An innovative approach based on real-world big data mining for calculating the sample size of the reference interval established using transformed parametric and non-parametric methods

• DOI: 10.1186/s12874-022-01751-1
• Publication Date: 2022-10-20T13:04:27Z
• Authors: Chaochao Ma; Li’an Hou; Yutong Zou; Xiaoli Ma; Danchen Wang; Yingying Hu; Ailing Song; Xinqi Cheng; Ling Qiu
• Keywords: Sample (material); Parametric model

Abstract Background Currently, the direct method is main approach for establishment of reference interval (RI). However, only a handful studies have described effects sample size on RI and estimation size. We describe novel when establishing RIs using transformed parametric non-parametric methods. Methods A total 3,697 healthy participants were enrolled in this study. adopted two-layer nested loop to determine RI, thyroid-related hormone as an example. The was selected calculation result width confidence (CI) upper lower limit both stably < 0.2 times RI. Then, we calculated via methods hormones. Results Sample sizes thyroid stimulating (TSH), required by establish 239 850, respectively. free triiodothyronine (FT3), thyroxine (FT4), (TT3) (TT4) all less than 120, while those more 120. Conclusion estimating corresponding open-source code has been developed available applications. established suitable most analytes, with evidence based hormones indicating that different are analytes variations.