Key factors driving cloud analytics in healthcare

While there are several advantages to deploying cloud-based solutions, the most important concern around the use of cloud models is information security. High profile data breaches at Anthem, Premera and other healthcare organizations have made CIOs understandably nervous about cloud hosting. Here’s a checklist of what to consider when deploying a cloud model for analytics.

9 cloud analytics
Credit: Thinkstock

In a white paper titled The State of Healthcare Analytics published last quarter, my firm identified some key factors that drive analytics adoption today in healthcare by drawing on some industry research and our own market observations:

-- Analytics is driven by margin pressures and a focus on improved outcomes

-- 65 percent of healthcare providers and 60 percent of healthcare payers plan to increase analytics spend in 2015

-- Data security concerns, Interoperability challenges and data scientist talent are impacting adoption rates

-- Operationalization of analytics in clinical and operational workflow is impacting ROI on analytics investments

-- New data from wearables, Internet of Things (IoT) are creating opportunities for improved insights

Our key observation was that value creation in analytics today is focused on a couple of areas: data integration and visualization, and the use of predictive models for managing clinical outcomes. We call this the “zone of activity.” We also concluded that the major headwinds included interoperability concerns, and the lack of a “plecosystem” or a platform ecosystem approach to derive benefits from advanced analytics by incorporating them into the clinical workflow in hospitals.

However, we also found that cloud adoption is increasing, and clients are looking for easy-to-deploy solutions that can deliver speed to value at low cost. A 2014 industry survey by HIMMS found that over 80 percent of healthcare organizations use some form of cloud services today.

Cloud based models offer several advantages over traditional software models, including shorter implementation cycles and lower total cost of ownership (TCO). However, there are several other considerations that go into determining if a cloud model is right for an organization.

The factors in favor include the following:

-- Faster deployment of analytics solutions and speed to insights: Cloud solutions can be spun up practically overnight and preconfigured solutions can deliver analytical insights through standard dashboards that often include predictive models and risk scores built in.

-- Improved cash flow and diminished IT support costs: Cloud solutions typically come in subscription-based models that eliminate the need for high upfront investments in hardware and software, and also eliminate ongoing support costs since all of this is assumed by the cloud solution provider. In addition, the solution can be scaled up or down based on data volumes, # of users, and other factors that provide a great deal of flexibility

-- Access to advanced predictive algorithms and data integration capabilities: Most healthcare organizations do not have in-house capabilities for developing the advanced analytics models or data integration infrastructure, especially if it involves structured as well as unstructured data from a wide variety of sources. Cloud based solutions can be an easy option to tap into the vendor’s capabilities while keeping the organization focused on improved patient care. Among other things, this also obviates the need for hiring expensive and hard-to-find data scientists

The most important concern around the use of cloud models is information security. High profile data breaches at Anthem, Premera and other healthcare organizations have made CIOs understandably nervous about cloud hosting, notwithstanding the robust infrastructures developed by the likes of Amazon, Microsoft, IBM and Google.

However, it is not just the cloud hosting provider, but the whole ecosystem that includes the analytics solutions provider, the data integration platform, and a host of other players that need to be able to demonstrate robust information security and compliance. HIPAA, which governs all of the data privacy and security around healthcare data, requires physical, administrative, and infrastructure controls that many cloud solution providers may not be fully aware of.

Data security, in particular, has to consider data at rest and in flight. Physical security requires access controls to facilities that host protected health information or PHI. Associates are required to undergo orientation and training to HIPAA before being allowed to work on the data.  

And finally, all parties need to ensure encryption and anonymization of data before it is transmitted to a cloud provider for analysis.

Many vendors of cloud solutions, especially those that are new to healthcare, fail to understand the implications of HIPAA, and fail to appreciate their responsibilities under a Business Associate Agreement (BAA) that may impose unlimited liabilities for data breaches.

Despite all this, cloud-based models offer a great deal of advantages to healthcare enterprises with limited resources and budgets looking to deploy analytics for improved clinical, financial and operational outcomes.

I will be presenting on the key considerations in selecting a cloud vendor for analytics at the Healthcare 2015  conference this week in Nashville. If you happen to be there, be sure to stop by.

This article is published as part of the IDG Contributor Network. Want to Join?

NEW! Download the State of the CIO 2017 report