I am teaming up with CIO David Chou (@dchou1107), an executive with more than 13 years of experience in the healthcare industry to bring you a series of blog posts that provide a unique perspective on some of the healthcare industry’s biggest trends and challenges. David has been named to several “Top Social CIO” and “CIOs to Know” lists. He is a visionary and resourceful leader with expertise in healthcare and digital technology and a proven track record of delivering innovative, state-of-the-art solutions.
In our first post we discussed the Evolving Role of the Healthcare CIO. Our second post, Embracing Digital Transformation took a look at how digital is impacting patient care and the overall mindset of healthcare organizations. In this interview we turn our attention to healthcare data and analytics. We have invited Priyal Patel, Solutions Architect and Consultant at Perficient, to share her healthcare analytics knowledge and insights gained from working with many largest healthcare organizations across the United States.
Before we get started with the interview I would like to invite you all to download our new healthcare analytics trend guide: 10 Healthcare Analytics Trends for 2016. In this guide we take a look at ten analytics trends healthcare executives need to be thinking about in 2016 and beyond. We identify technology strategies and solutions that will help healthcare organizations succeed in a data-driven, digital world.
KATE: What are the biggest motivators driving healthcare analytics?
DAVID: Rising healthcare costs, government regulations and incentives and value-based care initiatives are factors and motivators in analytics. Analytics is the number one priority for CIOs according to Gartner and it is also on top of the CEO’s mind. Healthcare is consolidating, margins are shrinking and reimbursement is declining. When I think of any business motto in providing the best quality of service at the lowest cost, the same rule applies in healthcare. We have to provide the best care for the lowest price and the only way to do that is to invest heavily in analytics. To combat rising healthcare costs, healthcare organization can use analytics and data to start putting together standards in the clinical treatment setting. This ranges from standardizing supplies, implants, and workflows to helping with the identification of patient costs. Every organization should leverage data to make informed decisions.
KATE: What are the biggest challenges/barriers healthcare organizations have when it comes to transforming big data into insights?
DAVID: There are two major challenges that I see currently – data governance and infrastructure.
Data Governance: The healthcare industry has gone through a massive transformation in the last few years with ACA and meaningful use. When we think about the transition from paper to electronic or the consolidation of multiple application systems, there are a lot of issues with the data governance. Data governance includes two key areas that will be the foundation for analytics.
- Data Dictionary – Do you have a solid definition of the data elements especially when you are pulling in data from disparate systems? It will be critical to have the data definition be consistent across the enterprise to ensure that you have optimal quality in the data.
- Data Access – How will your organization access the data and who will be allowed to access the data is a challenge. Imagine in an academic medical center where you have students, healthcare operators, researchers, and external stakeholders who all want access to the data. What is the model to allow access and who has the right to regulate the access for compliance measures?
There are more challenges than the two that I have highlighted for data governance but it is a starting point for an organization to think about when they are developing a data analytics strategy.
Infrastructure: The majority of healthcare providers are moving towards a consolidated healthcare IT platform which will help with identifying the source of truth but at the same time creates its own challenges. Once you have finally united all of the data from the disparate systems and put it in a single production system there are infrastructure challenges that must be addressed.
- IT Infrastructure – Do you have the right IT infrastructure to support the data growth? You can build the infrastructure internally on premise in your data center but this is also where I believe we can utilize cloud technology. Cloud computing can help alleviate the barrier of IT infrastructure. One of the biggest IT challenge when an organization decides to purchase an analytics software is building the IT infrastructure of servers with the accurate compute and storage capacity. The cloud model takes the IT infrastructure out of the equation because the cost of the cloud will be based on consumption and the organization can scale quickly in a cloud environment. Cloud can also decrease the implementation time of an analytics solution where the time of implementation can be completed in 24 hours if the application is a hosted cloud model. From a budgeting perspective if the organization prefers to shift the capital cost to operational than the cloud model works well from a financial perspective. Putting together the IT infrastructure will be a barrier but it can be alleviated with the use of cloud technology.
- Staffing and Resources – Do you have the right resources and data scientists to create the data modeling that will be used by the organization? One of the biggest challenges in healthcare is recruiting for data scientists. This is a very unique skill set for someone that can extract knowledge from a large set of complex data – both structured and unstructured. Unfortunately I do not have the magical answer for finding data scientists but I believe that healthcare will need to look for the resources in other verticals and start training their internal staff to become healthcare data scientists.
KATE: What does a successful data analytics strategy and organization look like?
PRIYAL: Visionary, realistic and enterprise focused. I often see healthcare organizations trying to implement analytics solutions without an enterprise strategy to manage the people, processes and technology. The absence of this enterprise strategy is in part due to the fact that analytics solutions are typically implemented in silos – IT departments oftentimes purchase analytics solutions without truly understanding the type of solution(s) they will need because they lack insight into the key business drivers and value that is expected to be delivered from the solution(s). From my experience, this deficiency in coordination and collaboration between IT and the business is one of biggest pitfalls to implementing analytics and leads to minimal focus and insufficient accountability. A successful data analytics strategy needs to always keep the end in mind – focusing on the enterprise view of how the solution will be used to deliver business objectives and use this as a guide to set priorities and develop an implementation roadmap. Once a plan is in place the organization can continue down the road of success by managing the timeliness, accuracy, availability, and accessibility of their data efficiently to drive realistic business value; use analytics to create a culture of innovation and collaboration; measure the success of the analytics solution(s) against expectations, both IT and business, and above all keep the focus on enterprise delivery to assure that analytics will be embraced as an asset to the growth of the organization, not a limitation.
KATE: How are the Internet of Things (IoT) and connected devices playing a role in healthcare data analytics?
DAVID: The Internet of Things (IoT) will be huge in healthcare and analytics. There are so many ways for healthcare organizations to use the data from sensors and wearables. One area that most healthcare organizations struggle with is patient engagement. We should be able to use the IoT data to improve patient engagement and allow the providers to proactively manage the care of a patient. If I have a history of diabetes and heart attack, the sensors from the IoT can monitor any irregular heartbeats to alert a clinician and wearables can be used to monitor a diabetic blood sugar. Other data that can come from the IoT are step trackers to track a patient’s activity monitor, medication reminders, and even home health monitoring devices. Even though we are seeing an enormous amount of technology, adoption in the healthcare setting will be slower than other verticals.
KATE: What are the benefits to advanced analytics and what role does it/will it play in healthcare?
PRIYAL: Advanced analytics seem to have catapulted to the top of healthcare executives wish list. Advanced analytics can increase the competitive edge by generating actionable insights to deliver on the healthcare triple aim – better care, better health, lower costs. There are many benefits to advanced analytics in healthcare. For example, predictive analytics can allow healthcare organizations to process and mine vast amounts of clinical, claims, pharmacy data and apply statistical algorithms to predict risk factors, allowing clinicians to proactively manage their patient populations.
Comparative analytic capabilities enable healthcare organizations to use internal and/or national industry benchmarks to increase their transparency to better monitor their clinical (treatment outcomes), financial (reimbursement) and operational (utilization and productivity) effectiveness and identify continuous areas for improvement.
Natural language processing (NLP) allows for voluminous unstructured clinical data (free text and speech) to be analyzed and extracted into meaningful formats, such as populating EHR fields and valuable analysis, like patterns in disease detection. Free text or dictated physician documentation sometimes houses extraordinary amounts of key clinical data that is oftentimes required to make informed decisions about a patient’s care. Unfortunately, it takes manual review and analysis to extrapolate this value. With appropriate NLP the time to value is drastically shortened.
Advanced analytics, even in its early stages within healthcare, has already proven its value in how we deliver care today. As this area continues to evolve within healthcare and its use more widely accepted, it will play a vital role on how we use accurate and meaningful data and timely analytics to continually improve our efficiencies and services to drive better access, quality and outcomes.
KATE: Are out of the box solutions offered through EMR solutions sufficient to support analytics needs?
DAVID: I don’t believe so even though the EMR solutions are getting better for supporting analytics. When I think of big data in healthcare, I believe most organizations are still far from it. The majority of the organizations still focus on “EMR data” so essentially they only have the data within their EMR. Big Data incorporates external data sources, which the current EMR solution does not provide nor should it be part of the EMR system functionality. The solutions from the EMR vendors are getting better but it is only good at solving problems within your EMR. If the organization’s goal is to use analytics to solve internal operations or clinical behaviors then the EMR solution can provide the analytics and data. As the healthcare industry is moving towards value-based care, it will be critical to learn more about the patient/consumer and the behaviors of the patient outside of the hospital therefore the EMR alone will not be able to provide that solution.
KATE: How are healthcare organizations leveraging tools/experience to reduce data integration times when it comes to analytics?
PRIYAL: As more and more healthcare organizations (health plans and providers) invest in business intelligence and analytics platforms to deliver enterprise value, the need to integrate the data from multiple, disparate systems, across the organization, into a single repository, such as an enterprise data warehouse (EDW), presents a significant challenge – how to consolidate and standardize this data to achieve a holistic view to drive meaningful analytics. The technical process necessary to accomplish this integration is known as extract, transform and load (ETL). The ETL process basically extracts data from multiple systems, transforms the data into the appropriate structure and then loads the data into the target, such as an EDW. It sounds simple but given the variability in data, formats and systems there are challenges, specifically time to value. Healthcare organizations are relying on 3rd party vendors and even their own experiences to find innovative ways to help alleviate some of the challenges and cost around data integration and specifically the ETL process. Healthcare organizations are utilizing “accelerators” to automate the population of an EDW, by providing standardized integration formats of data from typical sources, such as clinical, financial, operational and external systems. Using these standard templates helps remove the manual process of source mapping, reducing the data integration time by as much as 50% and thus decreasing time to value and reducing implementation costs. As an example, Perficient’s Health Analytics Gateway was created to compress the long lead time to load data into IBM’s Healthcare Data Model and subsequently into high-value analytics use cases by leveraging pre-built templates.
KATE: What are the biggest and most costly mistakes an organization can make in healthcare data analytics?
DAVID: The biggest mistake an organization can make is to purchase an enterprise platform before determining what problems they want to solve and look at the resources and tools internally. I have seen organizations produce thousands of reports and the feedback is still I can’t get the reports I need or the data that I need. I am sure that 80% of the reports needed are already built out. Before an organization spends millions of dollars on an enterprise data analytics platform I recommend that the organization answer the question of what problem are they trying to solve and then the team can focus on whether or not they have the right tools currently within the organization. Once the organization has the problem identified and the data analytics solution, it is critical to make the data available to the users at all times. The number one recommendation to avoid a costly analytics expense at the start is to identify the problem and make sure the data is accessible.
Follow me @KateDTuttle
