Agreement between a wrist-worn solar-powered watch and a chest-mounted light logger in quantifying daily time above 1500 lx in children: a pilot field comparison study

While time outdoors is known to reduce the risk of myopia onset in children, objective quantification across large-scale cohorts remains challenging. This pilot study aimed to assess the performance of a solar-powered wristwatch (SunGo TT120) against a chest-mounted HOBO light logger in measuring daily time above 1500 lx. In January 2026, eight children aged 7–12 years in Kaohsiung, Taiwan, wore a wrist-worn TT120 and a chest-mounted HOBO logger for 7 consecutive days. Daily TT120 values were read manually from evening dial photographs uploaded by parents via a mobile app featuring automated dial recognition. The primary outcome was the daily duration above 1500 lx. Statistical agreement was evaluated using Bland–Altman limits of agreement (LoA) adjusted for repeated measurements per subject, complemented by Spearman’s correlation coefficient (ρ) and the intraclass correlation coefficient (ICC). Among the ten enrolled children, two failed to complete monitoring, leaving 43 child-days available from eight children. The mean difference (HOBO minus TT120) was − 6.3 min/day (95% CI − 28.6 to 16.1), with 95% LoA ranging from − 153.3 (95% CI − 203.4 to − 117.7) to 140.8 (95% CI 105.2 to 190.9) min/day. Spearman’s ρ and ICC were 0.82 and 0.74, respectively, although within-child correlation remained weak ( r = 0.30). Upon removal of the child exhibiting the largest measurement discrepancy, the LoA were − 46.5 to 51.8 min/day. The automated program misread the hour hand in 3 of 54 photographs. In this small pilot study, daily duration above 1500 lx measured by a wrist-worn TT120 correlated strongly with the chest-mounted HOBO, accompanied by a minimal mean difference despite wide limits of agreement for individual days. The TT120 may describe average exposure in groups but cannot substitute for the HOBO in daily individual tracking. Larger studies incorporating a calibrated co-located reference standard and a validated automated reading are required.