Introduction
For John Showalter, Chief Product Officer at Jvion, healthcare is on the cusp of a major data-fueled revolution. With healthcare being one of the few industries where Artificial Intelligence (AI) is ready for deployment, Showalter foresees a rapid change that will improve patient care and health outcomes. One of the most interesting applications is the use of AI to augment clinicians’ capabilities and offer additional pathways for patient care by supplying data that not only helps them to predict what might, or will happen, with a patient to enabling what should to happen for an optimal outcome for each individual under a clinician’s care.
But before you imagine an Orwellian future for healthcare with machines determining who receives treatment and who doesn’t, it’s important to understand what prescriptive analytics are – and are not – and why experts like Showalter are advocating for healthcare systems to embrace this data-fueled revolution.
AI in Action: Prescriptive Analytics
“AI is poised to improve healthcare in myriad ways,” Showalter shared in our recent conversation. “In the past, we were limited to systems that were only able to report the news – using data to describe what’s already happened or what’s going on now. Today, we can use data to predict what will happen, but the next step, the one AI is enabling, is to use data prescriptively.”
Prescriptive analytics not only anticipates what will happen and when it will happen, but also why it will happen. Further, prescriptive analytics suggests decision options on how to take advantage of a future opportunity or mitigate a future risk and shows the implication of each decision option.
By using data prescriptively Showalter was clear to point out that this doesn’t mean the data will prescribe a therapy or a course of treatment, but instead be used to streamline patient care, stimulate an action or a need to change course in patient care. “Right now I’m seeing prescriptive analytics being used to great effect in oncology and in medical imaging,” he shared.
Showalter continued, “for the treatment of patients with cancer, information from their Electronic Health Record (EHR) is analyzed by the AI platform, which then prescribes an action, such as beginning a conversation about palliative care, for the clinician to consider. The clinician then evaluates the suggested action within the full context of the patient’s diagnosis and treatment plan, and then any decisions that are made are made between the clinician and the patient.”
In medical imaging the role of AI is even more pronounced. “Imagine a patient goes to their primary care physician (PCP) with a persistent headache,” said Showalter. “While the PCP doesn’t suspect a brain bleed, based on other clinical evidence they send the patient for a CT scan. For a non-emergent CT scan it will be at least 72 hours before a clinician will read the image. However with AI, a machine can read and triage images based on what the algorithm finds and flags as urgent conditions for immediate attention by the radiologist. This can take the diagnosis time from 3 days to 30 minutes, which in the case of brain bleed, is the difference between life and death.”
Building the Infrastructure and Managing Obstacles
In many industries the ability to implement innovative technologies is stymied by a lack of infrastructure and an over abundance of obstacles. But healthcare is a little different, according to Jeffrey Lowe, Business Development manager at NetApp, who has collaborated with Showalter on numerous engagements. “Healthcare organizations have embraced cloud storage and data management solutions to the benefit of the entire organization,” Lowe shared.
“For the IT team, moving to the cloud has enabled the retirement of legacy systems and the opportunity to leverage nearly everything As-a-Service,” Lowe shared. With limited resources – both financial and personnel – the IT team is now able to focus on their core responsibilities managing the unique needs of their organization and its patients’ data.
So what obstacles are standing in the way of these tools in a field that is, as Showalter has shared notes, primed for adoption?
The first obstacle is concern over the security of data in the cloud. “AI requires vast amounts of data to be stored in order to generate useful outcomes for activities like prescriptive analytics,” Lowe said. “While there’s always risk in storing data, whether it’s a sticky note on your monitor or patient data in a database, the major hyperscalers like Microsoft, Google, and Amazon spend billions of dollars annually to ensure that patient data is protected. Microsoft, alone, has over 3,500 cybersecurity experts who work 24/7/365 on security,” Lowe shared
Even with that secure foundation, the need to share data securely without violating HIPAA and other regulations is a top of mind concern. “With so much personally identifiable information (PII) and personal health information (PHI) attached to each electronic health record (EHR), healthcare records are prime targets for theft and often unwittingly exposed as they are shared between organizations, departments, and devices,” shared Showalter.
Looking beyond security Showalter identified a very interesting barrier to adoption. “The biggest barrier to adoption as I see it is actually the clinician,” Showalter shared. “Healthcare organizations have the tools – cloud and high-performance computing – they need; they’re both readily available and inexpensive.”
Clinicians, however, need training on decision support and application of insight. Taking clinician-directed prescriptive analytics from concept to an integral part of practice requires training. “Putting prescriptive analytics to work is not just about creating a checklist or a carrying out a specific set of tasks,” said Showalter. “However, this will ease as technology becomes more prevalent and as the next generation of clinicians come on board.”
Conclusion
Even with these obstacles, Showalter predicts that in the next five years there will be a rapid adoption of predictive analytics beyond today’s use cases in hospital medicine and medical imaging. “Within the next decade population health will be AI-driven, and there will be extensive use of predictive analytics in ambulatory medicine as well for risk mitigation, in particular.” This change will free up clinician time from routine tasks so that they can focus on the patient and high-value tasks like motivating behavioral change in patients.
More importantly, because AI is a cloud-driven solution, it’s not just leading healthcare systems in major cities that will benefit from this innovation. “Cloud has leveled the playing field in healthcare,” shared Lowe. “Regional hospitals in rural Washington can access the same services as a major research and teaching hospital in Seattle, but scaled to their needs. The simplification of access and the democratization of sophisticated IT will speed healthcare organizations to adopt new tools, like predictive analytics, more quickly than anyone anticipated.”
With prescriptive analytics addressing some of the biggest challenges faced by healthcare systems and being able to facilitate better care for patients, it’s no wonder that it’s poised to be one of the biggest disruptors in the healthcare field in the next decade. While no evolution is without its pain points, healthcare providers have the foundation in place, so the steps from innovation to execution are reduced. “With a bit of due diligence, ample training, and keeping clinicians in the loop, there are few barriers to fully realizing the benefits of prescriptive analytics,” Shoalwater concluded.
