Effectively Using Big Data in Healthcare

Newer technologies in healthcare are generating vast amounts of data. This data has to be captured, stored, curated and analyzed continuously which poses unique challenges for healthcare providers. This problem will be compounded with digitization, the Internet of Things and newer sensors. Big data can come to the rescue of healthcare providers in effectively managing data resources.

Big data is here to help

Healthcare data is produced from a large variety of sources such as electronic health records, diagnostics, imaging data, genetic data, clinical records, clinical trials, adverse events reporting, sensors, probes, wearable devices, etc. In 2020, worldwide digital healthcare data is expected to reach 25 exabytes as per a survey in Nature magazine.

Healthcare providers are increasingly seeking data-driven solutions to transform the way they operate their businesses. More and more providers are seeking evidence-based decision-making processes in healthcare, which can be compiled from aggregating individual datasets generated and analyzed by analytical algorithms harnessing big data. Healthcare providers are already under pressure from competition, regulatory, patient sentiment, and intimacy and brand aspects.

Big data-based analytical solutions and new healthcare domain specific solutions are increasingly available now in the market. These newer products offer greater insights and actionable intelligence into how healthcare providers are managing patient care, cost and outcomes keeping in view the vast data generated in healthcare which can be actively mined.

Big data based analytics solutions can be utilized to provide payment innovation, optimal use of available resources, cheaper diagnostics and remote care as well as for proactive identification of potential problems in patient care based on historical data and data pattern identification by big data based machine learning algorithms. In the near future, patient clinical records can become easily portable across providers bringing in greater accuracy, transparency in decision making and convenience on the part of patients to move from one provider to another.

Big data analytics can also enable providers to combine and correlate data from parts of the healthcare spectrum such as billing, claims, patient history, sentiment data, third party providers, pharmaceutical data, etc. to get a 360-degree full spectrum holistic view of patient data. Thus, healthcare providers must adapt big data-based analytics to realize quick returns in terms of better patient care and achieving cost efficiencies.

What to look for big data analytics solution in healthcare?

The prime objective of selecting a big data analytics solution must be improving patient care through effective data mining of clinical and other sources of patient-related data. This goal can be followed by the goal of achieving cost efficiencies.

Actually, these two goals are complementary in nature, as better patient care very likely will result in cost efficiencies for the healthcare provider. Big data analytics can be used to create knowledge-based expert systems, which can be employed to avoid costly medical errors and litigations.

A report of the Institute of Medicine Committee on the Quality of Healthcare in America estimated that hundreds of patients die each year due to medical errors. Big data-based analytics can be of immense help in reducing occurrences of such errors through identification of anomalies or patterns. Healthcare providers can develop coordinated approaches, which can provide equal and timely access to caregivers with accurate patient data anywhere in a secure manner to make the right clinical decisions.

Healthcare providers can also perform better capacity planning and long-term budgeting through big data-based predictive analytics techniques such as predicting hospital readmissions, reducing unnecessary hospitalizations, predicting epidemics probabilities, predicting staff demand and optimal allocation, which eventually results in better care for patients in the future, and at the same time, reducing cost via a phased proactive spending approach.

Healthcare providers can also engage big data-based prescriptive analytics, which can simulate high-tech interventions in patient treatment, simulating patient/subject reported outcome to proactively manage and reduce adverse event occurrences. Thus, providers have a wide choice of analytical techniques to improvise their processes.

Inferring knowledge from complex healthcare data sources as mentioned earlier in this article can pose unique challenges in establishing correlations and identifying trends and patterns. Understanding unstructured clinical records must be accompanied with the complete context of patient treatment.

For example, if an adverse event occurs, providers must be able to track drugs administered, dosage information, subject profile, patient medical history, prior medications, vaccinations, food/nutritional diet provided to the patient, etc. Another example of a challenge in processing unstructured data is in analyzing medical imaging/radiological data and identifying suitable biomarkers and potentially useful information from treatment purposes through image pattern identification algorithms.

Electronic Health Records (EHR) systems implemented by healthcare providers must also be able to communicate with existing systems within the healthcare provider IT landscape. For getting effective, actionable intelligence, EHR systems must contain accurate and complete profiles of patients, which should be available to all caregivers involved.

Healthcare providers involved in creating and maintaining EHR information should be also able to identify key predictor variables based on a combination of domain knowledge and selection of correct analytical algorithm to utilize EHR in disease prediction and progression effectively.

Currently, available EHR systems in the market offer varying degrees of compliance to standards, making interoperability between different EHR systems for highly mobile patients very difficult and complex. Data migration and data transformation from older systems into EHR systems can also provide unique challenges in losing contextual or dimensional information, which may be required by analytical algorithms.

Analytical insights derived from EHR should also be utilized, keeping in mind privacy and security of data in compliance with applicable rules and regulations. In one study cited on PubMed, implementation of a commercial computerized physician order entry system resulted in increased mortality rates due to delays in propagating information required in time-sensitive therapies. Performance and accuracy of the EHR system is very critical in avoiding such occurrences.

Data analytics on an EHR system should thus provide meaningful, verifiable use through coordination of clinical professionals, quality assurance and data and safety monitoring boards, IRBs, data scientists, ethicists, and information technology professionals, while avoiding data issues that can threaten patient care. Big data solutions also must address issues of privacy, veracity, and governance of highly protected and confidential data flowing and processed through the big data ecosystem.

A Gartner survey indicated that in the last year, more than 33 percent of big data-based insight functionality will be delivered via handheld devices. The traditional way to look at big data-based insights is to distribute data from within the enterprise. Mobile insights delivery can adapt with data from the outside in, making information situational or user-specific, which is ideal for catastrophe situations or claims investigation.

Healthcare providers are now proactively looking for cost-reduction opportunities. Cloud-based data storage and compute power provides an excellent mechanism that goes a long way towards that goal. Patient information in the cloud, however, must conform with all applicable regulations. Records should be shared only to authorized people on a need-to-know basis. Data privacy and security are of paramount importance.

For example, patient records in the cloud can be accessed by physicians directly without need to contact administrative staff or other physicians for transfer.

Big data analytics provides a very promising mechanism to provide deep learning and insights for improving patient care. Key issues such as data transportation, privacy, security and governance needs will, however, need increased focus and scrutiny.

About the author: Yash Sowale is an analyst with Capgemini.

 

This article was written by Yash Sowale from Information Management and was legally licensed through the NewsCred publisher network. Please direct all licensing questions to legal@newscred.com.