The Current Paradigm in Data Analysis to Identify Potential Fraud, Waste and Abuse in Medicare Part D HCCA Puerto Rico Regional Annual Conference - May 1, 2014 Olgamarién Castillo Aguiar, J.D. www.abarcahealth.com Anti Fraud, Waste and Abuse Program 2 Agenda 1 2 3 Objectives Medicare Part D Anti- Fraud Activities: Current Framework Data Analysis Models for Prescription Fraud 4 5 6 Examples of Risk Assessment Measures for Pharmacies Recommendations Data Analysis Toolbox 1
3 Objectives Understand the importance of data analysis as an integral part of the CMS Fraud Prevention Initiative. Participants will learn about publicly available methods to identify questionable billing in Medicare Part D and how to effectively incorporate this information in their monitoring processes and audit activities. Participants will learn how to recognize unusual trends in the billing patterns of pharmacies. 4 Costs of Health Care Fraud The FBI estimates that between 3% and 10% of all health care spending in the U.S. goes toward fraudulent billing. 1 Accordingly to the National Health Care Anti-Fraud Association, anywhere from $70 billion to more than $200 billion per year is lost because of health care fraud. 2 Retail pharmacies each billed Part D an average of nearly $1 million for prescriptions in 2009. Over 2,600 of these pharmacies had questionable billing. 3 Doctor Shopping activities of controlled substances resulted in about $148 million in Medicare Part D payments. 4 An average of $1.4 billions are spent in Medicare Part D in Puerto Rico each year. In this scenario, a conservative guessestimate of 3% for improper payments represents $42 million. Sources: 1 http://www.fbi.gov/stats-services/publications/financial-crimes-report-2010-2011/financial-crimes-report-2010-2011#health 2 http://www.nhcaa.org/docs/nhcaa/pdfs/member%20services/whitepaper_oct10.pdf 3 Office of Inspector General (HHS-OIG) and Department of Justice s (DOJ) U.S. Attorneys Offices (USAO) data. OEI-02-09-00600 4 United States Government Accountability Office. GAO-11-699 2
5 Potential Fraud Risks in Medicare Part D Billing for nonexistent prescriptions Billing for brand-name drugs but dispensing generics (Upcoding) Forging or altering prescriptions Invoice shortings Identity Theft (Prescribers, Beneficiaries) Doctor Shopping Overprescribing Dispensing expired or adulterated prescription drugs Prescribers and Clinics Pharmacies Wholesalers and Distributors Beneficiaries 6 From Potential Risks to Data Analysis Medicare Potential Risks Data Analysis Recoveries and Improper Payments 3
7 What is Data Analysis? A tool for identifying coverage and payment errors, and other indicators of potential FWA and noncompliance Questions: who, what, where, when, why, and how. Data analysis to transform questions into queries Query results are responses that help us identify trends, patterns or potential errors Indicators of potential fraudulent claims include: excessive billing amounts, higher per-patient costs, excessive patients per doctor, and higher per-patient prescriptions. 8 Compliance and Fraud Detection Fraud Management Life Cycle Mitigation Detection Prevention For detection purposes, fraud can be defined as an unexpected or rare event with significant financial impact. Added value for the Compliance program. Compliance must be the SME on the models of potential fraud and schemes detection. The technology expert s job is to figure out how to implement them. Let the SME s of Business Intelligence resolve the hurdles of data mining, statistics, models and algorithms. Some challenges include: Finding a needle in a haystack. Organizational barriers to convince those who do not normally work with data to integrate data analytics into their work. 4
9 CMS Anti-Fraud Initiatives Center for Program Integrity (CPI), established in 2010 to function as CMS' focal point for all national and State-wide Medicare and Medicaid programs and CHIP integrity fraud and abuse issues. Shift beyond pay-and-chase approach to a proactive model. Data driven fraud prevention. Anomaly detection and predictive analysis. Two Pillars approach: improper claims detection and provider screening. On June 2011, CMS implemented a predictive analytic technology called the Fraud Prevention System (FPS)to identify the highest risk claims in Medicare FFS for fraud, waste and abuse. Early Efforts on Data Analysis Medicare Part D Medicare FFS 10 Medicare Part D Anti-Fraud Activities: Current Framework Structure Activities Medicare Program Integrity Group Division of Plan Oversight and Accountability (DPOA) for Medicare Parts C & D National Part D Outreach & Benefit Recovery Education Integrity Audit MEDIC MEDIC Contractor Medicare Part D activities Data Analysis to Address Opioid Overutilization and Questionable Prescribing Patterns Monitoring Prescribers and Pharmacies Pharmacy Risk Assessments Sharing data for fraud detection Affordable Care Act requires the centralization of certain claims data from CMS Watch out for One Program Integrity (One PI System) 5
Regulatory Guidance for Fraud Detection in Medicare Part D 11 Retail Pharmacies with Questionable Part D Billing (HHS-OIG, 2012) Medicare Part D: Instances of Questionable Access to Prescription Drugs (GAO, 2011) Chapter 9- Part D Program to Control Fraud, Waste and Abuse (CMS, 2006) Schedule II Drugs: Inappropriate Medicare Part D Payments for Schedule II Drugs Billed as Refills (HHS-OIG, 2012) MEDIC Benefit Integrity Activities in Medicare Parts C & D (HHS-OIG, 2013) Prescribers With Questionable Patterns in Medicare Part D (HHS-OIG, 2013) Updated Version Chapter 9- Compliance Program Guidelines (CMS, 2013) 12 Compliance Program Requirements Element VI: Effective System for Routine Monitoring, Auditing and Identification of Compliance Risks 50.6.9 Use of Data Analysis for Fraud, Waste and Abuse Prevention and Detection Establish benchmarks to recognize unusual trends, changes in drug utilization over time, physician referral or prescription patterns Analyze claims data to identify potential errors, and provider billing practices and services that pose the greatest risk for potential FWA to the Medicare program Identify items or services that are being overutilized Identify problem areas at pharmacies, pharmacists, physicians, and other health care providers and suppliers 6
13 Data Analysis Models for Prescription Fraud Rules Based Detection Anomaly Detection Model Predictive Analysis Social Network Analysis Screen claims for known types of errors and improper payments Duplicates Sex/Age Drug Interactions Sanctioned Providers Deceased Plan Members/Prescri bers Prescriber Identifiers Identify claims outliers by comparing an individual pharmacy s behavior patterns during a period and against aggregated patterns of other pharmacies (peer to peer) Pharmacy Risk Assessment Based on past known cases Claim may become suspicious only when factors are considered as a whole; where independently, those factors may not be suspicious. Too Far Distance to identify beneficiaries who live too far away from their pharmacy and prescriber. Built on providers that are associated with identified linkages among potentially fraudulent subjects. Sanctions or Excluded Individuals Potential Issues for Inclusion in the Compliance Monitoring and Auditing Plan 14 Top Case Referrals and Investigations of Potential Part D Fraud and Abuse Associated With the NBI MEDIC s, 2010-2011 Attempts to steal beneficiary identity to obtain prescriptions, 471 Diverting prescriptions, 412 Doctor shopping or stockpiling, 352 Overcharging beneficiary for prescriptions, 283 Forged or altered prescriptions or other documents, 537 Inappropriate Billing, 631 Inappropriate Prescribing, 653 Source: Office of Inspector General (HHS-OIG) and Department of Justice s (DOJ) U.S. Attorneys Offices (USAO) data. OEI-03-11-00310 N=4,029 7
15 Examples of Risk Assessment Measures for Pharmacies Some of the industry measures applied to describe the billing patterns of pharmacies include: 1. Average amount paid per beneficiary 2. Average number of prescriptions per beneficiary 3. Percentage of prescriptions for abuse drugs 4. Percentage of prescriptions that were for brandname drugs 5. Percentage of prescriptions that were refills 6. Percentage of prescriptions that were for drugs not dispensed as written 7. Percentage of prescriptions that were for compound drugs Source: Office of Inspector General (HHS-OIG) OEI-02-09-00600. 16 Recommendations Partner with your PBM for FWA detection and investigations. PBM s possess subject matter expertise of pharmacy billing and contractual relationships with the pharmacies/providers billing prescription claims. Start with the known industry fraud indicators and data analysis, but be creative. There s always opportunities to improve! Analysis of the HPMS Complaint Tracking Module for trends and patterns of potential fraud and abuse. Analysis of potential fraud center on controlled drugs. Is important to look at other drugs with potential for diversion or prone to overbilling errors. Documentation is critical. 8
17 Data Analysis Toolbox Public Data Resources OIG Website: www.oig.hhs.gov Key Reports and Publications Retail Pharmacies with Questionable Part D Billing; OEI-02-09-00600 Basic Reports and Analysis Drug Utilization Reports Controlled drugs utilization by member High percentage of certain drugs/classes billed by pharmacy Commonly Diverted Drugs- Non Controlled 18 Questions Olgamarién Castillo Aguiar, J.D. Fraud, Waste, & Abuse Manager Abarca Health LLC (787) 523-1297 olgamarien.castillo@abarcahealth.com 9