-collection of specimens should be considered. Patients report that self-collection of vaginal swabs and rectal swabs in a clinical setting for Ct and Ng is preferred over healthcare provider (HCP)-collection because it offers increased patient autonomy, confidentiality and convenience.20–22 In addition to the increased acceptability, self-collection of vaginal specimens and extragenital specimens can be more accurate than HCP-collected specimens.23–28 As such, both the WHO and CDC recommend self-collection of vaginal swabs, first-void urine, pharyngeal, rectal and urethral swabs be available for Ct and Ng testing.17 29

Specimen self-collection at home or in other non-clinical settings refers to the process of collecting a sample, such as a vaginal swab or urine, at home or anywhere outside a healthcare facility without direct clinical supervision. The specimens are then mailed to a testing laboratory or dropped off at a healthcare facility or other drop-off location. This method can reduce logistical barriers like transportation issues and limited clinic availability and is useful for those who want private or more frequent testing, have had negative experiences with HCPs, or prefer greater control over their healthcare. During the COVID-19 pandemic, demand for self-testing and self-collection kits grew.30–33 When England’s National Chlamydia Screening Programme began offering internet-based testing in 2015, only ~5% of tests were conducted online; by 2022, 43% were.11 Internet-based testing programmes vary; some kits can be requested by a patients or providers, with some services offering direct treatment following a positive test result, while others link patients to local HCPs for follow-up testing and treatment.30–34 The cost of the self-collection kits also varies widely, depending on whether the service is privately-funded or government-funded.30–33

Prior systematic literature reviews evaluated at-home specimen self-collection compared with collection (either self or clinician) in a healthcare facility for Ct/Ng.35 36 In 2015, Fajardo-Bernal and colleagues concluded there was no significant difference between home self-collection and collection in a healthcare facility for a number of testers and positive tests,36 and Odesanmi et al suggested that home self-collection increased testing in women.35 Since these reviews were conducted, advances include more sensitive NAATs, new near point-of-care tests for Ct and Ng, new evidence on the acceptability of home self-collection, and increasing demand for autonomy and acceptance of self-collection. Therefore, we performed a current systematic literature review and meta-analysis to assess whether specimen self-collection at home or in other non-clinical settings for Ct and Ng testing increases testing uptake, Ct or Ng diagnosis, linkage to care and has high concordance compared with collection in a healthcare facility.

Research question and inclusion criteria

We aimed to address whether specimen self-collection at home for Ct and Ng testing should be offered as an additional approach to specimen collection in clinical settings.

Population

All sexually active individuals who were recruited to participate in the study by any method. We did not place any age restrictions for the studies included in our analysis, but some studies had specific age restrictions as indicated in table 1.

Intervention

The intervention was unsupervised, specimen self-collection outside a healthcare facility. Collection could be anywhere outside a traditional healthcare facility that would provide medical supervision, including inside the home. We refer to this mode of testing as specimen self-collection at home throughout the rest of the manuscript for increased readability and ease.

Comparison

The comparison was specimen collection within a healthcare facility—either self-collected or HCP-collected for Ct and/or Ng testing or screening.

Outcomes

Search strategy and screening process

To identify relevant literature in an unbiased approach for our systematic literature review and meta-analysis, we requested a literature search from the Stephen B. Thacker CDC library. We initiated the request on 10 April 2023, by providing rationale for the study, inclusion/exclusion criteria, primary/secondary questions and any seed articles we identified on our own. The results of the request were returned on 19 April 2023. Databases included Medline (OVID), Embase (OVID), Global Health (OVID), Cochrane Library, CINAHL (EBSCOHost), Scopus and ClinicalTrials. Search terms included, but were not limited to home collection, self-collection, chlamydia, gonorrhoea and mail-in testing (online supplemental file 1).

To assist with review management and record keeping, we used the review management tool, Covidence. Two authors, PGT and ACS, independently assessed eligibility via an initial title and abstract screen followed by a full-text review. The screeners discussed any disagreements and consulted EK, when needed, for resolution.

Data extraction and management

The authors developed and used a standardised data extraction form to extract data from all the included studies. To ensure the accuracy of extracted data, either PT or ACS extracted data while the other checked the information for accuracy.

Assessing the risk of bias

To assess the risk of bias (RoB) in included RCTs, the authors used the Cochrane Risk of Bias (RoB2) tool. The RoB2 tool evaluates bias in five domains: randomisation process, deviations from intended interventions, missing outcome data, measurement of the outcome and selection of the reported result. The overall RoB for each study was determined based on the judgements made for each of the five domains. Each domain is assessed for RoB as either low, some concerns or high. The authors followed the guidelines provided by the Cochrane Handbook for Systematic Reviews of Interventions to assess the RoB in the included studies. Non-randomised cross-sectional studies were evaluated using the ROBINS-I tool which evaluates bias across seven domains. PT and ACS independently assessed the RoB in each study and any discrepancies were resolved through discussion. The results of the risk assessment were plotted in R Studio using the robvis (V.0.3.0.9) package.37

Data analysis

To evaluate testing uptake and linkage to care at home in this meta-analysis, we calculated and reported pooled prevalence ratios (PRs) with 95% confidence intervals (CIs) using a random effects model using the package metafor38 in R Studio. To evaluate prevalence of Ct or Ng at home and in clinical settings we calculated and reported PRs with 95% CIs on an intent-to-treat basis. Heterogeneity among the studies was assessed using the I² statistic. Subgroup analyses based on sex and specimen type were conducted to explore the sources of heterogeneity. For cross-sectional studies that compared the test results between specimen self-collection at home and collection in a clinical setting from the same individual, we did not pool the results. However, we calculated the concordance (per cent agreement) for each study and described the results in a narrative format. Harms and issues associated with self-collection at home were described in a narrative format, and no pooled effect estimate was calculated. Finally, publication bias was assessed visually using a funnel plot and statistically using Egger’s regression test.

Quality of the evidence

We used the Grading of Recommendations, Assessment, Development and Evaluation (GRADE) approach to assess the quality of evidence based on the RoB, indirectness, inconsistency, imprecision and publication bias.

Patient and public involvement

Patients and public were not directly involved in this review; we used publicly available data for the analysis.

Included studies

The present literature review and meta-analysis encompass a comprehensive analysis of 19 studies, comprising 4 concordance studies and 15 RCTs from 1998 to 2024. One of the included published reports contributed data from two separate RCTs (table 1). Among the RCTs, 13 incorporated female participants,39–50 and 8 included male participants39–42 49–52; 2 studies specifically focused on MSM.41 50 The cumulative population size under investigation across all included RCTs was 62 369 individuals. These studies were conducted across various global regions, with five RCTs originating from the USA,47 48 52 52 three from Denmark42 43 51 and one each from the UK,41 Australia,50 France,40 the Netherlands,39 Norway,49 South Africa44 and Brazil.46 Within the RCTs, three studies primarily evaluated the retesting of individuals previously diagnosed with the infection.39 45 50 Two studies examined testing uptake in partners of index patients who had tested positive for either Ct or Ng.43 51 All but one of the RCTs focused on Ct testing,39–44 46–52 while five studies specifically investigated Ng testing.41 44 46 47 52 Furthermore, RCTs assessed the proportion of individuals who tested positive and subsequently received appropriate treatment.41 44 46 49

From the four cross-sectional concordance studies, two were conducted in Canada,53 54 and one each from the USA55 and Belgium.56 All these studies had the same participants collect the same type of specimen at home and in a clinical setting. These studies had a total of 906 participants; three studies exclusively included MSM54–56 and one exclusively included women.53 All studies evaluated concordance for both Ct and Ng infections, but each evaluated different specimen types.

Uptake of STI testing at home and in clinical settings

Fifteen RCTs measured the uptake of STI testing for those randomised to specimen self-collection at home in comparison to specimen collection in clinical settings.39–52 Overall, we found that significantly more men and women collected and returned specimens for Ct and/or Ng testing at home compared with those randomised to clinical settings (PR=2.61 (95% CI 1.81–3.77), I²=99%, figure 2A). Due to the high heterogeneity, we also investigated subgroup differences between men and women. There was no significant difference between men and women when evaluating uptake of testing (p=0.26); indicating this was not the cause of the high heterogeneity observed among the studies. However, men were over 3.5 times as likely to get tested for Ct or Ng at home compared with clinical settings (PR=3.68 (95% CI 1.64–8.27), I²=99%) while women were only over twice as likely to get tested for Ct or Ng (PR=2.18 (95% CI 1.46–3.26), I²=99%). Despite the overall high heterogeneity measured by the I² statistic, all but two studies demonstrated significantly more Ct/Ng testing for those randomised to specimen self-collection at home. For all studies that evaluated uptake of testing, we evaluated the RoB using the Cochrane RoB2 criteria. Most of the studies had a low RoB in all categories, while a few studies had some serious concerns relating to the randomisation process and adherence to the intervention (figure 2B).

Request permission

Proportion of positive Ct and Ng infections detected

14 RCTs reported the number of positive Ct tests.39–44 46–52 Meta-analysis of these RCTs suggested a significantly greater proportion of positive Ct tests in the individuals who collected specimens at home compared with the specimens collected in clinical settings with an overall PR of 1.61 (95% CI 1.10–2.35) (online supplemental figure 1A). Only five of the included RCTs provided data regarding the number of positive Ng tests detected.41 44 46 47 52 Moreover, there was a relatively limited number of infections detected overall, with 28 out of 2127 testing positive in the intervention group and 24 out of 2138 testing positive in the control group. The analysis of Ng tests did not reveal a significant difference in the proportion of infections detected between the intervention (at home) and the control (clinical settings) groups (PR=1.16 (95% CI 0.68–1.97), online supplemental figure 1B). We also evaluated the RoB for the proportion of positive Ct or Ng tests. Most studies demonstrated low RoB across all outcomes, but a few studies had some concerns and serious concerns for bias in the randomisation process, including deviations from the intended intervention and bias in the measurement of the outcome.

Linkage to care

Of the included RCTs, four provided information about treatment following positive Ct or Ng test results.41 44 46 49 Within those four studies, we found no significant difference in the proportion of treated cases between the intervention and control groups (PR=0.96 (95% CI 0.91–1.01), figure 3A), with 88.6% (195/220) of the intervention group treated compared with 93.9% (232/247) treated in the comparison group. The quality of evidence was high, suggesting that the use of specimen self-collection at home for STI testing provides appropriate linkage to treatment.

Request permission

Test concordance between self-collection in clinical settings and home collection

Four studies evaluated the test concordance (per cent agreement) for specimens collected at home compared with specimens collected in clinical settings by the same individual.53 55 56 Due to the small number of studies, and lack of overlap between anatomic sites, we did not pool any of the following results. In one study, conducted in Canada with 198 women aged 16–25, the Ct and Ng result concordance between the specimens collected at home and clinical settings was assessed for vaginal swabs, urine and these two specimens pooled together.53 Women who had not previously self-collected vaginal swabs were assigned to home collection first and were given a kit with two swabs and two tubes for vaginal self-collection and a urine collection jar.53 The order of collection of vaginal swabs and urine was randomised in both settings. Sampling in both settings was performed on the same day or within the week. Samples were transferred to the laboratory and tested within 24 hours for Ct and Ng using the Aptima Combo 2 assay on a Panther Instrument (Hologic, Inc). Specimen self-collection in the clinic and self-collection at home resulted in high test result concordance for all specimen types assessed. For Ct, concordance was 96.7% for vaginal swabs, 96.5% for urine and 96.3% for pooled specimens (table 2). The highest prevalence was observed using vaginal swabs for both Ct (15.2% home, 14.6% clinic) and Ng (6.6% home, 6.1% clinic). The lowest prevalence was observed using urine for both Ct (11.1% home, 11.6% clinic) and Ng (5.6% home and clinic). For Ct, one positive case was missed for at-home sampling for both urine and pooled specimens compared with clinical self-collection. One additional positive Ct case was detected using vaginal swabs collected at home compared with vaginal swabs collected in the clinic. For Ng, concordance was 92.3% for vaginal swabs and 100% for urine and pooled specimens (table 2). One additional positive Ng case was detected with vaginal swabs collected at home compared with vaginal swabs collected in the clinic.

Three studies evaluated the concordance between specimens self-collected at home and in clinical settings among MSM.54–56 In a study by De Baetselier et al in Belgium,56 MSM on PrEP collected urine at home using a standardised device and mailed it to a lab for testing using the Abbot CT/NG RT assay, then collected another urine sample at a clinic. Across 471 at-home specimens with matching clinic samples, there was high concordance for Ct (99.1%) and Ng (99.6%). One positive Ct sample was missed for at-home specimen collection, but three additional Ct cases were detected through at-home collection. For Ng, two additional cases were detected in the samples collected at home.

Two studies assessed concordance for extragenital specimens in the USA55 and Canada.54 In the USA study, clinic-based rectal swabs were self-collected, and pharyngeal swabs were collected by research associates and placed in the Xpert Vaginal/Endocervical Specimen Collection Kit. Extragenital specimens collected at home were mailed in within 48 hours of a clinic visit. For symptomatic individuals, they collected both sets of pharyngeal and rectal samples within the clinic to ensure treatment did not impact results. Specimens collected at home had a high concordance with those collected at the clinic (Ct: 95.0% rectal, 99.3% pharyngeal; Ng: 95.7% rectal, 97.2% pharyngeal). However, home collection missed 28% (N=9/32) of Ng infections and 37% (7/19) of Ct infections. In the Canadian study, MSM 16 years or older self-collected pharyngeal and rectal specimens at home and brought the specimens into the clinic where they were swabbed again, either by an HCP or self, depending on patient preference. Both specimens collected at home and in the clinic were tested using the Roche Cobas CT/NG assay on the Cobas 8800 system. Positive Ng results were confirmed using the PivNG assay V2 on the Roche Cobas omni utility channel. Concordance was high between the two settings (Ct: 96.0% rectal, 99.0% pharyngeal; Ng: 100% rectal, 99.7% pharyngeal). Home collection identified six Ct infections that were missed by specimens collected in the clinic. Of the results that were positive from clinic collection but not from the specimens collected at home, a majority had cycle threshold values higher than 35, indicating a low amount of bacterial DNA present in the specimen. Together, these results suggest that collection at home and in clinical settings are highly concordant, but some cases are potentially missed with collection in either setting.

Harms associated with at-home specimen self-collection

Only three of the included RCTs investigated potential harms associated with specimen self-collection at home versus collection in clinical settings.44 47 57 More people in the home specimen collection group reported some level of difficulty understanding the instructions for the kit compared with those in a clinical setting group.47 Importantly, 2.6%–17% of individuals randomised to the home self-collection group reported feeling pain or discomfort during specimen self-collection, compared with 12.3% of the clinical self-collection group.44 57 Finally, in one of our included studies55 there was a slightly higher proportion of specimens collected at home with invalid results (1.7%) compared with those collected in clinical settings (0.5%).

Summary

In this comprehensive literature review, we identified 19 studies, encompassing 15 RCTs involving a substantial cohort of 62 369 participants, alongside four cross-sectional studies. Our primary objectives were to evaluate if specimen self-collection at home and other non-clinical settings should be offered as an additional approach to STI testing based on testing uptake, prevalence of Ct or Ng, linkage to care, and the per cent agreement with collection in clinical settings. Offering specimen self-collection at home resulted in a significantly higher number of Ct and Ng tests, particularly among men which may be due in part to men in general are less likely to seek healthcare than women.58 A higher proportion of positive Ct cases was observed in specimens collected at home, likely due to asymptomatic individuals finding testing more convenient when provided outside clinics. There was no significant difference in the proportion of positive Ng cases between the intervention and control which might be due to the more pronounced symptoms in Ng infections compared with Ct infections59 and lower overall prevalence of Ng. Linkage to care was similar between specimens collected at home and in clinical settings. The four concordance studies showed high agreement in Ct and Ng results between home and clinical settings for self-collected vaginal swabs, urine and extragenital specimens.

Consistency

Our results for testing uptake at home are consistent with previously published meta-analyses.24 35 36 All three meta-analyses reported increased uptake in Ct/Ng for those assigned to self-collect at home. We identified and included three additional studies40 41 50 that reported on index patient uptake and one additional study that evaluated retesting behaviours.39 One previous meta-analysis demonstrated there was an increased likelihood of a positive Ct test at home across four studies,24 which we also observed here (online supplemental figure 1). However, previous reviews limited the analysis to individuals who tested rather than all people who were randomised.24 36 This type of analysis resulted in a significant decrease in positive STI tests at home compared with those who had specimens collected in clinical settings. This should be interpreted with caution as symptomatic people, who have the highest retest probability, may be more likely to choose to test in the clinical setting and expect to receive immediate treatment.45 50 60

Strengths

The inclusion of both RCTs and cross-sectional studies broadened the scope of our investigation, allowing for a more holistic assessment of self-collection for STI testing at home. This diverse selection of study designs enhances the validity and applicability of our findings and is balanced by our strict inclusion criteria and use of only peer-reviewed studies. The rigorous methodology employed for article selection, guided by a trained librarian, ensured the systematic identification of relevant studies, minimising the RoB in our dataset. Finally, the deliberate inclusion of studies from various socioeconomic contexts and encompassing populations with varied risk profiles, enhances the generalisability of our results and makes them relevant to a broader spectrum of healthcare settings and individuals. This diversity strengthens the depth and breadth of our discussion and the utility of our findings in informing public health strategies.

Limitations

Some limitations exist. Data on the use of self-collected extragenital (rectal and pharyngeal) specimens at home for both men and women are scant. Only one RCT provided rectal swabs along with urine for MSM (36), and two cross-sectional studies evaluated the concordance of rectal and throat swabs in MSM collected at home compared with these same specimen types collected in the clinic.54 55 In the US study, the missed cases were likely due to the difference in transportation method as specimens collected at home used dry swabs that were reconstituted at the lab, while specimens collected in the clinic were transported in a liquid transport reagent. This highlights an ongoing need to understand how specimen transportation impacts the specimen stability and test sensitivity. Despite the limited data included in our literature review, self-collection of extragenital specimens is acceptable to patients and often preferred.20 61 The often asymptomatic nature of rectal and pharyngeal infection means screening is an important pathway to interrupt disease transmission that cannot be ignored.62–64 However, more research is needed to understand the feasibility and accuracy of extragenital specimen collection at home. Additionally, a key gap is a lack of studies conducted during or after the COVID-19 pandemic, which has undoubtedly increased the acceptability of many telehealth models involving self-testing.65 66 It will be important to understand how the pandemic influenced acceptability and feasibility of self-collection at home for STI testing. Finally, information regarding uptake and positivity for transgender individuals is severely lacking, and this key population needs to be included in future research to appropriately provide guidance.

Our comprehensive review and meta-analysis present a nuanced but encouraging perspective on the use of self-collection at home for Ct and Ng testing. According to GRADE, this review provides modest evidence that offering specimen self-collection for Ct and Ng testing at home leads to increased uptake among both men and women across diverse settings, without loss to follow-up or treatment. While internet-based testing is a strategy already used in some countries such as the UK32 and Australia,33 it is a novel strategy in Canada54 and the USA (FDA Grants Marketing Authorisation of First Test for Chlamydia and Gonorrhoea with at-home Sample Collection | FDA). Specimen self-collection at home could provide a valuable strategy to increase partner testing and retesting as a way to reach those who need it most. This approach may increase health equity by addressing critical barriers to screening, allowing broader uptake and enabling the identification and treatment of more asymptomatic individuals who might unknowingly contribute to disease transmission. Notably, our findings highlight that high concordance with specimens collected in clinical settings can be achieved. As we look ahead, future research should expand to include transgender individuals, assess linkage to care and fully explore the viability of collecting extragenital specimens at home.

While specimen self-collection at home for STI testing has high acceptability and reliability, several concerns need to be considered. Specimen self-collection at-home kits need to be appropriately priced to ensure accessibility. Currently available collection kits in the USA for Ct and Ng can be hundreds of US dollars,31 which is prohibitive for many. The need for a mailing address and internet access may also be limiting to some. Additionally, for those living with others, privacy may be a concern. Allowing patients to pick up and drop off specimen self-collection kits in healthcare settings could help address some of these key issues. Working within a healthcare setting would also ensure the proper tests are ordered and linkage to care and communication of results is done in an appropriate and timely manner. Community pharmacies are well-positioned for this need as they are typically open longer hours and are often conveniently located. Finally, the Faculty of Sexual and Reproductive Healthcare and the British Association for Sexual Health and HIV in the UK have developed standards for web-based sexual and reproductive health services.18 These standards guide providers and help users understand what to expect from online service providers and the development of similar standards should be considered for implementation elsewhere.