Rethinking Healthcare Analytics
The healthcare industry has become data rich through a digital transformation with the advent of the Electronic Health Record. We now have a tremendous volume of data flooding in, but managing and analyzing that data remains a challenge. A challenge on top of that is the accelerated pace of change in healthcare. We also face the problem of capturing the right data to truly be able to impact patient outcomes. As we move towards a more patient centric model, we need ways to capture data that does not exist in the traditional healthcare system, such as biometric data, lifestyle choices, and environmental data. Venturing towards the horizon, the future holds potential disruptors that could really change healthcare. Healthcare IT needs to refocus and realign in order to tackle these new challenges.
The healthcare industry is very fragmented when it comes to data with each line of business application having its own database, data warehouse, and analytics tools
One of the challenges most industries face today is the rapid pace of change. Healthcare is struggling with this on multiple fronts, from shifting payer and reimbursement models, constant regulatory changes, and clinical integration challenges. We need faster modes of delivery and innovation, and technology is only part of the equation. We need to combine modern design patterns and architectures with new methodologies to meet this challenge. Leveraging data blending and automation tools along with an intuitive presentation layer will take out thousands of hours of manual labor and bring solutions to bear in a matter of weeks, not months. Using modern data architectures like Hadoop, NoSQL, and cloud services to augment or bypass traditional data warehouse design will access relevant data sets from more sources quickly.
On the methodology side of the equation, introducing concepts such as design thinking into the development lifecycle help identify the “why” for each solution, allowing for better targeted innovation. Principles from methodologies like Agile allows the creation of ideation frameworks that lead to solutions in a couple of weeks instead of months or years. As these methodologies propagate across the organization, we can realize faster time to value, improve efficiencies, and fail fast if necessary. This is in contrast to the traditional waterfall model approach pervasive in healthcare.
The healthcare industry is very fragmented when it comes to data with each line of business application having its own database, data warehouse, and analytics tools. Trying to integrate data across multiple vendors and domains is a very labor intensive and time-consuming process. We need modern architectures and design patterns to help us deal not only with the volume and velocity of data coming our way as its only going to increase exponentially, but also to help manage and analyze that data. For example, big data technologies like Hadoop and Spark allow us to ingest data from a variety of data sources (e.g., patient records, census information, environmental data, images, Internet of Things) and build data sets for analysis and pattern detection without the limitations of traditional data warehouse design patterns. This allows us to analyze our historical data to create a baseline of where we are and build a foundation for maturing into predictive and prescriptive analytics.
Evolving technologies such as big data, advanced analytics, and data visualization allow us to automate our data pipelines, build data sets and gain insight from our data. We can then take that insight and operationalize it to help propel the organization and healthcare forward. These technologies also help us pivot our focus from transactional systems to a patient centric model focused on outcomes or populations and all the data elements those entail, regardless of where that data comes from. Using these technologies, we can create predictive models based on a number of factors for improving patient care (e.g., early sepsis detection, diabetes risk, early heart failure detection) and optimizing our care delivery models (e.g., readmission rates, population health risk stratification, patient engagement).
One of the most disruptive technologies on the horizon is Artificial Intelligence. AI will create new models of care delivery, personalized medicine, and research. AI’s learning abilities are opening up new insights for hypotheses about basic care and basic operations. In the future it will allow for exploring countless molecular pathways for drug development addressing a wide spectrum of diseases. It will provide physicians with an augmented reality diagnostic tools that can detect disease states years before any symptoms develop. With the maturation of the Internet of things we will not only be able to collect and analyze that information but transform and tailor how those devices interact with our patients. We are still in the early phases of big data and advanced analytics but we are keeping an eye out for the disruptions that AI could potentially bring to the table.
Over the last five years, the technology landscape has undergone a substantial transformation. Business now has a wide array of tools at their disposal through technologies and platforms such as Cloud Services, SaaS, Advanced Analytics. They are looking at the best solution to align to their initiatives and business needs regardless of whether it comes from IT. IT needs to make the transition from running the infrastructure and traditional IT domains of Networking, Telecom, and Storage to serving as a strategic business partner. To do that, IT needs to bifurcate into a highly evolved and automated operations organization and an innovation center to effectively support business strategy. IT needs to better understand the full technology landscape and the business strategy to make effective recommendations to help enable drivers of the business.