From Knoesis wiki
Jump to: navigation, search
DisasterRecord Logo

Knowledge enabled personalized chat-bot system for self-management of asthma in pediatric population.


Asthma is a lung inflammatory disease that has prevailed more than 26 million people in the United States, out of which 6.1 million are children [ref]. It is the most chronic condition among the pediatric population[ref] accounting for 13.8 million missed school days each year [ref]. Though incurable, asthma can be controlled and managed by strict adherence to a medication care plan and by avoiding the triggers [ref]. However, due to a lack of consistent adherence to the asthma care plan, and inadequate information about the patient’s environment, asthma management can be challenging [ref]. These issues are further compounded by its multi-factorial nature, where every patient is sensitive to different triggers, and react differently even when exposed to the same trigger [ref]. This demands personalized care beyond the regular hospital setup to which the clinical professionals are limited. The advent of digital health monitoring technologies such as smart wearables and increasing adoption of mobile devices and low-cost sensors translate to an unprecedented amount of data being generated and collected on a continuous basis. It is now possible to monitor long-term medication adherence, exposure to environmental triggers, and asthma control level in real-time.


  1. Continuous monitoring of the patient’s asthmatic condition requires their active involvement in the process. However, due to the static user interface of current monitoring systems, poor patients compliance badly affects the quality and quantity of data collection.
  2. In asthma care, the effectiveness of treatment is affected not only by patient compliance but also by the inhalers usage techniques. If asthma inhalers are not used in a proper way, it has no effect on patients asthmatic condition which may lead to worsening of asthma. For the effectiveness of asthma treatment, patients need to be educated on proper inhaler techniques and encouraged towards medical compliance.
  3. Asthma is a multi-factorial disease; each patient reacts to both triggers and treatment differently. A general treatment approach such as a generic asthma care plan proves to be ineffective in such a scenario. A more personalized approach to monitor patients and their environmental factors are required.


Recent years have seen immense maturity in Artificial Intelligence (AI) research which has in part proliferated the growth of intelligent conversational systems, also known as chatbots. They are increasingly popular due to their capability of simulating human-like conversations with a user through speech, text, smart display, and multimodal communication [ref]. Contrary to static applications, it can understand user intents and choices through interactions and communicate accordingly. An exciting trend is that chatbot-assisted queries are 200 times more conversational than search, and users are demanding more human-style interaction [ref]. Such technology is rapidly gaining traction in the healthcare domain where professional care is limited [ref]. It is evident that current long-term healthcare monitoring demands a more ubiquitous solution. Chatbots are capable of delivering a more convenient and accessible approach through cost-effective mediums. Nonetheless, they are limited in their inherent ability to contextualize, learn interactively, and provide the proper hyper-personalization needed to hold a meaningful conversation.

By leveraging such technology and considering the challenges mentioned earlier, we introduce a knowledge-enabled personalized chatbot system kBot, that is intended to replace our use of the mobile app in the kHealth Framework. It is capable of interacting with a patient through a contextualized and personalized manner on hand-held devices (mobile phones and tablets), or as part of smart displays (e.g., Google Home Hub). It curates and contextualizes its asthma domain knowledge from different online sources such as Asthma and Allergy Foundation of America (AAFA) [ref], verywell health [ref], Mayo clinic [ref], and webMD [ref] as well as local inputs from our clinical collaborators. It then aggregates this knowledge with patients’ data such as symptoms and medication intake to deliver a personalized conversation experience. The first version of kBot reported here takes on a design approach that centers around addressing medication non-adherence issue in pediatric asthma management and assessing environmental triggers at an individual level. The ultimate goal is to bridge and simplify long-term real-time monitoring of asthma condition, alert on potential environmental triggers, and educate the patients on various asthma self-management skills. Our current implementation is limited to the Android ecosystem but can be easily adapted to iOS.


kBot Architecture

kBot follows a client-server architecture where the client is a lightweight front-end chat interface, and the back-end server is a standalone web application hosted in the cloud. Users could interact with kBot through the client application in both modes of communication: text and voice. A unique CHAT ID is assigned to each user profile during the initial user profile setup which works as a primary ID to identify the user. This ID is generated in client app using the MD5 hashing algorithm that takes a string of any length as input and encodes it into 128-bit fingerprint as hash output. It is used by the client to authenticate the communication request with the server and maintain the user session.

The client communicates with patients at least twice a day and logs all the patient-specific conversation history in the server. JSON [ref], a lightweight data-interchange format, is used for client-server mode of communication. The data of interest are then extracted from the conversation logs in the server. Concurrently, the server continuously monitors and collects environmental data, at different frequencies, through third-party weather Application Programming Interface (API) for the zip codes specific to each patient.


  • ReaCTrack Architecture
  • ReaCTrack Architecture

kBot uses asthma domain knowledge to conduct contextually relevant conversation. Asthma knowledge is manually extracted from different online sources such as Asthma and Allergy Foundation of America (AAFA) [ref], Verywell health[ref], Mayo clinic[ref], and webMD[ref] as well as inputs from our clinical pulmonologist. This information is curated to best represent the domain knowledge (context) and stored in kBot cloud server as a knowledge base in the form of NoSQL database as well as preloaded to DialogFlow as entities. Patients can, therefore, ask kBot on asthma-relevant questions to learn more about asthma zones [ref], symptoms, various triggers, medications (their usage and side-effects), and self-management skills.

Apart from asthma domain knowledge, kBot uses rich media contents such as images and videos to deliver and present information more effectively. Images of different asthma medicines and inhalers are used to help patients to quickly identify the various types of medicines, and video contents are used to educate on skills like how to use a Metered Dose Inhaler. The image contents are available from A Guide To Aerosol Delivery Devices by American Association for Respiratory Care (AARC) [ref], and video contents on inhaler techniques are sourced from Use Inhalers - interactive guidance and training [ref]. All the information and knowledge including the media contents are consulted and revised with our clinical collaborators to validate their authenticity and quality.


kBot takes a personalized approach to converse with patients, collect their health data, and help them manage it. Before kBot client is provided, the patient first consents on data privacy. A user profile is then initialized based on the patient's existing medical record. The patient data in the user profile is anonymized using a pseudonym to comply with HIPAA. This information is stored as a separate patient knowledge base and is continuously updated with data captured in patient-kBot conversations. This patient profile (knowledge base) is used inturn to add personalization in its response and its approach to managing the patient's asthmatic health. As the patient continues to converse with kBot, past captured data regularly updates the patient profile and is referenced to generate a more personalized and palatable experience.


A preliminary evaluation was conducted on kBot to assess technical viability, effectiveness, and usability. This evaluation primarily tries to measure how well the target population will accept this technology as a system for self-management of asthma. The criteria of the evaluation are chatbot quality, technology acceptance, and system usability. Chatbot quality is further divided into three categories: naturalness, information delivery, and interpretability. Naturalness focuses on how natural were the phrase and dialogues used by kBot during the conversation. Information delivery focuses on how well was the kBot able to provide asthma-related information to the patient through the conversation. Interpretability tries to measure if kBot was able to interpret the asthma-related data from user dialog correctly.

The evaluation involved eight domain expert (clinicians from pediatric pulmonology, and allergy departments) and eight non-domain experts (computer science researchers from Kno.e.sis center). The non-domain expert group participants were provided with background information of asthma and the patient scenarios prior to their participation in the evaluation. The reason behind conducting the evaluation on two different cohorts is to measure how well technology performs within a diverse group.

kBot evaluation result

The evaluation responses from both clinicians and researchers for each question of all four metrics (naturalness, information delivery, interpretability, and technology acceptance) are aggregated and averaged to get per metric mean scores. kBot received a mean score greater than 8 out of 10 for each of the four metrics from clinicians. Similarly, researchers rated kBot with a mean score better than 8.4 for all the metrics. A response better than 7.5 on 11 points Likert scale is equivalent to a score better than 4 on a 5-point Likert scale [ref]. The detailed score of each metric is shown in above figure.


Team Members

Related Publications

Dipesh Kadariya, Revathy Venkataramanan, Hong Yung Yip, Maninder Kalra, Thirunarayan Krishnaprasad, Amit Sheth. "kBot: Knowledge-enabled Personalized Chatbot for Asthma Self-Management". In Proceedings of the IEEE SMARTSYS Workshop on Smart Service Systems (SMARTCOMP 2019). IEEE, 2019.



  • This research has been partially supported by:
  • Grant Number: 1 R01 HD087132-01
  • Principal Investigators: Amit P. Sheth (Kno.e.sis, Wright State University)
  • Project Title: KHealth: Semantic Multisensory Mobile Approach to Personalized Asthma Care