Next-gen medical devices: Security, AI, rethinking design for patient experience

Medical devices are proliferating throughout the healthcare landscape, especially with the advent of the Internet of Things and the myriad new products that come with it.

The medical device arena is undergoing a lot of change, in fact, as new technologies and calls for greater security push manufacturers to upgrade their products — and the next generation of medical devices could see many new features and functions.

As medical devices continue evolving, the top health IT areas manufacturers and hospitals will be ramping up include security, clinical workflow integration, data management automation and patient experience.

Device security a key priority

It’s no secret that medical device security has become a key priority across the industry. The U.S. Food and Drug Administration in mid-April released the Medical Device Safety Action Plan, with the goal of making devices safer.

Similar to how computer networks are vulnerable to security issues, exposing medical device vulnerabilities could put patients at risk, compromise how a medical device functions or allow an unauthorized user to access a provider’s network, said John Schoew, a managing director at consulting giant Accenture.

More medical devices today are network-connected with the ability to transmit data wirelessly and often integrated within the hospital’s clinical workflow.

"With the clinical integration of connected devices, health systems will need to assess if changes are needed within their IT infrastructure."

John Schoew, Accenture

“What this means: While health systems have often seen this area as a standalone component, medical device risk must be addressed holistically with a coordinated governance approach across security, patient safety, supply chain and biomedical engineering,” Schoew said. “Business and security leaders are increasingly focused on interlocking security strategy and cyber defense capabilities to strengthen business resilience and brand trust that are so essential in today’s connected world.”

On another front, as EHR interoperability deepens across clinical functions, connected medical devices are becoming more common. As this happens, health systems must augment their technology and processes to establish integration of medical devices within the clinical workflow.

“With the clinical integration of connected devices, health systems will need to assess if changes are needed within their IT infrastructure,” Schoew said. “As with any new connected device, health systems should assess potential privacy and security risks such as PHI, data management/handoffs, confirm basic provisions such as encryption and credentialing, embedded functionality such as firmware updates, and how it connects to the existing IT infrastructure such as the EHR.”

And with a growing ecosystem of devices and applications within health systems, so grows the complexity and cost of system management. Implementing robotic process automation to scale processes in an efficient and cost-effective manner is proving effective for building resilience in complex technology environments and in taking out costs associated with routine processes and maintenance of systems, Schoew said.

Real-time analysis of device data

Brian Lawrence, chief technology officer at Hill-Rom Welch Allyn, which offers many technologies in categories including nurse communication, falls prevention, vitals, cardiopulmonary, retinal imaging and medical devices, said it is critical that companies in the medical device industry – and in healthcare settings generally – start to think differently about the data generated by these devices: What data is being captured? How is this information accessed? Are isolated parameters being studied but others that could be of tremendous value being ignored?

“The very first next-generation feature we need to look at is creating an ecosystem that enables the real-time capture, access to and predictive analysis of the rich data coming off our devices,” Lawrence said. “The average age of data entered into an EHR is five hours – too old to be of any predictive use for a patient who needs care now. Enabling real-time capture and access to key patient data is essential to helping patients get better care.”

"Enabling real-time capture and access to key patient data is essential to helping patients get better care."

Brian Lawrence, Hill-Rom Welch Allyn

The second next-generation area is sensing – and getting more out of the technology, Lawrence said.

“Our industry has created brilliant hardware, software and algorithms that can sense literally hundreds of patient parameters from basic vitals to more complex and subtle measurements,” he said. “Unfortunately, we often don’t make full use of the capabilities we already have and aren’t attune to what we could be doing technologically.”

An example: Most advanced hospital beds today can weigh a patient. The organization takes the relevant parameter, a patient weighs X number of pounds, and enters that into the EHR. And that same bed might also be outfitted with a bed-exit alarm. So the parameter studied is whether the patient is in or out of bed. But using the same weight-sensing technology perhaps with more advanced algorithms, the organization can capture the dynamic aspect of the patient’s movements in real time to anticipate an adverse event or otherwise ease their comfort and safety.

“In other words, with technology already available, but using a different perspective, sensing can offer variable, as opposed to static, data – and help caregivers use that data efficiently and effectively,” Lawrence said.

Total care: The patient experience

The third next-gen development is patient experience. Healthcare has to look at people as needing total care – physically and emotionally, including their ability to be comfortable, connected and generally empowered with a sense of control over their care environment.

“We need to design devices with simple patient controls that are easy to access; devices that are more comfortable and quiet – blood pressure cuffs and beds are two good examples,” Lawrence said. “Blood pressure cuffs that take measurements on the inflate cycle versus the deflate cycle are far more comfortable for patients while maintaining high accuracy. Beds designed with ergonomic controls to make patients feel empowered over their environment, less like prisoners and more like VIPs, incorporating conveniences that address today’s needs such as mobile devices, help patients stay connected, and even addressing issues like incontinence in discrete ways, preserve dignity.”

These seemingly little advances in patient comfort and satisfaction have a huge impact on the overall patient experience, translating directly to enhanced outcomes, he added.

A big dynamic in healthcare is the promise of data science, which is often lumped under headings like artificial intelligence and some of its tools, such as machine learning or deep learning. Healthcare executives can think of these tools in much the same way people thought about microprocessors in the 1980s. Suddenly, there is a new technology that allows designers to be significantly more agile in the development of capabilities.

“Machine learning is the CPU of the current decade, adding an entirely new range of capabilities: Auto interpretation of images and waveforms; automation of workflow; gamification as a replacement for traditional education/learning; and so on,” said Paul Mullen, GE Healthcare acute care general manager. ”But a key difference is that in the 1980s, only a handful of companies – Intel, Motorola, to name two – had the capabilities to produce microprocessors. Today, many organizations have machine learning capabilities.”

A new appreciation for health data value

Nearly every academic healthcare provider has on-campus production capabilities for new algorithms, Mullen added. That can shift the buyer-supplier relationship dramatically. Notably, some major healthcare providers are aligning themselves with large academic sites. GE Healthcare’s partnerships with UCSF and Partners are examples. But exactly how the collaboration and competition dynamic will play out is still an open question.

"Machine learning is the CPU of the current decade, adding an entirely new range of capabilities."

Paul Mullen, GE Healthcare

“Along with the explosion in algorithms comes a new appreciation for the value of healthcare data,” Mullen said. “There is little debate that the ultimate owners of data are the patients themselves, but the data owners, often healthcare providers, exercise a kind of proxy ownership. As these institutions discover the value of the data, the common refrain will be, ‘I’d love to share my data, but…’”

The “buts,” Mullen said, come in three flavors: “I’d love to share my data, but it cannot leave my country or region.” The Europeans are out in front on this and the implementation of General Data Protection Requirements adds new rigor to this trend. The second is, “I’d love to share my data, but I don’t trust anyone to manage privacy.” The U.S. Congress’ interrogation of Facebook CEO Mark Zuckerberg in April of this year highlights the generalized ignorance of modern data practices and the anxiety that comes with such uncertainty.

The third but is, “I’d like to share my data but my data is gold-plated, and you should pay a lot for it.” This posture will be most common among the academic powerhouses. None of the three of these are sustainable positions. Expect that the models for data sharing will change dramatically, Mullen said. A collection of data puddles may become much more common than huge data lakes, he predicted.

Rethinking design

Mullen also said that what constitutes an exemplary medical device will be completely reshaped.

“Today’s designers over-value signal quality: Another micron of image resolution, an improvement in ECG signal fidelity, a more sensitive measure of respiration – these are the reflexes of today’s device designers,” he said. “But consumer devices are probably good enough for many medical applications and many designers will find themselves frustrated as users choose lower traditional quality sensors for many applications.”

A mobile phone can read a urine dipstick as well as a traditional device, Mullen said. Today, one can buy a simple ECG device for one’s mobile phone, as well. These may not be good enough in traditional thinking, but they’re likely to have more value than today’s designers suspect.

“It may be helpful to consider the difference between diagnosis, screening and surveillance as three forms of testing — the first requiring the most traditional fidelity and the last requiring the lowest cost of deployment,” he explained. “When confronted with whether a cheap substitute might serve in the place of an expensive traditional device, designers and consumers will have to ask, ‘What would I have to believe in order to consider that ‘inferior’ solution?’ Where have we over-designed and what can we rethink?”

Twitter: @SiwickiHealthIT
Email the writer: [email protected]

Source: Read Full Article