The healthcare industry generates a large amount of data. The retention of this data is driven by the need to meet intricate compliance and regulatory requirements.1 Electronic health record (EHR) data is at times so massive that it can be difficult to manage with the usual data management tools that hospitals and physicians currently use.2 According to a study by the Institute for Health Technology Transformation, the U.S. healthcare system held over 150 exabytes of data back in 2012, and will soon reach a yottabyte (1024 gigabytes).3 This volume of data is roughly equivalent to 200 trillion feature length HD movies, which it would take more than 50 billion years to watch.
The desire to understand big data in healthcare is driven by the belief that this data can be used to improve a wide range of healthcare functions including decision support, disease monitoring, and community health management. The holistic potential of big data analytics can also be used to tackle the triple aim of enhancing patient experience, improving population health, and reducing healthcare costs.
This will be no easy task! The amount of data associated with a patient’s healthcare is not only overwhelming, but also diverse. It includes EHRs, vital signs, and social media posts from Twitter, blogs, and Facebook. In some cases it has also included news feeds and journal articles.4 The job of big data scientists is to find ways to understand the patterns in the data and synthesize the results in order to provide meaningful insights that doctors can use to make more informed decisions.
This post will explore three big data trends in healthcare as well as some barriers that still need to be overcome before the full potential of big data can be realized.
Three Top Trends
1. Healthcare Internet of Things (IoT)
This trend has the most potential to make a positive impact because of the sheer number of people that are affected. Healthcare IoT refers to the enormous number of interconnected smart devices, volumes of data that those devices collect, as well as how that information is analyzed and presented to help physicians take care of patients.
Smart devices can monitor all kinds of health-related data points from glucose levels and fetal heart tones to the electrical patterns in your heart. If any of these measurements fall out of the normal range then it would usually require a visit to the clinic, although data scientists aim to refine this process so that follow-up can occur over the phone. Other smart devices have been able to monitor whether a patient is taking their medicine by linking the device to smart medication dispensers. It would also be beneficial if these devices could communicate with pharmacies when refills are needed.
All in all, there are many ways in which IoT will improve upon current patient outcomes and reduce healthcare costs.
2. Predictive Modeling
Predictive modeling and analytics is a tool that more hospitals are using to improve patient therapies. Predictive modeling in healthcare uses data from EHRs to address clinical issues such as sepsis and congestive heart failure in order to reduce mortality rates. Like most illnesses detecting early warning signs is key to limiting possible complications and future hospital visits. However, it is easy for physicians to miss subtle indicators. Recently, Georgia Institute of Technology developed a machine learning algorithm capable of analyzing more data from a patient’s EHR than a physician would be able to. In doing so, the algorithm was able to better differentiate patients who had congestive heart failure and those who didn’t.5 The goal is to combine predictive analytics with more data about patients to enable doctors to notice subtleties that went previously unnoticed.
3. Limiting Fraud and Abuse
In my experience, one issue that isn’t discussed widely enough is the prevalence of fraud and waste in the healthcare system. Limiting these factors will have a substantial impact on reducing healthcare costs. Healthcare systems are detecting fraud by examining large datasets of historical claims and then using machine learning algorithms to identify anomalies. These patterns include issues such as a hospital’s overuse of services, patients receiving clinical services from separate hospitals in various locations simultaneously, and prescription medications for the same individual being filled by different providers across hospitals. Using big data analytics, the Centers for Medicare and Medicaid Services was able to eliminate approximately $211 million in healthcare costs arising due to fraud and abuse over a one year period.6
Barriers to Implementation
Any time you try to implement paradigm shifting changes into a well-established system there will be barriers to overcome. The two most prominent challenges to integrating big data analytics into the healthcare system are concerns regarding security and technological expertise.
When dealing with protected health information such as names, social security numbers, dates of births, and addresses you must comply with the Health Insurance Portability and Accountability (HIPPA) Act. Unfortunately, it is difficult to secure private patient data when dealing with big data. In all honesty, security has been almost an afterthought in dealing with big data because most hospitals have limited access to data scientists. However, if hospitals continue to push the capabilities of big data by opening access to data about a larger and more diverse cohort of patients then security concerns need to be at the forefront. In the meantime, healthcare providers should be discerning when choosing a big data vendor to work with, and shouldn’t assume that every data analytics firm will have adequate security protections in place.
The second barrier is expertise. While big data has numerous possible applications, it is often limited to being used for research because of the unique skills required to manipulate big data. Data scientists usually have a technical PhD and related experience. Because of the rarity of these individuals they are costly to hire and are usually only found at academic research institutions or large technology firms. The demand for data scientists continue to increase, especially in the banking and internet sectors.
Conclusion
Hospitals are moving toward more evidence-based medicine. These changes will require taking full advantage of all available patient data, analyzing the information using various algorithms, and translating the results into actionable insights.
By collecting data from EHRs and smart devices, healthcare providers stand to gain a more complete view of patients. This has the potential to assist clinicians with providing tailored care and improving patient outcomes. At the same time, this holistic view can decrease errors in administering drugs as well as reduce redundant testing on patients.
Kevin Anderson is a graduating medical student at Duke University School of Medicine and will be starting at LEK Consulting later this year. He’s most passionate about healthcare redesign, patient engagement, and the life sciences. His free moments are spent traveling and enjoying sporting events with his wife and daughter.
Image: Pexels
References:
- Raghupathi W. Data Mining in Health Care. In: Kudyba S, editor. Healthcare Informatics: Improving Efficiency and Productivity. 2010. pp. 211–223.
- Feldman B, Martin EM, Skotnes T. “Big Data in Healthcare Hype and Hope.” October 2012. Dr. Bonnie 360.
- Transforming Health Care through Big Data Strategies for leveraging big data in the health care industry. 2013.
- Raghupathi W, Raghupathi V. An Overview of Health Analytics. 2013.
- Beckett, Jamie. “How AI Can Predict Heart Failure Before It’s Diagnosed | NVIDIA Blog.” The Official NVIDIA Blog, 14 Apr. 2016, blogs.nvidia.com/blog/2016/04/11/predict-heart-failure/.
- Tahir, Darius. “Predictive Analytics Play New Role in Fraud Detection, but Critics Want More.” Modern Healthcare, 15 Feb. 2015, www.modernhealthcare.com/article/20150225/NEWS/150229947.
🔴 Interested in consulting?