A Novel Direct Observation Video Coding Tool Shows High Reliability For Muscle Strengthening Exercise

BACKGROUND: Regular muscle strengthening exercise (MSE) helps prevent type 2 diabetes, heart disease, osteoporosis, depression, anxiety and other diseases. However, qualification and quantification of this type of training remains challenging. Video coding is an increasingly valuable tool in both clinical and fitness settings. Video provides a permanent record of technique, exercise form, and adherence, all highly important for optimizing effectiveness and minimizing injury. Manual coding allows detailed in-depth analysis of exercise choice and performance along aspects such as modality and movement pattern, which afford an informative evaluation of the outcomes of MSE protocols. Developing video based coding relies on high levels of precision and consistency for research and exercise-based intervention planning. PURPOSE: The aim of this study was to test intra- and inter- rater reliability of a novel tool for video coding of MSE sessions. METHODS: A group of 39 participants, aged 18 to 64 years old (35.2 ± 12.4 years), provided up to 4 videos each for a collective 79 videos of their MSE sessions. For each video, coding was completed by two of the trained students, selected randomly. The number of exercise segments and the total duration of these segments were considered for analysis. Inter-rater reliability was established using Intraclass Correlation Coefficients (ICCs, & 95% CI), where ICC(2,k) estimated the overall reliability of the rater scheme and ICC(2,1) estimated the consistency of a single rater with the group. RESULTS: Coders identified 19.5 ± 10 segments per video, representing 9.8 ± 6.2 minutes of MSE. The ICC(2,k) for the count of exercise segments was 0.85(.7-.9), while that for total duration of exercise was 0.84(.7-.9), indicating good overall reliability of the rater scheme. In contrast, ICC(2,1) for the count was 0.74(.6-.8) and that for time was 0.73(.6-.9), reflecting moderate consistency of individual raters with the group. All p < .05. CONCLUSION: The results indicated that the novel video coding tool was reliable for use by trained students. This work advances the ability to track a detailed analysis of MSE, which will improve research and exercise-based intervention planning.