Prototype Development of an Expert System of Computerized Clinical Guidelines for COVID-19 Diagnosis and Management in Saudi Arabia

Abstract

The increasing number of COVID-19 patients has increased health care professionals' workloads, making the management of dynamic patient information in a timely and comprehensive manner difficult and sometimes impossible. Compounding this problem is a lack of health care professionals and trained medical staff to handle the increased number of patients. Although Saudi Arabia has recently improved the quality of its health services, there is still no suitable intelligent system that can help health practitioners follow the clinical guidelines and automated risk assessment and treatment plan remotely, which would allow for the effective follow-up of patients of COVID-19. The proposed system includes five sub-systems: an information management system, a knowledge-based expert system, adaptive learning, a notification and follow-up system, and a mobile tracker system. This study shows that, to control epidemics, there is a method to overcome the shortage of specialists in the management of infections in Saudi Arabia, both today and in the future. The availability of computerized clinical guidance and an up-to-date knowledge base play a role in Saudi health organizations, which may not have to constantly train their physician staff and may no longer have to rely on international experts, since the expert system can offer clinicians all the information necessary to treat their patients.

Keywords: COVID-19 management; clinical guideline; expert system; mobile track and evolution of expert system.

Figure 1

The system architecture. The framework includes five sub-systems. (i) The information management system includes laboratory tests, supportive care, treatment recommendations, and risk classifications. The health practitioners can request information from the database and the expert system. (ii) The expert system includes knowledge acquisition, knowledge verification and validation, and knowledge representation after the information is collected by a knowledge engineer. The knowledge can be collected from two resources: clinical guidelines and the adaptive learning system. (iii) Adaptive learning includes the revised knowledge base. This new knowledge can be collected from new clinical guidelines, radiomics analysis of chest images, and rules discovered by decision trees (CART) from the electronic health records of previous patients. This new knowledge will be verified by experts, and his/her feedback and permission will add it to the validated knowledge base after the knowledge is revised. (iv) The notification and follow-up system includes the alert watch and the response engine software to alert health workers and lab managers or radiologists about abnormal tests, while patients can also follow up with his/her symptoms automatically. (v) The mobile tracker system includes Google Maps and a phone number list of suspected cases. This system will display a Google map showing places that can be visited by healthy persons—i.e., where there have been suspected cases.

Figure 2

The prototype interfaces for the sub-system. The information management system: (a,b) patient information, (c) clinical data, (d) hospitalization, (e) epidemiological information, (f) contact exposure, and (g) expert system recommendation.

Figure 2

The prototype interfaces for the sub-system. The information management system: (a,b) patient information, (c) clinical data, (d) hospitalization, (e) epidemiological information, (f) contact exposure, and (g) expert system recommendation.

Figure 2

The prototype interfaces for the sub-system. The information management system: (a,b) patient information, (c) clinical data, (d) hospitalization, (e) epidemiological information, (f) contact exposure, and (g) expert system recommendation.

Figure 3

The prototype interfaces for the sub-system. The information management system: (a) add test results, (b) symptoms, (c) upload chest imaging.

Figure 4

Prototype interfaces for the sub-system. The notification and follow-up system include (a) clinical laboratory results, (b) symptoms, (c) chest images, (d) treatment history, (e) the adaptive learning system and the new clinical rules in the knowledge inbox, and (f) the mobile tracker system.

Figure 4

Prototype interfaces for the sub-system. The notification and follow-up system include (a) clinical laboratory results, (b) symptoms, (c) chest images, (d) treatment history, (e) the adaptive learning system and the new clinical rules in the knowledge inbox, and (f) the mobile tracker system.