Cancer misinformation is becoming an increasingly complex issue. When a person or a loved one receives a diagnosis of possible cancer, that person, family and friends will try to better inform themselves in this area of healthcare. Like most people, they will turn to their clinician for guidance and the internet to better verse themselves on the topic. But can they trust the information provided online? Are there ways to provide a quick evaluation of such information in order to prevent low-quality information and potentially dangerous consequences of trusting it? In the context of the UL Cancer Research Network (ULCan), this interdisciplinary project aims to develop the Health Information Portal for Patients and Public (HIPPP), a web-based application co-designed with healthcare domain experts that helps to improve people navigate the health information space online. HIPPP will be used by patients and the general public to evaluate user-provided web-based health information (WBHI) sources with respect to the QUEST framework and return a quality score for the information sources. As a web application, HIPPP is developed with modern extreme model-driven development (XMDD) technologies in order to make it easily adaptable and evolvable. To facilitate the automated evaluation of WBHI, HIPPP embeds an artificial intelligence (AI) pipeline developed following model-driven engineering principles. Through co-design with health domain experts and following model-driven engineering principles, we have extended the Domain Integrated Modelling Environment (DIME) to include a graphical domain-specific language (GDSL) for developing websites for evaluating WBHI. This GDSL allows for greater participation from stakeholders in the development process of both the user-facing website and the AI-driven evaluation pipeline through encoding concepts familiar to those stakeholders within the modelling language. The time efficiency study conducted as part of this research found that the HIPPP evaluation pipeline evaluates a sample of WBHI with respect to the QUEST framework up to 98.79% faster when compared to the time taken by a human expert evaluator.