Objective digital bottle cap monitoring in 2429 post-MI patients demonstrated that temporal irregularity of medication taking 'may serve as a practical behavioral marker for identifying patients at increased risk of future nonadherence,' with dose consistency strongly associated with adherence over time.
Key Findings
Results
Younger patients under 40 years of age had significantly lower medication adherence compared to other age groups.
Age <40 years was associated with lower adherence (OR 0.47, 99.9% CI 0.23-0.98).
Patients aged 40-50 years showed lower dose consistency (DCR 1.23, 99.9% CI 1.03-1.47).
The analysis included 2429 participants from the DAPA-MI RCT with 733,490 timepoints recorded.
Adherence was defined at the patient-day level as at least one bottle opening on a given day.
Results
Current smokers demonstrated both lower medication adherence and lower dose consistency compared to non-smokers.
Current smoking was associated with lower adherence (OR 0.67, 99.9% CI 0.51-0.87).
Current smokers also showed lower dose consistency (DCR 1.16, 99.9% CI 1.03-1.31).
Dose consistency ratio (DCR) was derived from the day-to-day variation in timing of the first recorded opening event.
Both adherence and dose consistency associations with smoking were assessed using mixed-effects regression.
Results
Patients with NYHA functional class III/IV had higher medication adherence than those with lower functional class.
NYHA functional class III/IV was associated with higher adherence (OR 1.48, 99.9% CI 1.18-1.86).
This finding suggests that greater symptom severity may motivate higher adherence behavior.
NYHA classification was among the demographic, clinical, behavioral, and contextual variables linked with digital cap data.
Results
Low 7-day average time variance in medication taking (less than 15 minutes) was strongly associated with higher adherence.
A low (<15 min) 7-day average time variance was associated with higher adherence (OR 1.99, 99.9% CI 1.78-2.23).
This was among the strongest associations observed for adherence in the study.
Morning medication use was also associated with both higher adherence and higher dose consistency (DCR 0.88, 99.9% CI 0.86-0.90).
These findings support the conclusion that temporal regularity of medication taking is a behavioral marker for adherence.
Results
Nonworking days and weekends were associated with lower adherence and lower dose consistency, respectively.
Nonworking days were identified as a feature associated with lower adherence.
Weekends were associated with lower dose consistency.
These contextual factors suggest that disruptions to daily routine negatively impact medication-taking behavior.
Missed-dose streaks and routine disruptions were also associated with worsening consistency over time.
Results
Serious adverse events were associated with lower medication adherence.
Serious adverse events were listed among features associated with lower adherence in the mixed-effects regression model.
The analysis linked digital cap data with demographic, clinical, behavioral, and contextual variables to assess adherence factors.
The study was a post hoc analysis of the DAPA-MI phase 3 double-blind placebo-controlled RCT.
Results
Patients with severe angina grading had higher dose consistency compared to those with less severe angina.
Severe angina grading was associated with higher dose consistency (DCR 0.83, 99.9% CI 0.74-0.94).
A DCR below 1.0 indicates lower day-to-day variation in medication timing, representing higher consistency.
This finding parallels the NYHA class III/IV adherence finding, suggesting symptom severity may reinforce medication routine.
Results
Early high dose consistency was associated with better adherence maintained over time, while missed-dose streaks were associated with worsening consistency.
Patients with early high dose consistency tended to maintain better adherence over time.
Missed-dose streaks and routine disruptions were associated with worsening consistency.
These temporal patterns were identified using 733,490 timepoints recorded by the digital bottle cap across 2429 participants.
The findings suggest dose consistency may serve as a predictive behavioral marker for future nonadherence.
Methods
The study used objective digital bottle cap data from a large RCT to quantify both adherence and dose consistency at scale, unlike studies relying on self-reported or prescription data.
The digital bottle cap recorded when study drug bottles were opened, providing 733,490 timepoints.
The DAPA-MI trial evaluated 10 mg dapagliflozin versus placebo once daily in adults hospitalized for MI.
This post hoc analysis included 2429 participants.
The authors stated this study demonstrated 'the value of objective, continuous digital medication adherence behavior monitoring within a large RCT.'
What This Means
This research suggests that how consistently patients take their medication — not just whether they take it — is a powerful predictor of long-term adherence after a heart attack. Using an electronic bottle cap that recorded every time a medication bottle was opened, researchers tracked over 733,000 medication-taking events across nearly 2,500 patients enrolled in a large clinical trial. This approach provided a much more detailed and objective picture of medication behavior than traditional methods like patient self-reports or pharmacy refill records.
The study found several factors linked to missing doses or taking medication at irregular times. Younger patients (under 40), current smokers, and days when normal routines were disrupted — such as weekends or non-working days — were all associated with poorer adherence or less consistent timing. Interestingly, patients with more severe heart symptoms (higher NYHA class or severe angina) tended to be more adherent, possibly because they were more aware of the importance of their medication. Taking medication in the morning and doing so at nearly the same time each day were both associated with better adherence overall.
One of the most practically significant findings is that patients who took their medication at consistent times early in the study tended to stay more adherent over the entire follow-up period, while those who experienced gaps or disruptions in their routine showed worsening consistency over time. This research suggests that monitoring the regularity of medication timing — rather than just counting missed doses — could be a useful early warning signal for identifying patients who may be at risk of stopping their medication, potentially allowing healthcare providers to intervene before adherence breaks down.
Nowojewski A, Elizarova O, Rutgersson A, Zarodkiewicz A, Jiang Y, Olsson A, et al.. (2026). Medication Adherence Behavior Assessed Using a Digital Bottle Cap: Retrospective Cohort Study.. Journal of medical Internet research. https://doi.org/10.2196/91299