Efficacious drugs for the treatment of HIV/AIDS have been widely available in high-income countries since 1996, and are becoming increasingly available in low-income countries. The life expectancy of people with HIV using antiretroviral therapy (ART) is now almost identical to that of people without HIV.1 Moreover, the risk of HIV transmission is reduced considerably for successfully treated patients.2 However, despite a marked reduction in side-effects and complexity of combination ART regimens over the past two decades, suboptimum intake of drugs (faulty execution) and premature discontinuation (non-persistence) of combination ART continue to compromise treatment effectiveness.3 Non-adherence can lead to poor patient outcomes, the development of drug-resistant virus, fewer treatment options because of drug resistance, and increased transmission risks of viral strains, including resistant ones.4, 5, 6, 7, 8, 9 Hence, supporting patients' adherence is an important objective from a patient and public health perspective, and essential for achieving the UNAIDS 90-90-90 targets.10
For the long-term success of combination ART and its consequent effect on the spread of HIV, suboptimal adherence has to be addressed before virological failure occurs. Although results from meta-regression analyses suggest that the quality of adherence support provided to patients has a large influence on viral suppression rates,11, 12 little direct experimental evidence shows that adherence interventions have a sustained effect on adherence and—more importantly—on viral loads and CD4 cell counts.13, 14 A Cochrane review did not identify any low risk of bias trials of HIV adherence interventions in high-income countries that provided evidence of intervention effects on adherence and clinical outcomes such as viral load. Two trials were identified in low-income countries, the results from which showed promising effects on viral load.15 Moreover, no evidence is available that shows effective HIV treatment adherence interventions yield benefits for society in terms of cost-effectiveness.16 Our updated search of the scientific literature did not yield additional evidence.
Research in context
Evidence before this study
We searched for effectiveness and cost-effectiveness evidence from trials done in high-income countries, with at least 12 months follow-up including a clinical outcome that focused on adult HIV-infected patients. Interventions had to promote autonomous behaviour (ie, directly observed therapy interventions were excluded) and treatment simplification studies (eg, once-daily versus twice-daily medication) were excluded. For evidence on effectiveness of the interventions we searched MEDLINE, PsycINFO, and Embase with no language restrictions for articles published between January, 2013, and October, 2016, using the terms (“HIV” or “HAART” or “cART” or “Antiretroviral”) and (“adherence” or “compliance” or “persistence” or “concordance”) and (“viral load” or “virologic failure” or “CD4”) in the title or abstract, and (“random*” or “clinical trial”) in all text, and (“2013” or “2014” or “2015” or “2016”) in the year. We identified 529 unique titles, of which 27 assessed an adherence intervention. Only one was an eligible trial, assessing the Managed Problem Solving (MaPS) intervention, which noted that MaPS improved adherence. A particular strength of the trial was the high consent rate; possible weaknesses were differential attrition and a missing data imputation method that deemed missing data to equal treatment failure. No cost-effectiveness analysis was reported. For evidence on cost-effectiveness of adherence interventions, we searched the same databases and date range as above with the terms (“HIV” or “HAART” or “cART” or “Antiretroviral”) and (“adherence” or “compliance” or “persistence” or “concordance”) and (“Cost Analysis” or “Cost Effectiveness” or “Cost Benefit” or “Cost Utility” or “Cost Minimi#ation” or “Economic Evaluation”) in the title or abstract, and (“2013” or “2014” or “2015” or “2016”) in the year. We identified 137 unique titles and abstracts, of which only one was an eligible study that reported the cost-effectiveness of a computer-delivered intervention to promote adherence to HIV medication (FL, USA). This assessment was, however, based on effectiveness data from a subgroup analysis in a short-term intervention feasibility study. Further limitations were that the effectiveness data was derived from self-reported adherence and did not line up with the effectiveness input in the economic model. Thus, these searches did not identify any adherence interventions from high-quality, long-term trials, and economic assessment that provided evidence of effectiveness and cost-effectiveness.
Added value of this study
To our knowledge, this multicentre, randomised controlled trial and economic model is the first to show that our adherence intervention Adherence Improving self-Management Strategy (AIMS) produced meaningful effects on viral load and was cost-effective in a high-resource setting, compared with treatment as usual. The findings from the study showed that HIV treatment adherence interventions can increase quality-adjusted life-years (QALYs) while saving resources, even when compared with medium-to-high-quality treatment-as-usual adherence support. Moreover, AIMS required few resources because it has been adapted to fit in routine HIV clinic services, which should facilitate implementation in routine care.
Implications of all the available evidence
HIV treatment adherence interventions, such as AIMS, can benefit patients, even in high-resource settings, and lead to gains in QALYs while saving resources. AIMS seems at present to be the only adherence intervention for which the effects have been replicated in consecutive trials. The economic evaluation also provided robust evidence on cost-effectiveness. Implementation of AIMS in routine clinical care is therefore recommended.
In 2003, we developed the Adherence Improving self-Management Strategy (AIMS), based on empirical literature, behavioural theories, and input from health-care professionals and patients.17 AIMS is a nurse-delivered, one-on-one behavioural intervention that incorporates adherence feedback from electronic medication monitors (Medication Event Monitoring System [MEMS]-caps; an electronic pill-bottle cap that registers date and time of bottle opening) and is designed to fit into routine clinic visits. After a successful pilot study17 that highlighted its acceptability, feasibility, and effects on adherence, we did a single-centre randomised controlled trial with treatment-experienced patients.18 Although powered to detect an effect on adherence (the primary outcome), this trial also provided tentative evidence of improved viral suppression rates (a secondary outcome). However, this trial had a homogeneous patient group and a short follow-up of 7 months. Showing clinically relevant effects on viral load in a high-quality pragmatic trial with a long follow-up, and a heterogeneous group of patients and HIV clinics, could provide conclusive evidence that AIMS is effective. Moreover, showing that AIMS is also cost-effective would be important for policy makers, as well as for adherence intervention research generally, given the scant evidence of the economic benefits of adherence interventions.
Effective HIV treatment adherence interventions should benefit patient and public health, and reduce health-care expenditures; yet, experimental evidence of these benefits is scarce. This report describes findings from our study that assessed the effectiveness of AIMS, and the results of a Markov model assessing the cost-effectiveness of AIMS over a lifetime horizon.
