Fraud Prevention Data Analytics and other Methodologies Paul Crowder, FICO 1 © 2014 Fair Isaac Corporation. This presentation is provided for the recipient only and cannot be reproduced or shared without Fair Isaac Corporation’s express consent. FICO Snapshot 2 2 Profile The leader in predictive analytics for decision management Founded: 1956 NYSE: FICO Revenues: $743 million (Fiscal 2013) Products and Services Scores and related analytic models Analytic applications for risk management, fraud, marketing Tools for decision management Clients and Markets 5,000+ clients in 80 countries Industry focus: Banking, insurance, retail, health care Recent Rankings #1 in services operations analytics (IDC) #6 in worldwide analytics software (IDC) #7 in Business Intelligence, CPM and Analytic Applications (Gartner) #26 in the FinTech 100 (American Banker) Offices 20+ offices worldwide, HQ in San Rafael, CA 2,200 employees Regional Hubs: San Diego (CA), New York, London, Birmingham (UK), Toronto, Johannesburg, Munich, Madrid, Sao Paulo, Bangalore, Beijing, Singapore © 2014 Fair Isaac Corporation. © 2012 Fair Isaac Corporation. Payment Integrity A Range of Approaches “Crawl” “Walk” “Run” “Fly” Fraud, Waste & Abuse (FWA) are ID’d by chance. FWA are ID’d post-payment with Enterprise BI FWA are ID’d post-payment via predictive analytics. Strategy & process may be formal, or ad hoc. There is a formal PI program of strategy, process & people. FWA are ID’d pre- and postpayment via rules & predictive analytics. A limited program of strategy, process, or people. 3 © 2014 Fair Isaac Corporation. People are otherwise engaged. There is a sophisticated PI program of strategy, process & people. FICO Client Case Study 1 Provider Abuse + Systemic Weakness A single claim line (CPT 99070, Supplies and Materials), approved for US$259 payment, scored high for Procedure Rate (unusually rapid repetition over time). On review, this claim line should have been automatically denied during autoadjudication, as it was for an expense that is considered to be part of Provider overhead expenses. ► The Scheme discovered that it was paying all of their expense procedure codes in conflict with the Scheme’s established policy. ► Further, the Provider is discovered to be acting with intent. ► The US$259 high scoring claim became a US$1 million case with part-time effort by one investigator over a 2 week period … 100:1 ROI. 4 © 2014 Fair Isaac Corporation. The Power of Data-Driven Predictive Analytics Queries/Rules Simple schemes and billing errors Known fraud and abuse patterns Predictive/Data-Driven Analytics Queries/Rules benefits above AND Complex fraud and abuse patterns Undiscovered fraud New and emerging issues Organized Fraud 5 © 2014 Fair Isaac Corporation. FICO Client Case Study 2 Clinically Unnecessary Care A dentistry provider scored high (aberrancy) for 5 reasons, including “High Member Day” ► Peers ► The averaged US$195 per member per day suspect averaged over US$1,700 per member per day ► Findings ►Stainless Steel Crowns were routinely installed on every tooth of every child treated by this provider. ►Clinically unnecessary in every case. ►Multiple provider identities ►3 years of abuse. ► Result ►A US$3 million case ►Imprisonment of the provider 6 © 2014 Fair Isaac Corporation. Preventive Measures ► Credential ► Employ (review) providers before you admit them to your scheme. a strong claims adjudication system. ► Score claims for aberrancy, post adjudication/pre-payment, manually review high scoring (suspicious) claims, & don’t pay the claims that you shouldn’t pay. ► Use integrated “force multiplier” technologies such as decision management software, Link Analysis & Investigational Case Management to rapidly review & decision findings, ID broader suspicious patterns & build prosecution-ready case documentation. ► Use pre- and post-payment findings to strengthen your claims adjudication results, & take action on identified systemic weaknesses and policy gaps. ► Application of (new) preventive measures is a change in process … success depends upon the strength of your relationships with your internal & external peers, customers, members & stakeholders. 7 © 2014 Fair Isaac Corporation. FICO Client Case Study 3 Mathematics are the Universal Language ► The Client: A Dutch Dentistry Scheme ►1.2 million beneficiaries ►€101 million paid Dental claims per year ► The Engagement: A for-fee FICO IFM Analytic Assessment ►FICO scoring of 3,200 Providers using 12 months of paid claims data. ►FICO Delivery of results that are “blind,” due to a language barrier between FICO’s analytic scientists & the client’s data. ► The Results: ►106 (3%) aberrant Providers. In the top 30, ID of all 12 known fraudsters & 12 new suspects. ►€$15 Million Savings from review of 20 high scoring dentists - 14 were found fraudulent (70% Hit Rate). ►250% ROI projected for Year 1 of the predictive analytics solution. ► The 8 Conclusion: Mathematics is the universal language. © 2014 Fair Isaac Corporation. The Recommended Approach for the Medical Schemes ► Rules that target known types of fraud and abuse ► Example: Claim System Edits ► Unsupervised models that score paid claims and providers to detect known, unknown and emerging problems pre- and post-payment ► Vigorous pre- and post-pay workflow for scoring/detection, review and investigation. ► Integrated software that maximizes efficiency and that makes analytic results actionable is key. ► Strong relationships with internal and external peers, customers, members and stakeholders. 9 © 2014 Fair Isaac Corporation. FICO Client Case Study 3 Providers Have Bills to Pay Too ► The Client: A commercial Medical Aid Scheme who scores claims daily, post-adjudication, but reviews claims “quick post-payment.” ► The Initial Claim Scoring Result: A claim line which ordinarily pays at US$250 scores high for High Paid Claim (an unusually high amount paid for the procedure) … at US$25,000 paid. ► The Initial Explanation: “A clerical error.” ► On Further Review: Submittal and payment for the same US$25,000 procedure, one time in each of the previous 2 years. ► The Finding: The provider was manipulating the Scheme’s auto-adjudication system, one time each year, for payment of his child’s university tuition bill. 10 © 2014 Fair Isaac Corporation. The Rise of Auto-Adjudication ► Who ► If processes their claims manually? not now, soon: Nobody ► Advantages of Auto-Adjudication ► For the Providers: Quick Payment ► For the Members: Limited level of involvement ► For the Medical Aid Scheme, payment of: ► The correct claims ► For the correct amount ► For your members ► For providers who are authorized to participate in your plan. ► The Problem with Auto-Adjudication: You will pay claims that, by policy, contract or design, should not be paid. 11 © 2014 Fair Isaac Corporation. FICO Client Case Study 4 Technology: Good, Technology: Bad ► The Client: A commercial Medical Aid Scheme who successfully promoted the use of “Baxter” machines for drug dispensing ► More accurate dispensing for chronic conditions ► Lower dispensing fees for use of automated dispensing ► The Initial Claim Scoring Result: Numerous pharmacy claims scoring high for Duplicate Class, Rate and Excess Days ► The Findings: ► Unnecessary normal dispense on the same day as the first Baxter fill. ► Normal + Baxter dispense to mask excessive dispensing of narcotics ► Excess supplies not prevented with use of Baxter ► Claiming for normal dispense of drugs that should have been (or were) dispensed with Baxter Results: €1.4 million in annual savings via clarification of policy & policing of abusive providers. ► The 12 © 2014 Fair Isaac Corporation. Questions? Confidential. This presentation is provided for the recipient only and cannot be reproduced or shared without Fair Isaac Corporation's express consent. 13 © 2013 Fair Isaac Corporation. THANK YOU Paul Crowder Pre Sales Consulting, FICO paulcrowder@fico.com Confidential. This presentation is provided for the recipient only and cannot be reproduced or shared without Fair Isaac Corporation's express consent. 14 © 2013 Fair Isaac Corporation.