Table 4.
Digital health randomized control trials in arrhythmia detection or management
| First author | Country | Age mean (SD) | Design | N | Race ethnicity | Duration | Inclusion criteria | Primary outcome | Results | Device | Intervention | Health care team interpretation |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
|
William et al (2018) [95] |
US | 68 years | Single-Center Non-Randomized | 52 | Not reported | NA | AF admitted for anti-arrhythmic drug initiation, 38–85 yrs, hx of paroxysmal AF | Sensitivity and specificity | 96.6% sensitivity and 94.1% specificity of 225 recordings | Kardia Mobile Cardiac Monitor coupled to Wi-Fi smart device (iPod) | 30 s recordings of lead I ECG + automatic analysis by KMCM algorithm | I-lead ECG reviewed by blinded electrophysiologist |
| Steinhubl et al. (2018) [96] | US | 72.4 (7.3) | RCT + prospective matched cohort | 2659 | Not reported | 4 months | ≥ 75 yrs, male > 55 yrs or female > 65 yrs* | Incidence of new AF diagnosis at 4 months immediate vs delayed group | 3.9% in the immediate versus 0.9% delayed group | iRhythm ZioXT | Stored data analyzed by an FDA approved algorithm | I-lead ECG adjudication by blinded committee |
| Guo et al. (2019) [97] | China | 54 | Prospective cohort | 187,912 | Not reported | At least 14 days of monitoring | ≥ 18 yrs and smartphone ownership | AF detection efficacy | PPV of 91.6% (91.5–91.8) | Smartphone plus smart wrist band | AF detection using PPG in the smartphone or wrist band | Confirmed by patient’s provider with use of ECG or 24-h Holter monitoring |
| Perez et al. (2019) [98] | US | 41 (13) | Prospective single group pragmatic study | 219,297** |
68% White 12% Hispanic 7.7% Black 6.2% Asian |
Median 117 days of monitoring | ≥ 22 years without prior AF diagnosis or AC use | Proportion of notified participants with AF on ECG patch and PPV of irregular pulse intervals | PPV 84% (0.76–0.92) | Apple Watch + iPhone | AF detection by app with irregular pulse notification algorithm | Confirmed by ECG patch worn for 7 days |
| Goldenthal et al. (2019) [99] | US | 61 (12) | RCT | 238 |
77% White, 3% Black 1% Asian 20%, 9% Hispanic |
6 months | AF patients who underwent RFA or DCCV | AF recurrence detection | Likelihood of recurrent significantly greater IT*** vs control (HR = 1.56, 1.06–2.3) | KardiaMobile + iPhone + cellular servce plan | Record ECG daily or with symptoms plus motivational texts | Confirmation was determined by patient’s provider |
| Koh et al. (2021) [100] | Malaysia | 65.3 (7.4) | RCT | 203 | Not reported | 30 days | ≥ 55 years without AF and ischemic stroke or TIA within the preceding 12 months | AF detection 30-day monitor KardiaMobile vs 24-h Holter | 9.5 vs 2% IT vs control (P = 0.024) | KardiaMobile | Use KardiaMobile monitor 3 times a day for 30 days | I-lead ECG adjudicated by blinded electrophysiologist |
| Lubitz et al. (2022) [101] | US | 74 (7) | Cluster RCT | 30,715 | 82.4% White, 5.3% Black, 2.2% Hispanic | 1 year | ≥ 65 years | New AF diagnosis at 1 year | 1.72% vs 1.59% IT vs control P = 0.38 | KardiaMobile + iPad | Screening AF at primary care clinic with tracings reviewed by cardiologist |
I-lead ECG reviewed by independent cardiologist Confirmation with 12-lead ECG determined by patient’s PCP |
*And any of the following diagnosis: prior stroke or TIA, heart failure, diabetes and hypertension, mitral valve disease, left ventricular hypertrophy, COPD on home O2, sleep apnea, history of pulmonary embolism, history of myocardial infarction or obesity
**450 returned patches
***IT = intervention