Harnessing Big Data Analytics to Deliver Optimal Care

Suman Bhusan Bhattacharyya,  Head, Health Informatics, TCS Member, National EHR Standardisation Committee, MoH&FW, Govt. of India Member, IMA Standing Committee for IT, IMA Headquarters, India

With the advent of wearable healthcare devices and remote monitors, Internet of healthcare things is now a reality. Using the power of big data analytics it is possible to monitor and intervene in health events to ensure maintenance of health in all its aspects. This article discusses the various issues involved.

With the advent of ‘electronics age’ in healthcare, humungous amounts of data of varied types from verifiable sources is continuously getting generated in ever-increasing volume. By analysing this ‘big’ data, for that is what it exactly is, the practice of preventive medicine and ensuring continued wellness is now moving from the academic world of debates and discussions to the real world of healthcare services.

Big Data in Healthcare

Before getting into big data in healthcare, it is necessary to try and understand the term ‘big data’. It is important to realise that not all data qualifies as ‘big data’. Big data is data having certain special characteristics. These are high volume, high velocity, high variability, and high veracity – the 4V’s of big data. In healthcare context, a good amount of data of various types from verifiable sources gets generated during a single patient encounter irrespective of the care setting (OPD, IPD, or emergency), even in non-electronic environments, at quite a reasonable rate.

With just the use of electronic medical records, this amount increases dramatically as more data gets recorded and retained. When data collected by an individual gets added to this, the volume becomes pretty high. By 2015 it was estimated that average US-based hospitals alone were generating 665 terabytes of patient data per year. This relates to volume.

Through the use of health monitors, medical devices and wearables, the rates at which the generated data get collected is also quite significant. This relates to velocity. The types of this data ranges from binary to alphanumeric text including audio and visual, and in structured, semi-structured as well as unstructured formats. This relates to variability. The sources of these data are known and trusted with lot of it being collected by those who have been authenticated prior to data collection like doctors, nurses, paramedics, etc. This relates to veracity.

Coupled with the availability of large quantities of electronically-processable data are a number of concomitant technical advances that are driving big data analytics. These advances include multi core processors, low power-consuming devices, low storage costs overall and high-speed local networking.

It is the very nature of big data that makes their analysis demand special consideration. The process needs to factor in the high volumes of varied types of data in varied formats arriving very rapidly in real-time. As a consequence, the traditional analytical processes are rendered impracticable, forcing special methods to be adopted. Special systems for data storage, data retrieval, data preparation and data analytical are employed to ensure this.

Data Sources

The sources of such data are many. To name a few, care providers supply data via Electronic Medical Records (EMR), Electronic Prescription or Order Entry (CPOE) systems, Medical Administration And Reconciliation Systems (MAR), Hospital Information Systems (HIS), and health monitors (used mostly in critical care settings). Increasingly, however, the patients themselves are proving to be substantial generators of such data through the use of data collecting agents in the form of wearables, home-based monitors and medical devices, and mobile health apps; not to forget their social media posts that include Twitter feeds, Facebook entries etc. The payers contribute through generation of insurance claims data and the government through generation of regulatory compliance data.

Data Exchange

Generation of data is one thing, but until that data is pooled prior to its processing, it is as good as being unavailable. Salvation in this regard comes from the ‘Internet’, or more precisely, the ‘Cloud’ that provides the necessary networking infrastructure that facilitates data exchange. Internet of healthcare things is expected to play a crucial part in this regard.

Internet of Healthcare Things

The Internet of things (IoT) is the inter-networking of ‘things’ represented by physical devices, vehicles (also referred to as ‘connected devices’ and ‘smart devices’), buildings, and other items embedded with electronics, software, sensors, actuators, and network connectivity that enable these objects to collect and exchange data. Typically, IoT is expected to offer advanced connectivity of devices, systems, and services that goes beyond machine-to-machine (M2M) communications and covers a variety of protocols, domains, and applications.

As of 2016, the vision of the Internet of things has evolved due to a convergence of multiple technologies, including ubiquitous wireless communication, real-time analytics, machine learning, commodity sensors, and embedded systems. This means that the traditional fields of embedded systems, wireless sensor networks, control systems, automation (including home and building automation), and others all contribute to enabling the Internet of things.

When these ‘things’ are healthcare-related, they get christened as ‘healthcare things’.

Healthcare Big Data and Analytics

Some of the different types of healthcare bigdata analytics that can be performed are as follows:

• In the area of health monitoring and intervention, where vital changes are monitored and alerts raised using signal processing, rule-based algorithms, etc., for proactive intervention at the bedside and at home

• In the field of population health management by facilitating targeted decisions to improve care and outcomes of chronically ill patient population

• Analytics for care management and transitions help in improving transitions of care by identifying high-risk patients and formulating alternative care plans

• Supporting healthcare consumer insight and engagement by creating a patient-focused view to enable targeted personalised marketing and clinical engagement strategies

• In the field of translational research by facilitating identification of the genetic basis for diseases to help clinicians provide personalised medical care

• By providing biomedical insights and facilitating search and discovery by helping accelerate data sharing in support of research, new product development and clinical trials.

Health Events Monitoring & Intervention

Probably the most important benefit of big data lies in real-time data processing and analytics, something that is almost impractical in other types of data analytics. The real-time aspect permits identification of the various indicators at an early stage. This leads to generation of appropriate proactive warnings that in turn leads to early intervention. This, in turn, helps in preventing more serious problems, averting crises and lessening morbidity and mortality, more frequently and sometimes considerably.

Data from the various monitoring devices at the patient-end get streamed in to various processing systems where using techniques such as signal processing, cluster analysis, pattern recognition, logistic regression, network analysis, etc., deviations from normal and potentially problematic clinical states get identified rapidly in real-time.

With the help of these alerts the care providers monitoring the patient are able to proactively intervene, thereby preventing health events, even stopping them before they occur or cause any significant or long-lasting, serious harm.

The care providers are able to get in touch with the patient/person and instruct them accordingly e.g., asking them to visit a healthcare facility or visiting them where they live (extremely useful in ageing population with chronic illnesses) or instructing them to carry certain tasks out. For example, lying down on bed, taking some medication, or avoiding some food items.

Health Management

Although not as dramatic as health intervention but equally as important, health management is significantly improved through the use of big data analytics. Early warning systems for possible outbreaks of endemics, long-term monitoring of patients with non-communication diseases through trends and analysis for treatment effectiveness and further planning – all become very useful to provide improved care impacting overall morbidity and mortality rates of the population.

Challenges

Big data inherently suffers from the same challenges as any patient data in terms of privacy, confidentiality and secrecy issues. However, due to its real-time aspect big data comes with a few unique challenges.

The first of these is under investment in the technology due to the uncertain return on investment. With all the stakeholders not being entirely sure of the accrual of ultimate benefits of many of its promises notwithstanding, it is quite understandable to discover that they are unwilling to commit themselves in any significant way. Too many technological promises have fallen flat far too often to provide any serious comfort to the concerned stakeholders. With increasing instances of use and successes in terms of lowering of morbidity and mortality coupled with ease of use, this situation should improve.

Next is the cumbersome nature of data sharing process, much of which resides in non-interoperable silos. Data needs to be aggregated first to be useful in any kind of analysis and interpretation. Although things are improving — with increasing number of Continua as well as non-Continua complaint devices in the form of wearables, monitors, devices, and electronic systems that are getting inter-connected through the use of many useful networking protocols like Bluetooth, Wi-Fi, GSM, ZigBee, etc., along with the increased instances of use of interoperable electronic systems — there still remains some way to go.

Lastly,  and perhaps most significantly, resistance to change remains the most difficult hurdle to surmount and it not just a case of convincing the doubting Thomas’ per se. Where one’s life and limb are concerned, who can be blamed to err on the side of caution? It is possible to address this issue, but it takes a whole lot of time and effort backed up with sound evidence of success. This will happen only through increased and proper use of the technology with patience and perseverance proving to be a key success factor.

Benefits

The benefits promised by big data analytics are many. These include optimised care, improved clinical efficiency, quality and outcomes, disease outbreak prediction, risk stratification, reduction in cost of care, reduced hospital readmissions, reduced fraud through pre-adjudication fraud analysis, etc.

Organisations Deriving Benefits

Some organisations that have been successful in deriving benefits are as follows.

• Kaiser Permanente® through HealthConnect™ achieved ~US$1 billion savings from reduced office visits and investigations
• Blue Shield® of California with NantHealth™ enabled delivery of more coordinated and personalised evidence-based care
• AstraZeneca® with HealthCore™ was able to conduct real-world studies to determine the most effective and economical treatments for some chronic illnesses and common diseases.

Conclusion

The world of healthcare, especially delivery, is changing significantly and in many instances rapidly with the various technological advances beginning to have a positive impact on the various underlying processes. In retrospect, this was inevitable since healthcare could not have continued to remain in isolation totally immune to the many changes permeating through to every other aspect of daily existence.

The encouraging news is that general awareness of big data and analytics is present and growing, as isolated cases of successful outcomes get reported. This causes the various stakeholders gather enough confidence in it to adopt it in their daily routine in ever-increasing numbers.

With emphasis on Smart Cities where digital health is an important component, big data and its associated analytics will help make a real difference to the health sector through improvement of the various health-related indicators and efficient addressing of various health processes and resources-related issues.

Author Bio

Suman Bhusan Bhattacharyya

Suman Bhusan Bhattacharyya is a practising family physician and a business solution architect for medical devices and healthcare IT applications with nearly twenty nine years of experience. He has worked for several IT MNCs in India and is currently Head of Health Informatics in TCS based out of Delhi-NCR region.

Currently, he is a member of National EHR Standardisation Committee, Ministry of Health and Family Welfare, Government of India, member of Healthcare Informatics Standards Committee, Bureau of Indian Standards and also member of Standing Committee for IT, IMA Headquarters. His main areas of interest include clinical data analytics particularly treatment protocol planning using predictive analytics, designing EHR& EMR systems, mobility applications and application of machine learning techniques in healthcare.

TOP