Increasing attendance at pre-booked sexual health consultations: a systematic review

Keywords: attendance, behaviour change, health services research, HIV, intervention, sexual health, sexually transmitted infections, systematic review.

When left untreated, sexually transmitted infections (STIs) can lead to poor health outcomes including infertility, chronic pelvic pain and an increased risk of human immunodeficiency virus (HIV).^1,2 Access to, and attendance at sexual health clinic appointments (SHCs) can provide advice on prevention as well as rapid testing and treatment to reduce STI transmission and harmful sequelae. Yet, despite being preventable and treatable, STIs commonly remain untested and untreated. The latest National Survey of Sexual Attitudes and Lifestyles found that 75% of individuals at risk of STIs do not attend SHCs.³ Moreover, those who do identify themselves to be at risk and book a clinic appointment, often still fail to attend.⁴ Such non-attendance can lead to inefficient use of staff time and wasted resources, as well as increased waiting times for other service-users⁵ and poor outcomes for individuals.

Healthcare services often implement techniques that remind patients about their appointment to reduce non-attendance, such as sending short message service (SMS) text messages.^5,6 However, literature examining factors associated with the use of SHCs, suggests that barriers and facilitators to attendance can also occur at individual and interpersonal levels, such as the embarrassment of a genital examination, lack of STI knowledge, and fear related to the stigma associated with STIs.^7,8 Further barriers can occur in healthcare service provision and delivery, including, for example, access to SHCs, healthcare provider characteristics and previous experiences or expectations of discrimination.^8–10 To date, only a small number of interventions to improve attendance at sexual health consultations have previously been investigated.¹¹ A greater understanding of what kinds of interventions can be implemented and which elements within these have the greatest effect on appointment attendance in the context of sexual health care is essential for optimising future intervention delivery.

The aim of this review was to identify the range and effectiveness of interventions implemented to improve attendance at pre-booked sexual health consultations. A secondary aim was to identify theoretical constructs and behaviour change techniques (BCTs) used within included interventions.

This review is reported in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement.¹² The review protocol was registered with the International Prospective Register of Systematic Reviews (PROSPERO) (CRD42020197465). The development of this review’s research questions and outcome measures were informed by an advisory group that consisted of the research team, sexual healthcare professionals and sexual health service-user representatives.

Studies were eligible for inclusion if they: aimed to increase appointment attendance at a pre-booked, sexual health consultation during which service-users and healthcare providers were able to interact synchronously (e.g. face-to-face, telephone, video-call); interventions were delivered through any mode (synchronously or asynchronously); used a community sample of individuals aged ≥13 years; included an outcome measure for attendance of any real-time sexual health consultation (including reduction in missed appointments); and any study design comprising an intervention evaluation (including randomised controlled trials (RCTs), non-randomised control groups, single-arm designs, retrospective or prospective cohort studies).

Studies were excluded if they were published before 2000 or not fully available in English. Also excluded were studies conducted in non-WEIRD (western, educated, industrialised, rich, democratic) countries. This was because heterogeneity in access to health care and populations was considered to reduce meaningful conclusions that could be drawn from the data.

Four databases (Web of Science; ProQuest; PubMed, and Scopus) were systematically searched from 1 January 2000 to 1 September 2021. Additional studies were identified through reference chaining and citation checking via Google Scholar. The search strategy was developed in line with the Population (community sample) Intervention (set of behaviour change techniques) Comparator (any) Outcome (attendance) Study design (any) (PICOS) framework.¹³ Boolean operators were used to adapt the search for each database and searches were limited to those available in English (see Supplementary File S1).

One reviewer (RC) screened titles and abstracts and three reviewers independently screened the full text of potentially relevant articles (RC, GH and CF). Data were extracted from the included articles on key study characteristics, including year of publication, country, study design and setting, recruitment information, sample demographics and intervention characteristics, including use of theory, mode of delivery and BCTs. Data on the use of theory, mode of delivery and BCTs were independently coded by the review team (RC, GH and CF). Discrepancies were resolved through discussion.

The Theory Coding Scheme¹⁴ was used to assess the extent to which theory had been applied within the intervention design. This 19-item checklist contains statements that are coded ‘yes’, ‘no’ or ‘don’t know’ based on the explicit description of theory within the article. Items 1–11 were used to assess whether theory had been mentioned in the study, whether theory had been used to select participants or tailor intervention techniques, and whether theoretical constructs/predictors were explicitly linked to intervention techniques.

Adapting an approach outlined by Webb and Sheeran,¹⁵ the intervention’s mode of delivery was subdivided and presented as two aspects: (1) intervention format (e.g. text message); and (2) intervention facilitator (e.g. digital).

Intervention content was coded for BCTs using the Behaviour Change Taxonomy (v1).¹⁶ This taxonomy contains 93 BCTs, clustered into 16 groups: Goals and Planning, Feedback and Monitoring, Social Support, Shaping Knowledge, Natural Consequences, Comparison of Behaviour, Associations, Repetition and Substitution, Comparison of Outcomes, Reward and Threat, Regulation, Antecedents, Identify, Scheduled Consequences, Self-Belief, and Covert Learning. BCTs were only reported as being used in an intervention when explicitly described in the paper.

The Mixed Methods Appraisal Tool (MMAT)¹⁷ was used to assess methodological quality of retrieved studies independently by three researchers (RC, GH, CF). An overall quality score was calculated after responding ‘yes’, ‘no’ or ‘can’t tell’ to five questions relevant to the study design. Discrepancies were resolved through discussion.

Meta-analysis was not possible due to heterogeneity in the intervention modes of delivery, outcome measures and participants. Instead, a narrative approach was used to synthesise intervention characteristics and outcomes, theoretical application, mode of delivery and BCTs. Data were presented in a tabular format. Interventions were considered effective if the SHC attendance outcome was reported to have significantly increased (P < 0.05) in the intervention group and, where available, was significantly greater than that in the control group. To ensure that the reported effectiveness of intervention components only reflected active elements in the intervention group, BCTs present in both the control group and intervention group were not coded.

The systematic search of the databases identified 615 articles, with 13 additional articles identified through reference checking. Of these, 13 articles met the inclusion criteria (Fig. 1). Of the 13 included articles, eight used non-randomised before–after study designs^18–25 and five used randomised controlled trials.^26–30 Five studies were conducted in Australia, five in America and three in the United Kingdom. All interventions were delivered within a clinical setting and one included community engagement.¹⁸ The studies reported a variety of outcome measures for attendance at SHC consultations: attendance at clinic appointments (n = 6),^{18,23,24,28–30} return visits to clinic (n = 1),¹⁸ and retesting rates (n = 7).^{19–22,25–27} The follow-up period for measuring participants’ attendance ranged from 1 month^29,30 to 12 months.^24,25,28 Further details about intervention characteristics can be found in Supplementary File S2.

Fig. 1.  Preferred reporting items for systematic reviews and meta-analyses (PRISMA) flow diagram of the systematic search and selection of articles.

Methodological quality was identified as moderate in 10 studies, with three rated as low quality^20,27,30 (Table 1). Within the included studies, intervention fidelity was often unclear^19,21,24,25 and four interventions were reported not to have been delivered as intended because SMS messages were not delivered to all participants.^{20,22,26–29} Furthermore, not all studies reported whether analysis controlled for confounders^18,20,22,24 and there were doubts about the representativeness of some populations studied.^20,27,28 Rutland et al. (2012)³⁰ was reported as a conference paper. Consequently, the information provided about the study is very limited and, therefore, increases the risk of bias.

Table 1.  Mixed-methods appraisal tool for included studies in the systematic review.

Eligible studies included four interventions that aimed to support patients’ attendance at HIV care appointments. Following a 6-month bidirectional texting intervention, Rana et al. (2016)²³ reported participants’ appointment adherence was an average of 79.1% of scheduled appointments, with 47% of participants achieving 100% attendance at their scheduled appointments. However, as no pre-intervention attendance data were reported, it is not known if the intervention significantly increased attendance. Tanner et al. (2018)²⁴ found that a combination of personalised messages sent through social media, SMS and mobile applications significantly decreased participants missed appointments from 68.4% in the 12 months before the intervention to 53.3% during the 12-month intervention (P < 0.04). Ingersoll et al. (2015)²⁸ also found that personalised bidirectional SMS messages improved missed appointments from 26.9% to 9% compared to 31% to 28% in the control group. Nevertheless, the findings were not significant (P = 0.12). Another study by Norton et al. (2014)²⁹ did not find the use of an SMS reminder and reminder phone call to significantly improve attendance rates compared to a control group who also received a reminder phone call (P = 0.42).

Two further studies explored intervention effectiveness on attendance in SHCs. Biggs et al. (2016)¹⁸ reported a significant increase in the number of Aboriginal people attending sexual health consultations following a peer-based, incentive-driven intervention (n = 313) in comparison to a historical control group (n = 83, P < 0.01). However, despite this increase, there was no significant difference in the number of return visits in the 12-month period following the first appointment (n = 169 (intervention group) vs n = 51 (historical control group); P < 0.31). Rutland et al. (2012)³⁰ found an SMS notification with a health promotional message increased reattendance rates by 15.2% (P = 0.032) compared to 8.2% with an SMS notification without a health promotional message (P = 0.36) and 4.5% in the control group.

Seven studies assessed intervention effectiveness for increasing attendance to repeat STI testing, including two studies that used SMS reminders in a sample of men who have sex with men (MSM). Bourne et al. (2011)¹⁹ found the use of an SMS reminder for repeat STI screening significantly increased retesting in MSM (64% attendance) compared to the comparison group without an SMS reminder (30% attendance – OR 4.4 [95% CI 3.5–5.5], P < 0.01) and the pre-SMS group (31% attendance – OR 3.1 [95% CI 2.5–3.8], P < 0.01). Similarly, Zou et al. (2013)²⁵ found the number of men who returned to the clinic to be significantly higher among men who had 3-monthly (89.5% attendance, P < 0.01) or 6-monthly (87.7% attendance, P < 0.01) SMS and/or email reminders compared to a concurrent control group (70.8% attendance).

Four further studies used SMS reminders for men and women. Downing et al. (2013)²⁶ found both SMS reminders (22.7% attendance) and SMS reminders plus financial incentives (29.17% attendance) to increase retesting rates compared to a control group (0% attendance; P < 0.04 and P < 0.04 respectively). Guy et al. (2013)²¹ reported retesting to be significantly higher in the SMS reminder group compared with the pre-SMS group (30% vs 21%; P < 0.04). Those in the SMS group were more likely to return than the pre-SMS group (OR 1.57 [95% CI 1.01–2.46]); however, SMS reminders did not significantly increase retesting when compared with a concurrent non-SMS group (30% vs 25%; P < 0.30). Nyatsanza et al. (2016)²² found that sending a personalised SMS reminder significantly increased re-attendance rates for testing (56% [95% CI 50–62%]) when compared to a non-personalised SMS group (33% [95% CI 28–39%]; P < 0.01). However, Burton et al. (2014)²⁰ did not find SMS reminders affected re-attendance rates when compared to a historical control group. Burton et al. (2014)²⁰ hypothesised that tailored or bidirectional SMS messages might have been more effective.

Malotte et al. (2004)²⁷ examined the effects of a variety of interventions including financial incentives, motivational counselling and a phone call reminder on re-attendance for repeat testing. Their findings suggested that reminder phone calls were most effective to increase client return visits (OR 18.1 [95% CI, 1.7–193.5]). Malotte et al. (2004)²⁷ recommended combining motivational counselling and phone call reminders to maximise re-attendance.

Theory was mentioned in three included studies (Table 2).^24,27,28 Ingersoll et al. (2015)²⁸ reported that the Information, Motivation and Behaviour Skills (IMB) Model of Adherence³¹ and Social Action Theory³² were the theoretical foundations for the intervention. However, how the theories were used to inform the development and application of the intervention were not reported.

Table 2.  Summary of intervention’s use of theory and theoretical base.

Tanner et al. (2018)²⁴ reported that intervention content was informed by Social Cognitive Theory³³ and Theory of Empowerment.³⁴ Examples of SMS messages sent to participants were provided, with the messages explicitly linked to constructs from Social Cognitive Theory (e.g. information, outcome-expectancies, self-efficacy, direct experience, vicarious learning, persuasion/social support, incentives) and Empowerment Theory (e.g. critical consciousness, action).

Malotte et al. (2004)²⁷ described motivational interviewing as commonly related to cognitive-behaviour theories, such as Social Cognitive Theory³³ and Theory of Reasoned Action.³⁵ However, the article did not report whether cognitive-behaviour theories were used to inform the motivational counselling used, which theories were used, or the extent of their use.

Seven modes of intervention format were reported (Table 3). The most used was SMS (n = 11), for which implementation varied in content and frequency (Table 4). Other interventions utilised individual, face-to-face sessions (n = 2), telephone calls (n = 1), email contact (n = 1), social media (n = 1), app-based messaging (n = 1) and letters (n = 1). A total of 10 interventions used only one intervention format, whereas the remaining three used a combination of two or more formats.

Table 3.  Summary of intervention’s modes of delivery.

Table 4.  SMS message reminder content and delivery.

Reflecting the intervention formats, digital facilitation was commonly reported (n = 11), with nine interventions facilitated via automated messaging, and three using bidirectional messaging.^23,24 One intervention sent reminder letters.²⁷ The following professionals facilitated in one intervention each: Aboriginal sexual health worker, SHC staff, interventionist, and a cyber-health educator.

A total of 19 BCTs were identified in the included interventions (Table 5). The number of BCTs used ranged from 0 to 14 (mean: 3.2). The most observed BCTs across the 18 intervention groups were credible source (n = 12) and prompts/cues (n = 10). The following BCTs were only observed once across intervention groups: social support (practical), social reward, self-incentive, reduce negative emotions, restructuring the physical environment, restructuring the social environment, focus on past success, and vicarious consequences. Although Nyatsanza et al. (2016)²² described use of both prompts/cues and credible source in the intervention group and Norton et al. (2014)²⁹ described prompts/cues in the intervention group, these BCTs were also reported in the control group and thus, they were not coded.

Table 5.  Behaviour change techniques¹⁶ identified in the interventions.

Within the 10 intervention groups found to have a significant increase in attendance behaviours, the number of BCTs reported ranged from 0 to 14 (mean: 4). The most frequently used BCTs described in effective interventions were: credible source (n = 8), prompts/cues (n = 8), and information about health consequences (n = 5). The following BCTs were solely used in interventions found to increase attendance at SHCs: social support (practical), instruction on how to perform behaviour, social reward, self-incentive, reduce negative emotions, restructuring the physical environment, restructuring the social environment, focus on past success, and vicarious consequences.

Eight intervention groups did not report a significant increase in attendance behaviours. The use of BCTs within these intervention groups ranged from zero to seven (mean: 2.3). The most commonly coded BCTs among intervention groups that did not report an increase in attendance were: credible source (n = 4), prompts/cues (n = 4), problem solving (n = 2) and social support (unspecified) (n = 2).

This review identified 13 interventions designed to increase attendance at pre-booked SHC appointments. Findings suggest that behavioural interventions can be effective at supporting appointment attendance in the context of sexual health. Across all included interventions, only one study explicitly linked theoretical constructs to the BCTs implemented. A total of 19 BCTs were identified within 18 intervention groups, of which the most common were: information about health consequences, use of prompts/cues, and information provided by a credible source. There were seven different modes of intervention delivery and six different intervention sources. SMS was the most frequently employed mode of delivery, with a digital (automatic) facilitator.

Consistent with literature relating to other healthcare settings,³⁶ the present review indicates that mobile health (mHealth) interventions have the potential to increase attendance rates at SHC appointments.¹¹ This review strongly suggests, however, that the content of the mHealth intervention is as important as the mode of delivery. Previous research suggests that appointment reminders may be more effective when combined with additional behaviour change strategies such as providing sexual health information, access to advice from healthcare professionals, and social and psychosocial support that can address knowledge deficits, low motivation and behaviour change.³⁷ This review also supports previous calls to tailor appointment reminder systems to a specific service or sub-population.³⁸ For example, appointment reminders could be combined with additional messages tailored to specific subgroups of patients, such as MSM (e.g. messages that target service-users perceived at risk of sexually transmitted diseases in MSM,²⁵ health consequences of missing HIV appointments²⁴) and culturally appropriate messaging (e.g. messages that address attendance concerns specific to stigma within a local community¹⁸). Nevertheless, for the additional messages to be effective, acceptable and engaging, it may be essential to co-develop such messages with service-users. Future research also needs to assess the cost-effectiveness of more complex mHealth interventions that provide a variety of content compared with simpler approaches, such as SMS appointment reminders. Furthermore, consideration should be given to patient communication and technology preferences due to the sensitivity of sexual health and service-user privacy concerns.²⁹ Thus, it may be beneficial to explore the effectiveness of mHealth interventions compared to alternative modes of delivery.

The identification of BCTs within interventions highlights the importance of using credible sources to provide information, implementing appointment prompts/cues, and providing relevant information about health consequences to increase SHC appointment attendance. However, credible sources and appointment prompts/cues were also identified in ineffective interventions and it is possible that the BCT taxonomy does not capture differences within their application or content, which may alter their effectiveness.³⁹ Alternatively, the credible sources and appointment prompts/cues may have interacted with additional BCTs used within the interventions. As theory and literature suggests that appointment attendance should be understood in terms of both practical (e.g. prompts/cues) and perceptual barriers (e.g. information about health consequences),^8–10,40 it is essential for interventions to combine BCTs that address both of these elements. Combining strategies in this way was recommended by Malotte et al. (2004),²⁷ who suggested using a telephone reminder alongside motivational interviewing to increase return visits. However, the effectiveness of combining BCTs could not be reliably tested within this review due to the low number of relevant studies, and this should be a priority for future research.

Behavioural science literature further suggests interventions based on theory or guided by theoretical constructs are more effective at changing health behaviours.⁴¹ The use of theory can help explain and predict specific behaviours in different contexts, populations and settings, highlighting which causal pathways should be targeted within interventions.⁴² However, despite recommendations for complex interventions to be theoretically underpinned,⁴³ only 1 of the 13 included interventions explicitly reported the application and implementation of theory,²⁴ whereas another two referenced the intervention as being theory-based without providing details.^27,28 Consequently, it is not possible to identify specific theoretical constructs that might mediate attendance to SHCs. Nevertheless, the review does highlight the importance of utilising and explicitly reporting theory in future evaluation of interventions that may, or may not, lead to increased SHC attendance to help enhance understanding of the causal pathways.

This review used a robust and systematic methodology that limited bias in the identification and selection of relevant studies. Three researchers independently extracted data from included studies, coded for theory and BCTs and assessed quality, which adds further rigour. Although the review updates an earlier systematic review on interventions to increase testing and clinic re-attendance in SHCs,¹¹ it further provides new knowledge on behavioural theory, mode of delivery and behaviour change components in existing SHC interventions, contributing to a deeper and more nuanced understanding for developing future interventions.

Nevertheless, the review has limitations. Variation between included studies (e.g. research designs, outcome measures, participant demographics) rendered it impossible to perform more complex meta-analyses. Second, some studies were unclear as to whether all appointments were pre-booked or whether the interventions motivated attendance to walk-in clinics. Furthermore, the inclusion of a conference paper provided limited information.³⁰ Such studies were included in order to continue building knowledge within this limited field. Third, reporting of intervention development and intervention characteristics, such as the theoretical underpinning and application of BCTs, was limited. Interventions may have used additional theories and BCTs that were not reported; however, this is recognised as a common issue in systematic reviews that report the use of theory and BCTs.^11,44 Future interventions should report theoretical application and use standardised measures for BCTs to support the development of a robust evidence base.

This review provides new knowledge about the types of interventions implemented to improve appointment attendance at SHCs, including the effectiveness of specific intervention components. Findings indicate that SHC appointment attendance can be increased when both practical and perceptual barriers are targeted through the inclusion of BCTs, such as providing information from credible sources, using appointment reminders and giving information about the consequences. This provides a focus for future research to assess combining BCTs to improve clinic attendance rates. mHealth interventions may offer an effective approach for delivering theoretically informed interventions to a wide population, but cost-effectiveness analysis is also needed before widespread implementation.

Supplementary material is available online.

The data that support this study will be shared upon reasonable request to the corresponding author.

Jonathan Ross reports personal fees from GSK Pharma and Bayer Consumer Care; ownership of shares in GSK Pharma and AstraZeneca Pharma; lead author of the UK and European Guidelines on Pelvic Inflammatory Disease; Member of the European Sexually Transmitted Infections Guidelines Editorial Board. He is an NIHR Journals Editor and associate editor of Sexually Transmitted Infections journal. He is treasurer for the International Union against Sexually Transmitted Infections and chair of charity trustees for the Sexually Transmitted Infections Research Foundation. The other authors report no conflicts of interest.

This study was funded by University Hospitals Birmingham NHS Foundation Trust.