P996 - NUTBOT–ASK ANNE: EVALUATION ON NUTRITION CHATBOT TO SUPPORT PATIENTS WITH MALNUTRITION

Linked sessions

P996

NUTBOT–ASK ANNE: EVALUATION ON NUTRITION CHATBOT TO SUPPORT PATIENTS WITH MALNUTRITION

W. L. Ng1,*, P. D. Tay1, J. M. Y. Tan1, Y. L. Chan1, C. C. Y. Tong1

1Dietetics, Sengkang General Hospital, Singapore, Singapore

 

Rationale: NutBot–Ask Anne (A NutritioN Expert) is a rule-based chatbot designed to complement dietitians' care for patients with malnutrition to address the challenges of limited dietitian manpower. Evaluation of the prototype is important to determine the potential for scaling up and incorporating as a complementary tool for patient care.

Methods: Participants are recruited from July 2024 to December 2024 at Sengkang General Hospital Singapore, including malnourished patients/their caregivers, dietitians (DIT) and healthcare workers (HCW). They were allowed one-off trial of NutBot on all 6 curated topics and completed user experience survey including satisfaction, practicality and perceived usefulness using a 5-point Likert scale (5 being very satisfied/strongly agree).

Results: 76 participants (13 patients/caregivers, 5 doctors, 16 nurses, 20 Allied Health Professionals, and 22 DIT) completed the evaluation. Patients reported the highest satisfaction (mean score: 4.46/5), followed by HCW (4.15) and DIT (3.82). Majority found the chatbot’s suggestions practical (patients: 92%, HCW: 90%, DIT: 78%) and would recommend it to others/patients with malnutrition (patients: 92%, HCW: 83%, DIT: 78%). Notably, 96% perceived NutBot to be useful in helping them or their patients cope better with nutritional care. Participants suggested areas for improvement including multilingual support, voice interaction, an interface with better flow, fewer clicks and repetition, inclusion of graphics over text, and expanded content such as cultural food options, meal planning and guidance for progression of condition.

Conclusion: NutBot prototype has demonstrated high satisfaction, practicality and perceived usefulness by various types of users as a promising complementary nutrition support tool. Artificial intelligence should be considered to enhance NutBot and reality application evaluation with larger number of patient/caregivers is required.

Disclosure of Interest: None declared