The global pandemic is pushing the healthcare system even harder to find ways to help hospitals efficiently address cost and streamline operations. From managing healthcare billing and the insurance process to providing a faster diagnosis of a serious disease, artificial intelligence (AI) has the potential to completely change how hospitals operate. MedTech Intelligence recently discussed some of the areas of impact with Jim McGowan, head of product at ElectrifAI.
MedTech Intelligence: How is AI helping hospitals manage healthcare bills and the insurance process?
Jim McGowan: The original areas within a hospital where AI created efficiency were in registration and insurance processing, most notably in revenue cycle management (RCM). RCM was envisioned as a seamless process across patient appointment and registration; claim coding and submission; payment reconciliation; and appeals. Over time these solutions grew so complex that parallel industries around Pay and Chase emerged, in which providers needed incremental support to capture all their revenue. With margins in the low single digits each dollar counts.
These RCM systems are rule based, which is antiquated AI technology. [Our] RevCaptureAi solution combats the limitations of these traditional revenue cycles with the dynamic intelligence of artificial intelligence (AI) and machine learning (ML) that track, analyze and generate insights about your missed charges. In a billion-dollar health system, just 1% of missed total charges adds up to $10 million in lost revenue. This is the opportunity.
Both providers and payers are implementing chatbots to more efficiently engage with patients/members by automating common support topics like confirming eligibility, getting claims/payment status, scheduling appointments and more. Machine learning is in the early stages of adoption. ElectrifAi has used machine learning to capture missed codes on hospital bills for [more than] five years, and building practical solutions to AI problems for [more than] 15 [years].
ElectrifAIs CEO Edward Scott discusses artificial intelligence and machine learning during the coronavirus crisis in Beating COVID-19 Is a Team SportMTI: How is the technology streamlining medication management? What is its role in managing procedures?
McGowan: Medication errors are still a significant issue in hospitals. EMR solutions were implemented to improve workflow and data capture for a complete patient view. These solutions have reduced adverse drug events (ADEs). Technology has been used to create many checks-and-balances within hospitals, which requires a double-check and scan of a barcode for each patient and medication to validate the drug was prescribed by a physician. There is continued work needed to capture the full patient history as these solutions are hospital system specific, do not include interoperability with the PBM data, and do not share with other hospital systems. Ultimately, a more complete patient system of record may be necessary to ensure that each system connects to each other to share data.
One of the areas where AI in healthcare has shown the most promise is in diagnostics, which can ultimately be leveraged in operating and emergency room settings. Right now, early diagnosis is one of the most important factors in the ultimate outcome of a patients care. AI deep-learning algorithms are being used to shave down the time it takes to diagnose serious illnesses. Our PulmoAi X-ray solution is an example of a tool that amplifies the work of radiologists, who leverage AI to triage cases as emergency rooms and ICUs overflow.AI is being used within healthcare for evidence-based recommendations. AI algorithms ingest collected vitals, lab results, medication orders and comorbidities and produce smarter triage tools.
We have seen growth in digital applications for mental health and virtual assistants to answer patient questions. As telehealth grows, I would not be surprised if the virtual assistants handle increasingly large volumes of questions, significantly greater than live operators. These bots are becoming much more important as the front-end to a telehealth call.
AI and Robotics for laser eye surgery and orthopedic surgeries are growing. AI-based visualizations are exploding in the market. AI is attempting to enter every facet of healthcare.
MTI: What factors should technology developers consider when designing AI solutions for hospitals?
McGowan: There are a number of important factors: Regulatory concerns, community demographics, fitting into existing workflows, technical proficiency of both the hospital personnel and consumers.
Healthcare is a highly regulated industry. HIPAA balances portability with privacy. This is for a very good reason, but has a lot of side effects, like complicating marketing efforts. You cant send an email to a patient telling her its okay to get the hip surgery she canceled when COVID-19 struck, because you cant guarantee someone else wont read it. If you send someone a reminder about their diabetes medication and are too specific in the email, what happens when that email is opened by someone other than the specific patient? Solutions that require you to log into a website to view the information was the evolution during the 2010s and continued to evolve with the growth in depth and sophistication of the mobile app solutions. Inappropriate sharing of data, even within a family, can create legal liability that hampers more specific and appropriate messaging.
When building solutions, AI can enable a very quick solution to the above concerns. Tools like robotic process automation (RPA) and chat bots have allowed providers to quickly create solutions that gather patient information and respond with an appropriate response, even in the patients preferred language. These more natural language conversations guide the patient to a choice without being overly and overtly intrusive.Most importantly, AI and ML people really have to deeply understand their craft if they want to influence medical decisions of any kind. Data science is not just technology development. It requires deep understanding of the problem domain being addressed, as well as statistics, inference, and logic. And data science without exceptional data engineering is useless. There is no magic inside the algorithms. If the data is bad, the results will be bad. Weve seen data systems where almost half the data is inaccurate. Let that sink in. Would you go to a doctor if half the facts in their medical books were wrong? AI solutions start with great data engineering.
Id like to talk directly to the C-Suite in the hospitals for a moment.
Lets discuss the elephant in the room: many hospitals are poorly run businesses, with razor thin margins and inadequate spending controls. These are not financially healthy organizations.
This year we saw 42 hospitals file for bankruptcyso far. All have two things in common: They all had revenue capture solutions, and they all couldnt pay their bills.
First, revenue capture doesnt address your problem: you need elective surgeries. Revenue Capture fixes leaks in your billing process. Hospitals dont go bankrupt because their billing process is too leaky. The revenue isnt coming in. The elective surgeries arent there.
Second, the revenue capture programs you do have use rules-based systems, and those dont work when the rules change. COVID-19 changed the rules. You needed a machine-learning based solution. Rules-based systems have been around since the 1950s. The world has moved on. We have a machine learning based revenue capture solution, and not one hospital using it has gone bankrupt. And still, that should not be your priority right nowthats just a part of getting healthy.
You need to restart elective surgeries. You need to manage your finances.
Customer engagement isnt optional for any other business, and it isnt optional for yours. Machine learning can help.
You also need to get control of your spending. Spend analytics is critical. Again, this is not optional for any business, hospital or not. Machine learning can help.
AIespecially machine learninghelps improve the health of the patient, the financial health of the hospital, and ultimately the health of the community. The pandemic should not be a reason to push off these technologiesits the reason you should embrace them today.
Artificial intelligence and machine learning are proving to be meaningful weapons in our arsenal during the coronavirus crisis.
Change is constant, and we continue to evolve.
A recent paper released by Duke University cites the promise of AI, but urges policy changes in order to bring AI-enabled clinical decision software to fruition.
Expanded designs that enable clinicians to leverage data in making healthcare decisions, but privacy challenges remain.
