Disparities in HIV Prevention and Treatment
Geographical disparities in HIV-related health care access and treatment adherence exist in the United States, with nearly 50 percent of new HIV infections and lower rates of pre-exposure prophylaxis (PrEP) utilization in the South.1,2 Many people with HIV (PWH) do not receive adequate care, despite the availability of antiretroviral therapy (ART). In fact, 76 percent of PWH received HIV care, 54 percent were retained in care, and 65 percent of those in care were virally suppressed in 2022.3
Rural communities face even more significant barriers to accessing care, including limited health care facilities and few primary care providers who are trained to address the comprehensive needs of patients with chronic conditions such as HIV.4 Some rural residents travel many miles to the nearest health care facility, while others without transportation experience significant gaps in clinic visits.
Structural inequities such as poverty and unemployment also contribute to disparities in HIV care access, poor retention in care, inadequate viral suppression, and exacerbation of comorbidities.5 According to data from the Centers for Disease Control and Prevention (CDC), adults residing in regions marked by significant poverty showed the highest rates of HIV diagnoses but the lowest rates of linkage to HIV medical care within one month after receiving a diagnosis, and the lowest rates of viral load suppression within six months of receiving an HIV diagnosis.6 These disparities may be linked to care accessibility, as about 10 percent of PWH reported travel times over an hour and nearly one in five travel more than 30 minutes.5,6
Stigma, discrimination, and social isolation further exacerbate these challenges, often deterring individuals from seeking HIV prevention and treatment services.7 These disparities highlight the urgent need for innovative and tailored interventions to expand access to ART for PWH and PrEP for those at high risk for HIV, with special considerations for rural and underserved populations.
The Role of Artificial Intelligence to Address Gaps in HIV Care
Artificial intelligence (AI) has emerged as a promising avenue to address gaps in HIV prevention and treatment services, which is critical to meet the UNAIDS 95-95-95 targets to end the HIV epidemic by 2030.8 AI tools have shown progress in rapid screening and early diagnosis, enhanced tracking of disease progression and treatment, and assessment for co-morbid conditions. One form of AI, machine learning, applies algorithms by analyzing large datasets (e.g., electronic health records) to identify patterns and predict risk factors.9,10 For example, four machine learning approaches performed better than multivariate logistic regression in predicting HIV and sexually transmitted infection among men who have sex with men.10 Also, machine learning with electronic health record data demonstrated accuracy in identifying people who were at high risk of HIV, but not yet using PrEP.9 Similar algorithms may predict factors related to viral load suppression and ART and PrEP adherence. These prediction tools can help health care providers initiate and tailor treatment plans to promote adherence and improve HIV-related outcomes.
Recent advances in AI-integrated wearable biosensors combining physiological monitoring (e.g., blood levels) with machine learning algorithms have shown promise for chronic conditions.11,12 AI-integrated wearable biosensors can be beneficial for continuous monitoring of ART levels, CD4 count, and viral load in real time with alerts and rapid results communicated to patients and health care providers through connected apps. Chatbots and virtual assistants, powered by AI, provide interactive platforms for delivering information on HIV prevention, testing, and treatment.13 Chatbots can also support ART and PrEP adherence by sending reminders, tracking medication intake, and sending motivational messages. For example, in a study examining the feasibility and acceptability of an AI chatbot promoting PrEP uptake in 14 men having sex with men, 13 (93%) rated the chatbot quality as high and all of the participants perceived that the chatbot helped reduce stigma.14
Despite the potential benefits of integrating AI in HIV prevention and treatment, there are several challenges that must be addressed to promote equitable access to digital health care technology for rural and underserved communities. Guided by Picker’s Patient-Centered Care Framework, this article discusses potential AI challenges within the context of HIV prevention and treatment that are unique to rural and underserved populations, including older adults and those with cognitive and sensory impairments. This article ends with clinical strategies for promoting equitable access to digital health care tools and inclusive AI adoption, preventing the amplification of existing HIV disparities in rural and underserved communities.
Patient-Centered Care Framework
The Patient-Centered Care Framework guides health care providers in understanding patients’ needs and preferences while encouraging patients to have an active role in their care and decision-making. Picker proposed eight principles of the Patient-Centered Care Framework:15
- Fast access to reliable health care advice.
- Effective treatment by trusted professionals.
- Continuity of care and smooth transitions.
- Involvement and support for family and caregivers.
- Clear information, communication, and support for self-care.
- Involvement in decisions and respect for preferences.
- Emotional support, empathy, and respect.
- Attention to physical and environmental needs.
The framework is relevant for integrating AI and HIV prevention and treatment services in rural and underserved communities, because it also focuses on equity access, personalization, care coordination, and communication.15 First, it is critical to address social determinants of health (e.g., transportation, language barriers, low health literacy) when designing AI tools to prevent the risk of amplifying disparities rather than reducing them. Second, data can be used to generate AI options based on patient preferences, and such personalization may improve use of AI and health outcomes for patients with cognitive and visual or auditory impairments. Third, AI integration with telehealth can be beneficial for patients living in remote areas with lack of transportation and/or limited mobility.
Assessments performed using AI tools can send alerts to health care providers to coordinate referrals for management of comorbid conditions, while enabling continuous monitoring during transition of care. Lastly, AI medical jargon can be converted to culturally and age-appropriate language, which can improve communication and address health literacy as a driver of health inequities. Table 1 summarizes the Patient-Centered Care principles and how AI can be applied to advance HIV care in rural and underserved communities, with equity safeguards and illustrative metrics.
AI Challenges in Underserved Communities
Despite the opportunities to integrate and embed AI tools in HIV care, there remains continued challenges for rural and underserved communities. More broadly, there are technology limitations (e.g., digital infrastructure and technology acceptance), workforce training and organizational support, socioeconomic challenges, and infrastructural deficiencies. Many rural communities lack reliable digital infrastructure (e.g., wireless fidelity (Wi-Fi), internet) necessary to support and implement these AI tools.16,17 Robust digital infrastructure is needed to implement AI tools with data processing for accurate diagnostics, treatment, and care adherence. Without these, individualized care will not be achieved with these tools to improve outcomes for this underserved population.17
Inherently, there are technology acceptance issues, too. Individual patients in rural areas may have little to no experience with digital technologies, resulting in digital literacy gaps.18,19 Additionally, trust issues may be present with data management concerning privacy with sensitive information and the lack of transparency to explainable AI (how data are used to inform machine learning algorithms). Trust and data security issues may limit an already marginalized population from seeking health care and accepting the AI technology interventions engineered for them.20
AI literacy gaps are not unique to patients but are found in health care professionals as well. Workforce training barriers are prevalent where health care professionals lack the educational training to effectively deploy the AI tools. Prior research identified workforce fundamental knowledge of AI was a major barrier followed by organizational support.21 Organizational support systems that facilitate workforce training in AI and related tools within rural communities can promote technology acceptance and help health care professionals leverage these tools to enhance HIV care.21
Socioeconomic challenges contribute to engaging AI technology opportunities in rural areas for PWH. Several variables exist where the high incidence of poverty and public transportation hinder this highly stigmatized population from seeking necessary health care,22,23 thus making it difficult to target preventive AI solutions for this community. Moreover, AI tools are often biased, so development and implementation must integrate culturally competent care strategies to promote inclusivity to encourage acceptance of these technologies.24
Infrastructural deficiencies and costs to isolated rural communities is another significant barrier to AI technology opportunities. As previously discussed, communities lack reliable digital infrastructure to deploy AI tools successfully.19,21 Without the needed infrastructure and costs to build the infrastructure, AI technologies will not come to fruition further, marginalizing these underserved communities.19
Clinical Implications for Equitable Access to AI Technologies
Integration of AI into HIV care, especially in rural and underserved communities, may improve health outcomes, appointment attendance, and medication adherence if implementation is intentional. Patients may have mistrust about and concerns about the use of AI in health care management, worries about loss of control in their care, and feel intimidated about new technology if benefits and limitations of this technology are not explained to them.25 Clinical staff should use easy-to-understand language when explaining AI and provide patients with an opportunity to ask questions about how AI will be used in their care.26 It is important that patients understand this technology can be used to identify trends and inform clinical decisions, but it is not used to replace shared decision-making between the patient and provider. Clinical staff and health care organizations should inform patients who will have access to data generated by AI and inform them when they are interacting with AI (i.e., chats that use generative learning).27 Transparency and education about the integration of AI may empower patients and increase comfort with the use of AI in their care.
AI can promote collaboration between interdisciplinary team members who provide care for people who have HIV in rural and underserved communities. Providers and pharmacists can leverage clinical decision support systems (CDSS), including machine learning algorithms, to assess ART use and identify strategies to improve medication adherence.28 If medication is not taken as prescribed or there are missed/late refills, virologic suppression of HIV may not be achieved. Providers and pharmacists may contact the patient, or delegate to another staff member, to ask follow-up questions to identify barriers for taking medication as prescribed. Interventions like pill organizers, financial resources, and use of mail order services may be beneficial for these patients.
Peer educators/navigators, social workers, and other clinical staff can use AI to identify individuals who are at risk for falling out of HIV care and people who are at risk for HIV. People who are identified as being at high risk for falling out of care may require additional contact by peer educators to remain engaged, or need additional social services like food, clothing, or housing assistance. Similar interdisciplinary collaboration using machine learning can help identify individuals who are at risk for HIV acquisition and improve medication adherence for persons taking PrEP. This information could be used by clinic staff, peer educators, or community educators to develop interventions – such as community-based testing, alternative delivery methods for medication, and appointments to decrease risk for HIV.
While AI can be beneficial for addressing barriers to HIV treatment and PrEP for rural and underserved communities, it is critical for health care organizations to promote use of AI-based privacy-preserving techniques that ensure protection of patient data. For example, access control and role-based authentication within health care settings and research institutions can help prevent unauthorized access to patient data. Whether integrating AI tools for clinician support and/or clinical care, health care providers should inform patients about the use of AI tools, including how their data will be used, stored, analyzed, and accessed by a third-party. For data sharing, AI tools should be compliant with the Health Insurance and Portability and Accountability Act (HIPAA) and include deidentified data and provider-patient communication regarding what data will be shared and how the data will be used across health care entities. Privacy techniques for AI use in HIV care for rural and underserved populations should also include use of encrypted devices and servers to further safeguard patient health care information. When leveraging AI to support HIV treatment and prevention strategies, health care providers should collaborate with local stakeholders and technology developers to create tools that reduce bias and ensure patient data reflects diverse demographic groups.
Building equitable technology access in resource-limited settings requires strong policies, infrastructure, and community-driven strategies. Collaborations between rural organizations, health care organizations, federal and state agencies, technology companies, and people who have HIV can help accelerate deployment and evaluation of digital infrastructure. In addition, policies are needed to allocate funds and resources to support AI integration into rural health care systems and reduce disparities in digital access. Reliable open-access HIV prevention and treatment education modules can reduce licensing costs, which may foster sustainable technology adoption for rural and underserved communities. In addition, community-driven strategies (e.g., shared digital access hubs and multi-sector community-based digital training programs) may help build AI technology infrastructure in resource-limited settings. Community participation in HIV technology development and integration should prioritize continuous feedback loops from diverse stakeholders to ensure technology accounts for social determinants of health and include accessibility features (e.g., screen readers) that support users with disabilities.
Conclusion
People living in rural and underserved communities continue to experience greater barriers to HIV treatment and PrEP access due to socioeconomic factors, scarce resources and trained health care providers, transportation challenges, stigma, distrust of health care systems, etc.4,6,29 AI offers transformative potential in closing gaps in HIV prevention and treatment, supporting global efforts to achieve the UNAIDS 95-95-95 targets by 2030. AI-driven tools may improve health outcomes for people in rural and underserved communities by predicting HIV risk, monitoring health metrics in real time, and reducing stigma through interactive platforms.
When implemented intentionally, AI can support predictive analytics, clinical decision-making, and interdisciplinary collaboration to address barriers in treatment and PrEP uptake. However, equitable implementation remains a critical challenge, particularly in rural and underserved communities where many experience barriers to digital access. Success depends on transparency, patient education, and strong privacy protections to build trust and prevent misuse of sensitive data. Equitable adoption requires policies, infrastructure, and community-driven strategies that prioritize accessibility, inclusivity, and bias reduction. By combining technology with patient-centered care and stakeholder engagement, AI can help close gaps in HIV prevention and treatment without amplifying existing disparities.
| Table 1. Patient-Centered Care Principles and AI Applications for HIV Care in Rural Communities | |||
|---|---|---|---|
| Patient Centered Care Principle | AI Applications for HIV Care (Rural & Underserved) | Equity Safeguards | Outcome Metrics |
| Fast access to reliable health care advice | Telehealth triage; asynchronous messaging; AI prioritization for mobile clinics; offline apps; Geospatial AI | Device & connectivity subsidies; low-bandwidth design; accessibility compliance | Telehealth utilization; appointment completion; geographic reach; reduction in disparities; time from diagnosis to first visit; ART initiation within 30 days; loss-to-follow-up reduction |
| Effective treatment by trusted professionals | Clinical decision support for ART changes; drug–drug interaction flags; resistance prediction | AI augments clinicians; HIPAA adherence, encryption, and consent management | Adverse event reduction; trust and perceived respect scores; complaint rates; engagement persistence |
| Continuity of care and smooth transitions | Predictive models for missed visits; dynamic reminder schedules; retention risk flags; AI triage; care navigation; medication compliance; automated referrals; interoperability tools linking clinics and providers | Bias audits; fairness constraints; outreach prioritization policies | Time to linkage-to-care; successful referral completion; care continuity rates; medication adherence; Increased patient quality of care/patient outcomes |
| Involvement and support of family and caregivers | Predictive analytics for caregiver burden; offline-first AI apps; AI-powered communication platforms (SMS, WhatsApp, voice bots); AI-driven personalized learning modules; mental health screening for caregivers | Consent management, caregiver scenarios in education and training; HIPAA compliant data-sharing | Caregiver participation in telehealth sessions; caregiver burnout and stress, caregiver knowledge pre-post education, patient and caregiver satisfaction |
| Clear information, communication, and support for self-care | Chatbots that simplify HIV info; plain-language med instructions; multimodal education (voice/text); translation | Readability testing; inclusive languages; guardrails to prevent hallucinations; transparent sourcing | Knowledge gains; health literacy scores; reduced misinformation reports |
| Involvement in decisions and respect for preferences | AI-powered preference capture, adaptive reminders, culturally tailored content, decision aids explaining ART choices & side effects; individualized risk calculators; visual tools | Community co-design; cultural validation; opt-in/opt-out controls; explainable AI; bilingual content; conflict-of-interest transparency | Patient satisfaction, trust in care, adherence aligned with preferences, decisional conflict reduction; ART regimen adherence; |
| Emotional support, empathy, and respect | AI screening for depression/substance use; stigma-sensitive messaging; crisis routing | Ethical algorithms; emergency escalation protocols; trauma-informed content | Mental health screening completion; linkage to behavioral health; decreased depressive symptoms scores on Patient Health Questionnaire – 9 |
| Attention to physical and environmental needs | Automated plain-language rewriting; pictograms; voice companions; tailored numeracy support | Readability thresholds; usability testing with target populations | Readability scores; task success; error reductions |
Shameka L. Cody, PhD, AGNP-BC, PMHNP-BC, FAAN, is an Associate Professor at The University of Alabama Capstone College of Nursing. She is a board-certified Adult-Geriatric and Psychiatric Mental Health Nurse Practitioner and the Principal Investigator of the NIH-funded REST Study which examines the efficacy of an evidence-based sleep intervention on insomnia and cognitive outcomes in older adults with HIV. Dr. Cody is an affiliate of the University of Alabama at Birmingham Center for AIDS Research, and she demonstrates an interdisciplinary clinic-to-community approach to eliminating HIV stigma and expanding access to behavioral health services in rural communities. With more than 17 years of clinical experience, she continues to provide advanced practice care to rural and underserved patients across Alabama.
