At work, most of us rely on a grab-bag of computer systems. Certain environments are especially burdened with complex, diverse systems; the military comes to mind. Health care certainly falls into that category, teeming as it is with systems to deal with scheduling, resource allocation, billing, security, and–of course–patient information. Many environments spend millions of dollars trying to get these systems to work together in order to save time and avoid errors.
Let me start with a small horror story (most people in the health care field could supply others). A psychiatrist friend of mine at a major, supposedly advanced Boston-area hospital had to issue a prescription for a patient. Even though the patient’s address was in one system, the psychiatrist had to fill out the address by hand on the prescription. You can guess what happened–she put in the address of the neighbor who lived next door to the patient. The neighbor was very polite and delivered the medication to the patient, but I can think of a dozen scary things that could have gone wrong.
Recently, I had a discussion of trends in integrating health care systems with Cheryl Rodenfels, CTO of Healthcare at Nutanix. With a strong background in finance and healthcare, Rodenfels previously worked at Scripps Health and then became CIO of Presence Health, a large Chicago integrated delivery network with 12 hospitals and 300 clinics. She divides clinical IT systems into four types, each of which gets a section below.
This level involves a sudden pivot. The organization has learned of a new need and has to act quickly. As an example, Rodenfels cited clinical trials, where the researchers are anxious to make use of a grant as soon as it’s awarded. In current institutions, the IT staff can’t ramp up quickly, and the frustrated researchers take matters into their own hands, setting up one-off solutions in the cloud. Another ramp-up situation comes with acquisitions of physician practices (very common in health care today), where the new practice must be given access quickly to the organization’s systems.
Progress in these situations consists of training IT staff in the use of modern tools, including cloud opportunities, to reduce costs and change quickly. This will allow the entire institution to innovate effectively and take advantage of opportunities such as research projects.
Healthcare institutions, like most businesses, have long used services for basic tasks such as customer management and payroll. Unlike the ramp-ups described in the previous section, these services tend to be spread across the whole organization. Healthcare organizations also tap external vendors for voice recognition and patient engagement through mobile messaging, which are increasingly important to many tasks throughout healthcare. Telemedicine, offered by numerous services, was on the increase even before the COVID-19 crisis thrust it into ubiquity.
The vendors of external services are usually poised to exploit the newest and most effective technologies, and adhere to standards when they are available. (For instance, every modern service offers a RESTful API, a very flexible way of connecting customers that is well understood by programmers.) The vendors are ahead of most healthcare providers technologically and offer basic IT functions such as high availability that relieve responsibility from overburdened healthcare IT staff.
However, the services still need to be integrated with other systems specific to healthcare, and this can turn into a burden shouldered by each clinical setting individually. We’ll turn next to these healthcare systems.
Core Clinical Tasks
Here we have the EHRs, PACS, document management systems, and other digital services specific to health care. Some follow idiosyncratic health care standards such as HL7, some claim to follow these standards but trip over subtle incompatibilities caused by ambiguities in the standards, and some are totally proprietary. The modern FHIR standard standard has recently provided welcome relief to system integrators, but it’s still evolving.
The systems are complex and have to remain in continuous use even as organizations try to upgrade them and convert data to more standard formats.
In addition to FHIR, clinical application vendors have been provided easier interfaces in order to follow their customers into the mobile age. Older versions of these clinical systems simply took input and displayed output on their desktop application. Now doctors expect to pull the apps up on a tablet, and their clients log in from home on a wide variety of devices. The vendors have stepped up to the challenge, and by following these common standards, they make integration easier too.
Legacy Systems and Devices
These consist of many types of critical systems scattered around hospitals: fetal monitors, IV pumps, infant security, and so forth. Because lives depend on them, they must be FDA-approved and are very hard to upgrade. Many other specialized systems are also part of the questionable legacies. Thus, Rodenfels says that every healthcare institution she has evaluated contains at least one critical system running Windows XP. The security implications of this dependence on obsolete software are disturbing. The right way to manage such systems is microsegmentation: wrapping application security policies around each of the legacy systems.
Some devices aren’t digitized at all. Although they presumably don’t have security problems, they may be unreliable in other ways–and forget about updating them.
Most clinical institutions have given up trying to integrate all these systems into their other activities. Like many organizations, they try to use modern tools and languages for new projects, such as telehealth and analytics. The legacy systems gradually disappear as they become obsolete.
What Does Nutanix Contribute?
Nutanix’s first offering was a software platform with virtual compute, storage and network capability. In other words, a client could place the tools it needed together with all the required operating system support into a package that would be easy to deploy on-site, essentially creating a private cloud. Nutanix also integrates with public cloud instances so that customers can manage their application workloads where it makes the most sense for their organizations. Virtual instance creation is a common task offered by many services, but it’s not always simple. In health care, EHRs and other specific clinical systems complicate the task by their unique interfaces, as I discussed before. Nutanix’s healthcare team works with the many application vendors to test and transition virtual instances to its platform, which saves the institution’s IT staff from having to do it. Nutanix calls its platform “hyperconverged infrastructure.”
Nutanix also covers a lot of their clients’ common needs, such as security, high availability, and backups. The company also provides a number of general services that all clients, including health care, may find useful. For instance, Beam analyzes a workload and compares it to how an organization is using private or cloud resources, to help the organization reduce waste and decide where to host its instance. Nutanix also also supports GPUs with vGPU technology. For healthcare, this makes a regular video monitors capable of providing diagnostic quality images, while costing a fraction of purpose built imaging displays.
I’m hoping that this article helps clinical IT staff classify their systems and anticipate the integration and upgrade problems each will call for. Nutanix seems to have gone through a lot of these headaches and overcome them. Connected health and analytics will make it more and more crucial to connect systems in modern ways–and the healthcare institutions that succeed will be the ones to survive.
This article is part of the #HealthIT100in100