A new survey provides some insight into how hospital and health system leaders are prioritizing healthcare technology investments for next year with strong indications that healthcare leaders are focused on investing in proven technology solutions that will have an immediate impact, and are proceeding cautiously with emerging technology like artificial intelligence (AI).
The survey, conducted by the Pittsburgh-based Center for Connected Medicine (CCM) in partnership with the Health Management Academy, reflects the opinions of healthcare C-suite leaders from 20 major U.S. health systems across the country. CCM is a collaborative health care executive briefing center and is operated by five partners — GE Healthcare, IBM, Lenovo Health, Nokia, and UPMC. The Alexandra, Va.-based Health Management Academy is a membership organization consisting of executives from the country’s top 100 health systems focused on sharing best practices.
CCM and the Academy conducted a quantitative survey of IT leaders, specifically CIOs, chief medical information officers (CMIOs) and chief nursing information officers (CNIOs) at about 25 leading health systems, followed by quantitative interviews with health system CIOs, CFOs and CEOs at 20 health systems about health IT trends for 2018 and how these trends fit into the overall strategy and priorities of their health systems. The corresponding survey report, “Top of Mind for Top U.S. Health Systems 2018,” focuses on five areas of health IT, namely cybersecurity, consumer-facing technology, predictive analytics, virtual care and artificial intelligence.
Healthcare Informatics spoke with Gary Bisbee Jr., Ph.D., co-founder, chairman and CEO of the Health Management Academy as well as Bryan Clutz, Ph.D., researcher director at the Academy and Melissa Stahl, research manager about the implications of the survey findings on future health system technology strategies and investments.
An overall emerging theme from the survey was that while health system leaders are excited about the prospects of emerging technologies such as AI and machine learning, yet the majority are proceeding cautiously on these technologies and continue to be focused, instead, on proven technologies and IT initiatives, such as enhancing existing electronic health record (EHR) systems, standardizing IT platforms and cybersecurity solutions.
“One thing that was surprising to me was the priority level around AI and machine learning; it tends to be a topic that health system leaders talk about frequently. When we reached out to our informatics executives and in talking to our CMIOs, CNIOs and CIOs, many indicated that implementing AI and machine learning technology solutions in 2018 was a much lower priority,” Stahl says.
In fact, the survey found that more than half of health systems currently use AI, but 63 percent of hospital IT executives ranked the implementation of AI solutions as a “low” or “very low” priority for 2018. Executives reported that AI is in its early stages where proving its value is difficult and the technology still needs refinement, but they expect the technology to have greater impact in the future, according to the survey report. And, responding health systems expect to spend an average of 2.6 percent of their IT budget on AI in 2018.
In the report, one CIO who was interviewed said: “In think that health care is still figuring out how to get value from data in general. We have use cases, but they are not ubiquitous. Until data is ubiquitous, it makes AI hard to prioritize. Additionally, it’s great to predict something but if you don’t have a corresponding intervention it doesn’t do much. It’s interesting, but still an experiment.”
The most common areas in which health systems have, or are planning to, implement AI technologies are clinical decision support (59 percent), population health (46 percent), disease management (42 percent), readmissions (41 percent), and medical costs/health plan (38 percent).
Health system executives report that they have implemented AI in more operational areas such as revenue cycle, billing, and scheduling, but have less commonly implemented in clinical areas. Health systems are starting to utilize AI for clinical areas such as readmissions and risk scores, however, this is commonly in a pilot stage and hasn’t been fully integrated. Only 4 percent of health systems currently use AI for cancer care, and only 8 percent have plans to add AI technology for cancer care in 2018, and cited cost as a factor. Even with these challenges and slow adoption, information executives anticipate AI technology will impact the use of unstructured data at their health systems in the near future (three to five years), according to the survey.
“Overall, a major theme that kept appearing when talking to health system leaders about the challenges of implementing solutions in all of these areas was the financial pressure that these health systems are under,” Stahl says. “When they talk about implementing predictive analytics or AI initiatives or even strengthening cybersecurity, cost constraints and resource constraints, and trying to find talent, those were overarching themes in healthcare, in general. Financial pressure is really seeping into all of these trends and making it hard for health system leaders to navigate implementing a lot of these technologies, even if they see value in implementing them, but there are competing priorities and a great deal of financial pressure that makes it challenging to do so.”
Bisbee adds, “The CEOs and CFOs want to see technologies that are ready to go and ready to implement. They’re thinking about resource allocation, and when it comes to wearables or when it comes to machine learning or AI or any of these specific technologies, they understand that these technologies are foundational to what they are interested in, but they want to see what is the technology that is going to make us more efficient, or improve care quality or lower cost. The technology executives are much closer to, ‘Okay, how can we integrate AI into our information systems?’ There is a little bit of a different level in how these various groups—the CEOs and CFOs and the information technology executives—look at technology.”
Looking at the use of predictive analytics in healthcare, more than half of survey respondents are using or plan to begin using genomic testing and data analytics as part of providing personalized medicine to patients. Those efforts are focused on oncology, anesthesia and pharmacogenetics.
“Across the health systems that we talked to, executives indicated an excitement around predictive analytics and seem ready to implement predictive analytics across the health systems as they are looking to use it to improve quality and safety, to reduce readmissions, to improve clinical decision support and to improve efficiencies,” Stahl says. “Many health systems are early on in their maturity, have done some pilots, are seeing some results there and are planning to scale it across their health system. We’re expecting to see a lot of growth in predictive analytics and acceptance of this within the next two years.”
Health system leaders cited resource allocation as a major challenge to implementing predictive analytics as well as organizational culture, standardizing the clinical practice and leveraging unstructured data.
The survey found that 92 percent of health systems plan to increase spending on technology to boost cybersecurity in 2018, while less than half (42 percent) plan to increase IT leadership dedicated to cybersecurity and 67 percent are adding cybersecurity staff.
Stahl says the majority of health systems have leadership around cybersecurity, such as a chief information security officer (CISO), and are now focused on investing in technology areas to strengthen current capabilities in the area of data security, such as vulnerability scanning and detection.
Health system executives also plan to invest in technologies to help improve overall strategy and response: 54 percent plan to invest in identifying threats (asset management, governance, risk assessment); 50 percent plan to invest in protection (access control, awareness and training); 50 percent plan to invest in detection (continuous monitoring, detection processes); 21 percent plan to invest in recovery (business continuity/disaster recovery planning) and 17 percent plan to invest in response (cybersecurity incident response and analysis).
Other areas executives reported planning to focus cybersecurity resources include security services, outside monitoring services, and retaining consultants for cybersecurity assessment and attack/breach response.
Clutz with Health Management Academy notes that there is a critical cybersecurity talent shortage in healthcare, which may factor into health systems’ plans to invest more in technology and other resources rather than internal staffing. “Trying to find the talent in this space is really difficult as health systems evaluate what they can outsource versus what they try to accomplish internally,” he says.
While health systems are highly focused on improving cybersecurity, many challenges remain for healthcare leaders. Approximately one-third of executives reported lack of talent (38 percent) and immature IT solutions on the IT Security Maturity Model (33 percent) as top challenges. Health systems also commonly listed competing priorities, medical devices, costs relating to cybersecurity operations and anticipating emerging threats as challenges to improving cybersecurity.
What could be welcome news to many cybersecurity experts, the survey found that only 17 percent of health systems reported that they have opened bitcoin wallets to be prepared to pay for a ransomware attack.
Consumer-Facing Technology Investments
While health system leaders recognize the potential of patient-generated data, mobile apps and wearables lagged patient portals and home monitoring equipment as sources expected to generate valuable patient-generated data in 2018. Less than a quarter of respondents expect wearables (17 percent) or mobile health apps (21 percent) to be sources of valuable patient-generated data in 2018. However, executives said they are planning for patient-generated data to make up a larger portion of a patient’s health record in the future.
Over half (54 percent) of responding health systems integrate patient-generated data into the EHR, with 46 percent integrating structured, useful data and 8 percent integrating unstructured data.
“In our discussion, healthcare leaders definitely indicated interest in utilizing data from mobile apps and wearables but it’s just not on the horizon for 2018; they haven’t really developed the infrastructure or ability to utilize that data in the short term.” Stahl says. “The benefit of patient portals and home monitoring is that it’s much more structured and standardized data and they can be ensured of the quality of the data, whereas with mobile health apps and wearables, executives indicated they were hesitant around the validity of the data and the unstructured nature and that presented challenges with them trying to integrate that data into their EHRs to use it in a way that provides value.”