Validity of Galaxy Watch for Estimating Energy Expenditure During Intermittent Running: Cross-Sectional Study.

No statistically significant differences were found between Galaxy Watch models and indirect calorimetry for energy expenditure measurement during intermittent running.

• The K5 indirect calorimetry system showed a mean EE of 213.60 (SD 43.04) kilocalories
• The GW6 measured a mean EE of 219.53 (SD 35.70) kilocalories
• The GW7 measured a mean EE of 202.67 (SD 47.42) kilocalories
• All pairwise comparisons yielded P>.05

Galaxy Watch models demonstrated good Spearman correlations and moderate intraclass correlation coefficients compared to indirect calorimetry.

• Spearman correlations ranged from 0.63 to 0.70 across comparisons
• Intraclass correlation coefficients ranged from 0.65 to 0.74
• These values were described as 'good' and 'moderate' respectively

Mean absolute percentage error for both Galaxy Watch models ranged from approximately 10% to 13% compared to indirect calorimetry.

• Mean absolute percentage error values ranged from 10.10% to 12.55%
• This metric was used to quantify the magnitude of estimation error relative to the gold standard

Bland-Altman analysis revealed wide limits of agreement between Galaxy Watch models and indirect calorimetry.

• Limits of agreement for all comparisons (K5 vs GW6 and GW7) ranged from -61.93 to 65.80 kcal
• This range of approximately 128 kcal indicates considerable individual-level variability in device accuracy

The study sample consisted of 148 healthy adults who performed a standardized intermittent treadmill protocol combining walking and running.

• Sample comprised 80 men and 68 women
• Protocol consisted of walking at 5 km·h⁻¹ for 1 minute alternating with running between 8 and 16 km·h⁻¹ for 2 minutes
• Running speed was based on participant preference
• Total exercise duration was 27 minutes
• Both GW6 and GW7 were used simultaneously with the COSMED K5 portable metabolic gas analysis system as the gold standard

Prior validation studies have demonstrated that wearable activity monitors tend to overestimate energy expenditure when compared with indirect calorimetry.

• This pattern of overestimation across varying exercise intensities and types was noted as a known issue in the literature
• Discrepancies between results from different smartwatch models and gold standard methods were described as 'particularly critical across varying exercise intensities and types'