Table 1.
Study characteristics (PPG, single-lead ECG and miscellaneous).
| Author, Year | Setting | Study Design | Device | Algorithm | Time Used | Gold Standard | Age (Mean ± SD) | Sex (Females %) | N | N in AF Group | N in Control Group (Type of Control Group) | Group of AFL | TP | TN | FP | FN |
Unclassified/ Uninterpretable |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Bacevicius et al., 2022 [39] | Inpatients and outpatients | Case–control selected cross-sectional study | Prototype of the wearable device | Automatic PPG-based algorithm (PPG) | 2 min. | 3-lead Holter | AF: 65.6 ± 11.2, SR: 67.3 ± 14.2 |
AF: 47.1% SR: 46.1% |
344 | 121 | 223 (SR) | Excluded from study | 114 | 216 | 7 | 7 | Excluded |
| Chang et al., 2022 [33] | Outpatients | Cohort selected cross-sectional study | Garmin Forerunner 945 smartwatch | Garmin Forerunner 945 smartwatch algorithm (PPG) | 24 h | 3-lead Holter | All participants: 66.1 ± 12.6 AF: 69.3 ± 11.5 Non-AF: 62.0 ± 12.9 |
All participants: 36.5% AF: 30.4% Non-AF: 44.3% |
200 | 112 | 88 (non-AF) | AF group | 109 | 78 | 10 | 3 | 0% |
| Chen et al., 2020 (A) [40] | Inpatients and outpatients | Case–control selected cross-sectional study | Amazfit Health Band 1S | RealBeats Artificial Intelligence Biological Data Engine (Huami Technology) (PPG) | 3 min. | 12-lead ECG | AF: 70.4 ± 11.5 Non-AF: 59.3 ± 14.8 |
AF: females: 43.3% Non-AF: females: 52.6% |
401 | 139 | 244 (non-AF) | Control group | 132 | 242 | 2 | 7 | 4.5% |
| Van Haelst et al., 2018 (A) [11] | Outpatients | Case–control selected cross-sectional study | Fibricheck | Fibricheck algorithm (PPG) | 3 min. | 12-lead ECG | All participants: 77.3 ± 8.0 AF: 78.8 ± 8.0 No AF: 75.9 ± 7.9 |
All participants: 57.4% AF: 51.1% Non-AF: 63.3% |
190 | 75 | 93 (non-AF) | AF group | 73 | 83 | 10 | 2 | 11.6% |
| Mol et al., 2020 [35] | Inpatients | Case–control selected cross-sectional study | iPhone 8 | Algorithm developed by Happitech (Amsterdam, The Netherlands) (PPG) | 90 s | Continuous electrocardiography | All participants: 69 ± 9 | All participants: 43% | 257 | 149 | 108 (SR) | Excluded from study | 139 | 101 | 2 | 3 | 4.7% |
| Sun et al., 2022 [34] | Outpatients | Case–control selected cross-sectional study | Industrial camera (FLIR BFLY-U3-03S2C-CS) | DCNN model (PPG) | 10 min. max or as much as tolerated | 12-lead ECG | All participants: 69.3 ± 13.0 AF: 74.3 ± 12.5 Non-AF: 67.8 ± 13.0 |
All participants: 46% AF: 51.4% Non-AF: 44.8% |
453 | 105 | 348 (no-AF) | Control group | 98 | 342 | 6 | 7 | Excluded |
| Tison et al., 2018 [30] | Inpatients | Case–control selected cross-sectional study | Apple Watch | Optimized Cardiogram app (PPG) | 20 min prior- and 20 min post-cardioversion | 12-lead ECG | All participants: 66.1 ± 10.7 | All participants: 16% | 51 | 51 | 51 (SR) | Excluded from study | 50 | 46 | 5 | 1 | 0% |
| Väliaho et al., 2019 (A) [36] | Inpatients | Case–control selected cross-sectional study | Empatica E4 wrist band | MATLAB® software version R2017b using AFEvidence or COSEn (PPG) | 5 min | 3-lead Holter | AF: 72.0 ± 14.3 years SR: 54.5 ± 18.6 years |
AF: 42.5% SR: 44.9% |
213 | 106 | 107 (SR) | Excluded from study | 102 | 105 | 2 | 4 | Excluded |
| Väliaho et al., 2019 (B) [36] | Inpatients | Case–control selected cross-sectional study | Empatica E4 wrist band | MATLAB® software version R2017b using AFEvidence or COSEn (PPG) | 5 min | 3-lead Holter | AF: 72.0 ± 14.3 years SR: 54.5 ± 18.6 years |
AF: 42.5% SR: 44.9% |
213 | 106 | 107 (SR) | Excluded from study | 101 | 105 | 2 | 5 | Excluded |
| Väliaho et al., 2021 (A) [41] | Inpatients | Case–control selected cross-sectional study | Empatica E4 wrist band | MATLAB® software (version R2017b) with a novel autocorrelation (AC) feature (PPG) | 1 min every 10 min, 20 min, 30 min, 60 min | 3-lead Holter | AF: 77.1 ± 9.7 SR: 67.3 ± 15.8 |
AF: 46.1% SR: 55.7% |
173 | 76 | 97 (SR) | Unclear | 75 | 79 | 18 | 1 | 0% |
| Väliaho et al., 2021 (B) [41] | Inpatients | Case–control selected cross-sectional study | Empatica E4 wrist band | MATLAB® software (version R2017b) with a novel autocorrelation (AC) feature (PPG) | 1 min every 10 min, 20 min, 30 min, 60 min | 3-lead Holter | AF: 77.1 ± 9.7 SR: 67.3 ± 15.8 |
AF: 46.1% SR: 55.7% |
173 | 76 | 97 (SR) | Unclear | 75 | 87 | 10 | 1 | 0% |
| Väliaho et al., 2021 (C) [41] | Inpatients | Case–control selected cross-sectional study | Empatica E4 wrist band | MATLAB® software (version R2017b) with a novel autocorrelation (AC) feature (PPG) | 1 min every 10 min, 20 min, 30 min, 60 min | 3-lead Holter | AF: 77.1 ± 9.7 SR: 67.3 ± 15.8 |
AF: 46.1% SR: 55.7% |
173 | 76 | 97 (SR) | Unclear | 72 | 94 | 3 | 4 | 0% |
| Väliaho et al., 2021 (D) [41] | Inpatients | Case–control selected cross-sectional study | Empatica E4 wrist band | MATLAB® software (version R2017b) with a novel autocorrelation (AC) feature (PPG) | 1 min every 10 min, 20 min, 30 min, 60 min | 3-lead Holter | AF: 77.1 ± 9.7 SR: 67.3 ± 15.8 |
AF: 46.1% SR: 55.7% |
173 | 76 | 97 (SR) | Unclear | 70 | 96 | 1 | 6 | 0% |
| Chan et al., 2016 (A) [32] | Outpatients | Cohort selected cross-sectional study | iPhone 4S | Cardiio Rhythm smartphone application (Cardiio Inc.) (PPG) | 51.3 s | single-lead ECG | All participants: 68.4 ± 12.2 | All participants: 53.2% | 1013 | 28 | 985 (non-AF) | Excluded from study | 26 | 963 | 22 | 2 | 0% |
| Dörr et al., 2019 (A) [37] | Inpatients | Case–control selected cross-sectional study | Gear Fit 2, Samsung | Heartbeats application (Preventicus GmbH, Jena, Germany) (PPG) | 1 min, 3 min and 5 min | single-lead ECG | All participants: 76.4 ± 9.5 AF: 77.4 ± 9.1 SR: 75.6 ± 9.8 |
All participants: 44.3% SR: 46.1% AF: 42.2% |
650 | 237 | 271 (SR) | Excluded from study | 222 | 266 | 5 | 15 | 21.8% |
| Dörr et al., 2019 (B) [37] | Inpatients | Case–control selected cross-sectional study | Gear Fit 2, Samsung | Heartbeats application (Preventicus GmbH, Jena, Germany) (PPG) | 1 min, 3 min and 5 min | single-lead ECG | All participants: 76.4 ± 9.5 AF: 77.4 ± 9.1 SR: 75.6 ± 9.8 |
All participants: 44.3% SR: 46.1% AF: 42.2% |
650 | 204 | 243 (SR) | Excluded from study | 191 | 235 | 8 | 13 | 31.2% |
| Dörr et al., 2019 (C) [37] | Inpatients | Case–control selected cross-sectional study | Gear Fit 2, Samsung | Heartbeats application (Preventicus GmbH, Jena, Germany) (PPG) | 1 min, 3 min and 5 min | single-lead ECG | All participants: 76.4 ± 9.5 AF: 77.4 ± 9.1 SR: 75.6 ± 9.8 |
All participants: 44.3% SR: 46.1% AF: 42.2% |
650 | 167 | 202 (SR) | Excluded from study | 156 | 198 | 4 | 11 | 43.2% |
| Lubitz et al., 2022 [31] | Outpatients | Cohort selected cross-sectional study | Fitbit device | Fitbit app (PPG) | 1 week | single-lead ECG | All participants *: ≥75 years: 9.7% 65–74 years: 33.2% 55–64 years: 37.4% 40–54 years: 16.6% 22–39 years: 6.1% |
All participants: 48.2% | 1057 | 340 | 717 (non-AF) | AF group | 230 | 706 | 11 | 110 | 0% |
| Badertscher et al., 2022 [42] | - | Cohort selected cross-sectional study | Withings Scanwatch | Withings Scanwatch detection algorithm (single-lead ECG) | 30 s | 12-lead ECG | All participants: 67 (54–76 years) | All participants: 48% | 319 | 34 | 285 (SR) | No comment | 19 | 247 | 3 | 6 | 13.8% |
| Bumgarner MD et al., 2018 [43] | Inpatients | Case–control selected cross-sectional study | Kardia Band | Kardia Band detection algorithm (single-lead ECG) | 30 s | 12-lead ECG | All participants: 68.2 ± 10.86 | All participants: 17% | 169 | 91 | 78 (SR) | In AF group | 63 | 37 | 7 | 5 | 33.7% |
| Campo et al., 2022 [44] | Inpatients and outpatients | Case–control selected cross-sectional study | Withings Scanwatch | Withings Scanwatch detection algorithm (single-lead ECG) | 30 s | 12-lead ECG | All participants: 67.7 ± 14.8 AF: 74.3 ± 12.3 SR: 61.8 ± 14.3 Other arrhythmias: 66.9 ± 15.2 Unreadable ECGs: 78.8 ± 12.5 |
All participants: 61.1% AF: 42% SR: 34.5% Other arrhythmias: 40% Unreadable ECGs: 75% |
258 | 87 | 155 (non-AF) | Control group | 77 | 144 | 11 | 10 | 6.2% |
| Chen et al., 2020 (B) [40] | Inpatients and outpatients | Cohort selected cross-sectional study | Amazfit Health Band 1S | RealBeats Artificial Intelligence Biological Data Engine (Huami Technology) (single-lead ECG) | 60 s | 12-lead ECG | AF: 70.4 ± 11.5 Non-AF: 59.3 ± 14.8 |
AF: 43.3% Non-AF: 52.6% |
401 | 150 | 25 (non-AF) | No comment | 131 | 249 | 0 | 6 | 3.7% |
| Cunha et al., 2020 [45] | Inpatients | Cohort selected cross-sectional study | Kardia® mobile | Kardia® mobile algorithm (single-lead ECG) | 30 s | 12-lead ECG | - | - | 129 | 22 | 78 (SR) | No comment | 20 | 76 | 2 | 2 | 22.4% |
| Desteghe et al., 2017 (A) [46] | Inpatients | Cohort selected cross-sectional study | AliveCor | AliveCor algorithm (single-lead ECG) | 30 s | 12-lead ECG | All participants: 67.9 ± 14.6 AF: 73.1 ± 12.2 SR: 65.1 ± 15.0 |
All participants: 43.1% AF: 51.8% SR: 38.3% |
265 | 22 | 243 (SR) | In AF group | 12 | 237 | 6 | 10 | 0% |
| Desteghe et al., 2017 (B) [46] | Inpatients | Cohort selected cross-sectional study | MyDiagnostick | MyDiagnostick algorithm (single-lead ECG) | 60 s | 12-lead ECG | All participants: 67.9 ± 14.6 AF: 73.1 ± 12.2 SR: 65.1 ± 15.0 |
All participants: 43.1% AF: 51.8% SR: 38.3% |
265 | 22 | 243 (SR) | In AF group | 18 | 229 | 14 | 4 | 0% |
| Desteghe et al., 2017 (C) [46] | Inpatients | Cohort selected cross-sectional study | AliveCor | AliveCor algorithm (single-lead ECG) | 30 s | 6-lead ECG | - | - | 113 | 19 | 94 (SR) | In AF group | 15 | 92 | 2 | 4 | 0% |
| Desteghe et al., 2017 (D) [46] | Inpatients | Cohort selected cross-sectional study | MyDiagnostick | MyDiagnostick algorithm (single-lead ECG) | 60 s | 6-lead ECG | - | - | 113 | 19 | 94 (SR) | In AF group | 17 | 90 | 4 | 2 | 0% |
| Ford et al., 2022 (A) [10] | outpatients | Case–control selected cross-sectional study | Apple Watch 4 | Apple Watch 4 algorithm (single-lead ECG) | 30 s | 12-lead ECG | All participants: 76 ± 7 | All participants: 38% | 125 | 31 | 94 (SR) | In AF group | 6 | 76 | 0 | 6 | 29.6% |
| Ford et al., 2022 (B) [10] | outpatients | Case–control selected cross-sectional study | KardiaBand | KardiaBand algorithm (single-lead ECG) | 30 s | 12-lead ECG | All participants: 76 ± 7 | All participants: 38% | 125 | 31 | 94 (SR) | In AF group | 26 | 68 | 5 | 1 | 20% |
| Fu et al., 2021 (A) [47] | - | Case–control selected cross-sectional study | Wearable Dynamic ECG Recorder | Amazfit CardiDoc application (single-lead ECG) | 60 s | 12-lead ECG | All participants: 59 ± 11.16 AF: 64.00 ± 9.38 SR:55.15 ± 11.01 |
All participants: 34% AF: 45.3% SR: 41% |
114 | 53 | 61 (SR) | Excluded | 47 | 61 | 0 | 4 | 1.8% |
| Fu et al., 2021 (B) [47] | - | Case–control selected cross-sectional study | Wearable Dynamic ECG Recorder | Amazfit CardiDoc application (single-lead ECG) | 60 s | 12-lead ECG | All participants: 59 ± 11.16 AF: 64.00 ± 9.38 SR:55.15 ± 11.01 |
All participants: 34% AF: 45.3% SR: 41% |
114 | 53 | 61 (SR) | Excluded | 50 | 61 | 0 | 2 | 0.9% |
| Fu et al., 2021 (C) [47] | - | Case–control selected cross-sectional study | Wearable Dynamic ECG Recorder | Amazfit CardiDoc application (single-lead ECG) | 60 s | 12-lead ECG | All participants: 59 ± 11.16 AF: 64.00 ± 9.38 SR:55.15 ± 11.01 |
All participants: 34% AF: 45.3% SR: 41% |
114 | 53 | 61 (SR) | Excluded | 50 | 61 | 0 | 2 | 0.9% |
| Van Haelst et al., 2018 (B) [11] | Outpatients | Cohort selected cross-sectional study | AliveCor | AliveCor algorithm (single-lead ECG) | 30 s | 12-lead ECG | All patients: 77.3 ± 8.0 AF: 78.8 ± 8.0 Non-AF: 75.9 ± 7.9 |
All participants: 57.4% AF: 51.1% No AF: 63.3% |
190 | 75 | 93 (non-AF) | In AF group | 73 | 83 | 10 | 2 | 19.3% |
| Himmelreich JCL et al., 2019 [48] | Outpatients | Cohort selected cross-sectional study | KardiaMobile | KardiaMobile (AliveCor, Inc.) algorithm (single-lead ECG) | 30 s | 12-lead ECG | All participants: 64.1 ± 14.7 | All participants: 46.3% | 214 | 23 | 191 (SR) | In AF group | 20 | 187 | 4 | 0 | 1.4% |
| Lown et al., 2018 (A) [49] | Outpatients | Case–control selected cross-sectional study | AliveCor | AliveCor (Kardia version 4.7.0) algorithm (single-lead ECG) | 30 s | 12-lead ECG | All participants: 73.9 ± 6.1 | - | 418 | 82 | 336 (non-AF) | In AF group | 72 | 332 | 2 | 2 | 2.4% |
| Rajakariar et al., 2020 [50] | Inpatients | Cohort selected cross-sectional study | KardiaBand | AliveCor Kardia application V.5.0.2 (AliveCor, Mountain View, CA, USA) (single-lead ECG) | 30 s | 12-lead ECG | All participants: 67 ± 16 AF: 76 ± 11 SR: 64 ± 17 |
All participants: 43.5% AF: 48% SR: 36% |
200 | 38 | 162 (SR) | No comment | 36 | 124 | 13 | 2 | 12.5% |
| Santala et al., 2021 (1) [41] | Inpatients | Case–control selected cross-sectional study | Suunto Movesense, Suunto, Vantaa, Finland (heart belt) | Awario, Heart2Save, Kuopio, Finland (single-lead ECG) | 24 h | 12-lead ECG | AF: 77 ± 10 SR: 68 ± 16 |
AF: 48% SR: 60% |
159 | 73 | 86 (SR) | No comment | 73 | 82 | 4 | 0 | 0% |
| Santala et al., 2021 (2) (A) [51] | Inpatients | Case–control selected cross-sectional study | single-lead Necklace-embedded ECG recorder (Including Movesense ECG-sensor, Suunto, Vantaa, Finland, Necklace-ECG) | Awario, Heart2Save, Kuopio, Finland (single-lead ECG) | 30 s | 3-lead Holter ECG | AF (years): 72.7 ± 14.1 SR (years): 61.5 ± 18.1 |
AF: 56.1% SR: 53.2% |
145 | 66 | 79 (SR) | No comment | 54 | 78 | 0 | 3 | 6.9% |
| Santala et al., 2021 (2) (B) [51] | Inpatients | Case–control selected cross-sectional study | single-lead Necklace-embedded ECG recorder (Including Movesense ECG-sensor, Suunto, Vantaa, Finland, Necklace-ECG) | Awario, Heart2Save, Kuopio, Finland (single-lead ECG) | 30 s | 3-lead Holter ECG | AF (years): 72.7 ± 14.1 SR (years): 61.5 ± 18.1 |
AF: 56.1% SR: 53.2% |
145 | 66 | 79 (SR) | No comment | 58 | 75 | 0 | 1 | 7.6% |
| Santala et al., 2022 [52] | Inpatients | Case–control selected cross-sectional study | Firstbeat Bodyguard 2, Firstbeat Technologies | Awario, Heart2Save (single-lead ECG) | 24 h | 3-lead Holter ECG | AF: 77 ± 10 SR: 68 ± 15 |
AF: 47% SR: 60% |
178 | 79 | 99 (SR) | No comment | 79 | 94 | 5 | 0 | 0% |
| Abu-Alrub et al., 2022 (A) [53] | - | Case–control selected cross-sectional study | Apple Watch Series 5® | Apple Watch Series 5® (single-lead ECG) | 30 s | 12-lead ECG | All participants: 62 ± 7 | All participants: 44% | 200 | 100 | 100 (SR) | Excluded | 87 | 86 | 1 | 7 | 9.5% |
| Abu-Alrub et al., 2022 (B) [53] | - | Case–control selected cross-sectional study | Samsung Galaxy Watch Active 3® | Samsung Galaxy Watch Active 3® (single-lead ECG) | 30 s | 12-lead ECG | All participants: 62 ± 7 | All participants: 44% | 200 | 100 | 100 (SR) | Excluded | 88 | 81 | 6 | 5 | 10% |
| Abu-Alrub et al., 2022 (C) [53] | - | Case–control selected cross-sectional study | Withings Move ECG® | Withings Move ECG® algorithms (single-lead ECG) | 30 s | 12-lead ECG | All participants: 62 ± 7 | All participants: 44% | 200 | 100 | 100 (SR) | Excluded | 78 | 80 | 3 | 2 | 18.5% |
| Wegner et al., 2020 (A) [54] | Inpatients | Cohort selected cross-sectional study | AliveCor Kardia ECG monitor | AliveCor Kardia ECG monitor algorithm (single-lead ECG) | 30 s | 12-lead ECG | All participants: 64 ± 15 | All participants: 38.4% | 92 | 27 | 65 (SR) | In AF group | 19 | 45 | 8 | 0 | 21.7% |
| Wegner et al., 2020 (B) [54] | Inpatients | Cohort selected cross-sectional study | AliveCor Kardia ECG monitor | AliveCor Kardia ECG monitor algorithm (single-lead ECG) | 30 s | 12-lead ECG | All participants: 64 ± 15 | All participants: 38.4% | 92 | 27 | 65 (SR) | In AF group | 15 | 39 | 3 | 0 | 38% |
| William et al., 2018 [55] | Inpatients | Case–control selected cross-sectional study | Kardia Mobile Cardiac Monitor | Kardia Mobile Cardiac Monitor (single-lead ECG) | 30 s | 12-lead ECG | All participants: 68.1 [42.6–85.6] | All participants: 32.7% | 223 | 80 | 143 (SR) | In AF group | 57 | 96 | 6 | 2 | 27.8% |
| Chan et al., 2016 (B) [32] | Outpatients | Cohort selected cross-sectional study | 1st generation; AliveCor Inc. | AliveECG application (version 2.2.2) (single-lead ECG) | 30 s | single-lead interpreted by cardiologists | All participants: 68.4 ± 12.2 | All participants: 53.2% | 1013 | 28 (AF) | 985 (non-AF) | In control group | 20 | 979 | 6 | 8 | 0% |
| Dörr et al., 2019 (D) [37] | Inpatients | Case–control selected cross-sectional study | AliveCor Kardia system | Heartbeats application (Preventicus GmbH, Jena, Germany) (single-lead ECG) | 30 s | single-lead interpreted by cardiologists | All participants: 76.4 ± 9.5 AF: 77.4 ± 9.1 SR: 75.6 ± 9.8 |
All participants: 44.3% SR: 46.1% AF: 42.2% |
650 | 319 | 331 (SR) | Excluded | 279 | 262 | 7 | 1 | 15.5% |
| Orchard et al., 2016 [56] | Outpatients | Cohort selected cross-sectional study | AliveCor Heart Monitor | AliveCor Heart Monitor algorithm (single-lead ECG) | 30 s | single-lead interpreted by cardiologists | - | - | 972 | 38 | 934 (SR) | Excluded | 36 | 844 | 8 | 2 | 8.4% |
| Leńska-Mieciek et al., 2022 [57] | Inpatients | Cohort selected cross-sectional study | Kardia Mobile portable device (AliveCor Inc., San Francisco, CA, USA) | AliveCor app. (single-lead ECG) | 30 s | Single-lead ECG interpreted by cardiologist | All participants: 64.44 ± 10.52 | All participants: 48% | 50 | 1 | 49 (non-AF) | No comment | 1 | 42 | 7 | 0 | 0% |
| Lown et al., 2018 (B) [49] | Outpatients | Case–control selected cross-sectional study | WatchBP | WatchBP algorithm (modified sphygmomanometer) | - | 12-Lead ECG | All participants: 73.9 ± 6.1 | - | 418 | 82 | 336 (No AF) | In AF group | 79 | 314 | 22 | 3 | - |
| Lown et al., 2018 (C) [49] | Outpatients | Case–control selected cross-sectional study | Polar H7 | PH7: (A Real-Time Atrial Fibrillation Detection Algorithm Based on the Instantaneous State of Heart Rate) (ECG data sensor) | - | 12-Lead ECG | All participants: 73.9 ± 6.1 | - | 418 | 82 | 336 (No AF) | In AF group | 79 | 330 | 6 | 3 | - |
| Lown et al., 2018 (D) [49] | Outpatients | Case–control selected cross-sectional study | Bodyguard 2 | BG2 (A Real-Time Atrial Fibrillation Detection Algorithm Based on the Instantaneous State of Heart Rate) (heart rate variability) | - | 12-Lead ECG | All participants: 73.9 ± 6.1 | - | 418 | 82 | 336 (No AF) | In AF group | 79 | 331 | 5 | 3 | - |
| Reverberi et al., 2019 [58] | Inpatients | Case–control selected cross-sectional study | consumer-grade Bluetooth low-energy (BLE) HR monitor, of the chest-strap type | RITMIA™ (Heart Sentinel srl, Parma, Italy) (HR monitor) | - | 12-lead ECG | All participants: 66.2 ± 10.7 | All participants: 21.5% | 182 ** | 99 | 83 (non-AF) | Excluded | 96 | 79 | 4 | 3 | - |
| Chen et al., 2020 (C) [40] | Inpatients and outpatients | Case–control selected cross-sectional study | Amazfit Health Band 1S (Huami Technology, Anhui, China) | RealBeats Artificial Intelligence Biological Data Engine (Huami Technology) (PPG combined with single-lead ECG) | - | 12-lead ECG | AF: 70.4 ± 11.5 Non-AF: 59.3 ± 14.8 |
AF: 43.3% Non-AF: 52.6% |
401 | 150 | 251 (non-AF) | Unclear | - | - | - | - | - |
(N: number of participants; AF: Atrial Fibrillation; SR: Sinus Rhythm; AFL: Atrial Flutter; SD: standard deviation; * age distribution; ECG: electrocardiogram; DCNN: deep convolutional neural networks; rPPG: remote photoplethysmography; min: minutes; max: maximum; app: application; s: seconds; TP: true positive; TN: true negative; FP: false positive; FN: false negative; x (): median age (interquartile range); x []: average age [min. age–max. age]; Väliaho et al., 2019 (A): testing the AFEvidence algorithm; Chen et al., 2020 (A): testing PPG device; Chen et al., 2020 (B): testing single-lead ECG device; Chen et al. (C): testing combination of PPG and single-lead ECG; Väliaho et al., 2019 (B): testing the COSEn algorithm; Väliaho et al., 2021 (A): testing device performance when time interval between every measurement is 10 min; Väliaho et al., 2021 (B): testing device performance when time interval between every measurement is 20 min; Väliaho et al., 2021 (C): testing device performance when time interval between every measurement is 30 min; Väliaho et al., 2021 (D): testing device performance when time interval between every measurement is 60 min; Dörr et al., 2019 (A): testing performance of device when recording for 1 min; Dörr et al., 2019 (B): testing performance of device when recording for 3 min; Dörr et al., 2019 (C): testing performance of device when recording for 5 min; Dörr et al. (D): testing AliveCor; Desteghe et al., 2017 (A): testing the AliveCor in the cardiology ward population; Desteghe et al., 2017(B): testing the MyDiagnostick in the cardiology ward population; Desteghe et al., 2017 (C): testing the AliveCor in the geriatric ward population; Desteghe et al., 2017 (D): testing the MyDiagnostick in the geriatric ward population; Ford et al., 2022 (A): testing the Apple Watch 4; Ford et al., 2022 (B): testing the KardiaBand; Fu et al., 2021 (A): testing the device in supine position; Fu et al., 2021 (B): testing the device in upright position; Fu et al., 2021 (C): testing the device after individuals climbed to the 3rd floor; Van Haelst et al., 2018 (A): testing PPG device; Van Haelst et al., 2018 (B): testing single-lead ECG device; Santala et al., 2021 (1): published in October 2021; Santala et al., 2021 (2) (A): published in May 2021 and testing the device between the palms; Santala et al., 2021 (2) (B): published in May 2021 and testing the device in the chest; Abu-Alrub et al., 2022 (A): testing the Apple Watch 5; Abu-Alrub et al., 2022 (B): testing the Samsung Galaxy Watch Active 3; Abu-Alrub et al., 2022 (C): testing the Withings Move ECG; Wegner et al., 2020 (A): testing the lead I; Wegner et al., 2020 (B): testing the novel parasternal lead; Chan et al., 2016 (A): testing PPG device; Chan et al., 2016 (B): testing single-lead ECG device; Lown et al. (A): testing AliveCor; Lown et al., 2018 (B): testing the Watch BP; Lown et al., 2018 (C): testing the Polar H7 device; Lown et al., 2018 (D): testing the Bodyguard 2 device; ** N refers to ECGs before and after cardioversion.