AI continues to be the topic of the moment in healthcare. Lots of attention has been given to big projects aimed at reducing clinical note taking including Nuance’s generative AI project and a similar one for the MEDITECH EHR. For all the buzz about improving dictation and removing burdensome work, there are also more varied uses to AI that can impact patient care in more direct ways than reducing provider burnout.
HealthcareITNews recently looked at healthcare technology company Premier, Inc., which helps to implement artificial intelligence analysis tools for hospitals. The company has been executing all sorts of AI projects that range from clinical trial selections to population health and predicting illness.
While AI can work with structured data, it’s usually in the unstructured information that things are overlooked. Unstructured data can appear in hospitals settings in a number of different ways. The most obvious may be forms that are faxed, scanned, and emailed to hospitals containing information like lab results or referral information. This data is always tricky because different organizations will send documents in dissimilar formats or using different terminology. Some hospitals have their staff manually enter data, many use automated data extraction software, and others may simply index the file to the patient’s medical record without entering all the data.
Unstructured data can also be found within a hospital’s own walls. While key pieces of discrete data collected on-site will be easy to access, trend, and analyze, the same can’t be said for a field like a physician’s notes. Qualitative text like this used to be confined to individual patient scenarios, but with the help of artificial intelligence, it can be unlocked for use in predictive and trending use cases.
One of these use cases is predicting dementia and Alzheimer’s disease. By analyzing these notes in conjunction with patients’ structured data, researchers were able to find terms that identified opportunities for intervention while cognitive impairment is still mild. AI is also being used to bring representative diversity to clinical trials, identifying candidates first, rather than health systems.
Population health projects that were too large to tackle are also coming into focus. AI was able to analyze 6.3 million patients’ COVD-19-era lung scans in just three weeks, identifying more than 150,000 patients with a pulmonary nodule that could potentially have lung cancer. This is a far cry from the 10,000 to 20,000 the researchers expected to find. The rapid turnaround will allow patients to receive preventative treatment and screenings that may have otherwise been delayed or missed entirely.
Now that the largest datasets can be analyzed more readily, it’s crucial that healthcare organizations are accurately ingesting all the data they can. This is especially true for departments dealing in specialty care. Organizations that draw from diverse geographies are more likely to be receiving non-interoperable documentation and are therefore going to need a strong data entry staff or an automation solution. If you’d like to learn more about how we can automate your incoming data and documents, please reach out today.
About the Author: Chris Mack
Chris is a Marketing Manager at Extract with experience in product development, data analysis, and both traditional and digital marketing. Chris received his bachelor’s degree in English from Bucknell University and has an MBA from the University of Notre Dame. A passionate marketer, Chris strives to make complex ideas more accessible to those around him in a compelling way.