Clinical Analytics at the Bedside
What do algorithms, predictive modeling and big data have to do with patient care? Plenty, it turns out.
The use of big data for predictive modeling is nothing new. Companies like Amazon.com have been using it quite successfully for years to predict consumer habits. With the implementation of electronic health records and the explosion of mobile devices and applications, a tsunami of data is now available in healthcare as well.
Meet IBM Watson Health
To better understand clinical analytics, it helps to know a little about machine learning — which is popularly known in connection with IBM’s Watson computer and its 2011 triumph on the game show “Jeopardy!”
Also known as cognitive computing, this technology learns as it goes, without human intervention, building an increasingly predictive knowledge base as more data is added and patterns are derived.
Watson’s supercomputing powers have become so effective that several years ago, it began moving beyond the game show circuit into healthcare. In 2015, IBM created an entire division for that purpose: IBM Watson Health.
The division recently announced another in a string of acquisitions: Truven Health Analytics, purchased for $2.6 billion. The goal of the acquisition? To enable Watson Health to access more patient data and “extend its leadership in value-based care solutions.”
With forecasts predicting that the clinical analytics market will grow 8–11 percent per year through 2020, these investments are obviously good for IBM’s bottom line.
Watson Health certainly isn’t the only player in clinical analytics, but it was one of the first to begin making a footprint in this sector. There are a growing number of vendors in this landscape. Most EHR systems have these functionalities built into them as well.
Six Practical Uses for Clinical Analytics
With the implementation of the EHR over the past decade or so, we now have at our fingertips a wealth of data for use in predictive analysis. Observational data (i.e., data that is collected during the provision of care) can predict complications before they occur. This enables early intervention and cost-optimized use of resources.
The authors of “Big Data in Health Care: Using Analytics to Identify and Manage High-Risk and High-Cost Patients,” a July 2014 Health Affairs article, examined six types of patient situations that could benefit greatly from the use of clinical analytics. Those six scenarios included:
1. High-cost patients
5. Adverse events
6. Treatment optimization for diseases affecting multiple organ systems.
The authors found that through the use of predictive systems, both short- and long-term benefits could be achieved. “EHRs make it possible to use models of diagnosis and care that combine thousands of disparate measurements to generate evidence in real time,” the authors concluded.
Much of the urgency to use clinical analytics can be attributed to the Affordable Care Act (ACA). With the ACA-mandated shift from fee-for-service to value-based payment models, healthcare organizations face new pressures to improve care at reduced cost.
Using collected patient data, computer algorithms are able to provide analysis and create predictive models upon which to evaluate risk and make recommendations. The goal is to more efficiently allocate resources while streamlining care.
Clinical Analytics in California
St. Joseph Health, which is based in Irvine, announced last September that it will be implementing MEDITECH’s Business and Clinical Analytics solution across its entire network of ministries.
Many providers throughout California already have access to analytical capabilities as part of their EHR systems, depending on their vendor and what functionalities are available with each software package.
Other clinical analytics initiatives that have recently made the news include:
• Identifying patients at risk for falls
A study of California nursing homes published in the June 2015 issue of the Journal of the American Medical Informatics Association found that EHR predictive analytics, combined with information from the CMS Minimum Data Set, can identify and flag patient fall risks.
• Making clinical trials more efficient
Originating at UCSF in 2014, the nonprofit Partnership to Accelerate Clinical Trials (PACT) is comprised of providers and pharmaceutical researchers at 11 California health systems and medical research institutions, including five University of California campuses, Stanford University and Dignity Health. The initiative uses analytical tools to identify patients who may qualify for clinical trials and help speed up the entire clinical trial process.
• Predicting sepsis risk
Researchers from UC Davis have found that EHRs can be used to predict sepsis risk and mortality based on simple, routinely gathered data points, including blood pressure, respiratory rate, temperature, white blood cell count and lactate level.
Big Data and Ethical Concerns
The potential benefits are certainly abundant, including helping nurses access real-time information that could optimize outcomes through evidence-based practice; exponentially increasing the speed at which research can be performed; and providing an overall decrease in costs.
However, a 2010 paper in the journal Biomedical Knowledge Management: Infrastructures and Processes for E-Health Systems, entitled “Ethical Issues of Health Management Predictive Modeling,” notes that there are a number of associated policy, ethical and legal issues which must be considered as well.
The most effective predictive analytics will link data from multiple sources, including clinical, genetic, outcome, claims and social data. Many new data sources are becoming available, including smartphones and social media applications. However, aggregating all this data for the purpose of clinical predictive analytics will require the adoption of standards and new methods of protecting patient privacy.
Since this field is so new and experiencing such rapid growth, the authors of a separate article in the July 2014 issue of Health Affairs, “The Legal and Ethical Concerns that Arise from Using Complex Predictive Analytics in Health Care,” offer a number of recommendations, including::
• Establishing protocols to address consent
• Developing standards for physician and institutional liability
• Requiring transparency regarding the use of predictive models.
Not “If” But “When”
“Predictive analytics promises to make dramatic changes in the way healthcare is practiced and delivered,” note the authors of the Health Affairs article. “The question is not if the healthcare system must be prepared to address the legal, policy and ethical challenges that predictive analytics raises, but when it will be ready to address them.”
Going forward, the technology to support clinical analytics will continue to improve. Healthcare systems will learn to make the most of what that technology has to offer. And platforms like Watson Health will move ever closer to the bedside to help support the quality of care and reduce the costs of making it happen.
Sue Montgomery, RN, BSN, CHPN, is a healthcare writer, editor and consultant specializing in end-of-life issues, digital health and bioethics.
This article is from workingnurse.com.