The mean percentage of days which participants maintained adherence to apixaban treatment was measured for each arm. Adherence to apixaban = number of units of adherence *100 / total number of eligible days for the time period from first dose date, up to 169 days. Unit of adherence: A 24-hour window where the treatment is taken as prescribed, ie, 1 tablet (5 mg or 2.5 mg, as appropriate) 2 times a day. If only one dose is missed in 24-hours, it is still considered as a unit of adherence. Adherence up to 24 weeks was calculated as the percentage of adherence units within that period. If a participant discontinued from the study before 24 weeks, the denominator time period was censored at the earlier of last dose date or discontinuation date for discontinuation due to reasons unrelated to participant adherence, such as withdrawn consent, or AE; otherwise, the period was censored at the minimum of 169 days and last dose date + 30 days.

Primary efficacy analysis set (Week 24), which consists of all randomized participants

The mean adherence to apixaban treatment during the first 24 weeks was measured between the standard of care (SOC) information and Additional Education Program (AEP) arms and expressed as a percentage. Adherence to Apixaban = number of units of adherence *100 / total number of eligible days for the time period.

Primary efficacy analysis set (Week 24) with available data at both study days 85 and 169

The mean percentage of days which participants maintained adherence to apixaban treatment was measured for each arm. Adherence to apixaban = number of units of adherence *100 / total number of eligible days for the time period from first dose date, up to 169 days. Unit of adherence: A 24-hour window where the treatment is taken as prescribed, ie, 1 tablet (5 mg or 2.5 mg, as appropriate) 2 times a day. If only one dose is missed in 24-hours, it is still considered as a unit of adherence. Adherence over 24 weeks was calculated as the percentage of adherence units within that period. If a participant discontinued from the study before 48 weeks, the denominator time period was censored at the earlier of last dose date or discontinuation date for discontinuation due to reasons unrelated to participant adherence, such as withdrawn consent, or AE; otherwise, the period was censored at the minimum of 169 days and last dose date + 30 days.

All treated participants in 24 to 48-week period. Participants that had not been using the EMD consistently throughout the study were excluded from the analysis.

Logit analyses were conducted on the Primary Efficacy Set to identify non-adherence predictors of 20% or more (vs. at least 80% adherence) at 24 weeks. In the Primary SOC group, alcohol use, Mini-Mental State Evaluation (MMSE) score, UK standard occupational classification, and type of atrial fibrillation were retained in the model (p-value <= 0.2). In the Additional Educational Program group, alcohol use, type of atrial fibrillation, age and Vitamin K Antagonists (VKA) status were retained in the model (p-value <= 0.2). Odds ratios are presented for predictors of non-adherence.

Primary Efficacy Set participants with evaluable data at week 24

AEs with onset date from day 1 through week 24 are included in this summary. AE=any new unfavorable symptom, sign, or disease or worsening of a preexisting condition that may not have a causal relationship with treatment. SAE=a medical event that at any dose results in death, persistent or significant disability/incapacity, or drug dependency/abuse; is life-threatening, an important medical event, or a congenital anomaly/birth defect; or requires or prolongs hospitalization. Treatment-related=having certain, probable, possible, or missing relationship to study drug.

All randomized participants. Participants in the safety analysis set were categorized according to the counseling actually received

Adverse events with onset date after 24 weeks are included in this summary. AE=any new unfavorable symptom, sign, or disease or worsening of a preexisting condition that may not have a causal relationship with treatment. SAE=a medical event that at any dose results in death, persistent or significant disability/incapacity, or drug dependency/abuse; is life-threatening, an important medical event, or a congenital anomaly/birth defect; or requires or prolongs hospitalization. Treatment-related=having certain, probable, possible, or missing relationship to study drug.

All randomized participants remaining in the study after week 24. All randomized participants. Participants in the safety analysis set were categorized according to the counseling actually received