ADDietCoach: A Personalized Virtual Diet Coach for Alzheimer's Disease

Introduction

As people worldwide are living longer, the pace of population ageing around the world is also increasing dramatically. Aging is related to many risk factors of noninfectious diseases, including Alzheimer's disease (AD) (Xia et al., 2018). AD is defined by cognitive function progressively worsen. It starts gradually with early signs including patchy memory loss and subtle changes in behavior. The disease is progressively proceeding; the patient may need aid on all aspect of personal care including their daily diet care. AD is one of the best-known aging‐linked diseases. According to Alzheimer's Association, approximately 5.8 million Americans of age 65 and above are living with dementia from Alzheimer’s in 2020; 80% of them are 75 years of age or older.

The majority of AD caregivers are unprofessional family members or friends who do not have any experience in dealing with such disease. One of the most important life aspects that informal caregivers need assistance in is how to provide healthy daily diet for AD patients. Appropriate nutrition is vital to maintain the body healthy and strong. For AD patients, poor nutrition can significantly raise the behavioral symptoms and cause weight loss (Pivi et al., 2012). Current epidemiological evidence suggests that an increase intake in saturated fatty acids (SFA) may have adverse impacts on mental functions, whereas increased polyunsaturated fatty acids (PUFA) and monounsaturated fatty acids (MUFA) may be protective against cognitive deterioration (Solfrlzzi et al., 2008). Present research on the effects of dietary components or food products on AD is inconsistent, mainly because people consume meals with complex combinations of nutrients; food items are likely to be synergistic causing difficulties in deciding which portion of the diet has a beneficial or adverse effect (Gardener et al., 2012).

There are lots of online websites and information about food and nutrition for AD care. However, it is not easy to find the most relevant and accurate information specifically to the patient’s special need. AD patient themselves may have difficulty to read and understand the content. AD caregivers are mostly informal caregivers of family members who normally have very busy schedule and may not have the time to read and study these materials. Mobile applications are also available for nutrition coaching e.g., Myplate (MyPlate Calorie Counter, n.d.), diet tracking e.g., Calorific (Calorific - What Do Calories Look Like?, n.d.), and meal planning e.g., emeal (Meal Planning Made Simple - EMeals, n.d.).

Despite the large number of applications and tools (Sarwar & Javed, 2019) (Javed et al., n.d.), they do not target specifically for people living with AD; the education, coaching and recommendations given by these applications tend to be generic without considering the uniqueness of each patient including their preferences, other health conditions, culture, and traditions; these applications do not consider caregivers’ conditions, such as their preference, their time limit to prepare a meal, and the economic status to afford the food items. To address the limitations in existing works, the authors propose a personalized AD diet coaching system. The system will respect the special socio-economic, cultural, ethnic, and geographical status, particularly to AD patients and their caregivers. By integrating the users’ profile with general clinical AD care guidelines, the system can make useful recommendations that are more appropriate to AD users. The main contributions of this work can be summarized as follows:

• 1.

  Define and reuse a set of ontologies to provide computational knowledge about food, nutrition, and AD care; the ontologies also describe patient and caregiver’s characteristics to provide evidence for personalization. Together the proposed ontologies provide knowledge foundation for personalized recommendation.

• 2.

  Collect, verify, and convert AD medical diet guidelines to logic rules which enable reasoning-based decision making and recommendation.

• 3.

  Implement a personalized virtual coach to provide context-aware and personalized diet education and recommendation to AD patients and their caregivers.