Abstract
Habit, a psychological process that automatically generates urges to perform a behavior in associated settings, is potentially an important determinant of medication adherence. Habit is challenging to measure because, as a psychological construct, it cannot be directly observed. We describe a method of using routinely available objective adherence data from electronic data capture (EDC) to generate a behavior-based index of adherence habit and demonstrate how this index can be applied.
Our proposed habit index is a "frequency in context" measure. It estimates habit as a multiplicative product of behavior frequency (generated from weekly percentage adherence) and context stability (inferred from time of nebulizer use). Although different timescales can be used, we chose to generate weekly habit scores since we believe that this is the most granular level at which context stability can be reasonably calculated.
A hallmark of habit is to predict future behavior, hence we used time series method to cross-correlate the habit index with nebulizer adherence in the subsequent week among 123 adults with cystic fibrosis (52, 42.3% female; median age 25 years) over a median duration of 153 weeks (IQR 74-198 weeks). The mean cross-correlation coefficient (
) between the habit index and subsequent adherence was 0.40 (95% CI 0.36-0.44). Adjusting for current adherence, the unstandardized regression coefficient (
) for the habit index was 0.30 (95% CI -1.04 to 1.65).
We have described a pragmatic method to infer "habit" from adherence data routinely captured with EDC and provided proof-of-principle evidence regarding the feasibility of this concept. The continuous stream of data from EDC allows the habit index to unobtrusively assess "habit" at various time points over prolonged periods, and hence the habit index may be applicable in habit formation studies.