Evaluating ChatGPT’s Multilingual Performance in Clinical Nutrition Advice Using Synthetic Medical Text: Insights from Central Asia

This dataset comprises ChatGPT-generated nutritional recommendations and sample diet plans for 50 distinct patient profiles. Each case study represents the patient profile with sociocultural background of ethnic groups in Central Asia and provides comprehensive data on age, gender, cultural and medical history, dietary patterns, anthropometric characteristics, disease-related functional indicators, biochemical and hematological parameters, and lifestyle behaviors. The LLM outputs include general nutritional recommendations and individualized daily meal plans based on Central Asian foods for each profile. At the same time, we are providing codes for translation into local languages.