As we live in the new world of quality, value-based care, we must be able to draw more insights and conclusions from ever-increasing amounts of information. We have the data, now we must put it to work. When we combine all of this data with machine learning, we are equipped to make smarter decisions. We have the power to transform healthcare – from the way we use electronic health records to the way we predict and deliver care.
A game changer for EHRs
Most EHRs are built on technology that is 20 or 30 years old. Generally, EHRs have kept up with rapid changes in healthcare by making incremental improvements over time. But it is challenging to retrofit EHRs to take full advantage of new innovations.
EHRs must do more than store data. They should be smart enough to deliver the right information at the right time, at the point of care. When an EHR is powered by machine learning, it can pre-populate information based on usage patterns and deliver preference reminders, constantly surveilling trends by user and organization to create opportunities for more effective care.
Plus, the power of machine learning surfaces information relevant to the encounter in real time, which helps improve quality and immediate interaction with the patient. Ultimately, this reduces the amount of time spent on documentation, helping address the problems of EHR fatigue and caregiver fatigue. That's all extremely important.
When EHRs can learn – gather and remember – what works best for each user, they can attain maximum efficiency.
A game changer for precision medicine
Precision medicine is an epiphany for clinicians and the patients they serve. Now, thanks to machine learning and AI, an individual's unique genetic makeup, environmental factors, lifestyle and family history can be factored into new protocols for an accurate diagnosis, personalized disease treatment and prevention planning.
Technology now has the power to bring the promise of genomics and precision medicine directly into the clinical workflow, while establishing a foundation for trial and research. This allows the industry to apply new genomic data models in a sensible way — to deliver the right information to the provider at the right time, while creating "research-ready" data to support a variety of objectives.
This can transform the way we care for a wide range of diseases and conditions – from cancer to hyperlipidemia to diabetes to renal disease to neurodevelopmental disorders. The ultimate goal is to drive better and more accurate diagnoses, treatments and outcomes — while simultaneously making this knowledge available for research and pharmacogenomics.
A game changer for population health, predictive modeling
Machine learning is also empowering us to analyze patient data at a level never before possible. We can now transform data into insights and actionable information.
Just think how a "data lake," where we are able to store millions of de-identified patient information to structure and to analyze data and study problems that are meaningful to health care, could transform diabetes care, for example.
We now have the power to compare things like blood sugar levels, body mass index, age and other risk factors and analyze treatment outcomes. Then, when clinicians are designing a treatment plan for a single patient, they can look to other similar patients and see which treatments worked well and identify other turning points that result in better, managed care.
This could be applied to the study of other areas of healthcare as well, including the opioid crisis. We can now couple information that is within the EHR with our "data lake" – and combine it with data that is available through public health mechanisms, such as PDMPs.
The goal is to develop algorithms to identify or even predict at-risk patients, and look at prescription patterns that most often lead to problems with abuse and overdose. Our research on this is still early, and we are just scratching the surface; it is clear that this is the direction in which we'll see excellent results.
The way of the future
Machine learning brings us an extraordinarily exciting set of capabilities today that didn't exist a decade ago. It enables computers to handle greater amounts of work than human beings can undertake, and will become increasingly important in this era of consumerization. It's making what we do better by improving the overall healthcare experience for both patients and providers.
Big Data & Healthcare Analytics Forum
The San Francisco forum to focus on utilizing data to make a real impact on costs and care June 13-14.
Paul Black is the CEO of Allscripts.