NEW YORK – In 1986, a person with lung cancer would be linked to approximately two types of disease. After the genome sequencing was completed in 2003, the medical community was able to find seven other molecular diseases associated to lung cancer and there are more to be discovered.
The access to new molecular data meant people had an increased chance of surviving lung cancer, whereas before the disease was considered a death sentence.
The original human genome project took 13 years to complete the sequence of chemical base pairs which make up human DNA at a cost of $3 billion. Today, this same process takes two days at cost roughly $15.
At the SAP Spotlight Tour earlier this week, healthcare professionals from CancerLinq, the Stanford University department of medicine and SAP’s own chief medical officer made a plea for more patient data to be made available in an effort to gain new insights for healthcare. The ultimate goal of this effort is to provide personalized medicine and treatment.
At the event, SAP introduced the SAP Foundation for Health and Medical Research Insights solution. The SAP Foundation for Health solution is an open, extensible scalable platform based on HANA that uses in-memory technology. This solution is able to process volumes of data in real time such as patient data, biomedical data, electronic health records, and information from clinical trials.
Greg McStravick, global general manager and head of SAP’s databases and technology group, said this data is industry specific for healthcare and can be digested from many different data sources.
“Doctors write…and the challenge is that it’s natural language data instead of data captured from sensors,” he said.
The end goal of SAP Foundation for Health, is for healthcare providers to develop personalized treatments from new drugs, devices, and services.
Dr. Clifford Hudis, member of the CancerLinq board of governors, provided the lung cancer data at the event in New York and noted real world patient information can be aggregated for the population, which can lead to improved care because the healthcare provider will get the data in real time.
CancerLinq is a non-profit subsidiary of the American Society of Clinical Oncology. CancerLinq is working with SAP on a cancer-data initiative developed by physicians using from millions of de-identified patient records.
The major problem encountered by CancerLinq is there’s no interoperability in digital health records today. “They do not talk fluidly,” Hudis added.
David Delany, the chief medical officer at SAP, said unstructured data or data in a free text format that’s outside the healthcare provider’s access are other stumbling blocks. “There are also too many siloed systems along with how the healthcare record is stored,” he said.
Carlos Bustamante from the department of medicine at Stanford University said big data in aggregate form enables the medical community to properly analyze rare events and draw conclusions from that.
There are other hurdles, such as routine clinical data, history of treatment and outcomes, the recent conversion to electronic records, and working with many disconnected systems.
“There is a high order of complexity that’s made the conversion to electronic records the same as the paper-based world. We have, shockingly, reproduced the pen and paper model,” Hudis said.
According to him, electronic health records are 10 years behind the digital world today. Electronic healthcare records, Hudis advocates, should work similar to the way Amazon suggests new book titles based on likes or the way Google anticipates the search term after three letters are inputed.
Today, approximately 1.7 million people in the U.S. are diagnosed with cancer and unfortunately one-third succumbs to it. Breast cancer is far greater, Hudis said, but there are a wide range of cancers such as basal skin cancer that does not lead to death. However, only three per cent of these people enrol in clinical trials, and because of this, medicine cannot advance this field.
“We are not learning from patients,” Hudis added.
Currently only 30 per cent of the world’s healthcare data is now digitized. “Think about the possibilities if more of that data is available,” McStravick said.
But even with only 30 per cent available data, access can be difficult due to a range several factors, McStravick said.
The biggest one is economic; in the U.S. alone 10 per cent of person’s income is directed towards the health industry and it’s going up to 17 per cent. Other factors include the lack of hyper-connectivity, the technology currently available, and big data issues such as handling unstructured data.
One example of this disconnect comes from The National Center for Tumor Diseases in Heidelberg, Germany. The facility works to deliver patient care for more than 4,000 people per year along with cancer research and prevention strategies; under the old system they were using required medical staff to search multiple databases, compile patient lists, print patient files and manually check if patients matched criteria for clinical trials by reviewing each file individually.
This resulted in increased complexity and time necessary for staff to properly analyze data. The National Center for Tumor Diseases jointly collaborated with SAP to produce the SAP Medical Research Insights solution, which is now the basis of SAP Foundation for Health.
The solution virtually eliminated the manual process, while reducing paperwork and the back and forth required between staff and the IT department. It also became a digital experience for the National Center for Tumor Diseases by enabling the staff to visualize patient data in real time. The staff can now search and filter multiple data sources through a single interface for areas such as medical history and deliver collections of data used for identifying clinical study participants.