HIV Medication Adherence #

Key Result: Single-tablet regimens reduce treatment discontinuation by roughly 13% compared with multi-tablet regimens: simplifying the behavior outperforms prompting it (peer-reviewed meta-analysis). BS-0073

Background #

Antiretroviral therapy works only if it is taken consistently, which makes daily adherence one of the highest-stakes recurring behaviors in medicine. The intuitive model of adherence failure, held by many clinicians, health systems, and app builders, is that patients forget. That model has a natural product response: reminders. SMS prompts, pillbox alarms, and adherence apps have been deployed at scale on the premise that memory is the limiting factor.

The evidence tells a different and more sobering story. For many people living with HIV, adherence breaks down not in memory but in context: stigma that makes each dose a potential disclosure, housing instability that disrupts any routine, food insecurity when a regimen requires eating, and the identity weight of a daily act that reasserts an unwanted label. A reminder arrives, and the barrier it was built to remove was never the barrier at all.

What actually drives adherence #

The behavior is taking medication every day, indefinitely, performed by patients whose circumstances vary enormously. Where the behavior fails, one or more specific factors is limiting, and the interventions that work are the ones aimed at the limiting factor:

• Simplifying the behavior has the strongest evidence. Single-tablet regimens improve adherence and reduce discontinuation relative to multi-tablet regimens across multiple meta-analyses (peer-reviewed). Fewer pills mean fewer decisions, fewer private moments required, and less regimen burden: the behavior itself gets easier for every patient, in every context.
• Reminders show mixed results because they target the wrong constraint. SMS reminder interventions produce inconsistent adherence effects in meta-analyses (peer-reviewed). Where a patient’s barrier is stigma, food insecurity, or structural access, a prompt changes nothing; where fit already exists, a prompt can help at the margin.
• The primary barriers are contextual and structural. Stigma, food insecurity, identity conflict, and access problems - not forgetting - dominate the barrier profile in many populations (peer-reviewed).

This is the sense in which reminders are configuration, not strategy. Configuration tunes a system whose fundamentals already work. Strategy changes the fundamentals. Choosing to develop and deploy single-tablet regimens was a strategic act: it redesigned the behavior. Sending a text about the old behavior was configuration, and the mixed evidence reflects exactly that.

Behavior Fit Assessment #

These ratings are analyst examples of a Behavior Fit Assessment describing the barrier profile where adherence fails; they are not measurements of any specific population. Dispositional Fit can be low for a population with persistent disclosure avoidance, treatment aversion, or sensitivity to the status consequences of being identified as HIV-positive. An unwanted self-label can intensify those tendencies, but it is not the definition of the dimension. Capability Fit is high for most patients because taking a prescribed dose generally requires little specialized ability or skill; cognitive, swallowing, or dexterity impairments should be assessed separately where relevant. Context Fit can be low when medication access or storage, housing stability, privacy, food, water, and routine-supporting conditions are unreliable. Side effects belong in a fuller person-side BSM diagnosis unless they demonstrably impair the physical or cognitive act of dosing; they should not be used as a proxy for missing skill. The practical use of the assessment is diagnostic: identify which factor is limiting for a given population, then design the intervention around that factor rather than defaulting to prompts. That is the discipline behind context engineering.

Results #

• Single-tablet regimens improve adherence and reduce discontinuation compared with multi-tablet regimens across multiple meta-analyses (peer-reviewed). BS-0073

• SMS reminder interventions show mixed results in meta-analyses; reminders alone do not address the primary barriers of stigma, food insecurity, identity conflict, and structural access (peer-reviewed).
• The strategic pattern across the evidence: interventions that simplify the behavior itself (fewer pills, fewer decisions) carry stronger evidence than interventions that prompt a behavior that does not fit (peer-reviewed).

Limitations #

Adherence measurement methods vary widely (self-report, pill counts, pharmacy refills, electronic monitoring) and effect sizes differ by method, so cross-study comparisons carry noise. Barrier profiles differ dramatically across populations: urban versus rural, food-secure versus insecure, disclosed versus undisclosed status. Adherence thresholds also vary by regimen. And because effective programs are usually multi-component, isolating which element drives an observed improvement is genuinely difficult; the single-tablet-regimen findings are valuable partly because the intervention is a single, clean change to the behavior itself.

Lessons #

1. Reminders are configuration, not strategy. They can support a behavior for which fit already exists; they cannot create fit. Deploying prompts against contextual and structural barriers is a category error, and the mixed SMS evidence is the predictable result, a pattern examined more broadly in Why Nudges Fail.
2. Find the limiting factor before designing the intervention. Dispositional, capability, and context fit can fail for different patients for different reasons. The intervention that works is the one aimed at whichever factor is binding, not the one easiest to ship.
3. The strongest lever is simplifying the behavior itself. Single-tablet regimens improved outcomes by removing burden for every patient regardless of barrier profile. When you can redesign the behavior instead of exhorting the person, redesign the behavior.

Sources #

• Single-tablet regimens and outcomes (PMC, meta-analysis)
• Interventions to improve ART adherence (Lancet HIV, 2017; includes SMS interventions) (PubMed)
• Global AIDS update (UNAIDS, 2024)
• Evidence Ledger: BS-0073