Strategies to Survive the Ai Revolution

Dale Kivi, MBA – Aquity Solutions

HIM is achieved with an ever-shifting balance of human effort and technology. The coming onslaught of Artificial Intelligence Ai applications, however, will dramatically tip that balance by the end of 2020. Although this movement will significantly improve detailed data capture and subsequent analytics, HIM professionals need to be wary of the volume and range of solutions about to hit the market.

The Ai movement is being called the 2nd industrial revolution. And just like its predecessor, the machine age, which accelerated large scale manufacturing with a fraction of the previous workforce, many vendors will try to cost justify their Ai offerings with forecasts of how many HIM positions they claim to eliminate.

A few years ago, AHIMA predicted the future of our profession would be in data analytics. As the industry makes a hard turn towards Ai, that future is now. To proactively defend our professional future, we need to show vendors, as well as our own C-suite, how well we truly understand the business of healthcare.

Make no mistake about it, thousands of vendors are currently developing products intended to prove any set of ‘if-then’ human decisions can be converted into software code that does it better, faster and (not necessarily) cheaper. Many will succeed, but even more can be expected to fail. (How many of the 492 inpatient or 1,932 ambulatory Stage 1 certified EHR vendors do you think are still in business?)

Sales will be driven by impressive demos and an abbreviated list of quality metrics measured within the scope of their targeted workflow process. What won’t be highlighted are the full range of everyday exceptions, downstream QA costs, and subsequent KPI impact. Our ability to identify those issues and quantify their impact will determine the cost of this shift, both financially and to our professional ranks.


Reality Check
The latest wave of document creation Ai tools feature CAC-type prompts to ensure physicians include comprehensive billing data. These applications present physicians with a list of potential coding/billing issues to resolve based on what they dictate AND score them on how well they resolve the prompts. Although this adds even more documentation responsibilities to physicians who would prefer to spend their time practicing medicine, it will significantly improve their overall efficiency.

Since this new Ai function identifies problems while physicians are still in the documents, the time to resolve those issues is minimized. As an added bonus, such in-process resolution significantly reduces the DNFB lag for cases that would have resulted in physician queries. Accordingly, between improved document integrity for RCM concerns, reduced physician query processing time, and corresponding dramatic DNFB gap reduction, this Ai application promises to deliver undeniable process advantages.

Unfortunately, it comes with a just as undeniable risk for how vendors calibrate the aggressiveness of their RCM prompts. If they don’t offer enough suggestions, the programs add no value. If they offer too many, you risk upcoding. One overly aggressive example resulted in Providence Health of California being charged with $188.1 M worth of Medicare upcoding due to their CDI program.

That case stems from unsupported secondary MCC references which inflated Medicare reimbursements. If the MCC references were resolved by physicians while creating the documents, versus through overly aggressive CDI recommendations after the fact, the problems would not have occurred. Accordingly, proactive RCM queries during document creation deserves industry-wide adoption, provided HIM professionals (not financially focused RCM staff) manage process integrity.


Looking Back to Move Forward
Just like the machine age, once Ai products start hitting the market, there’s no turning back. Consider the decimation of the MT workforce due to speech recognition. Suggesting a 60% labor force reduction over the past 10 to 15 years may be a low estimate. Promoters of Ai hope to have a parallel impact on coding over the next two to three years and it will be surprising if they don’t get there in five. Consequently, our profession is at a major turning point.

The vendor game plan is simple and predictable. First, outsourced vendors chip away at in-house staffing models through lower cost labor (weaker wages, benefits and/or offshore teams). Second, Ai vendors promise to reduce total staffing levels, but may not reduce total process cost if additional QA is required. Third, once the technology matures, outsourced vendors will bundle it to deliver dramatic savings that can’t be ignored. That’s what happened in transcription and that’s what’s presently unfolding in coding. ROI and billing functions should be expected to soon follow the same track.

As our profession braces for this inevitable (and yes, it’s inevitable) workforce reduction, we have a choice to make. Sit back and wait for it to be forced upon us or get out ahead of it and save our organizations the cost and quality pains inherent to less than mature technology. We can’t stop Ai vendors from releasing applications on the market and many will bring measurable data capture and analytics advantages. At the same time, we can, and should, make sure they deliver a positive return on investment based on full workflow cost and quality measures.


Survival Guide
HIM’s collective success will come down to how well we manage through the transition to an Ai driven marketplace. HIM is about managing process integrity, clinical and financial outcomes, and total process cost, not simply managing labor forces. To be competitive, we have to use the most cost-effective tools available, which includes vetting the tools that are offered.

To ensure proper vetting, we must define Ai pre-launch cost-benefit analytical tools that force vendors to validate their promised advantages before the big checks are written and their products are rolled out to the entire organization. If they won’t risk the price of their product through a controlled pilot, why should you? And make no mistake about it, the time and energy to run a controlled pilot is much less than rolling out and adapting to a not ready for prime time application across your entire workflow.

It’s also critically important to proactively acknowledge the Ai movement targeting HIM positions to your C-suite. By assuring them you are focused on managing process integrity, clinical and financial outcomes, and total process costs, not just protecting your existing labor force, we have a shot. If they think we are only looking to protect staff, we won’t stand a chance. We need to accept that the revolution is coming and assert ourselves as the ideal resource to manage through it with the least amount of pain.

By the end of 2020, clinical documentation and RCM efforts will require less people and more technology, yet only HIM professionals can ensure process integrity, clinical and financial outcomes, and total process costs. Moving forward, we must remain the stewards of health information management.