Development of risk assessment tool in predicting HIV treatment failure and the potential use of digital health technology in enhancing antiretroviral therapy adherence

Abstract

HIV treatment failure is a burden to a healthcare system, including economic, public health and clinical impacts. Managing individuals with treatment failure is expensive due to the need for additional medical interventions, increased hospitalisation and the use of second or third-line antiretroviral medications, which are often more costly. Patients who do not respond well to therapy face an increased likelihood of acquiring severe HIV-related diseases, opportunistic infections, a deterioration in overall health and an increased risk of mortality. The main objective of this research was to develop a risk assessment tool (RAT) to predict HIV treatment failure in HIV patients started on antiretroviral therapy (ART). The research consisted of four research projects. The first project was a retrospective cohort study conducted among 355 HIV patients to identify types and causes of drug-related problems (DRPS) in HIV treatment failure patients. Then, it was followed by a second project, which aimed to identify predictors of HIV treatment failure among HIV patients who started ART. Both findings in these projects were used to develop RAT in the third project. The third project consists of tool development and tool validation. The internal validation was used for tool validation. All these projects were conducted at the Infectious Disease Clinic, Hospital Sungai Buloh, Selangor and Hospital Tuanku Ja’afar, Negeri Sembilan, using secondary data from patients’ medical records. Adult patients older than 18 years old and who started ART were randomized for further analysis. Using the Pharmaceutical Care Network European classification version 9.1, the most common types of problems identified were related to treatment effectiveness (P1: n = 169, 59.93%; 1.84 DRPs per patient) and treatment safety (P2: n = 111, 39.36%; 1.21 DRPs per patient) in the HIV treatment failure group. The most common types of DRPs causes were patient-related (C7: n = 367, 69.11%; 3.99 DRPs per patient), followed by other (C9: n = 105, 19.77%; 1.14 DRPs per patient). Eight variables were identified as predictors associated with HIV treatment failure, including history of opportunistic infections, history of lost-to-followup, male gender, the use of reminders, duration of ART, isoniazid prophylaxis therapy, history of non-adherence and social support. In the final RAT model, four variables, including patients with a history of opportunistic infections (aOR 2.62; 95% CI 1.17- 5.87), a history of non-adherence (aOR 22.24, 95% CI 7.02-70.44), social support (aOR 0.23, 95% CI 0.06-0.84) and the use of reminders (aOR 0.09, 95% CI 0.03-0.23), were identified as significant predictors for HIV treatment failure. The final project was to identify digital health technology to enhance ART adherence since non-adherence was one of the significant predictors of HIV treatment failure. This systematic review and meta-analysis project identified that the combination approach of digital health technology could enhance ART adherence compared to the single approach. The developed RAT could predict patients with a high risk of HIV treatment failure and digital health approach can be one of the potential interventions to reduce the risk of treatment failure, and the RAT may also assist healthcare professionals to enhance HIV treatment and care.