Project Blog

Nov, 16th 2016

Precision medicine and Case Based Reasoning System
One of the main limitations in today‘s medicine is our understanding of the individual variations of diseases. Pharmaceutical companies runs numerous studies and programs every year. Such trials focus strictly on predefined questions. However, it documents only a moment of a disease. In return, health care industry heavily depends on in depth knowledge of diseases.

Studies or patient dossiers provide insight of a disease comparable to a pixels of a picture.

Revealing clinical picture capturing and describe all varieties of a disease is an essential step to find treatment solutions for a disease.
One of the main limitations in today‘s medicine is our understanding of the individual variations of diseases. Pharmaceutical companies runs numerous studies and programs every year. Such trials focus strictly on predefined questions. However, it documents only a moment of a disease. In return, health care industry heavily depends on in depth knowledge of diseases.

Great hope was associated with analysis of the human genome but it has failed to support the understanding of the most important diseases such cancer and dementia. What could be done alternatively?

In medicine, classic way to establish a big picture is to observe and conclude. Big data is an ideal basis to establish such dynamic picture. In other words bringing all kind of data together and process it by machine learning systems.
Precision medicine takes into consideration that each patient reacts differently to a drug depending on the clinical and environmental situation. The concept could be successfully demonstrated on singular examples such as Herceptin for Her-positive breast cancer. However, it could not be transferred to a brought clinical use as a reliable tool supporting treatment decisions is not available. Based on the results of the medical data processing project, Conceptual Process Solutions develops a so called case based reasoning system which allows health care professionals to identify the most efficient treatment solution and to avoid medical risks such as side-effects. The mentioned systems works on the basis of available data and turns it into a treatment recommendation which can be followed by the prescriber or not as the final decision stays with the treating physician. It is designed to ensure that all possible medical parameters including genetic information and newly developed tests can be integrated.

Drug Safety Automation


Drug Safety is an essential part of drug development as it is important to demonstrate the safe use of a drug before it can be marketed. For this reason, adverse events have to be collected and medically evaluated during the development and its brother use in the market. Central element herein is the collection of adverse events and its processing. Source data vary in their format and content. Currently data processing is done by data operator entering the data manually into special systems to be converted into uniform data sets which have to be forwarded to all health authorities who approved the drug in their country or region. The process bears several challenges as the process is time-depending and often extremely complex due to its logistic and the huge number of participants.
Conceptual Process Solutions developed in collaboration with the University of North-West Switzerland and the Commission for Technique and Innovation an automated drug safety system.
Concept is based on natural langue recognition in combination with machine learning systems. It is able to recognize the required information according to the CIOMS standards and to validate that information. In case of uncertainty an operator has to decide whether the data can be corrected based on the available data or a predefined query has to be sent to the reporter. Development goal was from begin on to reduce “human interactions” to an absolute minimum.
The system can either be used for data processing/ entering or for the quality control of processed cases. That means source documents can be automatically compared against a final case.