A scoping review of research validating precision wearable sensors for monitoring dairy cattle behavior
Research Poster Health & Life Sciences 2025 Graduate ExhibitionPresentation by Alana Lee
Exhibition Number 142
Abstract
For our scoping review, we assessed validation research that used precision wearable technology to record dairy cattle behavior to 1) quantify common statistical criteria used, and 2) set validity criteria and compare the literature to our standards. We used the PICO method to create a search method. Keywords were extracted from PubMed and Web of Science; 101 articles were sifted from 2955 articles. We assessed the literature using Power BI to extract precision, accuracy, bias, reproducibility (sample size, sensor type), and comparison method. Recorded dairy cattle behavior was sorted into 10 categories for consistency. We considered a study to have validity if there was high precision 85%, they reported no bias, there was reproducibility, a comparison method, and behavior was defined. One researcher manually reviewed papers that were reported as lacking bias measurements because multiple statistical techniques were used to assess bias. We found a very high reproducibility 93% (94/101) among studies. Activity behavior (61/101), feeding behavior (59/101), and resting (55/101) were commonly validated with wearables in dairy cattle. However, precision was rarely calculated 40% (40/101) among studies. For studies that reported precision, 90% (36/40) met our criteria for reproducibility. However, we observed that many precise studies failed to report a measurement for bias, 35% (14/40). Thus, a negligible level of research met our validity criteria 14% (14/101). We suggest that wearable validation studies for dairy cattle behavior need to measure precision, bias, and be reproducible to be considered valid.
Importance
This scoping review synthesized 101 studies from the past 11 years and found that only 40% (40/101) evaluated precision, with just 14% (14/101) meeting the full criteria for validity (precision 85%, repeatability, and bias evaluation). By identifying these gaps, this review advocates for improved validation standards. The societal impact of more precise, unbiased wearables includes better decision-making in dairy farming, leading to enhanced animal welfare, increased productivity, and more sustainable farming practices.