Abstract
Background
Despite benefits of oral anticoagulation (OAC), many individuals with diagnosed atrial fibrillation (AF) do not receive OAC.
Objective
The purpose of this study was to assess whether cardiac rhythm assessment for AF impacted use of OAC in patients with previously diagnosed AF.
Methods
VITAL-AF was a cluster randomized controlled trial conducted in 16 primary care practices assessing the efficacy of AF rhythm assessment with single-lead electrocardiogram in routine care. Patients 65 years and older were offered rhythm assessment at visits. In this secondary analysis, we evaluated rhythm assessment uptake and compared initiation and discontinuation of OAC in patients with previously diagnosed AF from intervention and control arms over 1 year.
Results
The study included 4593 patients with previously diagnosed AF (2250 intervention; 2343 control). In the intervention arm, 2022 (89.9%) completed rhythm assessment (median 2 visits with rhythm assessment) and 40.1% had ≥1 “Possible AF” result. Initiation of OAC was similar in the intervention (17.7%) and control (19.1%) arms but was influenced by the rhythm assessment result: higher with a “Possible AF” (26.1%; adjusted odds ratio [aOR] 1.62; 95% confidence interval [CI] 1.04–2.51), and lower with a “Normal” result (9.9%; aOR 0.45; 95% CI 0.29–0.71) compared to control. OAC discontinuation was similar in the intervention (6.3%) and control (7.2%) arms, with lower discontinuation with a “Possible AF” result (3.8%; aOR 0.51; 95% CI 0.32–0.81).
Conclusions
Including patients with previously diagnosed AF in a point-of-care rhythm assessment strategy did not increase overall OAC use compared to the control arm. However, the rhythm assessment result influenced both initiation and discontinuation of OAC.
Keywords: Prevalent atrial fibrillation, Cardiac rhythm assessment, Oral anticoagulation, Primary care, Pragmatic randomized controlled trial
Key Findings.
-
▪
Patients with previously diagnosed atrial fibrillation (AF) included in a point-of-care heart rhythm assessment strategy program completed single-lead electrocardiographic rhythm assessment at high rates, with 90% completing at least 1 over a 1-year period.
-
▪
Single-lead rhythm assessment did not impact the overall use of oral anticoagulation among those with previously diagnosed AF.
-
▪
The result of the rhythm assessment, however, did impact use of oral anticoagulation. Patients with a “Possible AF” result had a 62% increase in odds of oral anticoagulation initiation and 49% reduced odds of oral anticoagulation discontinuation.
Introduction
Atrial fibrillation (AF) is the most common sustained cardiac arrhythmia and is associated with increased risk of stroke.1, 2, 3 Oral anticoagulation (OAC) is effective for preventing strokes in patients with AF.4, 5, 6, 7 Despite clear benefits, about 40%–50% of individuals with a previously established diagnosis of AF (“diagnosed AF”) do not receive OAC.8, 9, 10, 11, 12, 13 Lack of physician awareness of diagnosed AF or of a guideline-based indication for OAC may contribute to the apparent underuse of OAC, in addition to concerns about bleeding risk.13, 14, 15 Moreover, among patients with diagnosed AF on OAC, discontinuation rates are high,16,17and patients who discontinue OAC have a 2- to 3-fold higher risk of ischemic stroke compared to those with continuous use.18, 19, 20
The proximate goal of AF screening programs is to identify undiagnosed AF. However, assessing cardiac rhythm may also encourage physicians to reconsider management of patients with previously diagnosed AF. The STROKESTOP (Systematic ECG Screening for Atrial Fibrillation Among 75 Year Old Subjects in the Region of Stockholm and Halland, Sweden) study reported increased OAC treatment in screening participants with known but untreated AF.21 We leveraged a large AF screening trial based in primary care practices to assess whether cardiac rhythm assessment led to new OAC prescriptions in patients with diagnosed AF. We hypothesized that rhythm assessment may increase OAC use by prompting reassessment of stroke risk and OAC eligibility or by identifying patients in whom AF was thought to have resolved.
Methods
VITAL-AF (Screening for Atrial Fibrillation Among Older Patients in Primary Care Clinics) was a pragmatic cluster randomized controlled trial assessing the efficacy of embedding AF screening with single-lead electrocardiogram (ECG) rhythm assessments into routine care for older primary care patients. The study methodology and primary outcomes have been described previously.22,23 The research protocol was approved by the Mass General Brigham Institutional Review Board. Participants provided informed consent to participate. The study was considered minimal risk, and a waiver of documentation of informed consent was granted. The research reported in this paper adhered to CONSORT guidelines.
Setting and participants
The VITAL-AF study was conducted in 16 primary care practices affiliated with Massachusetts General Hospital (MGH). Individuals aged 65 years or older who presented to a primary care practitioner (PCP) at a participating practice were included. Cluster randomization was at the practice level, with 8 practices assigned to the intervention arm (AF screening) and 8 to the control (usual care) arm. Practices enrolled patients over 12 months between July 31, 2018, and October 8, 2018. Group assignment of patients was based on the practice they first visited during the study period. To simplify implementation of rhythm assessment by intervention practice staff, we offered screening to all patients 65 years and older, including those with diagnosed AF.
This secondary analysis includes the subset of participants from the VITAL-AF study population who had diagnosed AF before their first (“index”) visit during the study period. Diagnosed AF was identified from the electronic health record (EHR) in 3 ways. (1) From a validated algorithm that required at least 2 International Classification of Statistical Diseases, Tenth Revision (ICD-10) codes for AF within the previous three years and an OAC prescription within the previous year (previously validated with positive predictive value 98.4%).22 These patients were considered to have diagnosed AF without further manual adjudication. (2) From the validated algorithm in (1) but without an OAC prescription in the previous year. (3) From an ICD-10 code for AF or atrial flutter or a 12-lead ECG with AF or flutter in the diagnostic statement during the 12-month study period. Patients identified by (2) or (3) were manually adjudicated by a clinical endpoint committee comprising 2 research nurses and a cardiologist. For patients identified by (3), manual adjudication also assessed whether diagnosed AF was transient in the setting of a secondary precipitant with a return to normal sinus rhythm within 3 months. Precipitants included cardiac surgery, acute myocardial infarction, acute infection, acute alcohol consumption, thyrotoxicosis, acute pericardial disease, acute pulmonary embolism, and other acute pulmonary pathology (eg, pneumothorax and bronchoscopy-related) ≤30 days before initial AF.24
Procedures
In intervention practices, AF rhythm assessment was performed by practice medical assistants as part of routine vital signs assessment. Patients with >1 visit could have rhythm assessment multiple times. Consenting patients placed their fingers on a single-lead AliveCor KardiaMobile ECG device (AliveCor Inc., Mountain View, CA; KardiaA1 version 1 algorithm) to conduct a 30-second AF rhythm assessment. Automated rhythm assessment result categories included “Possible AF,” “Normal,” “Unclassified,” and “No analysis (unreadable).” Following rhythm assessment, medical assistants documented the results in the EHR (Epic, Verona, WI) and were instructed to notify the PCP if a patient had a “Possible AF” result. All subsequent treatment decisions were at the discretion of the PCP. Patients in control practices received usual care during outpatient visits at the direction of their primary care provider.
Patient characteristics and outcome measures
Patient characteristics, diagnoses, cardiac utilization, procedures, and prescriptions were obtained from a central data repository at Mass General Brigham. We calculated the CHA2DS2VASc stroke risk score and ATRIA (Anticoagulation and Risk Factors in Atrial Fibrillation) bleed score for each patient.25,26
Outcomes included single-lead rhythm assessment uptake and results, initiation and discontinuation of OAC, and health care utilization. The distribution of single-lead rhythm assessment results was assessed at (1) all encounters; (2) the first encounter with rhythm assessment completed, in both cases using the final result if multiple assessments were completed at the encounter; and (3) the encounter with most “abnormal” result (“Possible AF,” “Unclassified,” “Normal” ranked from most to least “abnormal” with remaining tracings being “No Analysis”).
We used prescription dates and quantity to be dispensed to define periods on and off OAC. OAC included apixaban, dabigatran, edoxaban, rivaroxaban, and warfarin. Patients eligible for the OAC initiation outcome included those without evidence of OAC use in the 12 months before their index date. Patients were considered to have initiated OAC based on evidence of OAC use between the index date and the subsequent 12 months. Patients eligible for the OAC discontinuation outcome included those with evidence of OAC use on their index date. Among this population, patients were considered to have discontinued OAC if there was an interruption of OAC use within 12 months after the index date and no evidence of restarting OAC within 12 months.
Health care utilization outcomes included 12-lead ECGs, cardiac rhythm monitoring (Holter/patch monitors, implantable loop recorders), cardioversions, ablations, placement of a pacemaker or defibrillator, echocardiograms, inpatient admissions, emergency department visits, and cardiology visits. Twelve-lead ECGs were ascertained from ECG reports or Current Procedural Terminology (CPT) codes on the day of a primary care visit. Other utilization outcomes were assessed using CPT codes, whereas cardiology/emergency department visits and inpatient admissions were based on EHR encounters and discharge reports in the 12 months after a patient’s index date.
Statistical analysis
For descriptive data, we calculated mean ± SD, median [interquartile range], or number (percentage). Rhythm assessment results were reported among the intention-to-treat intervention group, with group assignment based on the patients’ first visit during the study period. We compared use of OAC and utilization outcomes using χ2 tests. OAC initiation and discontinuation stratified by rhythm assessment result in the intervention arm was compared to the control arm in adjusted logistic regression models controlling for CHA2DS2-VASc stroke risk score and ATRIA bleed score.25,26 In supplemental per-protocol analyses, OAC outcomes among patients completing rhythm screening were compared to all patients who did not complete rhythm assessment (all patients from intervention and control arms without a rhythm assessment) at 90 days and 365 days after the first visit with rhythm or the first visit for those not completing (Supplemental Tables S1 and S2). Patients who died during follow-up before events were excluded from OAC and utilization outcomes. Statistical significance was defined as 2-tailed P ≤.05.
Results
Over 1 year, 2250 individuals in the intervention arm and 2343 in the control arm with previously diagnosed AF had at least 1 visit at a study practice, corresponding to 6988 and 7397 encounters, respectively (Figure 1). In both intervention and control groups, the median number of visits per person was 3 [2–4]. Overall, mean age was 78.4 years, mean CHA2DS2-VASc score was 4.7, mean ATRIA bleed score was 5.0, and 67.8% were on OAC at baseline. Additional patient characteristics are given in Table 1. Patient features were well balanced between the intervention and control arms.
Figure 1.
CONSORT diagram. AF = atrial fibrillation.
Table 1.
Characteristics of patients with diagnosed AF in the intervention and control arms
| Intervention | Control | |
|---|---|---|
| Total no. | 2250 | 2343 |
| Age (y) | 78.3 ± 7.6 | 78.4 ± 7.7 |
| Age >75 y | 1411 (62.7) | 1486 (63.4) |
| Female | 927 (41.2) | 1008 (43.0) |
| Race/ethnicity∗ | ||
| White | 2003 (89.0) | 2070 (88.3) |
| Black | 52 (2.3) | 50 (2.1) |
| Hispanic | 34 (1.5) | 39 (1.7) |
| Other | 121 (5.4) | 147 (6.3) |
| Height (cm)∗ | 168.6 ± 11.5 | 167.8 ± 11.3 |
| Weight (kg)∗ | 81.6 ± 19.7 | 81.1 ± 19.8 |
| Body mass index (kg/m2)∗ | 28.5 ± 5.8 | 28.6 ± 6.0 |
| Systolic blood pressure (mm Hg)∗ | 129.2 ± 16.8 | 128.4 ± 17.0 |
| Diastolic blood pressure (mm Hg)∗ | 73.0 ± 9.3 | 72.1 ± 9.4 |
| Oral anticoagulant | 1511 (67.2) | 1601 (68.3) |
| Antiplatelet therapy | 196 (8.7) | 221 (9.4) |
| Rate control therapy, any | 2030 (90.2) | 2126 (90.7) |
| Antiarrhythmic therapy, any | 501 (22.3) | 537 (22.9) |
| Left atrial appendage occlusion | 20 (0.9) | 18 (0.8) |
| Ablation | 106 (4.7) | 94 (4.0) |
| Cardioversion | 239 (10.6) | 256 (10.9) |
| Antihypertensive | 2007 (89.2) | 2085 (89.0) |
| Hypertension | 2074 (92.2) | 2169 (92.6) |
| Myocardial infarction | 483 (21.5) | 530 (22.6) |
| Coronary artery disease | 1210 (53.8) | 1263 (53.9) |
| Diabetes mellitus | 717 (31.9) | 779 (33.2) |
| Heart failure | 1083 (48.1) | 1188 (50.7) |
| Previous stroke | 487 (21.6) | 517 (22.1) |
| Vascular disease | 1089 (48.4) | 1116 (47.6) |
| Anemia | 999 (44.4) | 1059 (45.2) |
| Previous hemorrhage | 1005 (44.7) | 1142 (48.7) |
| Renal disease | 703 (31.2) | 792 (33.8) |
| Liver disease | 410 (18.2) | 456 (19.5) |
| Current smoker | 71 (3.2) | 86 (3.7) |
| CHA2DS2VASc score | 4.7 ± 1.7 | 4.7 ± 1.7 |
| Low risk† | 55 (2.4) | 63 (2.7) |
| Moderate risk† | 736 (32.7) | 695 (29.7) |
| High risk† | 1459 (64.8) | 1585 (67.6) |
| ATRIA bleed score | 4.9 ± 2.8 | 5.1 ± 2.9 |
Values are given as mean ± SD or n (%).
AF = atrial fibrillation; ATRIA = Anticoagulation and Risk Factors in Atrial Fibrillation.
Variables with missing data: race : n = 77 (1.7%); height: n = 6 (0.13%); weight: n = 7 (0.15%); body mass index: n = 14 (0.30%); blood pressure: n = 5 (0.11%); similar distributions between the 2 groups.
Low risk: 1 (men), 2 (women); Moderate risk: 2–3 (men), 3–4 (women); High risk: 4+ (men), 5+ (women).
Rhythm assessment was completed at 4947 (71.7%) of 6903 eligible visits (median 2 [1–3] per patient) in the intervention arm. Among patients with diagnosed AF in intervention practices, 2022 (89.9%) completed AF rhythm assessment at least once during the study period. In the control arm, 65 patients (2.8%) completed rhythm assessment during subsequent visits to an intervention practice.
Single-lead ECG results
During the first visit with rhythm assessment, 31.5% of patients with diagnosed AF had a single-lead ECG result of Possible AF, 17.2% Unclassified, 43.7% Normal, and 7.6% No Analysis. Over the 4947 intervention practice encounters with rhythm assessment, the result was Possible AF (31.3% [n = 1550]); Unclassified (17.2% [n = 852]); Normal (43.1% [n = 2132]); or No Analysis (8.3% [n = 413]). In total, 40.1% of patients with diagnosed AF who completed rhythm assessment had at least 1 Possible AF result during the study period. Patients with diagnosed AF that was transient in the setting of a secondary precipitant (see Methods) were much less likely to have a Possible AF result, whereas older patients, those with a higher stroke risk, and those using OAC at baseline were more likely to have a Possible AF result (Table 2). Among patients completing rhythm assessment at a subsequent visit after a first Possible AF result (n = 494), 64.8% (n = 320) had a repeat Possible AF result, 14.6% (n = 72) were Unclassified, 12.1% (n = 60) were Normal, and 9% (n = 42) had No Analysis.
Table 2.
Distribution of single-lead rhythm assessment results in patients with diagnosed AF in the intervention arm at the first visit with completed rhythm assessment and according to the most abnormal result
| Possible AF | Unclassified | Normal | No analysis | |
|---|---|---|---|---|
| First visit with rhythm assessment | ||||
| All | 636 (31.5) | 348 (17.2) | 884 (43.7) | 154 (7.6) |
| Age (y) | ||||
| 65–74 | 201 (26.6) | 104 (13.8) | 410 (54.3) | 40 (5.3) |
| 75–84 | 305 (35.2) | 158 (18.2) | 305 (40.1) | 56 (6.5) |
| ≥85 | 130 (32.5) | 86 (21.5) | 126 (31.5) | 58 (14.5) |
| OAC at baseline | ||||
| No | 126 (19.2) | 106 (16.1) | 381 (58.0) | 44 (6.7) |
| Yes | 510 (37.4) | 242 (17.7) | 503 (36.8) | 110 (8.1) |
| Previous stroke | ||||
| No | 496 (31.3) | 273 (17.2) | 704 (44.4) | 114 (7.2) |
| Yes | 140 (32.2) | 75 (17.2) | 180 (41.4) | 40 (9.2) |
| CHA2DS2-VASc score | ||||
| Low risk∗ | 10 (21.7) | 5 (10.9) | 29 (63.0) | 2 (4.3) |
| Moderate risk∗ | 190 (28.7) | 91 (13.7) | 350 (52.8) | 32 (4.8) |
| High risk∗ | 436 (33.2) | 252 (19.2) | 505 (38.5) | 120 (9.1) |
| Transient AF in setting of secondary precipitant† | ||||
| No (n = 719) | 201 (28.0) | 124 (17.2) | 340 (47.3) | 54 (7.5) |
| Yes (n = 107) | 7 (6.5) | 18 (16.8) | 79 (73.8) | 3 (2.8) |
| Most “Abnormal” rhythm assessment | ||||
| All | 811 (40.1) | 382 (18.9) | 737 (36.4) | 92 (4.5) |
| Age (y) | ||||
| 65–74 | 256 (33.9) | 119 (15.8) | 358 (47.4) | 22 (2.9) |
| 75–84 | 373 (43.0) | 174 (20.1) | 288 (33.2) | 32 (3.7) |
| ≥85 | 182 (45.5) | 89 (22.3) | 91 (22.8) | 38 (9.5) |
| OAC at baseline | ||||
| No | 166 (25.3) | 143 (21.8) | 319 (48.6) | 29 (4.4) |
| Yes | 645 (47.3) | 239 (17.5) | 418 (30.6) | 63 (4.6) |
| Previous stroke | ||||
| No | 619 (39.0) | 298 (18.8) | 597 (37.6) | 73 (4.6) |
| Yes | 192 (44.1) | 84 (19.3) | 140 (32.2) | 19 (4.4) |
| CHA2DS2-VASc score | ||||
| Low risk∗ | 13 (28.3) | 4 (8.7) | 28 (60.9) | 1 (2.2) |
| Moderate risk∗ | 222 (33.5) | 110 (16.6) | 309 (46.6) | 22 (3.3) |
| High risk∗ | 576 (43.9) | 268 (20.4) | 400 (30.5) | 69 (5.3) |
| Transient AF in setting of secondary precipitant† | ||||
| No (n = 719) | 255 (35.5) | 136 (18.9) | 292 (40.6) | 36 (5.0) |
| Yes (n = 107) | 13 (12.1) | 28 (26.2) | 64 (59.8) | 2 (1.9) |
Values are given as n (%).
AF = atrial fibrillation; OAC = oral anticoagulation.
Low risk: 1 (men), 2 (women); Moderate risk: 2–3 (men), 3–4 (women); High risk: 4+ (men), 5+ (women).
As determined by chart review among a subset of the sample (n = 826).
OAC prescriptions
In the intervention arm, 28.2% (n = 635) had no evidence of OAC treatment in the previous year compared to 27.4% (n = 643) in the control arm. Of these patients, 26 from the intervention arm and 24 from the control arm died during follow-up and were excluded from assessment of OAC initiation. Among patients with previously diagnosed AF but not on OAC before their index visit, the proportion of patients who initiated OAC treatment over the subsequent 12 months was similar in the intervention (17.7% [n =108/609]) and control arms (19.1% [n = 118/619]) (P = .59). There was no significant difference in OAC initiation in the intervention and control arms by age, previous stroke, or CHA2DS2-VASc score, except among a small subset with a low-risk CHA2DS2-VASc score (P = .03) (Table 3). However, among patients with a Possible AF result, the proportion with OAC initiation (26.1%) was substantially higher compared to control patients (19.1%) (adjusted odds ratio [OR] 1.62; 95% confidence interval [CI] 1.04–2.51). In contrast, OAC initiation after a Normal result (9.9%) was lower (Figure 2A) compared to control patients (19.1%) (adjusted OR 0.45; 95% CI 0.29–0.71).
Table 3.
Initiation and discontinuation of oral anticoagulation in the 12 months after the first visit during the study period in intervention and control arms
| Oral anticoagulation initiation |
Oral anticoagulation discontinuation |
|||
|---|---|---|---|---|
| Intervention | Control | Intervention | Control | |
| All | 17.7 (108/609) | 19.1 (118/619) | 6.3 (92/1459) | 7.2 (111/1540) |
| Age (y) | ||||
| 65–74 | 17.0 (44/259) | 16.5 (41/248) | 5.7 (30/528) | 7.8 (43/553) |
| 75–84 | 20.6 (45/218) | 21.8 (55/252) | 5.6 (37/655) | 6.4 (43/670) |
| ≥85 | 14.4 (19/132) | 18.5 (22/119) | 9.1 (25/276) | 7.9 (25/317) |
| Previous stroke | ||||
| No | 18.2 (91/501) | 19.0 (100/526) | 6.8 (76/1119) | 7.1 (83/1165) |
| Yes | 15.7 (17/108) | 19.4 (18/93) | 4.7 (16/340) | 7.5 (28/375) |
| CHA2DS2VASc score | ||||
| Low | 10.7 (3/28) | 25.8 (8/31) | 12.5 (3/24) | 10.3 (3/29) |
| Moderate | 20.4 (52/255) | 18.5 (44/238) | 7.4 (32/431) | 6.8 (29/424) |
| High | 16.3 (53/326) | 18.9 (66/350) | 5.7 (57/1004) | 7.3 (79/1087) |
Values are given as % (n/N).
Figure 2.
Initiation (A) and discontinuation (B) of oral anticoagulation (OAC) in the 12 months after the first visit during the study period overall and by single-lead electrocardiographic (ECG) result. AF = atrial fibrillation.
In the intervention arm, 67.2% (n = 1511) had evidence of OAC use at baseline compared to 68.3% (n = 1601) in the control arm. Of these patients, 52 from the intervention arm and 61 from the control arm died during follow-up and were excluded from assessment of OAC discontinuation. Among patients with previously diagnosed AF who were on OAC at baseline, the proportion of patients who discontinued OAC treatment over the subsequent 12 months was similar in the intervention (6.3% [n = 92/1459]) and control arms (7.2% [n = 111/1540]) (P = .34). There was no significant difference in OAC discontinuation in the intervention and control arms by age, previous stroke, or CHA2DS2-VASc score (Table 3), although discontinuation was higher for younger patients (age 65–74 years) in the control arm (P = .06). Among patients with a Possible AF result, OAC discontinuation (3.8%) was lower compared to control patients (7.2%) (adjusted OR 0.51; 95% CI 0.32–0.81). A Normal result did not significantly impact OAC discontinuation (7.0%) compared to control patients (7.2%) (adjusted OR 0.98; 95% CI 0.65–1.47) (Figure 2B).
Supplemental per-protocol analyses of initiation and discontinuation of OAC comparing patients with rhythm assessment completed to those without rhythm assessment over 90 and 365 days showed similar results (Supplemental Tables S1 and S2). Over both 90 and 365 days, the rhythm assessment result influenced OAC initiation and discontinuation. A Possible AF result was associated with increased initiation and decreased discontinuation, and a Normal result was associated with decreased initiation of OAC.
Resource utilization
Utilization of cardiac imaging, cardiac monitoring, procedures, and health care visits was similar in the intervention and control arms (Figure 3). Intervention patients were slightly more likely to have a specialty visit to cardiology compared to control patients (74.1% vs 70.5% with at least one visit, respectively; P = .007). Results were similar when limiting the intervention arm only to patients who completed rhythm assessment, with a slight increase in cardiology visits (74.9% vs 70.5%, respectively; P = .001).
Figure 3.
Health care utilization in intervention and control arms in the 12 months after the first visit during the study period. 12L ECG = 12-lead electrocardiogram.
Discussion
In an AF screening trial among older individuals from primary care practices, >4500 individuals with a previously established diagnosis of AF were enrolled to simplify the implementation of widespread rhythm assessment in practices. Patients with previously diagnosed AF completed cardiac rhythm assessments with single-lead ECG at high rates, with 90% of intervention patients completing at least 1 rhythm assessment over the 1-year study. Fewer than half of patients with diagnosed AF who completed a single-lead ECG rhythm assessment received a “Possible AF” result. Single-lead ECG rhythm assessment did not impact the overall use of OAC among those with diagnosed AF compared to the control arm. However, patients with a “Possible AF” result had a 62% increase in odds of OAC initiation and 49% reduced odds of OAC discontinuation.
Although STROKESTOP, a previous large AF screening trial, suggested that a cardiac rhythm assessment program increased use of OAC in patients with diagnosed AF, it did not include results from a control group.21 Although 18% of patients with diagnosed AF initiated use of OAC in our study, this result did not differ between the intervention and control groups. However, the results of the rhythm assessment ECG did seem to influence use of OAC. Patients with diagnosed AF were more likely to be started on OAC if their rhythm assessment result was “Possible AF.” However, the rhythm assessment result was also influential in the other direction. Patients with diagnosed AF were less likely to initiate use of OAC if their result was “Normal,” which resulted in no overall increase in OAC initiation compared to the control arm. Patients with diagnosed AF already on OAC were less likely to stop OAC if their result was “Possible AF.”
Patients with previously diagnosed AF completed a single-lead rhythm assessment ECG at primary care visits at a high rate. Overall, 90% of patients with diagnosed AF in the intervention arm completed at least 1 rhythm assessment over the 1-year study period, and most patients had rhythm assessed at multiple visits. This was comparable to the rate of screening among patients without AF in VITAL-AF23 and was substantially higher than rates in other AF screening trials.27,28 Our uptake may be high because rhythm assessment was conducted by medical assistants, rather than research staff, and was integrated into routine vital sign assessments. Adding rhythm assessment to routine vital sign assessments creates additional work burden for medical assistants and could delay rooming the patient. In a busy primary care clinic, it still may be easier for clinical staff to conduct a quick single-lead ECG rhythm assessment on all patients above an age threshold rather than determining AF status from the EHR.
Most patients with diagnosed AF did not receive a result indicating “Possible AF.” At the first visit with a rhythm assessment, only 31.5% of patients with diagnosed AF had a “Possible AF” result. Overall, 40.1% of patients with diagnosed AF had a “Possible AF” result during any encounter during the study period. Utilizing EHR diagnosis codes may accurately identify patients who have diagnosed AF, but identified patients will have a mixture of persistent, paroxysmal, or transient AF. A 30-second single-lead ECG is more likely to identify patients who have persistent AF, a population in which the net benefit of OAC for those at elevated stroke risk is clear.29 Our results indicate that a large proportion of patients with diagnosed AF identified by EHR may have paroxysmal or transient AF. Among patients with persistent AF, cardiac rhythm assessment may lead to increased appropriate use of OAC. Previous interventions targeting individuals with diagnosed AF to increase use of OAC have had limited success,30, 31, 32, 33 potentially, in part, due to the heterogeneity in the populations identified. In an earlier trial, we found that electronic notifications to physicians did not increase use of OAC in patients with known diagnoses who were at elevated stroke risk.30 Physicians reported that they perceived the bleeding risk was too high or stroke risk was too low for the majority of patients not prescribed OAC.
The effect of including patients with diagnosed AF in an AF rhythm assessment program on health care utilization was not known. PCPs incorporated AF rhythm assessment results for patients with diagnosed AF into clinical workflow without increasing use of same day 12-lead ECGs. We observed a slight increase in cardiology visits among patients in the intervention arm to control patients but no meaningful differences in other utilization measures. These results are comparable to the mSToPS (mHealth Screening to Prevent Strokes) trial, which assessed health care utilization after screening by ECG patch in patients without a diagnosis of AF.34 In this population, they also observed an increase in cardiology visits but no increase in cardiac tests in patients completing screening compared to controls. However, they observed a significant decrease in emergency department visits and hospitalizations in the screened group.
Because it is challenging to quickly identify AF status, office-based cardiac rhythm assessment efforts are likely to include individuals with diagnosed AF. In our trial, a large proportion of patients with diagnosed AF were on OAC at baseline. In settings with lower rates of OAC use, our finding that a “Possible AF” rhythm assessment results in increased OAC initiation may be a strategy to improve OAC rates among individuals with higher stroke risk. For patients with paroxysmal AF who may not have an abnormal single-lead ECG result, a more continuous method of rhythm assessment will be needed to assess AF burden.35
Study limitations
This study was conducted within a single academic primary care network composed of a largely White population, so results may not be generalizable to other populations. Most patients did not have a rhythm assessment result indicating AF despite having a previous diagnosis of AF. The automated single-lead ECG algorithm has imperfect sensitivity, so it may have missed some true AF. Additionally, OAC and health care utilization were ascertained from the EHR and thus may be subject to misclassification. OAC initiation and discontinuation were based on prescription data from the EHR, so we are unable to determine whether patients filled prescriptions or complied with therapy.
Conclusion
Patients with previously diagnosed AF included in a point-of-care heart rhythm assessment strategy program completed rhythm assessment at high rates comparable to those without previous AF. Most patients with diagnosed AF did not have “Possible AF” on their brief rhythm assessment ECG, indicating that most AF patients in primary care practices have paroxysmal or transient AF. Overall, cardiac rhythm assessment in patients with diagnosed AF did not increase OAC utilization overall, but patients who had a single-lead ECG result of “Possible AF” had an increase in OAC utilization whereas those with a reading of “Normal” were less likely to receive OAC.
Acknowledgments
Funding Sources
The VITAL-AF study was investigator-initiated and funded by the Bristol Myers Squibb–Pfizer Alliance, which had no role in conducting the trial or analyzing data. All analyses presented were conducted by study investigators independent of the funders. Dr Ashburner is supported by National Institutes of Health (NIH) Grant K01HL148506. Dr Ellinor is supported by grants from the National Institutes of Health (1RO1HL092577, 1R01HL157635, 5R01HL139731); American Heart Association Strategically Focused Research Networks (18SFRN34110082); and European Union (MAESTRIA 965286). Dr Singer receives support from the Eliot B. and Edith C. Shoolman Fund of Massachusetts General Hospital.
Disclosures
Bristol Myers Squibb/Pfizer sponsored research support to Jeffrey M. Ashburner, Patrick T. Ellinor, Steven A. Lubitz, Daniel E. Singer, and Steven J. Atlas. Patrick T. Ellinor receives sponsored research support from Bayer AG and IBM Research; and served on advisory boards or consulted for Bayer AG, MyoKardia, and Novartis. Steven A. Lubitz is currently employed by Novartis Institutes for Biomedical Research; previously received sponsored research support from Boehringer Ingelheim, Fitbit, Medtronic, Premier, and IBM; and has consulted for Bristol Myers Squibb, Pfizer, Blackstone Life Sciences, and Invitae. Daniel E. Singer has consulted for Bristol Myers Squibb, Fitbit, Johnson & Johnson, Medtronic, and Pfizer. Steven J. Atlas has consulted for Bristol Myers Squibb, Pfizer, and Fitbit. All other authors have relationships relevant to the contents of this paper to disclose.
Authorship
All authors attest they meet the current ICMJE criteria for authorship.
Patient Consent
Participants provided informed consent to participate.
Ethics Statement
The research protocol was approved by the Mass General Brigham Institutional Review Board. The research reported in this paper adhered to CONSORT guidelines.
Footnotes
ClinicalTrials.gov Identifier: NCT03515057.
Supplementary data associated with this article can be found in the online version at https://doi.org/10.1016/j.hroo.2023.07.003.
Appendix. Supplementary Data
References
- 1.Go A.S., Hylek E.M., Phillips K.A., et al. Prevalence of diagnosed atrial fibrillation in adults: national implications for rhythm management and stroke prevention: the AnTicoagulation and Risk Factors in Atrial Fibrillation (ATRIA) Study. JAMA. 2001;285:2370–2375. doi: 10.1001/jama.285.18.2370. [DOI] [PubMed] [Google Scholar]
- 2.Wolf P.A., Abbott R.D., Kannel W.B. Atrial fibrillation as an independent risk factor for stroke: the Framingham Study. Stroke. 1991;22:983–988. doi: 10.1161/01.str.22.8.983. [DOI] [PubMed] [Google Scholar]
- 3.Wolf P.A., Abbott R.D., Kannel W.B. Atrial fibrillation: a major contributor to stroke in the elderly. The Framingham Study. Arch Intern Med. 1987;147:1561–1564. [PubMed] [Google Scholar]
- 4.Connolly S.J., Ezekowitz M.D., Yusuf S., et al. Dabigatran versus warfarin in patients with atrial fibrillation. N Engl J Med. 2009;361:1139–1151. doi: 10.1056/NEJMoa0905561. [DOI] [PubMed] [Google Scholar]
- 5.Granger C.B., Alexander J.H., McMurray J.J., et al. Apixaban versus warfarin in patients with atrial fibrillation. N Engl J Med. 2011;365:981–992. doi: 10.1056/NEJMoa1107039. [DOI] [PubMed] [Google Scholar]
- 6.Hart R.G., Pearce L.A., Aguilar M.I. Meta-analysis: antithrombotic therapy to prevent stroke in patients who have nonvalvular atrial fibrillation. Ann Intern Med. 2007;146:857–867. doi: 10.7326/0003-4819-146-12-200706190-00007. [DOI] [PubMed] [Google Scholar]
- 7.Patel M.R., Mahaffey K.W., Garg J., et al. Rivaroxaban versus warfarin in nonvalvular atrial fibrillation. N Engl J Med. 2011;365:883–891. doi: 10.1056/NEJMoa1009638. [DOI] [PubMed] [Google Scholar]
- 8.Gage B.F., Boechler M., Doggette A.L., et al. Adverse outcomes and predictors of underuse of antithrombotic therapy in medicare beneficiaries with chronic atrial fibrillation. Stroke. 2000;31:822–827. doi: 10.1161/01.str.31.4.822. [DOI] [PubMed] [Google Scholar]
- 9.Gladstone D.J., Bui E., Fang J., et al. Potentially preventable strokes in high-risk patients with atrial fibrillation who are not adequately anticoagulated. Stroke. 2009;40:235–240. doi: 10.1161/STROKEAHA.108.516344. [DOI] [PubMed] [Google Scholar]
- 10.Glazer N.L., Dublin S., Smith N.L., et al. Newly detected atrial fibrillation and compliance with antithrombotic guidelines. Arch Intern Med. 2007;167:246–252. doi: 10.1001/archinte.167.3.246. [DOI] [PubMed] [Google Scholar]
- 11.Hsu J.C., Chan P.S., Tang F., Maddox T.M., Marcus G.M. Differences in anticoagulant therapy prescription in patients with paroxysmal versus persistent atrial fibrillation. Am J Med. 2015;128:654.e1–654.e10. doi: 10.1016/j.amjmed.2014.11.035. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Hsu J.C., Maddox T.M., Kennedy K.F., et al. Oral anticoagulant therapy prescription in patients with atrial fibrillation across the spectrum of stroke risk: insights from the NCDR PINNACLE Registry. JAMA Cardiol. 2016;1:55–62. doi: 10.1001/jamacardio.2015.0374. [DOI] [PubMed] [Google Scholar]
- 13.Kakkar A.K., Mueller I., Bassand J.P., et al. Risk profiles and antithrombotic treatment of patients newly diagnosed with atrial fibrillation at risk of stroke: perspectives from the international, observational, prospective GARFIELD registry. PLoS One. 2013;8 doi: 10.1371/journal.pone.0063479. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Hsu J.C., Freeman J.V. Underuse of vitamin K antagonist and direct oral anticoagulants for stroke prevention in patients with atrial fibrillation: a contemporary review. Clin Pharmacol Ther. 2018;104:301–310. doi: 10.1002/cpt.1024. [DOI] [PubMed] [Google Scholar]
- 15.Piazza G., Karipineni N., Goldberg H.S., Jenkins K.L., Goldhaber S.Z. Underutilization of anticoagulation for stroke prevention in atrial fibrillation. J Am Coll Cardiol. 2016;67:2444–2446. doi: 10.1016/j.jacc.2016.03.515. [DOI] [PubMed] [Google Scholar]
- 16.Lip G.Y.H., Pan X., Kamble S., et al. Discontinuation risk comparison among “real-world” newly anticoagulated atrial fibrillation patients: apixaban, warfarin, dabigatran, or rivaroxaban. PloS One. 2018;13 doi: 10.1371/journal.pone.0195950. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Ruigómez A., Vora P., Balabanova Y., et al. Discontinuation of non-vitamin K antagonist oral anticoagulants in patients with non-valvular atrial fibrillation: a population-based cohort study using primary care data from The Health Improvement Network in the UK. BMJ Open. 2019;9 doi: 10.1136/bmjopen-2019-031342. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Rivera-Caravaca J.M., Roldán V., Esteve-Pastor M.A., et al. Cessation of oral anticoagulation is an important risk factor for stroke and mortality in atrial fibrillation patients. Thromb Haemost. 2017;117:1448–1454. doi: 10.1160/TH16-12-0961. [DOI] [PubMed] [Google Scholar]
- 19.Deitelzweig S.B., Buysman E., Pinsky B., et al. Warfarin use and stroke risk among patients with nonvalvular atrial fibrillation in a large managed care population. Clin Ther. 2013;35:1201–1210. doi: 10.1016/j.clinthera.2013.06.005. [DOI] [PubMed] [Google Scholar]
- 20.García Rodríguez L.A., Cea Soriano L., Munk Hald S., et al. Discontinuation of oral anticoagulation in atrial fibrillation and risk of ischaemic stroke. Heart. 2020;107:542–548. doi: 10.1136/heartjnl-2020-317887. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Svennberg E., Engdahl J., Al-Khalili F., Friberg L., Frykman V., Rosenqvist M. Mass screening for untreated atrial fibrillation: the STROKESTOP study. Circulation. 2015;131:2176–2184. doi: 10.1161/CIRCULATIONAHA.114.014343. [DOI] [PubMed] [Google Scholar]
- 22.Ashburner J.M., Atlas S.J., McManus D.D., et al. Design and rationale of a pragmatic trial integrating routine screening for atrial fibrillation at primary care visits: the VITAL-AF trial. Am Heart J. 2019;215:147–156. doi: 10.1016/j.ahj.2019.06.011. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Lubitz S.A., Atlas S.J., Ashburner J.M., et al. Screening for atrial fibrillation in older adults at primary care visits: VITAL-AF randomized controlled trial. Circulation. 2022;145:946–954. doi: 10.1161/CIRCULATIONAHA.121.057014. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Lubitz S.A., Yin X., Rienstra M., et al. Long-term outcomes of secondary atrial fibrillation in the community: the Framingham Heart Study. Circulation. 2015;131:1648–1655. doi: 10.1161/CIRCULATIONAHA.114.014058. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Lip G.Y., Nieuwlaat R., Pisters R., Lane D.A., Crijns H.J. Refining clinical risk stratification for predicting stroke and thromboembolism in atrial fibrillation using a novel risk factor-based approach: the euro heart survey on atrial fibrillation. Chest. 2010;137:263–272. doi: 10.1378/chest.09-1584. [DOI] [PubMed] [Google Scholar]
- 26.Fang M.C., Go A.S., Chang Y., et al. A new risk scheme to predict warfarin-associated hemorrhage: the ATRIA (Anticoagulation and Risk Factors in Atrial Fibrillation) Study. J Am Coll Cardiol. 2011;58:395–401. doi: 10.1016/j.jacc.2011.03.031. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Kaasenbrood F., Hollander M., de Bruijn S.H., et al. Opportunistic screening versus usual care for diagnosing atrial fibrillation in general practice: a cluster randomised controlled trial. Br J Gen Pract. 2020;70:e427–e433. doi: 10.3399/bjgp20X708161. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Uittenbogaart S.B., Verbiest-van Gurp N., Lucassen W.A.M., et al. Opportunistic screening versus usual care for detection of atrial fibrillation in primary care: cluster randomised controlled trial. BMJ. 2020;370:m3208. doi: 10.1136/bmj.m3208. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.January C.T., Wann L.S., Calkins H., et al. 2019 AHA/ACC/HRS focused update of the 2014 AHA/ACC/HRS guideline for the management of patients with atrial fibrillation: a report of the American College of Cardiology/American Heart Association Task Force on Clinical Practice Guidelines and the Heart Rhythm Society. J Am Coll Cardiol. 2019;74:104–132. doi: 10.1016/j.jacc.2019.01.011. [DOI] [PubMed] [Google Scholar]
- 30.Ashburner J.M., Atlas S.J., Khurshid S., et al. Electronic physician notifications to improve guideline-based anticoagulation in atrial fibrillation: a randomized controlled trial. J Gen Intern Med. 2018;33:2070–2077. doi: 10.1007/s11606-018-4612-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Cook D.A., Enders F., Caraballo P.J., Nishimura R.A., Lloyd F.J. An automated clinical alert system for newly-diagnosed atrial fibrillation. PLoS One. 2015;10 doi: 10.1371/journal.pone.0122153. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Silbernagel G., Spirk D., Hager A., Baumgartner I., Kucher N. Electronic alert system for improving stroke prevention among hospitalized oral-anticoagulation-naive patients with atrial fibrillation: a randomized trial. J Am Heart Assoc. 2016;5 doi: 10.1161/JAHA.116.003776. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Vinereanu D., Lopes R.D., Bahit M.C., et al. A multifaceted intervention to improve treatment with oral anticoagulants in atrial fibrillation (IMPACT-AF): an international, cluster-randomised trial. Lancet. 2017;390:1737–1746. doi: 10.1016/S0140-6736(17)32165-7. [DOI] [PubMed] [Google Scholar]
- 34.Waalen J., Edwards A.M., Sanyal A., et al. Healthcare resource utilization following ECG sensor patch screening for atrial fibrillation. Heart Rhythm O2. 2020;1:351–358. doi: 10.1016/j.hroo.2020.09.005. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Steinhubl S.R., Waalen J., Edwards A.M., et al. Effect of a home-based wearable continuous ECG monitoring patch on detection of undiagnosed atrial fibrillation: the mSToPS randomized clinical trial. JAMA. 2018;320:146–155. doi: 10.1001/jama.2018.8102. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.