Cognitive computing to support diagnosis of rare diseases

Since opening in 2013, more than 6,000 patients have contacted the centre to visit Prof. Dr. Jürgen Schäfer, an expert in rare diseases, who is also known as the ‘German Dr. House’, based on the character of the eponymous American medical television drama. Most of the patients he and his team meets with have year-long medical histories, which include a large number of unstructured data, such as laboratory tests, clinical reports, drug prescriptions, radiology findings as well as pathology reports.

“It’s not uncommon for our patients to have thousands of medical documents – leaving us overwhelmed, not only by the large number of patients, but also by the huge amount of data we have to review,” said Prof. Dr. Jürgen Schäfer, University Hospital Marburg. “This is especially challenging since our work is often like looking for the proverbial needle in the haystack – even the smallest piece information could lead to an accurate diagnosis.”

Currently, when the centre’s team of expert physicians meet a patient for the first time they conduct a thorough diagnosis by speaking with them and carefully reading their medical records. This is augmented by their own knowledge and experiences, medical journals, online websites and expert tools – a process, which can take several days for each patient.

Pilot phase at the ‘Centre for Undiagnosed and Rare Diseases’

RKA has teamed up with IBM on a 12-month pilot project where medical and technical scientists and designers from both companies are working together to develop a cognitive sustenance system for rare diseases. Its overall goal is to support physicians at analysing patient data and to help them to make decisions faster and safer.

“The amount of medical knowledge continues to explode to the point where it will double every 73 days by the year 2020. Therefore, the planned use of cognitive technology such as IBM Watson is intended to support our evidence-based and individually optimal treatment for each patient,” said Prof. Dr. Bernd Griewing, Chief Medical Officer, RHÖN-KLINIKUM AG. “We are developing an assistance system to facilitate the preparation and evaluation of existing patient information before and during a consultation with physicians. This will help our doctors in diagnosis decisions and selection of treatment options.”

Once the pilot begins, patients will fill out a digital questionnaire, developed by the medical staff in Marburg. This questionnaire ensures the data remains anonymous and sends it to Watson’s APIs in the IBM Cloud where the relevant information is extracted for the physicians.

Since the patient information will be provided in German, the cognitive assistant will use a natural language process algorithm for medical terminology developed by IBM Global Business Services (GBS) to correlate the German questionnaire to the corpus of English-based medical data to provide a differential diagnosis.

“Using cognitive computing we are building a list of hypothesis which we then present to the doctors, including their sources, for the ultimate data driven diagnosis”, said Dr. Matthias Reumann, healthcare scientist and technical leader of the project, IBM Research. “If one lab result is slightly out of the normal range, Watson can find it for further investigation.”

RKA is providing the required specialist information and medical knowledge as well as the IT system for the project and is accountable for data protection regulations.

Vision of RHÖN-KLINIKUM AG for an optimised patient management

The insight obtained from the project phase at the University Hospital will subsequently be analysed and used across other hospitals of the group. The system is expected to not only help physicians find a diagnosis, but also support the process of patient admission and provide the best possible patient routing. This will prevent cost- and time-intensive misdirection of patients to wrong specialists and treatment plans.

“Optimised patient management is crucial for both the nationwide medical treatment of tomorrow and the further development of our company. In this context digital solutions get more and more decisive,” said Jens-Peter Neumann, Chief Financial Officer, RHÖN-KLINIKUM AG. “Our cooperation with IBM is an important part of our innovation strategy. In addition, we are also focused on identifying and collaborating with start-ups in the IT-driven healthcare sector to strengthen our unique market position. Both medical and technical efforts and improvements serve our patients and create value for our company.”