Researchers and software developers are using Big Data analysis to try to predict the emergence of chronic illness and provide doctors with better diagnostic information.
Heart attacks are the number one cause of death and costly hospitalization in the United States, and a fifth of all US adults are likely to suffer one day from heart failure, a condition that is difficult to detect early on so as to take preventive measures. Now IBM has teamed up with a number of US hospital research centers with a view to developing Big Data analytics tools to help medical practitioners detect vulnerability to the condition. This tech initiative illustrates a firm trend towards building alliances between major pharmaceutical groups, research centers and software developers in order to exploit Big Data applications in the medical sector.
Identifying risk factors
The number of partnerships between software providers and hospitals is growing. It is developing a new class of diagnostic test to help design new treatments for autoimmune disease and cancer. For instance it uses a ‘multi-omic’ approach which analyzes a patient’s biological profile including genomic, proteomic, and metabolomic information. Wide data sharing will enable researchers to build predictive models that pick up common points from all the patient profiles registered in the available medical records. In similar vein, IBM Research has joined forces with medical network Sutter Health and integrated health services organization Geisinger Health Systems to develop an early warning system for predicting and detecting approaching heart failure. IBM is using an innovative system of Unstructured Information Management Architecture applications to carry out a targeted analysis of patients’ medical records, coupled with data mined from medical revues and case notes on previously treated conditions. In the absence of any single relevant indicator, the system detects ‘co-morbidity’ indicators – spotting potentially related medical conditions in the patient, such as hypertension or diabetes or a history of using particular medicines.
Helping doctors to analyze quantities of data
A sign of the advance of software applications in medical diagnosis is the forecast in a recent report by full service market research firm MarketsandMarkets that the healthcare sector is set to invest $5.4 billion in Cloud computing by 2017. In addition to its partnership with Sutter and Geisinger, IBM also recently teamed up with the Cleveland Clinic to develop the WatsonPaths and Watson EMR Assistant projects. The Watson EMR (Electronic Medical Records) Assistant project is designed to help physicians retrieve key information from patients’ medical records using the Watson computer system’s artificial intelligence skills. Watson has been trained to mine a wide variety of available data so as to offer ‘smart’ diagnosis and help improve care quality. For the moment the program is being used to train medical students but in the longer term it should be able to give direct help to doctors or even work more independently.