Purpose: Medication nonadherence predicts hospitalization in HF patients, but the phase of nonadherence most responsible is unclear (initiation, implementation, or discontinuation). Our purpose was to determine which phase is most responsible and to identify characteristics of patients in that phase.
Methods: Adults with chronic HF were enrolled from 3 clinic sites in the US. Medication adherence over 6-months was measured using medication-event monitoring system (MEMS). Medication initiation was defined as percent of time to first dose taken, discontinuation as percent of time after the last dose was taken, and implementation as the period between. Clinical data were abstracted from medical records by nurses. Trail Making Test B was used to measure complex attention; higher scores indicate worse attention. Higher scores on the American National Adult Reading Test (ANART) indicate higher premorbid intellect. Higher scores on the PHQ-9 indicate more depression. Higher scores on the Charlson Comorbidity Index indicate more chronic conditions. All analyses used adaptive modeling methods.
Results: Of the 218 patients with MEMS data, 6 (2.8%) did not initiate their medicines immediately (median 13% of time in study). All 218 patients had some implementation phase data with most (53.7%) completing the full study. 83 (38.1%) patients met criteria for early discontinuation (≥3.1% of the full study period). Early discontinuation was the best predictor of hospitalization (risk factors shown in table). Area under the curve was very good for this model (0.75).
Conclusion: Discontinuing the medication regimen risks hospitalization in adults with chronic HF. Complex patients with multiple chronic conditions, numerous medications, depression, problems with attention and analytic skills may consider stopping medicines if not assisted to integrate the regimen into their lives.