USPharm. 2014;39(8)(Pharm&Tech suppl):13-15.
The healthcare field, with pharmacy as one of its leading sectors, is overflowing with accumulated data. Whether they derive from electronic health records (EHRs), e-prescribing, external data feeds such as regional health information exchanges, or mobile devices, electronic initiatives are integrating data into healthcare systems. Meaningful data in the hands of end users such as pharmacists can enhance clinical decision support by ensuring that precise and complete information is available to make the best and most accurate decision about a person’s healthcare. This in turn shapes better outcomes with improved quality of care.
Consumers are also driving the demand for useful information, as they access real-time data on mobile devices to facilitate their healthcare decisions. Using data correctly can empower consumers with information about their options, and may lead to a better understanding of various healthcare choices. Ultimately, providing accurate and up-to-date information and guidance may help encourage patients toward using preventive measures to support wellness, making better healthcare decisions, and may in turn reward them for those efforts.
The Role of Data in Pharmacy and Healthcare
Pharmacists can take advantage of the healthcare data explosion to improve patient care and safety. By mining health and pharmacy records, pharmacists in some systems have uncovered patient safety issues that have been corrected, improving outcomes. For example, Express Scripts has built a rules engine with 3,200 drug rules and 2,203 medical rules that are used to routinely check prescriptions. Another example is mining big data—such as analyzing patient refill histories—to improve medication adherence.1,2
Health system and retail chain pharmacies will need to devote time and resources to accumulate and understand large volumes of data. Organizations that can successfully do this will realize the benefits of having access to a large volume and variety of real-time accumulated information, and be able to bring it together in a seamless and profitable way.
Big and Little Data in Healthcare
Healthcare professionals are starting to hear about “big” data and “little” data. To understand just why data big or otherwise are important to the healthcare landscape, consider how early adopters not specific to healthcare use big data to support their end goals: 30.1% use the data for operational analysis such as medical research; 29.7% use them to understand customer behavior; 20.6%, for transactional data; 18.8%, for service innovations; 17.0%, for analysis of machine or device data; and 11%, for nonanalytic workloads such as e-mail applications.3
Using big data to improve decision making and provide greater insights into disease states puts critical information more quickly into the hands of clinicians, who then can act on the data. This not only streamlines the diagnostic and treatment process, but also provides real-time opportunities to support a proactive approach to healthcare. For example, using sophisticated computerized decision support that incorporates knowledge of the patient, a pharmacist can spot potentially dangerous drug interactions, optimize patient management, and perform medication reconciliation.1,2
Drilling Down on Big Data
So what is “big data” anyway? Big data are vast amounts of data aggregated from a large number of sources, assembled into a massive data store and analyzed for patterns.3 The challenge of collecting all of this data is to identify, locate, analyze, and group it into specific pieces that are measurable and available for use. The results are accurate predictions based on a large volume of data that encompasses registries, data warehouses, and predictive modeling opportunities. Big data pool a large and disparate set of data and apply mathematics to derive associations, facilitate comparisons and generate insights that are otherwise not feasible using standard analytics. Such entities as reporting dashboards (e.g., reports on diabetic patients who are targets for glycated hemoglobin A, LDL cholesterol, and blood pressure measurements), scorecards, predictive modeling, data mining, and business intelligence all use big data. Big data require specialized analytical software to help gain insight through fast and consistent analysis of the information collected. Examples of using big data in healthcare would be looking for rare side effects among persons with a cluster of medical diagnoses who are using a specific medication, or analyzing data on households that have sedentary lifestyles as a predictor of diabetes.4,5
Accumulated big data are incredibly diverse; they reflect a landscape of data sources (product data, e-mail files, servers, office files, images, mobile devices, EHRs and health exchanges) and the form of delivery of the data—which can be structured (as in databases and transactional data), or unstructured (office documents, images, raw data on flat files). There is also a myriad of data about consumers themselves (healthcare consumers, clinical teams, clients, and business users). Big-data use enables correlations to be found across multiple and disparate data sources, predicting consumer behavior and financial risk and reward trends.3
The advantage of big data is their breadth—structured or unstructured, in the form of images, videos, and mobile-generated information. The disadvantage of big data is that they have to be “cleaned” and put into the context of what they are to be used for, so it is difficult to establish relationships between the data without proper context.4
Focus on Little Data
Little data are consumer-specific data, such as data that help patients trend and track their healthcare, perhaps by using a mobile device app to track blood glucose readings, or laboratory and radiology results to manage a disease state. Little data enable consumers to access information when they need it to support better decision making for improved health.6
Little data have the advantage of precise targeting opportunities to connect and empower the healthcare consumer. The data are geared toward a specific healthcare consumer and processed accordingly based on that modeling. For example, little data might encompass claims related to a hospitalization. Little data can create a snapshot of a consumer’s health and an opportunity to be proactive and promote wellness.6
Data Security, Privacy, and Standardization
Whether one is using big data or little data, the security and privacy of this collected information should be paramount. Analytics surveys recognize that there is always a risk around accumulating and securing data. One industry-wide survey revealed, for example, that 18% of databases do not encrypt even if the information is sensitive; 28% encrypt only some of the time; 20% admit to breaches or say they cannot be certain if a breach occurred; and 24% do no security at all.7 Therefore, in today’s healthcare landscape, database security is a concern.
Regardless of whether big or little data are captured, some generalizations can be made about data accumulation. First, the data collected must be accurate and reliable, or conclusions drawn from them will be false. Second, the security and privacy of the data captured and collected needs to be ensured, and access limited to those who need to know. Third, standards of data consistency need to be promoted to minimize discrepancies. Finally, facilitation of data retrieval must be uniform no matter where data are captured from (e.g., EHR, mobile device, social media platform), creating a seamless reporting system to support real-time use.4
Tapping Data's Potential
We now can accumulate large volumes of data using websites, social media, and formal information exchanges to provide insights and usable databases for retrieval. The convergence of mobile devices and social networking provides seamless opportunities to transfer data and use analytics to redefine and reshape the delivery of healthcare. And while this abundant data in or near real-time can ultimately improve health outcomes, streams of data sets containing complex and varied pieces are meaningless unless put in proper context. Time and resources need to be identified to enable healthcare organizations to use the data collected.3,8
The potential of using big data and little data analysis is enormous across every business aspect in healthcare. For example, an integrated healthcare delivery organization that wants to become an accountable care organization will need to analyze big and little data. Strategic decision making with a high degree of precision is now possible based on our ability to process complex, unstructured data into a cohesive message. This enables the creation of new best practices that can be implemented not only to stabilize chronic conditions such as diabetes, but also to catalogue and attempt to prevent disease states before they occur. New types of data-analytic software can analyze combinations of data from multiple sources to provide actionable information to the end user, such as drug-disease-state interactions to develop clinical rules that will improve the safety of prescribing certain medications.9,10
Conclusion
Pharmacists can play an active role in this effort to apply big data and little data to enhance healthcare. Data analytics are creating new opportunities and career paths for pharmacists, allowing them to combine the manifold skills they have acquired and put those skills to clinical use while developing new approaches to disease management.
REFERENCES
1. Express Scripts. Making big data actionable. October 10, 2013. http://lab.express-scripts.com/insights/industry-updates/making-big-data-actionable. Accessed June 29, 2014.
2. Hull J, Schueth AJ, Hein W. Health information technology: facilitating an evolution in health care. Specialty Pharmacy Times, February 18, 2014.www.specialtypharmacytimes.com/publications/specialty-pharmacy-times/2014/february-2014/Health-Information-Technology-Facilitating-an-Evolution-in-Health-Care. Accessed June 29, 2014.
3. Olofson C, Vesset D. Big data: trends, strategies, and SAP technology. August 2012. www.sap.com/bin/sapcom/en_us/downloadasset.2012-09-sep-26-13.idc-report--big-data-trends-strategies-and-sap-technology-pdf.html. Accessed May 29, 2014.
4. Awadallah A, Graham D. Hadoop and the Data Warehouse when to use which. www.teradata.com/Resources/White-Papers/Hadoop-and-the-Data-Warehouse-When-to-Use-Which/?type=WP. Accessed May 29, 2014.
5. Chaput JP, Leduc G, Boyer C, et. al. Objectively measured physical activity, sedentary time and sleep duration: independent and combined associations with adiposity in Canadian children. Nutr Diabetes. 2014;4:e117.
6. Planchart M. Little data to big data–one step at a time. Perficient Healthcare. August 6, 2012. blogs.perficient.com/healthcare/blog/2012/08/06/from-little-data-to-big-data-one-step-at-a-time. Accessed July 9, 2014.
7. Garey L. Big data brings big security problems. InformationWeek, May 23, 2014. www.informationweek.com/big-data/big-data-analytics/big-data-brings-big-security-problems/d/d-id/1252747?print=yes. Accessed May 29, 2014.
8. Figge HF. Technology and Adolescent Health: US Pharm. 2014;39(5):35-37.
9. Revolution Analytics. Advanced “big data” analytics with R and Hadoop. 2011.
www.revolutionanalytics.com/sites/default/files/r-and-hadoop-big-data-analytics.pdf. Accessed May 29, 2014.
10. Fania M, Peiravi P, Chandramouly A, et al. Predictive analytics and interactive queries on big data. 2013. https://software.intel.com/sites/default/files/article/486302/ra-predictive-analytics-and-interactive-queries-on-big-data.pdf. Accessed May 29, 2014.
To comment on this article, contact rdavidson@uspharmacist.com.
