Data Mining & Artificial Intelligence

Identify opportunities today. Predict opportunities for tomorrow

Healthcare Services

Ontash began working with HCRIS, a huge database provided by the Center for Medicare and Medicaid Services (CMS) back in 2001. Ontash client, Quality Reimbursement Services (QRS), needed a way to identify hospitals for marketing purposes. HCRIS contains cost report data for over 5,000 Medicare hospitals in the US and its territories over the course of the past 25 years. Each cost report contains hundreds of pages of information including: patients admitted/discharged over the course of a fiscal year, the types of services provided, the costs, the length of stays, insurance providers, unpaid bills, etc., etc.

Needless to say, the HCRIS database is huge – dozens of terabytes of confidential patient information. In order to analyze the data in a secure fashion, Ontash uploads the entire HCRIS database onto servers located in Microsoft’s cloud platform, Azure. Azure then, allows Ontash to upload unlimited amounts of data – expanding the number of servers needed automatically, seamlessly and securely.

Understanding the characteristics of “target” hospitals for QRS, Ontash developed programs and API’s to data mine HCRIS for providers that meet these characteristics. Further, with these applications in place, dashboards were created so that anyone could run reports, at any time, to identify “target” hospitals, calculate the Medicare reimbursements, show trends in reimbursements, flag years when payments are in jeopardy, or when they qualify for legal appeals and special government payments (DSH, 340B, Outliers) etc.

Client Hospitals Within 10% of 340B Threshold

“340B” is a program offered to hospitals that serve very poor populations. Drug manufacturers provide deep discounts to such hospitals, so they may pass the savings on to needy patients. Hospitals are eligible for the program, if they meet certain criteria (e.g. not-for-profit); and pass the threshold of Medicaid/SSI patients to total patients (DPP > 0.1175). Ontash mines the HCRIS database, with APIs to identify those hospitals that are eligible (A), displays DPP ratios for qualification (B), and calculates the number of Medicaid/SSI days needed to meet the threshold (C). Further, Ontash identifies those hospitals that already have a relationship with the client (D), so that the client may engage those hospitals to discuss the opportunities.

Ontash continues to refine its data mining & reporting services for QRS. As legislation changes the parameters for Medicare reimbursements, Ontash constantly revises the programs to extract necessary data. But Ontash data mining and reporting services today go beyond marketing, to provide “intelligence” that QRS shares with its clients, to strategically maneuver through the most advantageous path in creating cost reports to maximize government reimbursements.

Client Hospitals That Are Eligible for 340B, But Do Not Have A Program

Hospitals that qualify for a “340B” program, but do not currently have one, provides Ontash client, QRS, with other marketing opportunities. Here, the HCRIS database is mined to identify hospitals that are eligible for 340B (A), meet the qualifying threshold (B). This list is matched against a government database that contains all hospitals that currently have a 340B program (which are removed from the Ontash report). The remaining hospitals are candidates for QRS services to register and participate in 340B. Again, those hospitals that already have a relationship with the client (C) are identified for special marketing efforts.

Artificial Intelligence (AI)

Using AI techniques, Ontash has enabled QRS to tap a new and potentially huge new market: SSI Enrollment for hospitals. CMS reimburses hospitals that serve poor and underprivileged populations – because these individuals are less likely to pay for hospital services. The data CMS uses to determine whether a hospital qualifies for these reimbursements, involves patient stays that are covered by government supplemental pay and insurance programs (SSI and Medicaid). CMS uses the number of SSI/Medicaid patients days to determine whether or not the hospital qualifies for reimbursements. Unfortunately, many patients who use hospital facilities may be eligible for SSI/Medicaid, but don’t know they qualify and/or can’t afford the process necessary to gain benefits.

Identifying patients who are candidates for SSI/Medicaid, but don’t yet qualify, requires a level of sophistication in data analysis beyond data mining. It requires knowledge of “indicators” that point toward eligibility. These “indicators” are used in a “rule based decision making system.” In other words, Ontash utilizes Artificial Intelligence (AI) algorithms, applied to patient medical records to comb through tens of thousands of data variables to produce a manageable pool of prospective SSI/Medicaid candidates, for which a concerted effort will be expended to gain benefit status. The system improves itself over time by analyzing the accuracy of past decisions. The SSI Enrollment process often takes 2 years or more and requires hundreds of man-hours to accomplish. So, QRS depends heavily on the accuracy of Ontash AI applications.

A pilot program for SSI Enrollment, again hosted in the Azure cloud, and utilizing Ontash AI to identify candidate patients and has yielded $1.5 million for the hospital in its first year of operation. This technology will also be used to identify subsets of a patient population for intervention in order to improve healthcare and reduce costs.

Ontash Data Analytics In The Works

On a very granular level: Ontash is currently developing software tools that monitor a hospital’s patient population to predict which individuals are likely to cycle through hospital services at a high rate. These predictions may be used to intervene early, so treatment may be administered to break the cycle – improving healthcare and saving costs.

On a macro level: Ontash is developing software tools to forecast surges in patient demand, based on historical activity, weather, holidays and other factors that drive patient admissions. Using this kind of predictive modeling, hospitals will be able to optimize staffing for ER, rehab populations, special surgical procedures and even for day-to-day nursing rosters.

Health IT Services

Ontash uses its extensive experience (over 10 years) in providing Medicare reimbursement services to also provide core IT services to the Healthcare Industry.