If you think tax prepping is tough, try diagnosing your health

Tax season. That dreaded time of the year when you pull out that big yellow envelope that you've been filling up over the last year with tiny receipts and various statements from your mortgage company, bank, stock broker, etc.

Rather than enjoy a nice Saturday at the lake or on a golf course, you get to settle in with all those pieces of papers and try to figure out which ones are needed for your tax return and what details may help reduce your income tax bill.

This year the tax preparation company H&R Block is anticipating a nice uptick in business, thanks to their recently announced partnership with IBM Watson. H&R Block intends to help consumers get every last deduction and credit they're entitled to by using Watson's super computer brain to analyze tax returns. Watson has been transformed into quite the tax expert after being fed the government's 74,000-page tax code and plied with additional insights based on 600 million data points. And Watson will just get smarter over time as it processes more and more returns, thanks to the system's machine-learning capabilities.

Leveraging cognitive computing to maximize tax refunds is a perfect application for Watson. A tax preparer can enter absolute numbers into fields like income and exemptions, and "yes" or "no" for many, many other fields. Once all the values are entered, Watson can run its millions of algorithms and report the final tax bill. Everyone should have a high degree of certainty that Watson employed the most appropriate tax preparation method and calculated the minimum tax liability for each individual.

If only this technology worked as brilliantly for healthcare. What's different about healthcare is that not everything in medicine lends itself well to discrete values. Unlike the tax code, which is based on discrete values, EHRs contain a considerable amount of unstructured, yet critical, free text information. We obviously have thousands of CPT and ICD-10 codes, but sometimes those codes fail to capture certain nuances that are essential for understanding a patient's health situation. I can note in my EHR that a patient is having excruciating pain in her right upper abdomen and that she has a history of gallstones, and I can enter and store all that information in a structured format. But what if I also want to note some obscure observation that may or may not be relevant, such as the fact that she sleeps in a hammock in her backyard and only eats nuts and Brussel sprouts?

Another reason EHRs contain a fair amount of free text is because many clinicians prefer to dictate their notes rather than document details directly into an EHR. Free text information can be converted to a structured format using technologies like Natural Language Processing (NLP), but the downside is that some critical data will be missed. Though NLP continues to get better, even the most advanced NLPs are only 90-95% accurate – which may be acceptable if your goal is to calculate population risk, but it's probably not adequate when customizing a patient's cancer therapy.

In a perfect world, I think many of my clinical peers would love to dictate and then have a Watson-like computer auto-magically take all those words, run everything through its millions of algorithms and then spit out a precise diagnosis and customized treatment plan. While that would be amazing, we have a long way to go before the technology is that advanced.

Cognitive computing can only work its magic when data is in a structured format. That's why the H&R Block tax experts must input discrete financial information into Watson. While in a perfect world we'd love Watson to take our yellow envelope of receipts and statements and auto-magically produce our tax returns, that's just not how the technology works – today anyway.

In healthcare today, clinical data still must be in a structured format in order to take advantage of advanced machine learning technologies like Watson. To facilitate this, we need to make it easy for physicians to produce chart notes that are mostly structured. This requires documentation tools that are quick, intuitive, and easy-to-use.

Medicine – unlike the tax code – will never be 100% black and white and free text will never fully disappear. However, the easier we make it to create "Watson-ready" chart notes, the readier we'll be to take advantage of advanced machine learning technologies in healthcare.

About the author:
Jay Anders, MD is the Chief Medical Officer of Medicomp Systems. Dr. Anders supports product development, serving as a representative and voice for the physician and healthcare community that Medicomp's products serve. Dr. Anders spearheads Medicomp's clinical advisory board, working closely with doctors and nurses to ensure that all Medicomp products are developed based on user needs and preferences to enhance usability.

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.

© Copyright ASC COMMUNICATIONS 2020. Interested in LINKING to or REPRINTING this content? View our policies by clicking here.

 

Featured Webinars

Featured Whitepapers