The development of GCO was led by the BCCDC Online Sexual Health Services (OSHS) program, which consisted of a medical lead, program manager, business analyst, and epidemiologist. At the outset, a program of research was established between the OSHS and the Youth Sexual Health Team, a research unit at the University of British Columbia, based on an integrated knowledge translation model (ie, where knowledge translation principles are applied to the entire research process, with involvement of knowledge users as equal partners) [39], and drawing on a range of qualitative and quantitative research disciplines including public health and clinical research, epidemiology, and social sciences.

We identified 3 groups of stakeholders to engage in the development of GCO either on an ad hoc or continuous basis: internal stakeholders within the BCCDC; stakeholders within other PHSA agencies (eg, public health laboratory, privacy, information technology); and external stakeholders (Table 1).

We conducted a stakeholder analysis and developed a communication strategy for engaging stakeholders, including the development of fact sheets and standard presentations. Initially, we met with each stakeholder to provide an orientation to the concept of Web-based STBBI testing and GCO, and to discuss the nature of their involvement with GCO development. We provided ongoing updates about GCO development through a team blog and email bulletins, as well as seeking out opportunities to disseminate information about GCO through community agency networks, publications, and events [40-42]. We then established 3 stakeholder committees to guide the activities of, and provide updates about, GCO development:

1. Clinical Integration Committee (CIC): An internal decision-making group that met monthly to determine how GCO should be integrated with the provincial STI clinic at BCCDC, which would be responsible for clinical and public health follow-up of test results. The CIC was comprised of medical and nursing leads for the clinic, education and outreach programs at BCCDC.

2. Community Consultation Working Group (CCWG): An external advisory group that met biannually to ensure GCO would be a useful resource to community organizations and tailored to meet the needs of the populations they served. The CCWG was comprised of community organizations working in sexual health and STBBI prevention (including with youth and MSM).

3. Internet Services Committee (ISC): An advisory group that met monthly to receive updates about GCO and to guide its development, comprised of BCCDC, PHSA, and external stakeholders.

We reviewed published and gray literature to identify Web-based STBBI testing programs and evaluations of their impact, and to summarize barriers and facilitators of STBBI testing that are potentially mediated through Web-based testing. We then contacted program experts involved with comprehensive Web-based testing programs integrated with clinical services in San Francisco and Amsterdam, to learn about their experience with setting up these programs [14,15].

Based on this information and consultation, we developed a high level model for GCO to use in future consultations. This model described the program’s 3 main steps: create an account, complete a risk assessment, and print a laboratory requisition; provide blood and urine specimens (only; no oral, vaginal, or rectal swabs); receive test results (with paths described if negative, positive, or if there is a problem with the result) (Figure 1). This model did not include mailing of home test kits for human immunodeficiency virus (HIV) as these are not yet licensed for use in Canada. We also did not propose including self-collection of specimens at home as sending collected specimens potentially containing infectious agents by general mail is not permitted in Canada. In this high level model, positive results are delivered by a health care provider by phone, which is consistent with current practice at the provincial STI clinic in order to provide appropriate posttest counseling and ensure treatment and appropriate follow-up. In contrast, all negative test results would be viewed on the Internet, as is the case for many Web-based testing programs in order to reduce barriers to accessing test results (ie, by eliminating requirements to contact the clinic).

Given the importance of end-user input at early stages in the development of novel eHealth interventions, we conducted 3 research studies to ascertain the opinions of youth, MSM, and STI clinic clients about Web-based sexual health services and our Web-based testing model [32,34,43]. Overall, GCO was perceived as private, convenient, and providing greater control over testing. We elicited a number of concerns, including privacy of data and security of the app, reliance on outdated technologies (eg, printing), and anxiety at receiving a positive result on the Internet. In these studies, the participants proposed mitigation strategies for these concerns, many of which were incorporated into the GCO model (Table 2).

As a new testing paradigm, Web-based testing raised a number of questions regarding its relation to existing regulatory and practice requirements for physicians and nurses in BC (eg, can a physician order tests recommended through a Web-based app, and if so how, and is any liability assumed?). To answer these questions, we consulted with provincial practice regulatory bodies for physicians and nurses, and the national medical protection insurance agency. We established that GCO’s model, where a BCCDC STI specialist physician is the ordering physician with results reported to and followed up by provincial STI certified registered nurses and managed according to standard protocols, is an extension of acceptable clinical practice. We also developed policies to address identified clinical nursing practice gaps, including guidelines for providing nursing services on the Internet and for emailing clients.

As GCO is implemented in an existing sexual health care system where Web-based testing clients may also be accessing sexual health services at other clinics not operated by BCCDC, we conducted focus groups with local sexual health care providers to understand their opinions of GCO [44]. Providers perceived GCO as an inevitable evolution within the current system of care, with perceived benefits including shifting the locus of control from providers to patients, addressing testing barriers (eg, privacy concerns, clinic hours of operation), facilitating increased engagement in sexual health care (eg, including reminders for pap testing), and freeing up provider time and ability to see more complex patients. Providers also considered that these benefits may be offset by perpetuating existing inequities in populations GCO is trying to reach (eg, youth who do not have access to a private printer, MSM requiring swabs for diagnosis of oral or rectal STI) or predominantly being used by individuals who already have the resources necessary to access testing (eg, tech “savvy,” higher income). A number of potential personal or clinical harms were identified, such as anxiety at receipt of result notifications, repeated use by individuals at lower risk (“worried well”), misunderstanding test limitations, such as window periods, inadequate pre- and posttest counseling, and missed opportunities for education and prevention such as contraception. As with potential users, mitigation strategies for these harms were also proposed and incorporated into GCO (Table 3). At the same time, providers recognized that many of these harms could also occur within face-to-face clinical testing encounters.

The “techno-optimism” with which Web-based health interventions are viewed is tempered by the reality that their adoption is patterned along social gradients (eg, digital divides) [45,46]. New testing technologies such as GCO may reinforce or reproduce the relationship between social position and health status if only taken up by individuals who already have the resources needed to access clinic-based STBBI testing (eg, social capital, education) [47]. We conducted a preliminary health equity impact assessment (HEIA; screening, scoping) consisting of a literature review and expert consultations. In so doing, we identified ways in which GCO was likely to reinforce or circumvent health inequities in sexual health for historically underserved and marginalized populations with a higher burden of STBBI (including: youth; MSM; people from ethnocultural minorities; intersex, transgender or gender variant populations; Indigenous people; residents of rural areas) [48]. HEIA recommendations that were incorporated into GCO design included: collecting information on ethnicity and gender identity; avoiding normative and stigmatizing language and images; expanding testing options to include hepatitis C.

As flagged during consultations with end-users and sexual health care providers, we needed to determine how provider-delivered pretest counseling and the obtaining of informed consent could be translated to a Web-based app. We did so by: (1) reviewing the published literature to determine the effectiveness of alternate models of providing pretest counseling (eg, videos, written information), (2) reviewing national and provincial testing policies and procedures, and (3) consulting with a clinical ethicist, privacy advisor, legal counsel, and provincial nursing practice leads to determine the necessary app requirements in order to obtain informed consent on the Internet (eg, mandatory vs optional steps). We included a specific consent step on GCO that must be completed by users prior to printing a laboratory test requisition, and conducted individual interviews with end-users following their participation in usability testing of the GCO app to probe specifically about their perceptions of the consent webpage (which overall were favorable) [49].

As GCO was conceptualized as an extension of clinical STBBI testing services offered by BCCDC, we intended the assessment step and testing recommendations to mirror routine clinical practice as much as possible. We reviewed national and provincial guidelines for STBBI testing, treatment and scope of practice, and the BC epidemiology of STBBI in our target populations. To identify assessment questions that could be used for recommending specific tests and tailored educational messages, we reviewed existing Web-based risk assessment tools and the published literature to identify models or variables that were predictive of STI infection [50,51].

Our final set of assessment questions, testing recommendations, and tailored messages were evaluated through usability testing and consultation with clinical and community stakeholders. These included routine recommendation for all clients of urine testing for chlamydia and gonorrhea, and serum testing for HIV and syphilis (as these are generally routinely recommended for BCCDC STI clinic clients and MSM); serum testing for hepatitis C was also recommended for MSM (optional) and individuals with a history of injection drug use. Clients can opt out (deselect) any of the recommended tests. Assessment questions were also designed to (1) provide tailored recommendations for additional testing or prevention interventions that may be indicated (including HIV postexposure prophylaxis, emergency contraception, and need for oral and/or rectal swabs for chlamydia and gonorrhea testing), and (2) to determine the recommended frequency of testing in order to set up testing reminders that are sent by email to GCO clients (ie, 3, 6, or 12 months).

As the incorporation of user feedback is a key component of app development, we conducted testing of the website with potential end-users at 3 separate points during the development process. All usability testing was done in-person with participants recruited through Web-based advertising and from attendees of the Provincial STI Clinic at BCCDC. The first round of testing, performed on a prototype with 10 participants, was designed to observe how users interacted with the app and to validate the risk assessment; outcomes informed numerous changes to the app user interface. The second round of testing, with 8 participants, was performed on a functional but incomplete version of the website to test the overall functioning and content of the app; results led to a redesign of the homepage and changes to the website information architecture. The final round of usability testing, with 14 participants, was completed on a fully functional version of the website to test different options for presenting information via the homepage and the informed consent page; feedback informed the final homepage design and validated the informed consent process.

The final model and flow diagram for GCO [52] is shown in Figure 3. The app screenshots and a video walk-through are included in Multimedia Appendices 1 and 2, respectively.

In brief, potential users (clients) of GCO can create an account if they have an access code (included with promotional materials, or provided by the BCCDC STI clinic or a local sexual health care provider) or have provided an email address, which is then used by BCCDC staff to send an invitation to create an account. During the account creation process, the client provides an email address to serve as their login (which is subsequently verified) and chooses a password. Both mandatory (name, date of birth, sex; not verified) and optional information (phone number, first 3 digits of postal code, ethnicity) are collected. Clients are advised to use their real name or initials although without verification pseudonyms are possible, a model consistent with low-threshold clinic-based testing services offered through BCCDC in order to reduce barriers for clients with privacy concerns. Clients are asked to provide consent to be contacted for research purposes, and must indicate their understanding of the terms of use and privacy policy. While there is no minimum age requirement for use of GCO, individuals less than 19 years are defined as children in BC legislation and can consent to their own medical care if capable. All clients indicating an age less than 19 years of age are recommended to seek testing at an STI clinic in order that capacity to provide consent can be assessed. However, as the majority of clients using GCO less than 19 years of age are expected to be capable of providing consent they are not barred from proceeding.

Clients then complete the first part of the assessment, where the questions are used to identify clients who have symptoms or are a contact to someone with an STBBI. These clients are subsequently recommended not to test through GCO but to seek standard clinical STBBI testing as other tests or immediate treatment may be indicated; clients do have the option to acknowledge the recommendation and continue with Web-based testing. The second part of the assessment includes questions used to recommend hepatitis C testing and provide tailored educational messages. Next, the GCO app provides test recommendations and educational messages as described above. Clients can deselect tests if desired, proceed to the consent page to indicate their understanding of key pretest counseling messages, and then print their laboratory test requisition (which is saved for later printing if needed).

Clients present their laboratory test requisitions, on which their name has been replaced with a unique GCO client code, at a private laboratory specimen collection site and provide urine and/or blood specimens. Specimens are shipped to the PHSA Public Health Laboratory for testing, with results reported to the BCCDC STI clinic for appropriate management. All results are entered into the GCO app by clerical staff who then trigger the app to send a notification email to the client that their results are ready to view. If all results are negative, the client can see their results in their GCO account and there is no interaction with clinic staff. If any of the results are invalid (for example, a problem occurred with the specimen), clients will view any negative test results and see a notification to call the BCCDC clinic to arrange for retesting for the invalid result. If any of the results are positive, the client can see only a message directing them to call the BCCDC clinic for their results; at the same time, if the client has provided a phone number a BCCDC clinic nurse will attempt to contact the client directly. For each client, a testing history is maintained on the app with dates of testing and test types; test results are not retained within GCO after 1 month as a privacy and security precaution. Testing reminders are set for all clients at 3 or 12 months based on the degree of sexual risk reported on the assessment questions; clients can opt out or change these settings.

Preparing for implementation of the GCO service involved numerous activities related to privacy and security, IT support, operational protocols, reporting, communications, and final validation. We collaborated closely with PHSA Privacy, Risk Management, and IT Security teams to conduct a thorough privacy impact assessment and security threat and risk assessment that included testing of the app’s security model by an external vendor (penetration testing); recommendations from these assessments were incorporated into the final app. A privacy policy and terms of use specific to the program were developed, along with administrative policies around access control, auditing and data reporting, and all were vetted with PHSA Privacy and IT Security. A new software vendor was contracted to provide ongoing app support in partnership with PHSA IT, and extensive documentation was completed to establish internal IT support around client services, network, hardware, and software maintenance. Clinical procedures for the management of GCO clients were developed in consultation with clinical and clerical team leads, and multiple training sessions were conducted to familiarize staff with procedures and the app itself. A series of reports were developed to routinely monitor testing volumes, uptake of the service and drop-off of clients at different points in the testing process. Promotional materials (ie, sign-up sheets, posters, wallet cards, and brochures) were designed and a communications package was created, including email templates for the launch announcement, sample content for social media, a one-page overview of the program, and frequently asked questions. The last step prior to the official launch of the service was a comprehensive final validation (beta-testing) of the app, specimen collection and transport, specimen testing, and clinical procedures using actual clients; where possible, we used feedback from the testing to improve the overall user experience.

The pilot phase of GCO is fully funded by the PHSA, with costs shared between the BCCDC (costs of program operation, app revision, specimen collection) and the BC Public Health Laboratory (BCPHL) (costs of laboratory testing).

We made 2 major changes to the GCO model following implementation, reflecting further refinement based on the findings of the planning and development phases:

In the fall of 2015, we revised the risk assessment questions in tandem with an update of our preliminary HEIA in order to improve their appropriateness for clients of diverse gender identities. We also made revisions to prospectively collect the necessary variables to validate clinical prediction rules (CPR) developed by our team for urine testing for chlamydia and gonorrhea, and blood testing for HIV. These CPR were developed using test results from over 30,000 asymptomatic clients at STI clinics across BC [55,56]. Once these are further validated using prospective GCO data, we aim to include these CPRs prior to promoting GCO outside of our target populations, where our current model for recommending tests may not be appropriate.

In February 2016, we added self-collected rectal and throat swabs for chlamydia and gonorrhea testing. Swabs are recommended for MSM clients if reporting giving oral or receiving anal sex during the assessment step, women if reporting receptive anal sex (as per routine clinical practice). While always intended for future versions of GCO, we accelerated the inclusion of swabs as their absence was flagged as a clinical risk during provider consultations and a potential exacerbation of health inequities for MSM. Clients are given self-collection kits at the private laboratory collection sites when providing urine and/or blood samples, and clients can either self-swab on-site or take home for self-collection and return. Self-collection instruction guides (Multimedia Appendix 3) were developed through reviewing examples found on the Internet with 2 focus groups of potential users, and pilot tested with 12 users who found the guides easy to understand, sensitive to various genders and sexual identities, and conducive to successful self-collection [57].