In our fall newsletter, I asked, “Has AI really arrived in healthcare revenue cycle?” I shared that as a company born in a digital age, Salud recognizes the need to collect, store, and analyze clinical and financial data that can help predict outcomes and improve processes that affect cash.
However, I questioned then and continue to question whether the industry does revenue cycle artificial intelligence well – yet. The pathway to meaningful AI and machine learning requires smart design, lots of data and continuous focus, and I am not sure how many are hitting that trifecta. With all of the pressure on revenue cycle today, we need to be taking steps now to ensure we are ready when the tools and processes of AI that are common in other industries make their way into everyday use.
You might start by completing this question: “Wouldn’t it be great if we could use artificial intelligence to …?” How many revenue cycle professionals would answer almost reflectively, “… to get the patient’s bill paid.” Unfortunately, that is not possible today, but the promise of AI does apply to individual steps in the end-to-end cycle that over time will come together so payment starts to “figure itself out” – automatically, accurately and in a timely manner.
AI has near-term potential throughout the revenue cycle, including areas such as patient access, health information management and utilization review. It has applications in value-based reimbursement systems, consumer-driven care models and population health. Efficient revenue management is the common thread through all of this. Salud’s commitment to assist in the health of the communities that our clients call home, as well as our calling and mission align well with our clients’.
While you may have a robust and aggressive plan for applying AI throughout your organization, Salud’s vision is to be the national model for the delivery of revenue cycle services, so we focus on improving efficiency, accuracy and the patient experience. We seek the information we need from electronic health information and from outside the healthcare space and then apply analytics and trending data to bring meaning and quality execution to this vision.
Rich data for analytics in electronic transactions
Consider dissecting and expanding on the data contained in the 837 transaction set, established to meet HIPAA requirements for the electronic submission of healthcare claim information that broadly falls into four categories:
- Patient demographics
- Patient conditions/reasons for treatment
- Services provided
- Prices for services
Depending on the problem you are trying to solve, the 837 is rich with data for healthcare analytics (and AI). This data also reside in your EMR, but we find most providers struggle to extract and store data as concisely as it can be found in electronic data interchange 837 files.
The 835 transaction set, aka the Health Care Claim Payment and Remittance Advice, is the electronic transmission of healthcare payment/benefit information. It contains some of your data, but adds payer information that may inform advanced analytics; most importantly, payments and denials. Once again, you may be storing some 835 data in your EMR system.
If you have trouble aggregating demographic, service, and payer data and extracting it from your systems, Salud recommends you begin now to store these electronic files, as they are rich with data you likely need for your future work with AI.
Data not captured in your information systems
Salud looks within and outside of the industry to find information that our providers don’t have, but that should be part of our AI framework. We’ve been impressed with providers partnering to share data in health information exchanges.
Of course, the industry still lags way behind banking, social media, pharmaceutical companies, and even pizza chains, all of which already collect and understand our habits, struggles and capability/propensity to pay. Earlier this month it was reported by Atlantic magazine that one political party has “3,000 data points on almost every voter in America, and they use those data points to determine how exactly to pitch their message.” Salud sees parallels in utilizing data points in determining how exactly to “pitch” care and the account resolution process most seamlessly.
AI requires a dedication to collecting and archiving meaningful data. Those who would like to move the needle on the delivery of revenue cycle services need to advance current capabilities in this area. From this rich(er) base of data, what can then be mined and designed are the most efficient workflows and exception processing systems for your operation.
Salud’s digital infrastructure already helps simplify processes, drives results and enhances patient experience. With the proper balance of data and human ingenuity, we look forward to building on our platforms so that sophisticated AI will soon connect data, predict outcomes and prescribe real world, powerful solutions. Until such advanced capabilities become available, Salud is focused on pre-AI, akin to machine learning analytics, while simultaneously acquiring and storing the most meaningful data from which to mine optimal workflow and liquidity.
We are excited to partner with providers looking for the right revenue cycle model for the digital age where many or all of these AI and machine learning concepts may already be in the discussion or planning stage.