Anybody working with big data would clearly know that these humungous data sets are all set to create a revolution and change the face of the healthcare business. All that the industry needs to do is to leverage the potential of big data and accelerate its way of functioning. This would mean realising the capabilities of big data and using it in the best possible way.
Though big data shows abundant promise in the healthcare domain, there are several roadblocks hampering its progress and development. It could be the laws which are focused on protecting the information of the patient or simply put- the scarcity of technical talent. Therefore the search for certified and skilled big data scientists who could work across diverse industries and specialist in the healthcare domain is much needed.
Machine learning and big data in hospitals
Big data has been looked as a boon in several hospitals in Paris. A trial run on forecasting the rates of patient admissions have been undertaken as this could pave way for a better and efficient allocation of resources. All this is to ensure that the patient has a comfortable and satisfactory experience. Four hospitals which constitute the Assistance Publique – Hopitaux de Paris (AP-HP), have entered into the data crunching foray wherein information externally and internally have been analyzed thoroughly. The data includes hospital admission records of several years and predictive analysis about the patients entering and exiting.
Built on an open source platform called the Trusted Analytics Platform (TAP) humungous data is analyzed for probable patterns using time analysis techniques. The software also offers an environment which is ideal for collaboration and development. Intensive data crunching methods predict the admission rates at different times through pattern study. Along with these cutting-edge science techniques, machine learning tool is implemented. These help to understand the nature of algorithms that offer future trending information when data from the past are given as inputs. The performance of the software is also to make sure that scalability is reached and the ability of the algorithms is implemented to work over distributed systems.
Big data reducing healthcare fraud, waste, and abuse
Fraudulent activities, waste and abuse in healthcare are being tackled efficiently using Big data analytics. Using predictive analytics, The Centres for Medicare and Medicaid Services vouch on predictive analytics as it has led to the prevention of a loss of over $210.7 million in a single year. Similar was the case with United Health care which took to a predictive modelling platform. The software helped them to spot claims that were inaccurate and false. This exercise generated a 2200% return on investment for the company.
The process of fraudulent identification begins by looking for unstructured data sets coming with historical claims. Intelligent machine learning algorithms comb through voluminous data and then scan for anomalies and patterns. Healthcare organisations can see through patient records and billing to identify anomalies. This could be over utilisation of a hospital’s services in short time intervals, patients using the healthcare services from various hospitals in multiple locations at the same time, or similar prescriptions for the same patient filled in various areas.
The Centres for Medicare and Medicaid Services rely on predictive analytics to allocate risk scores to certain claims that are specific. The risk scores are assigned to providers too to recognize the billing patterns and claim discrepancies that are otherwise difficult to detect. Rules-based models indicate few charges on their own. Anomaly models focus on suspicions based on factors that seem doubtful. The predictive software models also compare charges against a fraud profile. Graph models are employed for detecting fraud claims which raise suspicion on the basis of the providers’ relationships since fraudulent billers coordinate in tight networks.
Unified data architecture –Geisinger Health System
The association of Geisinger health system and big data is a noted example. In the year 2015 Geisinger Health System pioneered an IT system called a Unified Data Architecture (UDA). The platform allows the integration of big data into the current and existing data analytics and management systems. The UDA is greatly used to monitor and keep a note of the patient outcomes, do an in-depth analysis, map patient’s genomic sequences with clinical care, and to visualize the abundant healthcare information across a large group of patients and with the provider’s networks. It has been noted that Geisinger’s UDA has been one of the most successful system as it has been found to be the largest practical application with respect to big data in healthcare. The system consists of thousands of CPUs which processes and delivers hundreds of terabytes of data per hour.
Geisinger was one of the early players in the big data revolution as they had forayed way back in 1996 by adopting a full-featured Electronic Health Record (EHR). Through this doctors can enter patient information and initiate support and care. Though this was a complex procedure as it involved retrieving meaningful data aggregated from many sources, several hospitals still have units of interconnected auxiliary systems which pull out the data in silos. These traditional databanks had trouble to accommodate new types of data which were unstructured or in the form of free-text patient notes.
With such drawbacks, it was time for a new technology to be implemented as the EHR was only producing an incomplete picture. A platform was required to track patients when they visited other clinics and completely different healthcare systems. Through this process, individuals leave their digital footprints wherever they interact, right from the grocery store to the usage of the smartphone and its apps. With patients’ providing a nod to access the data, integration of data happens easily. Thus the UDA offers a common data space for quick integration of data from chosen and filtered internal and external sources. The UDA is undoubtedly a powerful system as it has the capacity to process huge volumes of data from diverse sources, and also the ability to integrate and store these large data volumes. Thus the platform has aptly filled the gap between traditionally based healthcare systems to the modern ones. Data integration provides a 360-degree picture as hospitals get to see information exchanged across departments thereby providing a detailed, longitudinal insight of the patient.
Mark Zuckerberg’s contribution for Big data
The latest entry into big data is Facebook founder Mark Zuckerberg and Priscilla Chan who have donated $10 M to advance health through big data. The lead of this project at UCSF, Atul Butte believes the concept of data recycling and tapping the potential of trillion points of data. This would, in turn, lead to cost-effective approaches and newer initiatives in drug discovery and treatment of diseases.
With several companies embracing Big data to harness its potential, the future of healthcare does look futuristic, cost effective, advanced and highly patient –friendly.