Inter‐device concordance of Veggie Meter—A reflection spectroscopy to measure skin carotenoids

• Published in: Journal of biophotonics 2023-08, Vol.16 (8), p.e202300071-n/a
• Main Authors: Obana, Akira, Asaoka, Ryo, Takayanagi, Yuji, Gohto, Yuko
• Format: Article
• Language: English
• Subjects: Carotenoids; Coefficient of variation; Correlation coefficient; Correlation coefficients; Error analysis; inter‐device error; intra‐class correlation coefficient; linear regression; reflection spectroscopy; skin carotenoids; Skin tests; Spectroscopy; Systematic errors
• ISSN: 1864-063X; 1864-0648
• DOI: 10.1002/jbio.202300071

The Veggie Meter® (VM) is a refraction spectrometer that detects skin carotenoids. We tested the variability of the two modes (single‐scan and averaging modes) of four VMs of three different versions (VM‐1, 2, 3, 4) with 92 healthy volunteers. Both the modes exhibited a high intra‐class correlation coefficient (ICC); however, the averaging mode had a significantly lower coefficient of variance than the single‐scan mode. The Bland–Altman analysis showed a systematic error between VM‐1 and other three VMs. The actual errors between the VM‐1 and the other three VMs in the averaging mode were 7.4%, 10.4%, and 11.8% of the median of VM scores, but the errors could be reduced to 2.8%, 6.3%, and 7.0% by compensating VM scores using the regression equations. Averaging mode had higher accuracy than single‐scan mode. The reliability of the VMs was validated by the small coefficient of variation and high ICC. The error was improved using linear regression compensation. The Veggie Meter® (VM), a reflection spectroscopy to measure skin carotenoids, can assess vegetable intake very easily. We validated four VMs of three different versions and found them to be highly reliable. We proposed regression equations to further reduce inter‐device errors.