The Development of an Ontology to Facilitate Personalised Patient Education

Background: Traditional patient education involves the distribution of standardised printed material to all patients diagnosed with a condition. Nevertheless, the effectiveness of this education may be hampered by the patient’s inability to engage with the material. Personalisation provides a means to overcome this concern by providing education that is tailored towards the patient’s characteristics, needs and health objectives. Personalising education renders it more understandable for the patient and can reduce a feeling of being overwhelmed or confused by a volume of unfamiliar terms and images. Semantic web technologies provide a means for linking and reasoning on data that facilitates adaptation of the content and presentation of patient education.Objective: This study aimed to develop an ontology and identify which patient data was necessary to facilitate a personalised patient education experience.Methods: A literature review was undertaken to investigate patient education experiences. Subsequently it was decided that diabetes and obesity would be chosen as the case studies. Decreasing levels of physical activity, along with increases in unhealthy dietary habits, have contributed to a worldwide increase in diabetes and obesity. The main aspects of management may include medication, an increase in physical activity and following a healthy diet. Patient education is therefore a significant component of healthcare provision for patients with these conditions. The literature review was followed by small focus group discussions to consider use case scenarios and identify which patient datums can be used to tailor education material. The ontology development methodology included gathering diabetes and obesity related terms from relevant literature.Results: A web based framework for providing personalised patient education has been developed. This included an ontology (built using the Web Ontology Language) that captures information about patients who have diabetes and who are possibly obese.The ontology contains over 300 classes and associated relationships. The range of information modelled includes treatments specific to diabetic patients such as medications and health checks, along with more generic treatments for diabetes and obesity such as lifestyle changes. The ontology describes short-term and long-term health risks for each condition and a range of symptoms. Terms relating to healthy nutrition and sporting activities were also represented. The ontology also captures personal information for each patient including age, gender and literacy level, along with health attributes such as Body Mass Indicator and fitness level.The educational material will address symptoms, health concerns and treatments that are specific to each patient. For example, for a patient with type 2 diabetes, the information provided could relate to treatments such as cholesterol checks and weight management, while for another patient with the same condition the education may concentrate more on diabetic medications. Tailoring the education in this way ensures that the education is focused on the needs of the patient and can change dynamically in response to changes in their medical status. Moreover the patient is not overwhelmed by information that is not directly applicable to their needs. The content may include text and images that are personalised to the age group and gender of the patient. This will render the education more familiar and accessible. The content may also include context-aware suggestions for self-management of diabetes and obesity that are specific to the patient’s preferences and lifestyle. For example, if a patient wishes to become more active, information about preferred sporting activities that are taking place close to their home would be provided.The educational material can also be tailored to the health literacy of the patient. A patient’s health literacy can affect their ability to comprehend health information. Health literacy can be affected by contextual factors such as a change in the patient’s level of anxiety or medical condition. The ontology captures information about the patient that can be used to determine their literacy level and can infer whether any pertinent contextual factors exist. The framework thereby provides a means to adapt to contextual changes by dynamically correlating a patient’s health literacy level with textual information. This ensures that the education is comprehensible and useful for the patient in their current situation.Evaluation of the framework will focus on assessing whether a personalised approach is more effective when compared to standardised generic printed materials. This will be executed using a cross sectional randomised control trial that will assess user learning of the standardised and personalised education. To measure user engagement a number of additional apparatus may be used, e.g. eye tracking the user’s eye gaze.Conclusions: This study developed a semantic ontology which can be used to seamlessly provide education that is tailored to the needs of each individual patient and therefore may improve the patient’s understanding and self management of the disease (e.g. diabetes).