By Michael Atkins, VP of Technical Services, PointClear Solutions
Twitter: @PointClearHIT
Machine learning (a.k.a. AI, deep learning, and cognitive learning) is poised to reshape the world in which we live. An emerging tool that uses both structured and unstructured data to infer patterns, relationships, and rules, it’s already rocking the retail, entertainment, financial services, and transportation industries (among others). In fact, many of us experience machine learning daily in the form of targeted online advertisements, voice-to-text functionality, ride share apps, and remote check deposit.
But what about machine learning in healthcare?
No surprise, the healthcare industry has been a bit slow to embrace machine learning – and for a number of valid reasons. But, mark my words, in the not-so-distant future, it will transform healthcare. Because when we are systemically able to leverage data found in images, vital signs, blood tests, biopsy results, medical histories, medication histories, physician notes, genomic profiles, epidemiological data, and even medical research papers, we will be positioned to better:
- Diagnose (faster, more accurate, more accessible!)
- Predict outcomes (who’s at most risk for X or Y!)
- Provide follow-up care (reducing readmissions and optimizing patient flow!)
- Tailor treatments
Machine learning will make healthcare more intelligent, safe, efficient, and cost-effective for us all.
Preparing for the Future
As PointClear Solutions’ Technical Services lead, I’m often asked what hospitals, physician practices, payers, population health companies, and others can be doing now to prepare for a future where machine learning is the norm, rather than the exception. Here are a few suggestions…
- If you haven’t already done so, adopt an IT infrastructure based on modular and open architecture principles that make it easy to add or update components and integrate machine-learning solutions (as plug-in functions) to your EHR.
- Develop internal policies and mechanisms to transfer health data in and out of legacy systems securely, so that you can build, test, and deploy machine-learning algorithms at a later date. This approach positions you to take advantage of industry innovations, while reducing the risk of machine learning-system obsolescence and avoiding the cost of custom integrations.
- Recognize that EHRs often don’t capture outcomes of care or treatment goals in a standardized format – and solve for it. Without standardized endpoints, it’s impossible to “train” machine-learning algorithms to sufficiently explain the variability in outcomes that can be translated into better tailored diagnostic or treatment processes.
- Enable and enforce data standardization across your organization. 80% of the estimated 665 terabytes of data hospitals produce each year is unstructured, but with appropriate tagging, this problem is solvable.
- Advocate – among policy and decision makers in federal agencies and leaders of medical specialty societies – for efforts to identify and encourage the adoption and capture of standardized outcome measures in EHRs. Such outcome measures should be salient to specific clinical conditions and allow comparisons across conditions.
This article was originally published on PointClear Solutions and is republished here with permission.