Inter-rater agreement of trained and untrained observers in estimating platewaste using digital method

Most hospitals in Indonesia use real-time visual methods to measure the amount of inpatient plate waste. However, research showed the reliability of the digital method in estimating plate waste in hospital and school food service settings abroad. This study aimed to investigate the inter-rater agreement of two trained observers and mix-trained and untrained observers in measuring inpatient plate waste using the digital method. A total of 104 trays of inpatient lunch were collected. Each food item (rice, animal protein side dishes, plant-based protein side dishes and vegetables) was photographed before and after patient consumption. A trained observer weighed each item of food as gold. Then, two trained observers estimated the amount of plate waste from each food item image. In addition, an untrained observer also has the same procedures as the trained observer did by the digital method. This study showed the kappa of the trained observers was 0.84 for rice, 0.85 for animal protein side dishes, 0.77 for plant-based protein side dishes, and 0.65 for vegetables. On the other hand, the kappa of the trained and untrained observer was 0.62 for rice, 0.53 for animal protein side dishes, 0.67 for plant-based protein side dishes, and 0.49 for vegetables. In conclusion, the agreement between the two trained observers was higher. Meanwhile, the agreement between a trained observer and an untrained observer tends to be lower.