Table 1.
Characteristics of the included randomized controlled trials that used biometric monitoring devices (BMDs) for measuring outcomes (n = 75)
| Characteristic | Trials, no. (%) |
|---|---|
| Study design | |
| Parallel | 56 (75) |
| Cross-over | 15 (20) |
| Cluster | 3 (4) |
| Factorial | 1 (1) |
| Intervention assessed | |
| Non-pharmacological | 57 (76) |
| Pharmacological | 18 (24) |
| Number of patients randomized, mean (SD) | 144 (322) |
| Region of the primary author | |
| Europe | 36 (48) |
| North America | 20 (27) |
| Asia | 15 (20) |
| South America | 2 (3) |
| Africa | 1 (1) |
| Oceania | 1 (1) |
| Funding | |
| Public | 64 (85) |
| Private or mixed funding | 7 (9) |
| Not reported | 4 (5) |
| Medical condition investigated in the study | |
| Diabetes | 25 (33) |
| Cardiac and vascular diseases (incl. hypertension) | 21 (28) |
| Chronic obstructive pulmonary disease | 6 (8) |
| Cancer | 4 (5) |
| Insomnia | 4 (5) |
| Renal disease | 2 (3) |
| HIV/AIDS | 2 (3) |
| Obstructive sleep apnea | 2 (3) |
| Osteoarthritis/osteoporosis | 2 (3) |
| Psychiatric disorders | 2 (3) |
| Multiple sclerosis | 1 (1) |
| Psoriasis | 1 (1) |
| Rheumatoid arthritis | 1 (1) |
| Obesity | 1 (1) |
| Spinal cord injury | 1 (1) |
| Outcomes measured with BMDs per trial, mean (SD), no. | 6 (8) |
| Type of sensor used* | |
| Inertial measurement unit sensors | 43 (57) |
| Electrochemical sensors (including continuous glucose monitoring) | 21 (28) |
| Pressure sensors (including smart cap bottles) | 6 (8) |
| Electrodes | 4 (5) |
| Temperature sensors | 2 (3) |
| Optical sensor | 2 (3) |
| Management of missing BMD outcome data | |
| Unclear | 26 (35) |
| Exclusion of patients with missing outcome data | 25 (33) |
| Multiple imputation | 8 (11) |
| Use of models robust for missing data | 7 (9) |
| Last observation carried forward | 2 (3) |
| Value inferred by investigator | 2 (3) |
| Missing values considered as failures | 2 (3) |
| Other | 3 (4) |
| Reporting on management of incomplete BMD outcome data | 24 (32) |
*Exceeds 100% because some trials used multiple sensors