Reliability, validity, usability, and feasibility of the k-force dynamometer in patients diagnosed with ICU acquired weakness

The paper details the design and testing of extracting Mel-Frequency Cepstral Coefficients (MFCC) from a speech signal on an ARM Cortex-M0 System-on-Chip. A MFCC algorithm in C language was written which was compiled and uploaded onto the memory of the SoC. The aim of the project is to optimize the code to reduce program memory, keep a high accuracy and lower computation time. The optimized MFCC implementation resulted in a C MFCC implementation taking 17420 Bytes of program memory which cepstral coefficients have an accuracy of 83.98% compared to the MFCCs obtained by the mfcc function in MATLAB. Methods: The study was conducted using a prospective single-centered pilot feasi-bility study, which took place from February 2022 until July 2022 in the ICU of Gelre Hospitals. This study was divided into three sub-studies. The technical validity was as-sessed through the comparison of calibrated weights of a fixed mass that were employed in the center of the K-force grip and the readings the k-force grip generated, using a the Pearson correlation coefficient (R), Bland Altman plots, and Ordinary Least Squares regression (OLS) as statistical analysis. ICU-patients diagnosed with ICU-AW were in-cluded to measure their hand grip strength through intra- and inter-rater measurements in the reliability study, assessed through the Intraclass Correlation Coefficient (ICC). Be-sides, healthcare professionals were asked to assess the usability and feasibility of the device through a survey, which included the System Usability Scale, and the thinking aloud method in the usability and feasibility study. Results: The R and the OLS regression for the technical validity showed a perfect lin-ear association between the fixed weights and the readings and no type of bias could be identified. The ICC calculations of the 7 patients included in the reliability study demon-strated excellent intra-rater reliability with a ICC of 0.875 and 0.961. However, with an ICC of 0.511 and 0.397, the inter-rater reliability was rated as low to moderate, but was not found to be statistically significant. Additionally, the device was deemed feasible in the ICU environment and the usability was rated excellent by the 11 healthcare profes-sionals with a SUS score of 87.0. Conclusion: The K-force grip can be seen as valid, useful and feasible in ICU-AW patients. Further research with an increased sample size needs to be done to gain better insights into the reliability of the device.