Validation of a hand-mounted wearable sensor for scratching movements in adults with atopic dermatitis.

To the Editor: Currently, itch is evaluated primarily through subjective surveys. Technologies used as objective measures of scratching, the natural reflex to itch, show variable levels of performance.1Yang A.F. Nguyen M. Li A.W. et al.Use of technology for the objective evaluation of scratching behavior: a systematic review.JAAD Int. 2021; 5: 19-32Abstract Full Text Full Text PDF PubMed Scopus (4) Google Scholar We recently reported the development of an advanced acoustomechanic wearable sensor (ADAM) that is able to detect acoustomechanic signals generated by scratching via the fingers or wrist in a pediatric cohort with atopic dermatitis (AD), with high accuracy.2Chun K.S. Kang Y.J. Lee J.Y. et al.A skin-conformable wireless sensor to objectively quantify symptoms of pruritus.Sci Adv. 2021; 7eabf9405Crossref Scopus (19) Google Scholar Herein, we report the development of a new AI-algorithm that accurately quantifies scratching behavior and key sleep metrics on a single wearable device for adults with AD. A cohort of healthy adult subjects performed scratch and non-scratch activities while wearing the ADAM sensor. The data were then used to develop an algorithm for scratch detection. Patients with AD were recruited and consented to monitor nocturnal scratching behavior at home by wearing the ADAM sensor on the dorsum of the dominant hand. Manually-annotated infrared camera footage was scored by 2 authors as the ground truth for scratch events. Sleep quality was analyzed using a modified sleep-monitoring algorithm published previously from wrist-mounted actigraphy systems.3Lee J.M. Byun W. Keill A. Dinkel D. Seo Y. Comparison of wearable trackers' ability to estimate sleep.Int J Environ Res Public Health. 2018; 15: 1265Crossref PubMed Scopus (57) Google Scholar The number of hourly scratch events was determined for each sleep night by dividing the total number of scratch events by the patient's total sleep time (TST). We performed ANOVA and pairwise t-tests for key outputs across AD severity (mild, moderate, and severe). A total of 11 adults participated with mild to severe disease (validated Investigator Global Assessment [vIGA] 2.5 ± 0.7). A total of 73 nights (457 total hours) were analyzed. When compared to visually-observed scratching via infrared camera, the new AI-algorithm detected scratch events with 93% sensitivity, and 100% specificity, with an average Fleiss kappa of 0.84 compared to manually-coded infrared camera outputs.4Fleiss J.L. Cohen J. Everitt B.S. Large sample standard errors of kappa and weighted kappa.Psychol Bull. 1969; 72: 323-327Crossref Scopus (1062) Google Scholar Scratch events per sleep hour increased from 0.6 to 4.6 and 12.6 for mild, moderate, and severe disease with all pairwise comparisons statistically significant (Fig 1). When assessing the relationship between scratch events per hour and TST, there is a clear inverse correlation (R = 0.69). When comparing scratch events per night with TST, a similar inverse correlation is seen (R = 0.54). With increased scratch events per sleep hour, there is a direct correlation with increasing wake after sleep onset (WASO) (R = 0.64) (Fig 2). Despite its prevalence and impact on quality of life, itch has proven technically challenging to measure accurately with wrist-mounted systems. The flexible and miniaturized nature of the sensor allows for dorsal hand placement to accurately capture both finger and wrist scratching. Given that itch's primary reduction in quality of life is driven by worsening sleep,5Boozalis E. Grossberg A.L. Püttgen K.B. Cohen B.A. Kwatra S.G. Itching at night: a review on reducing nocturnal pruritus in children.Pediatr Dermatol. 2018; 35: 560-565Crossref PubMed Scopus (13) Google Scholar a single ADAM sensor is convenient in capturing both. Beyond utility for clinical trials, broader technology deployment to patients would enable costs that would resemble popular consumer fitness trackers. Additional software features may also link nocturnal scratch metrics to adherence, to medications, or changes in environmental conditions to empower patients and change behavior. This study describes the development and validation of an advanced wearable sensor for the accurate assessment of scratching behavior and sleep in adult patients with AD. This technology may be useful as a clinical outcome assessment for conditions where itch is prominent. K.S.C, M.K., D.R.S., and S.X. are employees in Sibel Health, a private company with a commercial interest in the technology. K.O. and A.I. are employees in Maruho Co, Ltd, a private company providing funding. S.X. is an inventor on a patent application related to this work filed by USPTO (no. PCT/US2019/018318, filed on February 15, 2019). The authors declare no other competing interests.