This study evaluates the effectiveness of search engines, large language models, and retrieval-augmented generation in accurately answering health-related questions.

• Search engines answered 50-70% of health questions correctly
• LLMs showed competitive performance but struggled with up-to-date information
• RAG approaches improved LLM performance by providing relevant context
• Combining multiple methods yielded the best results overall

Implications for healthcare: Understanding these limitations is crucial for developing reliable digital health information tools that can combat medical misinformation while providing accurate guidance to patients and providers.

Evaluating Search Engines and Large Language Models for Answering Health Questions