| Autor: |
Ibrahim MS; Faculty of Medicine, Medical Campus, Universiti Sultan Zainal Abidin, Kuala Terengganu 20400, Terengganu, Malaysia., Naing NN; Faculty of Medicine, Medical Campus, Universiti Sultan Zainal Abidin, Kuala Terengganu 20400, Terengganu, Malaysia., Abd Aziz A; Faculty of Medicine, Medical Campus, Universiti Sultan Zainal Abidin, Kuala Terengganu 20400, Terengganu, Malaysia., Makhtar M; Faculty of Informatics and Computation, Gong Badak Campus, Universiti Sultan Zainal Abidin, Kuala Terengganu 20300, Terengganu, Malaysia., Mohamed Yusoff H; Faculty of Medicine, Medical Campus, Universiti Sultan Zainal Abidin, Kuala Terengganu 20400, Terengganu, Malaysia., Esa NK; Faculty of Medicine, Medical Campus, Universiti Sultan Zainal Abidin, Kuala Terengganu 20400, Terengganu, Malaysia., A Rahman NI; Faculty of Medicine, Medical Campus, Universiti Sultan Zainal Abidin, Kuala Terengganu 20400, Terengganu, Malaysia., Thwe Aung MM; Faculty of Medicine, Medical Campus, Universiti Sultan Zainal Abidin, Kuala Terengganu 20400, Terengganu, Malaysia., Oo SS; Faculty of Medicine, Medical Campus, Universiti Sultan Zainal Abidin, Kuala Terengganu 20400, Terengganu, Malaysia., Ismail S; Faculty of Medicine, Medical Campus, Universiti Sultan Zainal Abidin, Kuala Terengganu 20400, Terengganu, Malaysia., Ramli RA; Faculty of Medicine, Medical Campus, Universiti Sultan Zainal Abidin, Kuala Terengganu 20400, Terengganu, Malaysia. |
| Abstrakt: |
During the initial phase of the coronavirus disease 2019 (COVID-19) pandemic, there was a critical need to create a valid and reliable screening and surveillance for university staff and students. Consequently, 11 medical experts participated in this cross-sectional study to judge three risk categories of either low, medium, or high, for all 1536 possible combinations of 11 key COVID-19 predictors. The independent experts' judgement on each combination was recorded via a novel dashboard-based rating method which presented combinations of these predictors in a dynamic display within Microsoft Excel. The validated instrument also incorporated an innovative algorithm-derived deduction for efficient rating tasks. The results of the study revealed an ordinal-weighted agreement coefficient of 0.81 (0.79 to 0.82, p -value < 0.001) that reached a substantial class of inferential benchmarking. Meanwhile, on average, the novel algorithm eliminated 76.0% of rating tasks by deducing risk categories based on experts' ratings for prior combinations. As a result, this study reported a valid, complete, practical, and efficient method for COVID-19 health screening via a reliable combinatorial-based experts' judgement. The new method to risk assessment may also prove applicable for wider fields of practice whenever a high-stakes decision-making relies on experts' agreement on combinations of important criteria. |
