Medication nonadherence - definition, measurement, prevalence, and causes: reflecting on the past 20 years and looking forwards.

In 2003, Sabate's World Health Organisation report defined medication nonadherence as a phenomenon where individuals' behaviour does not correspond to prescribed treatment recommendations from their healthcare provider. This concept of nonadherence evolved beyond a categorisation of patients as adherent or nonadherent. Rather, nonadherence varies within the same individual and treatment over time, and between treatments and individuals. The type and patterns of nonadherence are key determinants of outcome with individuals with the same percentage nonadherence having different outcomes depending on their pattern of nonadherence. Often the poorest clinical outcomes occur in individuals who do not initiate medication or discontinue early, but much of the nonadherence literature remains focused on implementation. This paper provides a nuanced discussion of nonadherence which has been enabled in part by the growing availability of technologies such as electronic nonadherence monitors, new biomarkers for adherence and greater access to 'big data' (e.g., on prescription refills). These allow granular assessment of nonadherence that can be linked with biophysical markers captured using technologies such as wearables. More validated self-report measures have also become available to profile nonadherence in research and practice. Together, in-depth data on dosing and clinical measures provide an opportunity to explore complex interactions between medications, therapeutic effects and clinical outcomes. This variation in measurement and definition means that there is a more fine-grained understanding of the prevalence of nonadherence and a greater recognition of the prevalence of nonadherence, with growing evidence suggesting that approximately a fifth of patients do not initiate treatment, of those initiating treatment approximately 30%-50% of patients do not implement their treatment as prescribed and that, over long follow-up periods in some conditions 80%-100% of patients discontinue. There is potential too to better understand causes of nonadherence. New behavioural models synthesise determinants of nonadherence previously considered separately. Frameworks like the COM-B (considering individual capability, opportunity, and motivation factors) and MACO (focusing on Medication Adherence Contexts and Outcomes) emphasize the multifaceted nature of nonadherence determinants. Greater focus on dynamic processes with interplay between individual, social, and environmental influences is needed. Addressing these complexities could lead to more effective and personalised support for patients.