Effective Knowledge Engineering Decision Support System for Neccessity Optimization

Abstract

Focusing on health systems can improve health outcomes now, and will also have a realistic approach in the future. Clinical examination forms the basis for medical decision-making. In the past, physicians used to rely on their clinical skills to diagnose a patient. Nowadays advanced technologies have the potential to replace part of this process and in this context the cost of healthcare has been rising. Growing economies have kept keen eye on this issue and creating new polices whereas in developing countries like Pakistan, it is not yet emphasized at all. In this thesis, research has been qualified in two paradigms, both qualitative and quantitative. A mixed method approach has been adopted. Firstly, quantitative research was carried out in which retrospective data of radiology investigation was collected over a 12-month period in a tertiary care hospital in Pakistan. A total of 5,610 tests were ordered by different physicians including x-ray, ultrasound and others but only MRI and CT-scans were included in the present study. The reasons for ordering MRI/CT scans were identified because these tests are generally expensive tests. After that qualitative research has been carried out, expert interviews were taken to obtain the basic knowledge of radiology based testing and respective necessity optimization information. Subsequently, expert’s interviews helped to create a decision support system that can fit to optimize the need of necessary and unnecessary orders as per using software development lifecycle protocols. MRI/CT Scan orders were analyzed on the basis of symptomatology and viii conclusions drawn in each gender. Data showed that 204 orders were placed due to user defined complaints without prior physician’s clinical evaluation to justify the test requirement. Important to note, emergency patient’s data were not included in necessity optimization. Only outpatient facility orders were considered in this study, emergency setting or monitoring orders were not considered. Rest 136 orders have justification of being prescribed in their patient notes. Most of the findings of the MRI/ CT scan showed normal results. On average more than 50% of the orders made were unnecessary, thereby adding extra financial burden to patients and the overall healthcare system. With the help of these results, qualitative data have been analyzed and responses of the specialists coded into the Nvivo system for running sequential queries. As per the analysis, a knowledge base decision support system has been designed and tested in a real-time environment. By analyzing results, experts suggested that careful clinical judgment supported by history, general physical and proper examination can save lots of unnecessary radiation exposure and cost. Standards of treatment need to be stressed more than by merely investigating and placing improper orders. A prototype decision support system has been designed as per the expert’s knowledge and tested in a real time environment. This research helps in modifying and enhancing the healthcare system and reducing economic burden in an already financially constrained country like Pakistan and should be also useful for other countries.