Fitness App Statistics: Adoption, Adherence & Outcomes

21% of US adults said they regularly wear a smartwatch or fitness tracker

Survey of June 3-17, 2019. Use was 31% in households earning $75,000 or more against 12% in those earning under $30,000.

Women were more likely than men to regularly use a smartwatch or fitness tracker: 25% versus 18%

Same survey.

58.2% of US mobile phone users had downloaded a health-related app, and fitness and nutrition were the most common categories

National survey of 1,604 mobile phone users. App users tended to be younger, more educated and higher-income.

45.7% of people who had downloaded health apps had stopped using some of them

The main reasons were the burden of entering data, loss of interest and hidden costs.

In randomized trials, smartphone activity apps increased daily steps by 477 on average, a difference that was not statistically significant

Meta-analysis of six of nine eligible trials (1,740 participants in all). Programmes shorter than 3 months, and apps targeting activity alone, worked better.

Smartphone step-counting apps were within -6.7% to +6.2% of observed steps, while wearable devices ranged from -22.7% to -1.5%

14 adults walked 500 and 1,500 steps on a treadmill at 3.0 mph, each twice, carrying phones and wearing devices.

Six of seven wrist-worn devices measured heart rate with a median error below 5% during cycling, but no device estimated energy expenditure within 20%

60 volunteers of diverse age, size, skin tone and fitness, measured against telemetry and indirect calorimetry while sitting, walking, running and cycling.

Device error was higher during walking, in men, at higher BMI and with darker skin tone

The Apple Watch had the lowest overall error for heart rate and energy expenditure and the Samsung Gear S2 the highest. Treat calorie readouts as rough guides.

Sleep-wake devices detected sleep well (sensitivity 0.91-0.98) but wake poorly (specificity 0.37-0.65) against polysomnography

Eight people over four nights in a sleep laboratory. Classifiers built on Apple Watch and Oura Ring heart-rate and motion data reached specificities of 0.41-0.82.

Compared with the lowest step quartile (about 3,550 a day), mortality hazard was 40% lower at about 5,800 steps, 45% lower at 7,800 and 53% lower at 10,900

Meta-analysis of 15 cohorts with 47,471 adults and 3,013 deaths over a median 7.1 years (hazard ratios 0.60, 0.55 and 0.47).

Mortality risk kept falling with more steps up to about 6,000-8,000 a day in adults aged 60 and over, and 8,000-10,000 in younger adults

Stepping pace added little once total steps were accounted for.

Across trials of digital self-monitoring, more tracking was linked to greater weight loss in 74% of reported comparisons

Systematic review of 39 randomized trials. Digital tracking was used more consistently than paper in 21 of 34 comparisons; few programmes reached tracking on 75% of days.

Over a 12-month program, only consistent food trackers (more than 66% of days) lost significant weight, about 10 lb

Participants in a diabetes prevention and management program, grouped by how often they tracked. Rare and inconsistent trackers' weight followed holiday and summer swings.

eHealth weight-loss programmes produced 2.70 kg more weight loss than control groups and 1.40 kg more than minimal interventions

Meta-analysis of 84 randomized trials in adults delivered through websites, email, text messages, apps and devices. Evidence for maintenance and prevention was insufficient.

Postmenopausal women given a Fitbit increased moderate-to-vigorous activity by 62 minutes a week over 16 weeks, though the difference from a pedometer group was not significant

51 inactive women with a BMI of 25 or more. The Fitbit was worn on 95% of days, and steps rose by 789 a day.

Multi-component programmes with an app were more often effective (8 of 13 studies) than stand-alone app interventions (5 of 14)

Systematic review of 27 studies of apps for diet, physical activity and sedentary behaviour; the authors call the evidence modest.