It is easy for health professionals to miss adherence problems because patient self-reports of adherence tend to be exaggerated (20,21) due perhaps both to a recall bias and a desire to please the provider and avoid criticism. Some patients have been known to dispose of their medication before a scheduled check on their adherence to it so as to appear to have adhered (22). Inadequate adherence coupled with biased reporting is ubiquitous across medicine (23). Conversely, patients who report problems with adherence are rarely trying to mislead their providers (24).

In addition to the misreporting of adherence by patients, estimates of adherence made by health care providers are also usually over-optimistic (25,26).Moreover, providers of health care are not able to predict very accurately which patients will adhere. Many providers believe that factors associated with socioeconomic status, such as lack of education and poverty are good predictors of nonadherence. However, predictors of adherence vary greatly across populations and settings and no one factor has been consistently associated with nonadherence across all studies (27).