Three-stage intelligent support of clinical decision making for higher trust, validity, and explainability.

The paper presents a conceptual framework for building practically applicable clinical decision support systems (CDSSs) using data-driven (DD) predictive modelling. With the proposed framework we have tried to fill the gap between experimental CDSS implementations widely covered in the literature and solutions acceptable by physicians in daily practice. The framework is based on a three-stage approach where DD model definition is accomplished with practical norms referencing (scales, clinical recommendations, etc.) and explanation of the prediction results and recommendations. The approach is aimed at increasing the applicability of CDSSs based on DD models through better integration into decision context and higher explainability. The approach has been implemented in software solutions and tested within a case study in type 2 diabetes mellitus (T2DM) prediction, enabling us to improve known clinical scales (such as FINDRISK) while keeping the problem-specific reasoning interface similar to existing applications. A survey was performed to assess and investigate the acceptance level and provide insights on the influences of the introduced framework's element on physicians' behavior.

Assessing Acceptance Level of a Hybrid Clinical Decision Support Systems.

We present a user acceptance study of a clinical decision support system (CDSS) for Type 2 Diabetes Mellitus (T2DM) risk prediction. We focus on how a combination of data-driven and rule-based models influence the efficiency and acceptance by doctors. To evaluate the perceived usefulness, we randomly generated CDSS output in three different settings: Data-driven (DD) model output; DD model with a presence of known risk scale (FINDRISK); DD model with presence of risk scale and explanation of DD model. For each case, a physician was asked to answer 3 questions: if a doctor agrees with the result, if a doctor understands it, if the result is useful for the practice. We employed a Lankton's model to evaluate the user acceptance of the clinical decision support system. Our analysis has proved that without the presence of scales, a physician trust CDSS blindly. From the answers, we can conclude that interpretability plays an important role in accepting a CDSS.

Machine Learning-Based Factor Analysis of Carbohydrate Metabolism Compensation for TDM2 Patients.

Type 2 diabetes is one of the most common chronic diseases in the world. World Diabetes Federation experts predict that the diabetes patients' number by 2035 will increase by 205 million to reach 592 million. For health care, this diabetes type is one of the highest priority problems. This disease is associated with many concomitant diseases leading to early disability and high cardiovascular risk. A severity disease indicator is the degree of carbohydrate metabolism compensation. Decompensated and subcompensated carbohydrate metabolism patients have increased cardiovascular risks. Therefore, it is important to be able to select the right therapy to control carbohydrate metabolism. In this study, we propose a new method for selecting the optimal therapy automatically. The method includes creating personal optimal therapies. This kind of therapy has the highest probability of compensating carbohydrate metabolism for a patient within a six-month. The method includes models for predicting the results of different therapies. It is based on data from the previous medical history and current medical indicators of patients. This method provides high-quality predictions and medical recommendations. Therefore, medical professionals can use this method as part of the Support and Decision-Making Systems for working with T2DM patients.