Non-adherence to treatment: a methodological approach to compensating for its impact in chronic infections, using tuberculosis as a model disease

Non-adherence to treatment costs hundreds of billions of dollars globally per year and is particularly problematic in chronic diseases, which require lengthy treatment. In communicable chronic diseases, poor adherence risks the development of drug resistance and an extended period of transmission.

National and international guidance on adherence frequently tacitly assumes that all medication doses must be taken for successful treatment. Clinical trials and observational studies often define adherence with simplistic thresholds e.g. 80% of doses taken. The reality is complex, requiring granular analyses of the relationship between adherence and treatment outcomes.

This work programme (currently funded by the Medical Research Council, £1.3mn) will determine how adherence patterns should influence our approach to treatment in order to maximise favourable outcomes, starting with using tuberculosis as a model disease and then moving into other areas. The objectives of this work are:

1) Describe statistically how and when non-adherence occurs and the relationship between adherence and outcomes.

2) Determine how adherence patterns influence the temporal development of drug resistance.

3) Adapt a mathematical modelling framework to assess the impact of adherence on optimal dosing.

Our findings will aid clinicians treating patients, behavioural scientists planning interventions to improve adherence, pharmaceuticals companies designing drugs to maximise ‘forgiveness’ for non-adherence, triallists designing treatment regimens, and guide policymakers on interpreting trial results in the light of operational efficacy.