Artificial intelligence (AI) is currently one of the hottest trends in business. And for good reason. AI has the potential to increase profitability an average of 38% and bring an economic boost of $14 trillion across 16 industries by 2035, a recent analysis suggests. This includes healthcare.
There is no shortage of possibilities for AI-supported solutions in healthcare. They range from improving doctor-patient communications to recognizing disease in diagnostic imaging. With its immense data-reporting demands, healthcare quality nears the top of this list. In fact, implementing value-based care is a concern keeping many U.S. health system CEOs up at night.
Numerous factors are increasing the need for highly efficient quality data management. These include the growth of clinical data registries, the adoption of pay-for-performance by commercial payers, and trends towards increased focus on patient-reported outcomes. As a result, hospitals must either direct more staff time toward quality reporting or find innovative approaches to meeting these demands.
Just as it has done for other industries, AI’s ability to automate traditionally manual tasks and create truly unique insights may yield unprecedented cost-savings for hospitals: up to 50% of data management expenditures. However, a giant computer furiously spinning mountains of health records into quality data “gold” is not an accurate depiction of how this happens. Rather, to get the most out of AI, our industry must quickly learn what problems this technology is best suited to tackle – and what it is not. There are important lessons to be learned from past information technology rollouts in healthcare.
In recent years, many hospitals struggled to submit complete electronic clinical quality measures (eCQMs) reports on time. A Joint Commission survey conducted in conjunction with two leading hospital associations found that 78% of hospitals were not ready for the 2017 eCQM reporting period. Around the same time, CMS released findings from an inpatient quality reporting validation pilot program evaluation highlighting workflow issues and processing procedures as critical barriers to successful quality reporting. This teaches us that even the most promising information technologies will only reach their full potential when paired with the proper clinical expertise and ongoing support.
Natural Language Processing (NLP) is a type of AI technology that enables a computer or software to understand human language and information patterns natively – versus ‘structuring’ data in a way that makes it possible for the software to consume. An obvious use for NLP is leveraging it for quality data reporting – since up to 80% of data in EHRs can be unstructured information, such as physicians’ notes and additional comments. This unstructured data tends to hold rich clinical insights. NLP has the power to analyze entire EHR databases, and can be trained to hone in on clinical data and documentation practices that are specific to each hospital or group of clinicians.
When properly deployed, NLP in healthcare quality will set the stage for numerous benefits, including:
• Speedier and more-accurate reporting, which can help hospitals work toward achieving maximum reimbursement and minimal payment penalties.
• Saved time and resources that can be redirected to take on additional clinical data registries and other quality improvement programs.
• Minimal disruption to staff. In other words, clinicians do not have to drastically change how they enter information into EHRs and instead can focus on delivering and improving care.
• Fewer staff hours needed for manual data abstraction.
• Improved patient-reported outcomes and satisfaction due to adherence to quality measures.
One solutions provider currently working to establish proper deployment of AI in healthcare quality, and provide NLP-supported solutions to hospitals, is Q-Centrix, where I serve as chief operating officer. Q-Centrix is using NLP to suggest to its clinical expert data abstractors what information within EHRs should be included in a hospital’s quality reporting.
Bakersfield Heart Hospital in California is one facility benefiting from Q-Centrix’s NLP-supported solutions. Analyses since the solution’s launch shows a collective improvement in data abstraction efficiency. Specifically, data specialists who typically took longer to abstract data experienced a nearly five-minute improvement in abstraction time thanks, in part, to the NLP technology. Considering the hospital does thousands of these transactions each month, the potential cumulative improvement is immense.
There is a myth that once NLP launches, it operates and improves on its own, but technology designed to behave like a human must be guided like one.
Many hospitals have large, diverse and unique sets of information systems in use, so their workflows and documentation practices are also unique. One constant, however, is the need for a cohesive combination of technology and people. For NLP to work, it must be guided by clinical experts familiar with each institution’s particular combination of technology and workflow. Continuous performance audits and tuning are critical to ensuring the desired output. In other words, clinical experts are the key to driving this new era of automation. With their support, NLP can achieve its potential.
For hospitals seeking more efficient and accurate quality data reporting, success will be a matter of how and when they choose to integrate AI technologies in these processes, not if. Modern technology, lean processes, and – most importantly – elite people comprise the necessary formula for success.
With the right approach, hospitals can recoup 25% to 50% of resources directed toward quality data management. This can translate to cost savings and time that an institution can redirect to programs it cannot get to today, and that can more directly and positively impact patient care – an outcome we all want to see.
As Q-Centrix’s chief operating officer, Patrick Herguth, is an expert on efficiency and technology deployment for process performance improvement. He led a thought leadership session on quality measurement strategy at the Fall 2017 IHI Change Conference.
The views, opinions and positions expressed within these guest posts are those of the author alone and do not represent those of Becker's Hospital Review/Becker's Healthcare. The accuracy, completeness and validity of any statements made within this article are not guaranteed. We accept no liability for any errors, omissions or representations. The copyright of this content belongs to the author and any liability with regards to infringement of intellectual property rights remains with them.