Big Data

Rambam Health Care Campus has established itself as a pioneering force in the growing field of Big Data for medical applications, and the hospital is a leader in this field both in terms of data collection and data analysis. Rambam established a computerized database of medical records already in the early 2000s and today the database contains over 20 years’ worth of detailed data culled from patients treated at Rambam during this extensive period of time.

“We have one of the broadest electronic databases in the world, and since the system was designed in-house by our Computer Department, we can make adjustments very quickly, as required,” explains Dr. Ronit Almog, Director of Rambam's Epidemiology Unit and biobank, and Rambam’s chief data scientist.

Moreover, the hospital was the beta site for MDClone, a platform for organizing, accessing, and protecting medical data. In the last three years, Rambam has been influential in perfecting this sophisticated Big Data tool and making the platform as user-friendly and effective as possible.

Rambam’s database contains 2.5 million unique patient health records, representing 25 million unique visits to the ER and to the hospital’s different departments and out-patient clinics. Many different types of data are entered into the central database every day, both structured and unstructured, such as MRI and CT scan images, EKG strips, EMR tables, bloodwork results, and text from medical reports.

Huge efforts have been made to create valid, complete and accurate standardized data. Since a large share of the data is entered by humans and is therefore susceptible to error, Dr. Almog and her team constantly seek to validate the data in order to ensure that answers to queries will be as precise as possible. They ensure that different types of medical data are entered correctly and proactively identify potential problems, using data cleaning techniques to look for mistakes.

In addition to the system’s efficient and trustworthy collection of medical data – a system that is constantly being tweaked and improved by Rambam’s IT staff – excellent tools have also been developed for retrieving and analyzing the data. Substantial resources have been invested to assure that the system is user-friendly, so that doctors and researchers can easily retrieve high-quality data by themselves.

One of the advantages of Rambam’s Big Data operations is its capbility to immediately convert personal data into synthetic data, thereby protecting patient confidentiality. The hospital’s management has mandated the use of anonymous, synthetic data to ensure no risk of privacy issues. “Studies comparing real data files and synthetic files show that the results are similar,” notes Dr. Almog, adding that, “in recent years, we have developed excellent tools for retrieving data from our central database, which is connected to all of the hospital’s data.”

Rambam doctors and researchers can thus easily receive answers to such queries as: “During the past five years, what were the lab results of cardiology patients on day two of treatment with a particular drug?” Or: “Do we have enough patients that meet three specific criteria in order to conduct a clinical trial?” These and many other questions can be answered quickly and accurately thanks to Big Data analysis of the hospital’s vast databank.

Rambam engages in Big Data and AI collaboration with start-ups and industry, as well as with academic institutions, HMOs, the Health Ministry, and other hospitals. Researchers from outside Rambam can access the data by partnering with Rambam researchers and receiving approval. Both internal and external users benefit from Rambam’s high-quality medical data, both synthetic and real, which is unique in Israel.