Design

A behaviourally focused cross-sectional exploratory qualitative approach, conducted as part of the SEQUENCE Digital Programme (https://www.sequencedigital.org.uk/).

Applying a behavioural lens

High quality applied behavioural science requires a considered understanding of the specific behaviour(s) that are intended to be changed by an intervention [45, 60]. Within the broad behavioural system of ‘accessing and using online sexual healthcare’ we identified seven distinct yet interconnected behavioural domains (see S1 Fig). Here, we focus on the first of these two domains that we consider to constitute the digital doorway: 1) seeking online sexual health information and 2) seeking online sexual health support.

Participants

Inclusion criteria were: 1) never ordered/used or struggled to order/use an STI self-sampling kit (to recruit participants of lower digital literacy); 2) aged 16+; 3) sexually active; 4) had phone and internet access to enable data collection; 5) lived in the UK; and 6) spoke English well enough to participate in an interview. Further, using PROGRESSPlus (PROGRESS: Place of Residence, Race/Ethnicity, Occupation, Gender/Sex, Religion, Education, Socio-economic Status, Social Network; Plus: e.g., Age, Sexual Orientation, Disability) [61, 62], we developed a sampling framework (see S2 Table) to purposefully recruit participants from a range of underserved populations. In line with Braun and Clarke [63], we did not seek to meet data saturation. Instead, prior to recruitment, a sample of 35 was agreed as appropriate to meet the sample targets and sufficient Information Power [64]. After 35 interviews, we reviewed the data and were satisfied that Information Power has been attained. See Table 1 for the characteristics and demographics of the final sample.

Download:

• PNG
  larger image
• TIFF
  original image

Table 1. Participant self-reported socio-economic demographic characteristics.

https://doi.org/10.1371/journal.pone.0315049.t001

Recruitment

Representatives from five regional National Health Service (NHS) Trusts/Boards (i.e., organisational areas), two non-governmental organisations (NGOs), and one community college across England and Scotland referred interested potential participants to the research team. One NGO served people with disabilities and learning difficulties, the other, people who identify as LGBTQI+ and Muslim. The community college served people with low educational attainment living in a deprived area. We then telephoned each referred participant to verify eligibility according to our inclusion criteria, screen them against the target sampling frame, collect demographics and information on internet use (see S3 Table), and schedule an interview (phone, video, or face-to-face). The sample was monitored closely throughout recruitment by checking the demographics of potential participants against the existing sample and target sampling frame. As sample targets were met, we selectively recruited only people of characteristics and demographics for which targets had not yet been met and liaised with NHS and NGO representatives about targeting relevant potential participants.

Materials

We composed a questionnaire-based assessment of eligibility to take part in the study, demographics (based on the PROGRESSPlus framework [61, 62]), and experience and skills using technology and the internet (S3 Table). In line with inclusive guidance [65, 66], the questions asked regarding socio-economic demographics, such as gender identity and ethnicity, were open questions to capture how participants self-identified.

We also developed an interview topic guide (S4 Table) and supporting visual aids (e.g., S2 Fig) to explore participants’ barriers (e.g., what makes it difficult, what are the drawbacks) and facilitators (e.g., what makes it easy, what are the benefits) to both seeking online sexual health information and support. Example topic guide questions were “What has made it difficult for you to search for sexual health information online?” and “What would make it easy for you to use an email or text service to get sexual health support online?”. Alongside the topic guide questions, we asked follow up questions to participants’ responses, for example, “You said that a barrier to searching for sexual health information online was privacy, can you tell me more about that?”. Barriers and facilitators could be either experiential (i.e., actually experienced) or hypothetical. The visual aids depicted the two behavioural domains of seeking online sexual health information and support. Respectively, these visual aids showed search engines with the typed words “sexually transmitted infection (STI) symptoms” for ‘seeking online sexual health information’ and examples of a sexual health live chat and SMS text exchange with a healthcare provider (HCP) for ‘seeking online sexual health support’.

Participant and patient involvement and engagement

For material development, we consulted public and patient involvement and engagement (PPIE) representatives (N = 12) of diverse ages, genders, ethnicities, sexual orientations, religions, and experiences of disability, learning difficulties, and digital STI healthcare. The representatives offered intersectional perspectives and advice on our questionnaire-based assessment of participant demographics and internet use (n = 5) and the interview topic guide and visual aids (n = 7) to be used within data collection. Consent to share PPIE representatives’ demographic information was not obtained.

Procedure

One-to-one participant-led semi-structured interviews (duration range 38–82 minutes, M = 60 minutes) were conducted remotely (video call n = 7, phone n = 23, or face-to-face n = 5, all by JMcL, between 15.10.21 and 18.03.22). Prior to the interview, we provided participants with an information sheet, consent form, and the visual aids by email or WhatsApp for remote interviews and hard copy for face-to-face interviews. At the beginning of the interview, for remote interviews, participants provided verbal informed consent, recorded (by JMcL) using an encrypted recorder, and for face-to-face interviews, participants provided written informed consent At the end of the interviews, all participants were offered a shopping voucher (value £30) and were provided with a list of sexual health resources (S5 Table) either by email, WhatsApp or hard copy.

Analysis

For RQs one and two, using NVivo (version 12), we conducted inductive thematic analysis at semantic level to identify barrier and facilitator themes for each of the behaviours, following Braun and Clarke’s [63] five steps (see Fig 1). First we familiarised ourselves with the data (step 1) then we systematically described data using brief summary barrier or facilitator statements (step 2). We then grouped similar summary statements to identify barrier and facilitator sub-themes (step 3) and grouped similar sub-themes to generate higher level barrier and facilitator themes, gaining insights into patterns across participant sub-groups (step 4). Finally, we reviewed the themes to ensure they matched the original quotes and finalised their names (step 5). Each of the steps were initially conducted by an experienced health psychology researcher (JMcL) then audited by an expert behavioural scientist (PF). Disagreements on the describing, grouping, and naming of themes were resolved through discussion until consensus was reached. An inter-disciplinary team of clinicians and clinical researchers, human computer interaction specialists, sociology researchers, and public health experts also had oversight of the analysis. Overall, the themes provide a rich description of what was learned from analysing the whole dataset, reflecting all participants. Insights into the sub-population patterning of barriers and facilitators has also been provided; however no formal analysis across participants groups was conducted for this (e.g., approaches such as framework analysis [67]).

Download:

• PNG
  larger image
• TIFF
  original image

Fig 1. The thematic analysis process, following Braun and Clarke’s [63] five steps, using examples from data for seeking online sexual health information.

https://doi.org/10.1371/journal.pone.0315049.g001

In line with Braun and Clark [63], thematic analysis was selected for the analysis, as it offered identification of patterns of important issues to participants, enabling us to determine the key barriers and facilitators to seeking online sexual health information and support across diverse underserved populations. Additionally, inductive analysis (i.e., themes are derived bottom up from the data without trying to fit it into a pre-existing coding frame or the researcher’s analytic preconceptions) was chosen, as the goal of the research was to specify participant-led barriers and facilitators. Neither the topic guide nor the analysis were approached with pre-conceived notions of the data or results. Further, semantic level analysis (i.e., themes are identified from the explicit or meanings of the data, not anything beyond what a participant has said) was selected as this offers an objective description of the data, rather than subjective interpretation of underlying assumptions or deeper meanings within the data.

For RQ three, using the BCW approach [45], barrier and facilitator themes were mapped onto appropriate components of the COM-B Model [68] (by JMcL and checked by PF). The COM-B model posits that behaviour is determined by ‘Capability’ (physical and psychological attributes of a person), ‘Opportunity’ (physical and social attributes of a person’s environment), and/or ‘Motivation’ (a person’s reflective and automatic mental processes). The COM-B components were then matched to relevant Intervention Functions (nine broad categories of potential interventions to change the capability, opportunity and/or motivation to engage in a behaviour) [45] (by PF, audited by JMcL and JMacD). Subsequently, we drew on our collective expertise and the Behaviour Change Technique Taxonomy version 1 (BCTTv1) [69] to operationalise the Intervention Functions into recommendations (conducted by PF and JM, audited by JMacD). The recommendations were reviewed by an interdisciplinary team including sexual health clinicians, public health researchers, and human computer interaction specialists. All BCW analyses were conducted by health psychology researchers who have completed the BCTTv1 training (https://www.bct-taxonomy.com/). Discrepancies in BCW coding were resolved through discussion between JMcL, PF and, latterly, JMacD.

Ethics

Written ethical approval for this study was granted by the East of England—Cambridge South Research Ethics Committee (REC) (reference 21/EE/0148) and Glasgow Caledonian University REC (reference HLS/NCH/20/045).

Participants

Participants (N = 35) (see Table 1) ranged in age from 18–70 (M = 34 years) and were diverse, representing several underserved populations: 51% (n = 18) lived in the most deprived areas of the UK (as defined by the Index of Multiple Deprivation [70]); 51% (n = 18) had no higher (i.e., university) education; 40% (n = 14) were of a minoritised ethnic group, of which, five were from an ethnic group other than White; 23% (n = 8) did not speak English as their first language; 49% (n = 17) had a mental or physical illness or condition lasting 12 months or more; and 29% (n = 10) had a learning difficulty. The majority of participants reported owning a digital device (e.g., mobile phone or laptop) to access the internet (n = 30, 86%) and using the internet every day (n = 29, 83%) for a wide range of activities, most frequently, social media (n = 22, 63%), work including research (n = 17, 48%), streaming TV shows or videos (n = 10, 29%), searching the internet for information (n = 9, 26%), and checking news (n = 5, 14%) and emails (n = 5, 14%). Over half of the participants described themselves as having ‘high’ level skills using the internet (i.e., digital literacy) (n = 20, 57%), 12 reported ‘medium’ (34%), and 2 reported ‘low’ (6%). Over a third (n = 12, 34.3%) had never searched for any sexual health information online, none had ever used a live chat or email or text exchange service for getting sexual health support, and few had used other online sexual health services such as booking an appointment (n = 4, 11.4%) or ordering medication through a private (non-state-funded) clinic (n = 1, 2.9%). Moreover, the majority had never ordered a postal STI self-sampling kit (n = 24, 68.6%) and a few reported having tried and struggled to order (n = 3, 9%), or use (n = 7, 20%), a self-sampling kit for STIs and blood borne viruses. See S6 Table for an overview of participant data regarding experience and self-rated skills of using the internet.

RQ1) What are the barriers and facilitators to seeking sexual health information online among underserved populations?

Table 2 details the nine barrier and eleven facilitator themes to seeking online sexual health information, with indicative data extracts, and their corresponding COM-B components.

Download:

• PNG
  larger image
• TIFF
  original image

Table 2. Barrier and facilitator themes and corresponding COM-B components for seeking online sexual health information among underserved populations.

https://doi.org/10.1371/journal.pone.0315049.t002

RQ2) What are the barriers and facilitators to seeking online sexual health support among underserved populations?

Table 3 details eight barrier and eight facilitator themes to seeking online sexual health support (via a synchronous live chat or asynchronous email or SMS text exchange), with illustrative extracts, and corresponding COM-B components.

Download:

• PNG
  larger image
• TIFF
  original image

Table 3. Barrier and facilitator themes and corresponding COM-B components for seeking online sexual health support among underserved populations.

https://doi.org/10.1371/journal.pone.0315049.t003

RQ3) What evidence-based and theoretically informed recommendations can be made to enhance seeking online sexual health information and support among underserved populations?

The BCW analysis identified potentially useful Intervention Functions to assist underserved populations in seeking online sexual health information and support: ‘Education’, ‘Persuasion, ‘Training’, ‘Modelling’, ‘Enablement’, and ‘Environmental Restructuring’ (see S7 Table for definitions). From these Intervention Functions, we proposed a range of more specific recommendations outlined below. Tables 4 and 5 provide an overview of the BCW analysis of barriers and facilitators and proposed recommendations for online sexual health information and support, respectively.

Download:

• PNG
  larger image
• TIFF
  original image

Table 4. Intervention content from Behaviour Change Wheel to address barriers and facilitators to seeking sexual health information online.

https://doi.org/10.1371/journal.pone.0315049.t004

Download:

• PNG
  larger image
• TIFF
  original image

Table 5. Intervention content from behaviour change wheel to address barriers and facilitators to seeking online sexual health support.

https://doi.org/10.1371/journal.pone.0315049.t005

Barriers and facilitators to passing through the digital doorway

Key barriers to people from underserved populations seeking online sexual health information and support were a lack of awareness of their availability and familiarity with using them; privacy concerns; and the perception of as inadequate to meet varied and complex sexual health needs and inferior to traditional information and support services (e.g., in-person appointment or phone call with a HCP). For seeking online sexual health information, specifically, barriers were navigating the overwhelming volume of information and different sources of and the perceived lack of inclusivity and relevance of information, particularly on government websites. For seeking online sexual health support, specifically, barriers were chatbots and automated responses; asynchronous communication; and wait times for responses.

Important facilitators to seeking online sexual health information and support were the perceived benefit of increased privacy compared to traditional services; the provision of video and/or audio options as alternatives to text; the presentation of information and support in a simple way–such as step by step and without jargon; and in-person assistance, for example, from key workers, family members, NGO staff, or GPs. For seeking online sexual health information, specifically, facilitators were inclusivity and personal relevancy of sexual health information, particularly for people from LGBTQI+ populations, religious backgrounds, and injecting drug users. For seeking online sexual health support, specifically, facilitators were synchronous communication; the perception of online support as acceptable for simple tasks such as appointment booking, general sexual health advice, and signposting to services; and delivery by qualified personnel.

An important consideration here is that none of the participants had ever used an online sexual health support service, despite their availability [e.g. 17–22], thus these barriers and facilitators were exclusively hypothetical. Additionally, barriers and facilitators differed in intensity and nature among subgroups; for example, privacy and confidentiality concerns were expressed by most participants but for different reasons and thus would likely differ among those not represented in this study.

Some of the barriers and facilitators are novel to this study. To the authors’ knowledge, low awareness of online sexual health information and support has not been reported previously and was common, even among participants who reported using the internet ‘all the time’. Additionally, the perceived inferiority and inadequacy of online sexual health information compared to traditional, such as in-person or phone, sources of sexual health information was novel to this study. Further, the unanimous rejection of chatbots and automated responses for sexual health support, due to the perceived complexity of sexual health, was novel to this study. This is in contrast to previous research showing that chatbots for sexual health were overall perceived to be useful for ongoing care or anonymous sex education (by samples of largely white heterosexual adult women) [59, 71]. However, beliefs about the inferiority of online services may change, as generative artificial intelligence becomes more commonplace and trusted [72].

On the other hand, many of our barrier and facilitators findings are consistent with previous literature, offering a novel perspective from a diverse sample of underserved populations. Privacy benefits and concerns [38, 57, 71, 73–75] and the value of video and audio options and simplicity [e.g. 29, 38, 75] have been previously reported amongst a range of populations. Further, the overwhelming volume of sources and information, need for inclusivity and personal relevancy of information, and preferences for online support to be delivered synchronously have been reported previously among sexual minority women [76] and African American youth [77], and young people [38, 78, 79]. Moreover, concerns about the adequacy of online information and support has been reported previously [71, 75, 80] and the acceptability of online support for simple tasks is in line with research regarding HCPs’ views of chatbots for sexual health [75]. Overall, the consistency of these findings with research with other populations indicates that addressing the barriers and enhancing the facilitators from this current study will improve online sexual health information and support for many beyond this sample.

Recommendations for opening the digital doorway to wider sexual healthcare

We identified a range of theory informed Intervention Functions to improve access to the digital doorway to wider sexual healthcare, including ‘Education’, ‘Persuasion’, ‘Training’, ‘Modelling’, ‘Environmental Restructuring’, and ‘Enablement’. First, advertising and promotion of services that provide online sexual health information and support and HCP’s endorsing online services and supporting patients to use them is vital (‘Education’ and ‘Persuasion’). Together, these findings resonate with wider research on telehealth per se [81], highlighting the importance of marketing and communication activities to increase awareness of online services to enable equitable access. Additionally, interventions could include online services offering audio-visual forms of communication, such as text-to-speech or cartoon animations. Moreover, our recommendations include the provision of inclusive and personally relevant information and support, such as information about sexual health relevant to those of diverse sexual and/or gender identities, religious backgrounds, and drug use (‘Environmental Restructuring’ and ‘Enablement’). This is in line with previous research showing that tailoring of information can positively influence acceptability of interventions and address barriers to care [82–84]. Further, interventions could include training HCP’s on how to use and promote online sexual health services effectively, in line with previous research regarding reducing digital inequalities [81]. Further, professional endorsement of high-quality online services for sexual health information and support is also known to be important [85]. Such endorsements should stress the services’ credibility, useability, and details of its functionality (e.g., trained professional will interact in real time). Interventions could also include embedding optional training for service users regarding how to use online services (‘Training’ and ‘Modelling’).

Implications

Within intervention development, context is critical for the likely success of the intervention [45, 86, 87]. As such, while the findings of this study are solely derived from participants living in Scotland and England, we are confident that they should be transferable to countries with similar healthcare settings and digital infrastructure, such as other countries in the UK not represented in this sample (i.e., Northern Ireland and Wales) or Australia, Canada, and New Zealand. We are less confident about the transferability of the findings to low-and-middle-income countries, given the lack of public health and accessible sexual health infrastructure in some countries [88, 89]. Research is needed to explore the barriers and facilitators to seeking online sexual health information and support and to develop recommendations to overcome the barriers and enhance the facilitators for people from underserved populations in low-and-middle-income countries.

Regarding service specific operationalisation and implementation of our recommendations, intervention development guidance [86] highlights the centrality of involving of stakeholders to co-produce culturally appropriate real world intervention content and help secure their effective delivery, given their investment of time and resource. For improving online sexual health information and support for underserved populations, our BCW analysis has delivered the first step of stakeholder engagement, offering practical theory-and-evidence-based recommendations (i.e., Tables 4 and 5). However, for each recommendation, the operationalisation and implementation for specific settings requires thorough consideration by service providers and other relevant stakeholders (e.g., healthcare providers, public health organizations, NGOs, government health departments, digital health platform developers, legal and regulatory bodies, funders and donors, educational institutions, and media and communication experts). Such stakeholder engagement should also consider affordability, practicability, efficacy, acceptability, side-effects, and equity [45] and address the operationalisation of our proposed recommendations to ensure final intervention content aligns with local circumstances and their wider legislative socio-cultural contexts. This could also involve the use of implementation science tools, such as Normalisation Process Theory [87, 90, 91], to ensure relevant issues of context are rigorously considered for maximum likelihood of success.

Strengths

The PPIE activities with diverse representatives of underserved populations ensured the development of inclusive study materials that were fit for purpose. Moreover, our recruitment strategy of working with clinics and NGOs focussed on the recruitment of participants via in-person rather than online settings. This facilitated the recruitment of people who may struggle to use online sexual health services. In addition, use of the PROGRESSPlus framework enabled the recruitment of a diverse sample of people from a range of underserved populations whose needs are typically unmet by existing services and whose perspectives are often overlooked (e.g., minoritised sexual, ethnic, and religious groups, those living in the most deprived areas). The diversity of our sample supports the transferability of the findings to a wide range of people who may struggle to use existing online sexual health information and support services.

Further, our use of in-depth participant-led qualitative interviews and rigorous inductive thematic analysis conducted by trained and highly experienced qualitative researchers allowed for the collection of rich data and identification of important evidence-based barriers and facilitators to seeking online sexual health information and support. Moreover, the rigorous auditing of the thematic analysis by an experienced behavioural scientist and a larger inter-disciplinary team enabled the development of a comprehensive set of themes, many of which were found to be consistent with previous research. In line with Braun and Clarke [63, 92], we demonstrate a transparent and comprehensive account of the rigorous thematic analysis and subsequent BCW analysis which appropriately answered the research questions. In addition, use of the BCW enabled the development of theory informed recommendations for improvements to seeking online sexual health information and support. As the BCW draws on the cumulative learning from multiple theories of behaviour change [66], it enabled us to specify potentially useful intervention content beyond that suggested by our participants.

Limitations

We sought to recruit a diverse sample of people from underserved populations with low digital literacy. However, while the sample was highly diverse from a range of underserved populations, it remains unclear whether the sample accurately represented people of lower digital literacy. The data we collected indicated that participants experienced difficulties using online sexual health services. However, a small majority self-reported as having high internet skills based on regular use of the internet to complete simple and familiar online tasks (e.g., social media, video streaming). Thus, a limitation of this study was that we did not use a validated measure of digital literacy. Although measures were available [93, 94], none were suitable, as they were too long and too involved, thus had the potential to limit participant engagement with the study. Research is needed to develop a validated short scale measure of digital literacy.

Further, no formal validation of the themes generated from the thematic analysis was conducted. While approaches are available, there is debate and no consensus regarding the best approach for this in qualitative research [95–97]. Additionally, in line with Braun and Clarke [63, 92], no inter-rater reliability of the thematic analysis was conducted, due to the highly reflexive nature of the analysis. Instead, a rigorous audit of the thematic analysis was conducted by a behavioural science expert and a broad inter-disciplinary team. However, the consistency of the barrier and facilitator themes identified in this study with previous research does suggest the reliability of the study and validity of the findings.

Moreover, the barriers and facilitators and subsequent recommendations presented here are based on the specific experiences of the participants recruited for this study and may not be generalisable across all underserved populations. However, the consistency in our barrier and facilitator findings with previous research with more specific populations suggests that the results of this study and the recommendations generated are transferrable and relevant for a broad range of underserved groups in the national settings mentioned above. Nonetheless, further research is needed to validate the findings in other contexts and confirm their applicability across diverse underserved populations. Lastly, as this study used the PROGRESSPlus framework for purposive sampling of people from diverse underserved populations, some specific demographic factors were not captured, such as urban/rural location and distance to sexual health services. Therefore, further research is needed for such underserved populations that were not included in this study.