Dr. BellyWell – keep your belly well



Dr. BellyWell – keep your belly well…  

According to the WHO excess body weight is one of the most serious public health challenges of the 21st century. The prevalence of obesity has tripled in the last two decades and has now reached epidemic proportions. Overweight is responsible for a large proportion of disease in Europe and probably Worldwide.  

The S3 Guidelines of Obesity show that bariatric surgery has come to be one main solution for helping obese patients to effectively lose weight and prevent comorbidities of obesity. Although several studies show that there is an issue with poor longterm outcome, digestive problems & sometimes even bad quality of life after surgery 

Since gastric bypass surgery is the current gold standard of bariatric surgeries, the information and assistance given by Dr. BellyWell will focus solely on this form of bariatric surgeries.  

Who is Dr. BellyWell and how do we plan to achieve adequate support in the pre- and postoperative care of gastric bypass patients?  

Dr. BellyWell is an application, which assists patients receiving gastric bypass surgeries. The focus is on giving those patients valuable information before their scheduled intervention and to primarily support their recovery after the surgery.  

To monitor the users’ progress and give suitable advice for different scenarios, Dr. BellyWell will need regular updates about the users’ daily routines and meals. For ease of access and a user-friendly experience, Dr. BellyWell utilizes a voice activated speech interface, using natural language processing. The voice service of choice for this project is Amazon Alexa. On the one hand due to Amazon Alexas great distribution worldwide and in Europe, on the other hand because of Amazons great documentation for developers and the active developer community. Many Amazon Alexa enabled devices in circulation mean a large potential group of users who can benefit from an application like Dr. BellyWell.  

How does it work?  

In order to give relevant and evidence-based advice, Dr. BellyWell needs to have access to a database containing medical data about gastric bypass surgeries and currently used guidelines for patient care before and after gastric bypass surgeries.  

To keep the advice relevant for each user, Dr. BellyWell needs data input by its users. Such as eating patterns, types of meals, current weight and possible discomforts. To keep the compliance rate as high as possible, the regular input of data by the patient needs to be as easy and non-disruptive as possible. Hence the use of natural language processing. Dr. BellyWell will decide on what to ask the users depending on the time of day and the number of previous conversations.  

This daily information is stored into Dr. BellyWell’s database, which then triggers different possible scenarios depending on the user’s weight trend. If the user loses weight according to the guidelines, Dr. BellyWell will encourage the user. If the user gains weight or loses weight to fast, Dr. BellyWell will react accordingly and inform the user to change his or her eating behavior. Or in extreme cases, advice the user to contact his or her physician.  

decicion tree for weight input

Additionally, to gastric bypass patients there are further groups who can benefit from the utilization of Dr. Bellywell. Doctors, therapists and insurance groups will be able to profit from adequate care of gastric bypass patients. Doctors and therapists will have access to the users’ data and can customize and optimize the treatment accordingly, if the users decide to share their data with doctors and/or therapists. If doctors and therapists can use their time with patients more efficient, insurance groups will save money in the long term. 


Out database of choice is the Amazon provided DynamoDB. It is planned to utilize different tables in the same database to store user relevant persistent data. 

Lambda Function

The lambda function provides the logic of the skills backend. It can be coded in different languages including Python and JavaScript, we use the latter.

The AWS console provides the database and logic for our skill, the invocation of the skill, intents and slots are configured in the Alexa Skills Kit developer console. 

The invocation is the skills name by which the user prompts Amazon Alexa to start the skill, in our case the user could for example state “Alexa, open doctor belly well.” to start using the Application. 


user interaction & backend

Dr. BellyWell was evaluated by 6 healthcare professionals who were shown an instructional video on how to use the speech assistant. After watching the video, the test users perfomed a task on Dr. BellyWell (e.g. saving their current weight).
Following the test of the application, the test users filled out a questionnaire about their thoughts on the speech assistant.

Using descriptive statistics, the findings were analyzed and visualized with Microsoft Excel.

results mean overall – Dr. BellyWell

Although none of the test users use speech assistants in their clinical environment, they could imagine using speech interfaces to assist in the care for their patients or to enable the patients for a higher amount of self-care. The test users also concluded that Dr. BellyWell could improve the long-term success of bariatric bypass surgeries. They also showed high confidence in possible improvements in extramural post-operative care of the patients. The test users were a bit sceptic about possible risks for patients. Misuse of the speech assistant or neglection of appointments with a health professional could have dire consequences. These results show that, on the one hand an interactive speech assistant, given it provides evidence based and scientifically verified information, could improve the post-operative progress of bariatric bypass patients as well as heighten the rates of long-term success of sustainable weight loss. On the other hand, a speech assistant can only be an addition to care by real life health professionals and should not be used without proper supervision.

Future Enhancements

  • More content for pre- and post Surgery 
  • Add background Information for different kinds of surgery