Health-care CEOs recognize that the cloud offers more than data sharing or infrastructure savings. They believe it can solve an age-old problem of connecting stakeholders, and move health care toward a streamlined delivery model. Yet, while health-care CEOs may have an idea of the end results, they may be hazy on what it takes to actually deploy that first mission-critical health-care system in the cloud.
Here are eight effective questions that will help health-care CEOs set the stage for a successful cloud transition.
What are we trying to accomplish? “The cloud” means many things. It can be a way to outsource computing infrastructure in a scalable and elastic manner; a model for how to build and deliver hosted applications; or a means to enable health information exchanges (HIEs), connecting stakeholders of care (payers, providers, patients) through an integrated, collaborative access point. You need to know what you’re trying to do and which approach makes sense. Are you trying to reduce operational expenses? Deliver products to market faster? Connect your people so they can better collaborate? Connect your applications so they can better interoperate? Some or all of the above?
Do we know our users? We need to give people what they want. A great application is one you deploy and conclude was a great choice. A bad application is one people can’t tolerate or, worse, resist or don’t use. Sadly, few organizations do usability analysis and testing. They design for what they think users want versus what they know users need. This is particularly important for new health-care delivery models such as the patient-centered medical home, where patients and providers might need a meaningful and productive online relationship. Users are your customers. What you give them must meet their needs and improve their ability to do their job.
Do we have a road map? Are you considering moving existing applications to the cloud or using SaaS applications from suppliers? Or perhaps a hybrid approach makes sense, where your applications are augmented by third-party cloud services. Is your business requiring you to support emerging health-care payment models such as episode-of-care (i.e., bundled payments)? Do you need to support payers and providers interacting electronically through HIEs? A road map will help you rationalize the financial aspects of tool and data delivery, support for such models, and understanding of costs and savings. It will also help you know when you should expect a ROI. Lastly, it will help stakeholders understand when they can benefit from new services and features that will be available.
What’s our deployment model? There are several cloud deployment models available. You need to understand what these are, how they work, what’s appropriate, and what your options are. You need to know how they differ in terms of risk, performance, and cost. And you need to know what direction your CIO is recommending and why he believes it’s best. This decision depends on how broadly you plan to integrate with other organizations (payers, providers, consumers, HIEs), how much data you’re willing to share, privacy levels required (especially as it pertains to protected health information), compliance requirements like HIPAA, and your risk tolerance.
What service levels do we need? It’s essential to talk about security, privacy, performance, and availability requirements. For instance, are your prospective service providers aware of HIPAA regulations? Are they equipped to respond to HIPAA audits? Can they comply with the HITECH act for health-care information security? These help you know whether the service providers you’re considering are up to the task. Unless your organization will manage its own infrastructure, you’ll be depending on third parties for basic cloud services. Many providers offer service and pricing levels to meet customer needs around disaster recovery, site redundancy, backup and recovery, and management. You need to know what service levels you need and whether the vendors can meet them.
Are we developing to the test? Building applications and then testing them is one thing. Behavior-driven development (test-driven development) is something else entirely. You can bet every system that ever failed was, in IT’s view, adequately tested. With behavior-driven development, you work from the outside-in, starting from a business goal, describing the goal with scenarios and user stories, and then defining the test before writing a line of code. You define how the application must behave, the behavior you expect in the test, and then you code to the test. The process is repeated until acceptable standards are met. When IT doesn’t understand the business goals and success criteria, they can get to the end zone and discover their assumptions were wrong or the design flawed even if traditional testing is done early and often. Engaging your users, such as providers and patients, in the process can help validate your assumptions, give users an opportunity to experience your solution early, and to provide you with valuable feedback to ensure a relevant, high-quality cloud solution that meets expectations.
Are we testing against load or volume? Every cloud system is expected to handle a specific number of simultaneous users, transactions, and data volume, and should be designed to scale as factors change. You need to test anticipated load and anticipated volume. Large systems often need contribution and collaboration across many parts. This is increasingly important as health-care platforms such as EMRs (electronic medical records) become compliant with meaningful use regulations, expanding the integration and size of the health-care information ecosystem. Knowing each part performs as you expect is important. But you also need to know that all the parts, once assembled into a whole, function properly. This integration testing is critical to ensuring success.
How should we deliver our solution? Consider a managed rollout (phased rollout). It’s “managed” in that you don’t just turn it on. You establish a way of phasing populations onto the system. You might phase it in city by city or state by state, small health-care provider networks, gradually ramping up the volume. You evaluate what’s happening at each phase to ensure the system can handle the volume. You can tweak things along the way, handling issues on a small scale, and avoiding overwhelming your support staff and partners. You learn and adjust by phase as you rollout until you’re fully up and running.
Given how important the cloud is to the future of health care—indeed, end-to-end integrated and connected systems in the cloud are the future of health care, as are connected stakeholders–CEOs can no longer afford to treat IT as a black box. The discussions can’t end with budgets, features, and timelines. CEOs need to know the crucial questions to ask and what answers to demand. Because today, you might have just one shot to get it right.
Date: March 4, 2014