SHARE: Selecting and sharing Hospital dAta for Research and Evidence-based medicine

Enhancing Personalized Healthcare and Clinical Research through Interoperable Data Management: A Focus on Symptoms Management

BACKGROUND

In healthcare organizations and hospitals in particular, patient medical data are stored in a multitude of heterogeneous systems and formats. These systems are often not interconnected, and information is not transferred across interfaces. The lack of flexibility and the difficulties involved in current data systems not only impede the quality of health care providers’ work and ability to deliver personalized health care but also limit research opportunities with large amounts of data.

Existing research emphasizes the critical need for interoperability between patient information systems to enhance personalized healthcare and facilitate clinical research (1–3).

To help solve this problem at the Cantonal Hospital of Baselland (KSBL) we have developed a new concept for data selection and extraction based on an innovative approach to data management. We will use the DAWIMED software developed by ID-Berlin (main company) and ID-Suisse (Swiss subsidiary) to enable interoperability between patient information systems and make routine clinical data accessible for clinical research and quality management.

This project aims to use this innovative approach to improve personalized healthcare and clinical research, with a specific focus on symptom management.

The project includes two clinically driven research studies:

1.    Assessment of the Management of Patients Admitted with Dyspnea as Their Chief Complaint

2.    Machine Learning to Predict the Cause of Dyspnea Based on Parameters Available at Admission

RESEARCH STUDIES

1.    Assessment of the Management of Patients Admitted with Dyspnea as Their Chief Complaint

Dyspnea, or shortness of breath, is a prevalent symptom leading to numerous emergency department (ED) visits and hospital admissions. It can result from various conditions, including cardiac and pulmonary diseases, as well as other less common causes. Dyspnea's diverse causes necessitate a comprehensive and prompt assessment using medical history, physical exams, biomarkers, and radiological evaluations. Despite its prevalence, there is limited data on its frequency and causes in Switzerland, and patient data is often fragmented across unstructured paper documents and digital systems. Identifying relevant patient data from mixed formats (paper and digital) is complex and often requires time-consuming manual searches.

The study aims to:

·        Describe the management of patients presenting with dyspnea at the KSBL ED.

·        Assess adherence to the medStandards algorithm for dyspnea management.

Significance: Understanding and improving the management of dyspnea can enhance patient outcomes, reduce in-hospital mortality, and streamline ED processes.

2.    Machine Learning to Predict the Cause of Dyspnea Based on Parameters Available at Admission

Dyspnea has multifactorial causes, ranging from common conditions like respiratory infections and heart failure to rarer causes like anaphylaxis or metabolic disorders. The wide variability in patient demographics, comorbidities, and clinical presentations complicates the diagnostic process. Distinguishing between cardiac and pulmonary causes is particularly difficult due to overlapping symptoms. Accurate and rapid diagnosis is crucial for effective treatment but is often challenging due to the symptom's urgency and diverse etiologies.

The study aims to:

• Identify parameters collected during the initial ED assessment (e.g., symptoms, vital signs, medical history) that predict the underlying cause of dyspnea.
• Develop predictive scores or digital tools to assist physicians in making prompt and accurate diagnoses.

Significance: By leveraging machine learning, the study seeks to enhance diagnostic accuracy, reduce the need for extensive testing, and improve the timeliness and effectiveness of dyspnea management in the ED. This approach can lead to better patient outcomes and more efficient use of resources.