Adapting Digital Health Interventions for the Evolving HIV Landscape: Examples to Support Prevention and Treatment Research

Abstract

Purpose of Review

While many HIV prevention and/or treatment digital health interventions (DHIs) have shown feasibility and acceptability, fewer have indicated efficacy, and only a subset have been adapted for new contexts. Adaptation is a key element of pragmatic implementation science research. Adaptation is cost effective and time efficient compared to new development. Leveraging adaptation can lead to accelerated scale-up and enhanced public health impact. Considering the value of adaptation, the purpose of this piece is to present examples of DHI to DHI adaptation sequences to inform future HIV prevention and/or treatment research.

Recent Findings

From an examination of recent academic articles (11/01/2016-10/31/2021), we identified adaptation sequences that included an original DHI, plus at least two adaptations. Four models are presented herein; examples consist of adapted DHIs for new population, health outcome, geography, or a combination thereof.

Summary

Adaptation is a promising scientific approach to expeditiously respond to the evolving HIV landscape. We present examples of DHI adaptations alongside considerations for each type of adaptation; we also present adaptation challenges with responsive strategies. We suggest when conducted with attention to rigor (leveraging adaptation frameworks, community engagement, and tailoring content), adaptation is a powerful tool to pragmatically address the HIV epidemic.

INTRODUCTION

To date, human immunodeficiency virus (HIV)-related digital health interventions (DHI) have made notable impacts on both prevention and treatment continuums of care.^1–4 People living with HIV (PLWH) often perceive DHIs to be more convenient, protective of confidentiality, destigmatizing, and accessible as compared to face-to-face interventions.^5,6 Considering that groups disproportionately affected by HIV (e.g., sexual minority men, transgender women, racial minorities, etc.) have experienced institutional mistreatment, intersectional stigma, and racism, leading these groups to be skeptical of research and healthcare systems,^7,8 the adaptation of private and secure DHIs to address the HIV epidemic is warranted and presents an opportunity to promote rapid scale-up and broad distribution.^9,10

While the development of new DHIs is justified when existing infrastructure does not already exist, we suggest that adaptation of previously developed and tested DHIs for new targets is a more advantageous approach. Adaptation is efficient, cost-effective, and when conducted with care, scientifically rigorous.^10,11 Since development of new mobile applications or DHI platforms is costly and time-consuming, and the routinization process (time it takes from discovery to regular use or clinical practice) is about seventeen years,¹² during which time the epidemic will have undoubtedly be transformed by new therapeutics and shifting demographics, we posit that intervention adaptation offers an expeditious process to facilitate agile and responsive research through changing landscapes and across varying contexts.^5,13

In light of the benefits of adaptation, value of tailored clinical care, and adaptation being a pragmatic and rigorous process enabling responsiveness to evolving HIV prevention and treatment considerations,¹⁴ we discuss considerations for DHI to DHI adaptation and provide examples (informative models) of HIV-focused DHIs that have been adapted multiple times. We hope that consolidating this information will assist in the development and refinement of future DHI research that aims to improve outcomes across the HIV prevention and treatment continuums of care.

ADAPTATION, MODIFICATION, AND CORE FEATURES

Adaptation is the process of modifying an intervention without contradicting its core features or internal logic.^13,15 Modifications are intended to achieve a better fit between the intervention and new contexts or targets. Modifications can include planned changes; these are anticipated modifications that are made before releasing the newly adapted intervention, and responsive changes. Responsive changes are modifications that address feedback ascertained during implementation or acceptability testing. While making modifications during the adaptation process, core features should remain intact.

Core features are elements that bolster effectiveness of the original intervention and therefore must be retained to maximize the likelihood that the adapted DHI will continue to be impactful.¹⁶ Core features reflect industry standards and regulations, population preferences, and behavior change or access to care strategies. While core features are not necessarily standard across DHIs and platforms, studies suggest there is some uniformity and consistency in core features. For example, privacy and data security are core features that are part of industry standards and are of paramount importance when collecting health-related data.^14,17,18 Core features reflecting population preferences, for example, of youth living with HIV and youth at elevated risk for HIV-exposure, include social interactivity, bi-directional communication, inclusion of tailored information, gamification, and youth-specific language.^14,17,18 Known core features that promote behavior change or improve access to care are linkage to services and non-intrusive medication adherence reminders.^14,17,18

METHODS

Digital health (including mHealth and eHealth) encompasses the provision of health services via technological modalities.¹⁹ We define DHIs as multicomponent systems that exclude text-only, standalone auto-calling, or single element technology-based or online strategies. DHIs appropriate for adaptation, replication, or tailoring will have an extensible infrastructure with strong underlying security and safeguards for data privacy.⁵

To identify recent examples of successful HIV-related DHI to DHI adaptations, we searched PubMed from 11/01/2016 to 10/31/2021 (5 years), using the following keywords and Mesh terms in combination: HIV/HIV infections or acquired immunodeficiency syndrome (AIDS); pre-exposure prophylaxis (PrEP); smartphone or smart phone or mobile phone or mobile app or mobile application or mHealth or mobile health or digital health; adapt*. Research studies and review articles were included in our evaluation. Among indexed articles, we narrowed our consideration to studies that reported on the adaptation process of HIV DHIs that utilized smartphone or tablet applications. We excluded publications that did not detail the intervention adaptation process, did not include HIV prevention and/or treatment as a primary outcome within the original DHI, were not written in English, and/or detailed the adaptation process of a face-to-face intervention into a DHI. While attention was given to articles published in the prior eighteen months, there were situations where an adaptation occurred recently, but the original DHI was developed prior to five ago. We retained these examples, because they were illustrative of how well-crafted DHIs can be adapted over time to be responsive to new developments in HIV prevention and/or treatment. Once we identified sequences of adaptations, we conducted a simple online search for the original DHI. We found two originating parent DHIs had dedicated websites containing information to support ongoing adaptations.

Our intent was not to conduct an exhaustive systematic nor scoping review, rather to identify informative examples of translational HIV digital health research that included adaptation from one DHI to another DHI. As part of this process, we considered numerous academic research articles and selected diverse candidate pieces to highlight. We believe these examples can assist HIV researchers frame DHI adaptation studies. We present four DHI platforms that resulted in at least two subsequent adaptations across multiple populations, HIV statuses, health conditions, and geographies (see Table 1).

Table 1:

Original DHI to Adapted DHIs

Original DHI	Adaptation	Population Demographics	Age (years)	Location	Status	Outcomes
1. Thrive With Me1		Adult men who have sex with men (MSM)	18+	United States (US)	HIV+	Viral suppression
	YouTHrive2	Youth	15-24	US	HIV+	Viral suppression
	TechStep3	Transgender and gender non-conforming youth	15-24	US	HIV−	Condomless intercourse, PrEP uptake
2. PositiveLinks4		Adults	18+	US	HIV+	Retention in HIV care
	PositiveLinks5	Spanish-speaking Latinx adults	18+	US	HIV+	Retention in HIV care
	MOCT6	Adults with tuberculosis and substance use	18-64	Russia	HIV+	Linkage to and engagement with HIV care
	PositiveLinks7	Adults with hypertension	18+	Uganda	N/A	Blood pressure control
3. AllyQuest8		Young MSM	16-24	US	HIV+	ART adherence
	P3: Prepared, Protected, emPowered9	Young MSM and transgender women who have sex with men	16-24	US	HIV−	PrEP adherence
	P3-Thai10	Young MSM	16-24	Thailand	HIV−	PrEP adherence
4. HealthMpowerment11		Young Black MSM	18-30	US	Status Neutral	Condomless anal intercourse
	HMP Stigma12	Young Black and Latinx MSM and transgender women	15-29	US	Status Neutral	HIV testing, viral suppression
	PrEPresent13	Young MSM	16-26	US	HIV−	PrEP uptake
	MASI13	Adolescents with perinatal-acquired HIV	14-21	South Africa	HIV+	ART adherence
	Vuka+13	Adolescent girls and young women	16-24	South Africa	HIV−	PrEP adherence
	Spark (PrEP Romania)13	Adult MSM	18+	Romania	HIV−	PrEP adherence

Bolded text indicates areas of adaptation from original intervention

ACCEPTIBILITY TESTING

While approaches to adaptation vary, most examples included in Table 1 describe a user-centered design approach informed by qualitative data collection, with iterative usability testing.²⁰ Often through the use of focus groups or in-depth interviews, investigators solicit end-user feedback to inform content modifications (e.g., to directly address new outcomes, ensure cultural relevance, etc.), feature enhancement (e.g., to increase engagement, expand support, provide linkage to care or resources, etc.), and address accessibility (e.g., limited Internet availability, health literacy, etc.). Once initial modifications are made, usability testing is typically performed to capture end-user feedback on the adapted DHI’s design, functionality, acceptability, and ease of use.²¹ Even though acceptability has often been proven on the original intervention, conducting acceptability testing of the adaptation increases the likelihood of DHI relevance, satisfaction, and acceptability in the new population of interest, or related to the new outcome or context. Including acceptability testing during adaptation can also uncover unanticipated technical issues that occurred during adaptation.

ADAPTATION FOR RAPID RESPONSE

The HIV epidemic is rapidly evolving with shifting demographics of disproportionately affected populations in the United States and abroad.^22,23 Reducing persistent HIV inequities requires efficacious DHIs are tailored to multiple populations, socioecological contexts, and individual constructs -- sometimes sequentially and sometimes concurrently. While there has been limited scale-up to date,²⁴ the examples provided in Table 1 offer models of rapid adaptation to address these disparities.

Example:

The first sequence in Table 1 illustrates how Thrive with Me was sequentially adapted to YouTHrive and then to TechStep.^25–27 The original DHI, Thrive with Me, was developed for adult men living with HIV; YouTHrive revised Thrive with Me tailoring it for youth living with HIV, and thereafter TechStep was adapted to reduce sexual risk while increasing PrEP uptake among transgender youth. This series transformed the original DHI first for age (from adults living with HIV for youth living with HIV) and then for gender and status (from youth living with HIV to HIV-negative transgender and gender nonbinary youth), all in less than three years.^25–27 This rapid transformation illustrates adaptation’s potential for accelerated responsiveness.

ADAPTATION FOR POPULATIONS AND OUTCOMES

The HIV epidemic in the United States is concentrated among underserved sub-populations;²⁸ thus, when a DHI is effective in promoting behavior change or improving outcomes in one group, core components and theoretical underpinnings of the successful intervention may be transferable to another group that also experiences harmful structural forces, warranting adaptation of the original intervention. For example, DHIs may be adapted for age, sexual orientation, gender identity, race, and ethnicity. When adapting for new demographic profiles, interventionists should consider how the adapted DHI will look and feel, including changes to content, language used, and imagery.^14,17,18 Further, given that many DHIs address distinct or singular outcomes, adaptation provides a pathway to address different outcomes along the same continuum, different outcomes on different continuums (prevention as compared to treatment), or provide greater flexibility required to account for changes in evolving science (e.g., approval of long-acting injectable PrEP or “2–1-1” dosing).

Example:

The third example in Table 1 illustrates such progress. AllyQuest is a DHI designed to improve antiretroviral therapy (ART) adherence among 16–24 year old, young men who have sex with men (MSM) who were living with HIV in the United States.²⁹ AllyQuest was significantly adapted to Prepared, Protected, emPowered (P3) to promote PrEP among young men and transgender women who have sex with men which was then modified for the Thailand context retaining its focus on HIV prevention among youth but narrowing the population to young MSM.^30,31 While the original DHI and both adaptations focused on HIV-related outcomes in sexual and gender minorities aged 16–24 years, there were shifts in continuum of care outcomes, population sexual identity, and location or geography.

ADAPTATION FOR GEOGRAPHY

North to South global development to address HIV epidemics, more often than not, adapt a DHI developed and tested in the United States for use in low and middle-income countries (LMIC). When engaging in North to South DHI adaptation, researchers should consider the types of mobile phones that are commonplace in the new setting. Older phones with less memory are routinely used in resource-limited international settings; thus, DHI adaptations may opt to limit application size and processing heavy features.^32,33 Although wireless fidelity (Wi-Fi) is accessible in major metropolitans in high income nations and those who live in these cities may have unlimited data on their phones, features are more restricted in rural communities and in LMICs.³⁴ In resource-constrained settings, data plans and Wi-Fi tend to be financially costly meaning few have all-day, every-day Internet connectivity; therefore, DHI adaptations for LMICs should examine if it is possible to adapt online intervention components to be available offline. Language used within the DHI, even when adapting from an English-speaking context to another English-speaking setting, should be evaluated for understandability and acceptability in the new context. Ongoing management of hardware, software, and databases in the proposed setting, including servers, system security, and data privacy must be attended to.²¹

Examples:

All four original DHI examples in Table 1 were created in the United States;^{25,29,35–37} three of the four led to DHI transformations for delivery in resource-limited global settings.^21,31,38,39 PositiveLinks was adapted to MOCT in Russia and retained the name PositiveLinks when adapted to address blood pressure outcomes in Uganda.^21,35,38 AllyQuest was adapted from addressing an HIV treatment outcome in the United States to P3-Thai (P3 adaptation as a step between AllyQuest and P3-Thai) targeting HIV prevention in Thailand,^29–31 and HealthMpowerment (HMP), a status neutral DHI platform, was transformed for HIV prevention and treatment in South Africa and HIV prevention in Romania.^36,39

ADAPTATION CHALLENGES AND RESPONSES

The same challenges inherent in adaptation of in-person interventions apply to the adaptation of DHIs, namely insufficient attention to ensuring cultural relevance, incorrect assumptions made during the adaptation process leading to reduced effectiveness, and a lack of rigor and transparency.^40,41 To address these challenges, researchers can leverage validated adaptation frameworks to guide the process and engage in DHI co-creation through community engagement.

Frameworks for adaptation include, but are not limited to ADAPT-ITT, intervention mapping, IM-Adapt (Intervention Mapping for Adaptation), an extension of intervention mapping, and the Accelerated Creation-to-Sustainment (ACTS) model.^42,43 ADAPT-ITT is a pragmatic framework that guides the adaptation process through eight sequential steps from assessment to testing.¹⁵ Intervention mapping, used both in new development and adaptation, is a planning approach that incorporates theory and prior evidence to inform the creation of change objectives within the intervention or adaptation.⁴⁴ IM-Adapt offers step-by-step guidance to explore which features and strategies should be retained or revised.⁴² ACTS is an integrative model for design, assessment, and sustainability that can inform adaptation. By separating DHI service components from the technological elements, ACTS parses features for adaptation.^43,45 Although, adaptation frameworks have historically been applied to in-person interventions or when adapting an in-person intervention for mobile delivery, frameworks are growing in use and are being applied more frequently to DHI to DHI adaptations.

Thoughtful community engagement, where representatives’ voices are elevated, maximizes the likelihood that an intervention will resonate within new contexts and with new populations.⁴⁶ Beneficiaries and community partners, can be included in formal community advisory boards. Beneficiaries and community partners should be involved in all aspects of the adaptation process beginning at the planning stage, though implementation, and concluding with dissemination and scale-up.⁴⁶ Soliciting feedback from community members on intervention adaptation can also lead to more innovative and creative methods for engagement and recruitment;⁴⁷ deep community engagement have the potential to produce insights about true lived experiences that can inform the conduct of pragmatic trials in real-world settings.

LIMITATIONS

This piece is not a comprehensive assessment of the literature; we did not search National Institutes of Health Reporter or conference databases. It is likely that recently funded studies and adaptation projects that concluded in the past year were not captured. We selected prominent DHI adaptation sequences to feature; however, it is possible that there are other DHI sequences that could have been included.

CONCLUSION

Current approaches to new DHI development are highly inefficient. When core features of a new endeavor match those within an existing platform, we recommend that limited resources be allocated to customizations and adaptation (e.g., new features, behavioral change strategies, tailored content, etc.) rather than building anew. Further, even when conducting an adaptation, researchers often create content de novo. We suggest that adapting or tailoring existing HIV prevention and/or treatment content can reduce significant, redundant effort. Future work to develop more efficient content management systems that disseminate content broadly and leverage advances in machine learning to improve inefficiencies are needed.^48,49

In sum, we offer readers pragmatic models through examples of HIV-related DHI to DHI adaptations that can inform future studies in HIV and HIV-adjacent research. We suggest that DHI adaptation -- leveraging validated infrastructure, adaptation frameworks, and thoughtful community engagement -- is an efficient scientific approach to expeditiously address HIV-related outcomes across new contexts, populations, and outcomes.

KEY POINTS.

• Digital health interventions, due to their ability to protect confidentiality and reduce stigma, have become foundational to HIV prevention and/or treatment research.

• The implementation science-informed process of adaptation enables researchers to leverage existing, validated infrastructure to address new health outcomes, populations, and settings, making intervention adaptation cost-effective and time efficient.

• Adaptation research is enhanced when applying rigorous frameworks, such as ADAPT-ITT or IM-Adapt, and engaging community members in co-creation to guide study processes.

• To facilitate rapid response to new HIV treatment modalities and prevention options, researchers should make their digital health interventions freely available for adaptation and to promote scale-up.