Big data in the healthcare industry
Healthcare has long focused on the individual. Only in the last 200 years have researchers begun to pool patients into groups to investigate the health of populations. Big Data offers healthcare the opportunity for a paradigm shift: Instead of thinking from the patient level up, there will now be enough good data to look at whole populations to extrapolate what will happen to an individual.
Genomics: From the Lab to the Clinic
Arguably no single factor will transform healthcare as much as personal genomics. In the coming decades, almost every patient will have their genomes analyzed. Through the expanding use of prenatal genetic screening, many patients are already partially sequenced before birth. Harnessing genetic data is an opportunity that healthcare providers should not miss because of its potential for future research. While some enterprises outsource, hospital system Providence Health & Services recently partnered with the health information company NantHealth to provide sequencing to each of its 22,000 patients diagnosed with cancer each year. Analysis of genetic data is becoming so sophisticated that scientists can have almost instantaneous results to indicate which markers are increasing and decreasing in expression, providing a real-time profile of a patient’s physiological changes based on external factors such as infection or stress. The standard 20-variable blood chemistry panel could realistically be replaced with one that tests for tens of thousands of known markers. This data will tell physicians if a person with genetic risk factors for disease development is starting to manifest illness.
How to Turn Promising Data into Something Useful
Healthcare is probably one of the most data-intensive industries around. Basically, there are four main sources generating all this healthcare data: Medical care providers, public and private payers, ancillary service providers – from pharmacies to laboratories –, and healthcare consumers. The challenge is not just in storage and access, but in making this data usable.
One physician or healthcare team would not be equipped to look at the results of a thousand markers and make insightful inferences. That level of analysis can only be conducted using a computer program. Implementation of these systems is expensive because of software and staffing costs. Hadoop, which many healthcare companies rely on to support their Big Data efforts, is an open-source software platform created to handle large datasets. Critical for information security and access, the software is housed across a global computing framework and is designed to work with many collaborative users.
Growing Need for Computerized Decision Support
The global clinical decision support system market is estimated to rise rapidly in the coming years and with good reason – the systems help to reduce costs and improve quality and clinical outcomes.
According to a report by MarketsandMarkets, the Global Clinical Decision Support System (CDSS) market is estimated to surpass US$550 million by 2018, at a compound annual growth rate (CAGR) of close to 10 percent between 2013 and 2018.
Factors That Influence Growth
The main drivers of growth in this market include rising budgetary pressure to reduce healthcare expenditures, growth of an aging population, rising incidences of various diseases resulting from medication errors, a growing need to integrate healthcare IT solutions, improved quality of care and clinical outcomes, and some favorable government initiatives. In concrete terms this means that CDSS, for example, helps to reduce readmission rates by as much as 50 percent. Furthermore, CDSS can lead to improvements in the quality of care, with a confidence interval of 0.67-0.99 and greater patient satisfaction. Factors that are hindering growth, however, include rising incidences of data breach and loss of confidentiality, the high costs of maintenance and service, a shortage of qualified IT professionals, and the expense of CDSS solutions.
A High Rate of Growth Worldwide
Geographic analysis reveals that North America is the largest contributor to the global market and can expect the highest rate of growth. Europe is in second place, which is attributed in part to improving economic conditions and to initiatives by the European Commission, such as the eGovernment Action Plan 2011-2015 to support and complement information and communication technologies (ICT), including e-health. Asia (China, India, and South Korea), and Latin America (Brazil) are also poised to grow at high double-digit CAGRs.
Taking Big Data Analysis to the Bedside
Universal access to data is crucial for a healthcare delivery system that increasingly implements telemedicine and data-driven care protocols at the bedside. In a pilot study of postcolorectal surgery cases, the Mayo Clinic cut complications by half, decreased patient stay, and saved US$10 million by using a program that identified best care practices, then measured and monitored those metrics in real time.
Big Data is also being captured and analyzed at the bedside. Analytic strategies can identify a patient’s risk of hospital readmission and divert staffing and resources to help prevent it. At SickKids Hospital in Toronto, Canada, infants in the neonatal intensive care unit wear biosensors that collect data thousands of times per second. These biosignals are uploaded and processed in real time for the fastest possible identification of hospital-acquired infections. The hospital can begin treatment 24 hours sooner than if physicians waited until traditional biometrics indicated an infection.
Changes in Future Patient Care
Big Data will create a convenient, real-time healthcare experience for patients. Insights gleaned from that data will improve the quality and accessibility of care, and help foster a spirit of cooperation and research between patients and providers.
A patient’s electronic health record (EHR) will form the hub of patient care. Instead of manually entering data, medical devices will automatically upload the generated data to the EHR, adding convenience and reliability. Data can also be combined with lifestyle devices that monitor exercise, sleep cycles, and heart rate. This gives physicians a more cohesive picture of a patient’s overall health status.
Beyond the EHR
In our digital society, every person generates hundreds of data points through the use of credit cards, loyalty cards, social media, and geo location. This information, which is widely available through third-party vendors, offers a unique view of customers’ health choices. It could tell a provider, for example, whether a former smoker has bought cigarettes recently. Harnessing this data will unlock opportunities for intervention and lasting behavioral change.
Long a buzzword in the field, Big Data will make personalized medicine a reality. Future patients will have their complete medical data, including their sequenced genomes, stored within their EHR. This will increase the effectiveness of matching treatments to individual patients, improve patient safety, and eliminate duplicate care.
Limitless Research Opportunity
The ultimate goal of harnessing data is not just to streamline healthcare but also to drive innovation. Some analyses can spot local and regional trends. The University of Pittsburg Medical Center, for example, uses its homegrown data system to track flu outbreaks. It then alerts doctors who are likely to see patients affected by these patterns.
Challenge: All-in Participation
Patients are wary of releasing their health information without knowing how it will be used. A reported 72 percent of U.S. adult social media users think that the information could be used against them when they try to take out insurance. The best way to encourage patient participation is transparency. Healthcare companies should prepare clear, concise, and informative consent materials to help patients understand how their data will be collected and why it will be useful.
Free Data, Endless Research
With access to a data supply that is growing so rapidly that it seems infinite, the real power of Big Data lies in the insights that can be gained from data sources, ranging from gene expression profiles to prescription rates and census data – many of which are free.
These analyses have already proven useful both in measuring the safety of drugs and in finding new uses for established ones. Healthcare system Kaiser Permanente’s informatics department was instrumental in identifying the link between the arthritis drug Vioxx and cardiac complications, simply through analysis of their own patients’ outcomes.
Atul Butte’s lab at Stanford School of Medicine discovered that the anti-seizure drug topiramate could be effective in treating the inflammatory bowel disorder Crohn’s disease and found genes that contribute to diabetes through control of immune cell receptors located in fat.
These ideas were not generated through the formulation of a hypothesis, followed by systematic testing. They were anomalies that emerged from previously generated data. The next step for these discoveries will be to take the potential treatments, which have been generated by millions of data points collected from thousands of people, and start testing them to see whether they will become useful to individual patients.
Read the second section on smart use of Big Data.
Healthcare is probably one of the most data-intensive industries facing several challenges and opportunities. Storage and access, but also making this data useful to reduce costs and improve outcomes being some of them.
1Healthcare IT Connect, http://bit.ly/1zP0QPd
2IBM, McKinsey Global Institute, http://bit.ly/1qaeJ9t
3McKinsey Global Institute, Big data: The next frontier for innovation, competition, and productivity, 2011, P. 103
4IBM, McKinsey Global Institute, http://bit.ly/1qaeJ9t
5BM, McKinsey Global Institute, http://bit.ly/1qaeJ9t
6Deloitte Development LLC, http://bit.ly/1rDtDYp (all references last accessed on 08/29/14)