The push toward healthcare interoperability and electronic prior authorization is meant to ultimately benefit patients and providers. However, providers will have to navigate these changes and integrate technologies to help meet coming requirements, streamline documentation, reduce administrative burden and continue to prioritize excellent patient care.
Here two leaders from MCG Health answer questions about emerging healthcare technologies, prior authorization, and the future of clinical documentation. Interviewees include:
- Conor Bagnell, senior vice president of product management at MCG Health
- Bernadette Minton, the vice president of data science solutions at MCG Health
Question: CMS recently proposed a new rule for prior authorization that sparked a lot of discussion in the healthcare community. What do you think are some likely benefits for hospitals or provider-led ACOs if that rule becomes final?
Conor Bagnell: First of all, I do believe that it is only a matter of time before the rule reemerges and becomes final. The current costs and frustrations with the existing phone and fax-based system are just too great. I think the interesting thing to watch will be the timeline for implementation. Once it is implemented, I do think there are quite a few benefits for hospitals and provider-led ACOs. I would also say there are benefits in there for patients and even payers too once they have gotten through the work needed to support the rule.
The first major benefit will be that staff at the hospital will be able to see whether a procedure or test requires prior authorization and what the requirements are for requesting them for any given patient. As the staff member is looking at the medical record, they will be able to see the requirements right then and there without having to navigate to another system. Should a specific intervention require an authorization then the next major benefit will be that the staff member can initiate an authorization request without having to transpose or copy information from one software system to another. These benefits will result in lower administrative costs for hospitals and provider-led ACO's.
A very nice secondary benefit of streamlining and standardizing the authorization processes will be that turnaround times on authorizations should decrease. This will benefit patients and their families who often experience stress when an authorization request takes days to process while they wait to schedule their procedures. Payers will also benefit from streamlined processes because data from the EMR is included within the request. Frequently denied requests occur because payers are not receiving the information needed to grant the authorization. By streamlining these processes every stakeholder will receive value. There is a lot to like about this new rule.
Q: Although the recent CMS proposed rule for prior authorization was put on hold for further review, adoption of the HL7 FHIR® standard is coming at some point. What are some solutions that might help facilitate that transition for healthcare providers?
CB: To some degree, adoption of the HL7 FHIR standard is already here. I do think we are only beginning to scratch the surface in terms of the potential benefits of having these standards within the systems that hospitals use every day. In many ways a good analogy would be the early computers. Early computers were the tools of tinkerers and hackers until the first "killer apps" were created. In those days, the electronic spreadsheet was that killer app. Today the large EMRs have done a nice job of implementing their support of the HL7 FHIR standards and now you are starting to see the first killer apps emerge that are being built on those standards.
We at MCG have been able to build solutions for our hospital clients on top of the HL7 FHIR standard that just would not have been possible even four years ago. One solution is Indicia for Admission Documentation which can take information from within the EMR and compare it against evidence-based guidelines to help the user by ensuring care decisions are medically appropriate and the documentation is sufficient for subsequent reviews with payers. This tool greatly reduces the time needed to review a case while increasing the confidence in the resulting documentation. Applications like this will become the killer apps" that make the HL7 FHIR standard foundationally important for innovation going forward.
Q: How is a leader in evidence-based medicine like MCG Health adapting to address the need for actionable data insights as well as developing new AI technologies?
Bernadette Minton: Health systems and payers are continually looking for ways to do more with less. They want to increase their throughput, to find ways for their staff to get more done in less time. And of course, they are looking for ways to reduce the friction and the administrative burden related to submitting claims and for prior authorization activities. MCG has long been recognized as a leader for its evidence-based content to address these areas. MCG also has a strong history of utilizing data and benchmarks to augment its evidence-based content. For example, MCG specifies a goal length of stay in its guidelines to help hospitals compare their performance to this benchmark and to give them an actionable goal to aim for.
MCG has been working on ways to bring data insights into the provider staff's workflow through the use of software tools. Artificial intelligence plays an important role in this process. For example, one way MCG has used AI to help with actionable data insights is with our Indicia for Effective Focus solution. The goal of this software is to reduce the need for utilization management staff to prioritize and reprioritize their work as patients come and go throughout the day. Indicia for Effective Focus uses an AI algorithm to prioritize the daily work, along with some well-designed workflow features to keep the current, most important tasks front and center, unencumbered by tasks that are "in waiting mode." The AI algorithm uses natural language processing technology to understand the admission diagnosis and combines this with predictive analytics to assign a priority category to each case. For observation and inpatient cases that are most likely to be in the correct status — these all fall to the bottom of the priority list. At the top of the priority list are those cases with a high probability of needing a status change. This priority ordering ensures that the utilization management staff is focusing their early and best efforts on those cases most likely to benefit from their increased attention and time. In this way, AI helps to ensure that the utilization management staff are practicing at the top of their license, putting their main focus and time on those cases most likely to benefit from their attention, and making it possible to spend less time on the "slam dunk" cases.
Most, if not all, companies developing new AI solutions are working closely with client innovators or early adopters. The clients provide the use cases and the data, and the companies bring the product development know-how plus AI expertise. This allows for the employment of lean startup principles for fast learning, pivots, and shorter time to market. As a side-bar, "Lean Startup" is a book by Eric Ries, that I highly recommend. MCG is no exception; we've been fortunate to work with many clients in developing our AI solutions, and these partnerships have fostered a fast pace of innovation and delivery that we hope to do even more of in the future.
Q: Although there has been some apprehension in the provider community about using AI, can you share some thoughts on where AI can be applied to solve some fundamental hospital challenges?
BM: AI obviously has a lot of application areas. There is a huge potential for AI to help with clinical decision support, perhaps even some clinical decision-making in the more distant future. The most advanced AI technologies in use today are highly sophisticated and scalable pattern-recognition engines. Super-sophisticated medical reasoning will occur much further in the future — if ever. If we think about it that way and ask, "where are the places that pattern-recognition is useful?" The answer is in clinical applications such as the reading of radiology scans. There has been a lot of success with AI-based radiology scan reading. And there have also been some challenges when the algorithms used for one piece of equipment in one healthcare setting don't translate well to other settings, so as exciting as it is to think about using AI to help with clinical work, I'm actually quite excited about the potential for AI to help with hospital operations.
A hospital is a large, multi-faceted system where time is of the essence and complexity can be overwhelming. AI can automate mundane, high-volume, repetitive "sub-tasks" that consume valuable human time. Using AI to automate these sub-tasks frees up human experts to spend more of their energy and mental creativity on tasks that require human expertise. MCG is conducting pilots with NLP technology to scan a radiology report and identify phrases potentially related to indications for admission; a human being then makes the final assessment about whether a given indication is met or not. This is an example of how NLP technology may reduce the total time needed for a human expert to employ their expertise by pre-processing part of the work for them. It allows the humans to do more of what they do best and to do what only they can do. There are many such examples of using AI to help with operational tasks, including scanning for relevant changes to patient records in real time and helping ensure hospital staff is working from the right guideline for a given patient. These are all examples where AI can automate sub-tasks of the entire day's workload for a clinician. More and more coverage of these sub-tasks by AI will allow clinicians to devote more focus to tasks that rely on clinical reasoning — something AI cannot do with any large degree of sophistication.
Learn more about MCG's solutions for hospitals.