Using Analytics to detect and prevent Healthcare fraud Copyright 2010 SAS Institute Inc. All rights reserved.
Agenda Introductions International Fraud Trends Overview of the use of Analytics in Healthcare 2 2
Introductions Chris McAuley, Director, Security & Intelligence Practice Chris.McAuley@SAS.com +44 7747 100189 (m) 3
Introductions 4
Copyright 2011, SAS Institute Inc. All rights reserved. Copyright 2010 SAS Institute Inc. All rights reserved. 5
Industry Data Points Copyright 2010 SAS Institute Inc. All rights reserved.
EXTERNAL VIEWPOINT COST OF HEALTH CARE FRAUD GLOBALLY 7 Bloomberg Businessweek Research Services 2011
Perspective on Cost of Health Care Fraud EHFCN Newsletter, March April 2010 http://www.ehfcn.org/newsletter/2010/q1-2/articles Estimated global dollars associated with health care fraud ( 160 / 180 / $260 billion each year) is enough to: Provide clean, safe water around the globe Bring malaria under control in Africa Provide the Diphtheria, Tetanus and Pertussis vaccine to all 23.5 million children under one years old who are currently not immunized (2.5 million die each year from diseases preventable by vaccines) AND quadruple the budget of the World Health Organisation and UNICEF (the United Nations Children s Fund) with more than 100 billion left over enough to build more than 1,000 new hospitals at developed world prices 8
Scope of the Problem Europe Country In Bn Germany 13,016 France 10,576 UK 8,554 Italy 7,021 Spain 4,328 Netherlands 2,687 Belgium 1,664 Sweden 1,527 Austria 1,394 Estonia 1,261 Greece 1,078 Poland 900 Portugal 839 Finland 722 Ireland 709 Hungary 398 Romania 235 Slovakia 168 Bulgaria 97 Lithuania 79 Latvia 57 Cyprus 48 Equates to approximately 5.5% of healthcare spend across Europe Czech Republic NA 9
The SAS Fraud & Financial Crimes Platform Copyright 2010 SAS Institute Inc. All rights reserved.
Trend in Health Care Fraud Management The current SIU standard Pay and Chase Claim Receipt & Data Integration Adjudication Integration / Claim Edits Adjudication Processing Claim Payment Fraud Detection & Alert Generation Alert Triage & Case Management Moving analytics and fraud detection upstream in the claims lifecycle to become proactive, versus pay and chase Step 1: Pre-payment Fraud Detection Claim Receipt & Data Integration Adjudication Integration / Claim Edits Adjudication Processing Fraud Detection & Alert Generation Alert Triage & Case Management Claim Payment 12
Solution Components Alert / Case Management Real-Time Event Processing Near Real Time Event Processing Compliance Advanced Hybrid Data Analytics Environment Data Integration + Data Quality + Data Enrichment 13
Enterprise Healthcare Fraud Environment Business Analytics Framework SAS Fraud Framework Healthcare Solutions (sample) Healthcare Claims Fraud Provider Fraud Pharmacy Fraud Waste Control Disease Mgmt. Prevention Detection Detection & Alert Generation Social Network Analysis Alert Management Case Management Data Quality & Integration Analytics Business Intelligence & Reporting 14 14
Common Types of Provider Fraud, Waste, Abuse Overutilization Upcoding False Claims Unbundling Billing for Non-Covered Treatments Fraudulent Dates of Service Waiver of Co-pay Free Medical Service Kickbacks Phantom Providers Misrepresenting Medical Records Billing for services not medically warranted, to receive insurance payments, or falsifying diagnosis to justify medically unnecessary procedures. Using a code for a more expensive treatment than what was performed. Billing for services not performed or supplies not provided. Improper submission of separate claims for services that should be combined under a global fee. Billing for non-covered treatment as though they were covered treatment (e.g. experimental not covered by insurance plan). Falsifying the date to avoid contract limitations on eligibility or payment maximums. Waiving coinsurance or deductible to accept insurance as payment in full, and then inflating charges to insurer. Free service to patient, then billed to insurer, to entice ongoing other treatments. Providers receiving cash payments in exchange for driving business to certain ancillary providers (e.g. labs). Unlicensed providers posing as physicians. Falsifying the medical records to justify services that were not provided or not warranted. 15
Spectrum of Analytical Techniques Past Future Mission Improvement / OUTCOMES Business Focus & Decision Support Value Operations & Administration / OUTPUTS Alerts Optimization Predictive Modeling Forecasting/extrapolation Statistical analysis Query/drill down Ad hoc reports Standard Reports What s the best that can happen? What will happen next? What if these trends continue? Why is this happening? What is happening? What exactly is the problem? How many, how often, where? What happened? Predictive Analytics Descriptive Analytics Now What? So What? What? Reactive Response Type Source: Adapted from Competing on Analytics: The New Science of Winning (Davenport / Harris) Proactive 16
Hybrid Fraud Detection Approach Rules Set-up rules to filter fraudulent transactions Anomaly Detection Detect individual and aggregate abnormal patterns SAS Fraud Framework Advanced Analytics Perform knowledge discovery, data mining, predictive assessment Social Network Analysis Perform knowledge discovery through associate linkage analysis Text Mining Unlock the power of unstructured data within reports and staff notes 17
Advanced Analytics are Required Using a Hybrid Approach for Fraud, Waste & Abuse Detection Enterprise Data Suitable for known patterns Suitable for unknown patterns Suitable for complex patterns Suitable for associative link patterns Providers Members Rules Anomaly Detection Predictive Models Social Network Analysis Facilities Claims Rules to filter fraudulent claims and behaviors Examples: Detect individual and aggregated abnormal patterns vs. peer groups Examples: Predictive assessment against known fraud cases Examples: Knowledge discovery through associative link analysis Examples: Referrals Fraud Flags Financials 3 rd Party Data CPT upcoding / correct coding Value of charges for procedure exceeds threshold Daily provider billing exceeds possible Ratio of $ / procedure exceed norm # procedures / provider exceeds norm # patients from outside surrounding area exceeds norm Like upcoding behavior as known fraud provider Predicted diagnosis does not match actual Like provider/network growth rate (velocity) Provider association to known fraud Linked members with like suspicious behaviors Suspicious referrals to linked providers Hybrid Approach Proactively applies combination of all 4 approaches at member, provider, facility, and network levels 18
SAS HIGH- PERFORMANCE ANALYTICS END-TO-END CAPABILITIES 19
Analytic Engine Analytic Approach: Unsupervised Methods Use when no target exists Examine current behavior to identify outliers and abnormal transactions that are somewhat different from ordinary transactions Include univariate and multivariate outlier detection techniques, such as peer group comparison, clustering, trend analysis, etc Avg. Number of PCS Services Submitted Provider is not only an outlier, also shows extreme variation for average number of services submitted per attending provider 22
Analytic Engine Analytic Approach: Supervised Methods Use when a known target (fraud) is available Use historical behavioral information of known fraud to identify suspicious behaviors similar to previous fraud patterns Include parametric and nonparametric predictive models, such as generalized linear model, tree, neural networks, etc Fraud Scores Incomes # of previous investigations Predicted Fraud Scores 23
Analytic Engine SAS Social Network Analysis Network scoring Rule and analytic-based Analytic measures of association help users know where to look in network Net-CHAID for local area of interest (node) in the network Density, Beta-Index (network) Risk ranking with hypergeometric distribution, degree, closeness, betweenness, eigenvector, clustering coefficients (node) Modularity (sub-network) 24
How do we help Investigations? Copyright 2010 SAS Institute Inc. All rights reserved.
Why use analytics for Healthcare? Provides the ability to apply Rules, Predictive Models, and Anomaly Detection on linked data More prioritized Fraud, Waste, & Abuse cases identified Including both previously undetected entities and networks and extensions to already identified cases Reduction in false positive rates Hybrid approach reduces false positives by up to 10+ times over traditional rules-based approaches Improved analyst / investigation efficiency Each alert takes 1/2 1/3 of the time to investigate due to data aggregation and visualization Provides alert logic and suggested path to initiate investigation Significant increase in ROI per analyst / investigator 26
What products are we going to discuss? Alert / Case Management Designed to improve investigator productivity and reduce claims investigation elapse times Real-Time Event Processing Intraday Event Processing Compliance Designed to identify criminal gangs, internal fraud and fraudulent Claims Data Integration 27
How are SFF and ECM related? SAS Fraud Framework (a.k.a. Fraud Network Analytics ) Enterprise Case Management Financial Crimes Monitor Social Network Analysis Alert Triage Cases Case Management & Reporting Investigation Results 28 28
Triage Dedicated triage team (could even be a separate unit from the SIU) will research the lead and look at all data points Plan of Action is done (details the steps investigator should follow on each claim and what documentation should be reviewed) Triage team will typically look at the first few claims that come in on that case to determine if case warrants opening After trial period, open cases are turned over to SIU for mainstream investigation activity Then, SIU investigator owns the case and keeps plan of action up to date. 29
Typical Steps in the Investigation Varies by Case Request documentation Medical records Invoices Second surgical report Operative notes Interviews (patients or provider) Surveillance and outside consultant reviews Referral to law enforcement Final letter showing reason code for outcome or request for recovery never use the word fraud unless prosecution has occurred. Prepare for prosecution (if applicable) 30
SAS Visual Analytics explorer 31
SAS Visual Analytics Explorer Company Confidential - For Internal Use Only Copyright 2012, SAS Institute Inc. All rights reserved.
SAS Visual Analytics Explorer Company Confidential - For Internal Use Only Copyright 2012, SAS Institute Inc. All rights reserved.
SAS Visual Analytics Explorer Company Confidential - For Internal Use Only Copyright 2012, SAS Institute Inc. All rights reserved.
SAS Visual Analytics Explorer Company Confidential - For Internal Use Only Copyright 2012, SAS Institute Inc. All rights reserved.
SAS Visual Analytics Explorer Company Confidential - For Internal Use Only Copyright 2012, SAS Institute Inc. All rights reserved.
SAS Visual Analytics Explorer Company Confidential - For Internal Use Only Copyright 2012, SAS Institute Inc. All rights reserved.
Copyright 2010 SAS Institute Inc. All rights reserved.