3 Reasons that the next generation of clinical informatics will drive real change

You don’t have to look far and wide to see signs of how the broader healthcare ecosystem is embracing innovations in clinical informatics as an avenue towards the transformation of patient care.

Earlier this year, drug maker Roche purchased for $1.9 billion Flatiron Health, a cancer-focused health analytics company that is using big data to improve oncology research and development. In June, Medidata acquired cloud analytics company Shyft for $195 million to build a data platform that helps life science companies streamline the way they develop pharmaceuticals and bring them to market. Further, in an article published earlier this year, new Novartis CEO Vas Narasimhan discussed the company’s need to “unleash the full potential of today’s scientific advancements and sophisticated analytical tools that make it possible to uncover valuable information from large-outcome trials to inform patient care.”

These are just a few among many potential examples that illustrate that healthcare organizations (HCOs) are making substantial investments in improving their data assets, infrastructure and technology. These investments in enhanced clinical informatics have become part of the core strategy of many HCOs, mirroring what other industries have been doing for years.

Having started several ventures in healthcare analytics going back to 2006, I’ve become increasingly excited to see this uptick in activity around the adoption of new and promising approaches to data and artificial intelligence to improve patient care. Over the coming years, as innovation continues to accelerate, we expect that the following three trends will play an integral role in shaping the future of clinical informatics and, more importantly, HCOs’ ability to utilize informatics to improve health outcomes.

Integrating patient-level data from disparate sources is key to realizing the full value of clinical informatics. Due in large part to the HITECH Act of 2009 that mandated many providers’ adoption of electronic medical records (EMRs), the healthcare industry is awash in data. As a result of rapid advancements in machine learning and artificial intelligence, these massive data sets can now be used to generate much deeper and actionable insights than were possible just a few years ago.

The challenge for many HCOs is that EMR data isn’t enough. To fully develop a 360-degree view of patients, HCOs need to integrate a broad spectrum of data types ranging from pure clinical information to social determinants of health data. We’re already seeing signs of this in oncology, where researchers and data scientists have made huge strides in integrating clinical, molecular, treatment, and outcomes data on a massive scale, for example. As time goes on, expect to see more incorporation of data from outside the traditional health sources, such as geo-demographic, mobile, social media, behavioral and real-time data, providing a more granular understanding of the patient.

Artificial intelligence (AI) must be developed with context-specificity. Artificial intelligence holds tremendous promise in reducing costs and variation across care, but AI systems are only as strong as the human knowledge they’re built upon. Clinical informaticists should remain humble, acknowledging that our current understanding of disease and human biology is fairly limited and that we need to continually drive towards discovering new insights without relying on our own hypotheses or biases. Today, we have seen some examples where players in the space have seen disappointment in the effectiveness of AI largely due to this issue.

One key lesson learned from large-scale AI projects is the difficulty in implementing analytics on free-form clinical text in a fully automated fashion. This is in part due to how difficult it is to generalize the application of AI in a way that is meaningful across different contexts. In this case, the problem is that the same words are used in different ways in different contexts. As a result, context-specificity is an integral component of developing text analytics for clinical information. This same dynamic applies to areas beyond clinical text, as well.

Common standards for healthcare data exchange are needed to overcome interoperability issues. HCOs are collaborating to share more electronic information than ever before, but that doesn’t mean they’re doing it efficiently. Though it seems like the health IT industry has been discussing this problem for years, too much data remains isolated in silos, with interoperability problems among different systems preventing HCOs from effectively sharing patient data. Indeed, it is still common practice for test results to be faxed or downloaded in such disparate formats that clinicians struggle to access all the data that could be useful in making decisions about patient care.

Despite progress towards data exchange standards such as FHIR, the health industry still needs to establish a standardized, portable way of accessing patient data to help providers make the most informed treatment decisions.

As the field of clinical informatics continues to evolve, the ability to leverage data to derive deeper and more unique insights will become a competitive differentiator for companies that successfully drive disruption in the market. Today’s innovators in data and analytics are likely to be tomorrow’s leaders in transforming patient care.

Copyright © 2024 Becker's Healthcare. All Rights Reserved. Privacy Policy. Cookie Policy. Linking and Reprinting Policy.

 

Featured Whitepapers

Featured Webinars