A food-based intervention in aged care that improved protein and calcium intakes in residents to recommended levels was associated with reduced falls and fractures and prevention of weight loss and decline in nutritional status1,2. Learnings from this were developed into a food service training program so chefs and cooks could improve food provision in line with the intervention. While effective in improving food intake of residents, a potential limitation observed was food delivery at mealtimes. To address this limitation, we developed and piloted a training program for care staff to improve mealtime delivery. The initial training was provided to seven care staff with feedback guiding the next iteration. Prior to conducting the next training session, mealtime observations of nine care staff were performed by research staff at three lunches and dinners, based on the CHOICE+ Mealtime Practices Checklist3, that assesses the dining environment, meal set-up and relationship-centred practices. Positive mealtime practices were considered consistent if they occurred 75-100% of the time, relatively consistent if occurring 51-74% of the time or needing improvement (<50%). From these mealtime observations and key takeaways from a focus group session with care staff, the training content was finalised and the training delivered to nine additional care staff. The training focused on challenges of sensory impairment in residents when eating and implementing strategies to support residents with their meals. Positive mealtime practices were observed for cleanliness (100%), standard menu displayed (100%), positive body language (100%) and respectful practices such as sitting at eye level and facing the resident when assisting with eating (100%). Time taken for all residents to receive their meal ranged from 3-17 minutes. Meal temperatures were not recorded. Positive mealtime practices requiring improvement (occurred < 50% of the time) include having a pleasant aroma in dining room (only occurred 44% of the time), not administering medications during mealtime (only occurred 22% of the time), texture-modified meals displayed on menu (0% never occurred), texture-modified meals not mixed together (only occurred 33% of the time) and residents offered second servings (0% never occurred). While some mealtime observations requiring improvement were operational (i.e. displaying texture-modified menu), the training aligned with others identified (i.e. mixing texture-modified meals together). Combined with care staff feedback, key learnings were not mixing pureed foods together to preserve flavour and taste, describing the meal to residents to encourage protein consumption, and improved knowledge of protein-rich foods to offer. The pilot training received positive feedback from care staff and support from management. A larger randomised trial will validate the effectiveness of the training on improving food intake and mealtime experience for residents.