Contract Performance Report
Medicaid Fraud and Abuse Detection System

FY2002, Second Quarter
March 1, 2002 - May 31, 2002

printable document - adobe acrobat/pdf format)
download required viewer

TABLE OF CONTENTS:

1. INTRODUCTION
2. SCOPE OF WORK
3. EVENTS
4. OPEN (PENDING) ITEMS
5. PERFORMANCE MEASURES
6. RISK MITIGATION

I. INTRODUCTION

As of December 30, 1997, the Texas Health and Human Services Commission (HHSC) has used neural network and learning technology to detect fraud, abuse, or waste in the Texas Medicaid program that may otherwise go unknown.

After HHSC's receipt of state and federal approval of the project, HHSC competitively acquired development and operational support of an automated Medicaid Fraud and Abuse Detection System (MFADS). The MFADS contract was awarded to Electronic Data Systems (EDS) Corporation. At this time, HHSC is into Option 2 of the base contract, which allowed for an extension through August 31, 2002.

The competitive procurement of a new MFADS contract, to be effective no later than May 1, 2002, and run through August 31, 2007 is complete. The contract has been awarded to EDS and was finalized on May 1, 2002.

This report covers MFADS' activities for the third quarter of state fiscal year 2002 and activities associated with the transition to the new MFADS contract. The MFADS has proven an effective tool to detect fraud, abuse, or waste not identified by other systems or processes. The MFADS has also proven to be an effective research tool for investigators and staff within the Office of Investigations and Enforcement (OIE) at HHSC and the Office of the Attorney General (OAG).

HHSC has seen a great increase in the productivity of its investigators as the result of the MFADS research tools. Investigators are now able to run queries specific to their cases and receive returns within a few minutes, or hours for very large queries. Before the implementation of MFADS, investigators had to request a report from the claims administrator contractors. Depending on urgency, workload, and staffing factors, the report would take anywhere from one to two weeks.

Another added benefit of the MFADS tool involves the ability to present massive amounts of information in a clear, graphic format, such as graphs and charts that condense provider's history and compare the provider's patterns and performance with his/her true peers. These charts and graphs have proven very valuable in settlement discussions, administrative hearings, and in preparing cases for referral to the Office of the Attorney General.

back to top

II. SCOPE OF WORK

The day-to-day operations of the MFADS include:

1. Vendor loads and refreshes data from current sources on a quarterly basis except when instructed by HHSC to load data on a monthly basis. Monthly loads are done within the limits of contracted project resources dedicated to the agreed upon Scope of Work.
2. Current targeted detection queries run on a schedule provided by HHSC.
3. Vendor conducts Business Objects™ user training on an as needed basis, not to exceed one session per quarter. Subject matter experts and vendor staff develop no less than 8, and no more than 10, targeted detection queries per option year. As directed by HHSC, ad hoc services and reports may replace queries.
4. Existing targeted detection queries are revised and updated as needed and as directed by HHSC.
5. Vendor is required to develop and update the master MFADS work plan that addresses scheduled work activities under the contract.
6. Spyder Analysis Webstation™ configuration, installation, and appropriate training is given by vendor staff as needed to accommodate new staff.
7. Vendor provides ongoing support for model results to include ad-hoc analysis, conference calls, e-mail communications, and limited on-site visits by staff from HNC Software, San Diego, California, the subcontractor for the MFADS project.
8. Yearly update of existing models (includes execution on most recent data).
9. Site visits, not to exceed 12 trips per year, by HNC personnel.
10. One of the following per year:
    • Development of one new model
    • Expand current generic MD model (type 20) to include DO (type 19) providers, to include:
      • Model development
      • Installation into Spyder Webstation™
      • Ongoing analytical support
      • Final report

Effective May 1, 2000, HHSC and EDS entered into a contractual agreement to amend the original contract to include the following enhanced services and deliverables. The enhanced services and deliverables are listed below, including status of completion of each item.

Data Retention - Hardware Pricing - Complete

A fourth year of source data was added to the MFADS platform.

Licensing Board Data - Complete

HHSC-OIE staff has made initial contact with the Texas State Board of Medical Examiners, the State Board of Nurse Examiners, and the State Board of Examiners of Psychologists staff to discuss process for receiving data from these boards for inclusion into the MFADS platform.

Based on meetings with these entities, as well as the cost and availability of data, a decision was made to add only the Board of Medical Examiners licensing file data to the MFADS platform.

HHSC-OIE staff has approved the fields to be loaded and tracked in history.

The Texas State Board of Medical Examiners data including the history was deployed into BusinessObjects™ on December 14, 2001.

Modeling - Complete

The dental, physician and pharmacy models for Year 2000 data have been delivered.

The Year 2001 models will be run against 6 months of data to allow results to be deployed in February 2002. This shortened time period allows HHSC-OIE staff time to review and work the results prior to the end of the contract in August 2002.

The physician, dental and pharmacy model for the first half of 2001 were released on February 25, 2002.

Utilization Review - In Progress

The technical and management reviews of the Claims Management System (CMS), the Prior Authorization System (PAS), and the Service Authorization System (SAS) at the Texas Department of Human Services.

HHSC-OIE has entered into a service agreement with DHS to implement automation changes needed and has submitted requests for the data to the source agency, DHS.

Hospital review software has been updated for identified Compass21 changes. Testing is in progress.

Weekly JAD sessions to document requirements for the hospital utilization review component began February 21, 2002.

Monthly Data Refreshment - Complete

This requirement has been implemented.

Medical Transportation - In Process

MFADS EDS staff are currently reviewing test files that have been received. It was decided that all relevant information would be loaded, as the volume is small.

HHSC staff is working with Medical Transportation staff to resolve outstanding data related questions.

Eligibility - Complete

Fully implemented during the third quarter of FY2001.

Enhanced Case Management

Phase I - Complete

All requested changes to the case management system, except for the RIMS auto-posting feature, have been implemented. Open cases previously tracked on the MPI Access database were cleansed and converted into Spyder™ Case Management during September and October 2000.

Phase II - On Hold

The RIMS auto-posting feature is ready to be implemented and tested by the MFADS EDS staff. However, this has been placed on hold until Compass 21 RIMS issues could be resolved.

Effective May 1, 2002, EDS was awarded the new contract for MFADS. The contract allows for a transition period of May 1 - August 31, 2002. Transition related activities and status are reported below.

Contract

The contract was signed and became effective May 1, 2002.

New Platform

• Delivery of the Hardware was accomplished on May 24, 2002
• Installation - currently underway
• Convert Source Data
• Load Source Data
• Convert Case Management Data
• Load Case Management Date
• Convert TQ Result Sets
• Convert Model Result Sets
• System Application Installation - Spyder Application WebStation
• System Application Installation - BusinessObjects
• Vendor Testing
• State User Acceptance Testing

Additional BusinessObjects™ Licenses

Order has been submitted for nine (9) additional licenses.

Documentation

• System and Network Configurations
• Preliminary Documentation -
  • Spyder WebStation - delivered on May 31, 2002
  • Case Management - delivered on May 31, 2002
  • BusinessObjects - currently available via on-line help
  • System Documentation
• Final Documentation -
  • Sypder WebStation
  • Case Management
  • BusinessObjects
  • System Documentation
• Training Documentation -
  • Spyder WebStation
  • Case Management
  • BusinessObjects

Training

• Vendor Staff Training - Ethics
• State User Training
  • Spyder Webstation
  • Case Management
  • BusinessObjects Beginner
  • BusinessObjects Advanced
  • Business Miner

back to top

III.   EVENTS

System Activities

• Made modifications to resolve closed cases with open activity segments as a one time clean-up of case management
• Corrected provider license information in Spyder and BusinessObjects
• A-Claims fix file loaded - fix file received and included in the monthly load of March data
• Non-Provider bubble help updated and the new Non-Provider file loaded
• Load of February, March and April 2002 data was completed
• Purge of February, March and April 1998 data was completed

Query/Ad Hoc Activities

• Completed ad hoc for MPI - multiple dental services for same day, same client
• TQs 1, 2, 3, 5, 7, 10, 12, 16, 18, 22, 31, 32, 35 and 42 were executed against new claims data and deployed in Spyder and BusinessObjects
• New dental related query 52 was developed, validated and released into production

Model Activities

• Deployed final runs of Physician and Dental Models for first six months of 2001
• Deployed Pharmacy Model for 2000 and first six months of 2001
• New Model Tracking report developed and deployed in Spyder

Number of cases opened, as well as dollars identified and/or recovered from cases identified by the models are included in the performance measures chart in Section V of this report. As of this reporting period there are 38 open cases that resulted from the various models, as follows:

Case Activities

back to top

IV.   OPEN (PENDING) ITEMS

1. Claims Management System (CMS) - long-term care program - data extract.
   
   Status: Request for data submitted to the DHS, the data source agency. Data request for NHIC CMS is being finalized.
     
2. EDS was requested to develop a plan to archive the purged claims data for historical access by HHSC staff. This data will not reside on the MFADS platform.
   
   Status: CLOSED. The new MFADS contract provides for a larger platform that will build up to six (6) years of data - four years of current data and two years of historical data.
   
3. Claims with "A" status indicating they have been adjusted continue to appear in the MFADS database.
   
   Status: EDS is conducting discussions with the source of the data. Analysis is underway regarding the October 2001 reconciliation file that EDS staff received February 19, 2002. The fix file was received and loaded during the weekend of April 13, 2002 and affected claims were identified. This item is remaining open at this time until it can be fully confirmed that the problem is completely fixed.
   
4. All available provider license information is not being displayed in MFADS.
   
   Status: CLOSED. Problem was identified in the load of the provider license information. The load script has been corrected and implemented during the March 2002 load of the provider data.
   
   

back to top

V.  PERFORMANCE MEASURES

• Performance Criteria

Performance criteria for MFADS were identified in the Request for Offers, the MFADS Project Management Development Plan, and the MFADS Initial Risk Analysis. These performance criteria include measures for the automation system, as well as the business processes. The performance criteria requirements for the automation system were carried forward into the results-based contract with the vendor. The vendor's performance is measured against these criteria. Payment of contract fees is dependent upon the vendor meeting or exceeding these performance measures.

Projected efficiency gains range from 60% of total project cost in fiscal year 1998 to 179% in fiscal year 2002.

There are some factors outside the control of HHSC and/or the vendor that may impact the development and implementation of MFADS and the achievement of some of the performance criteria. These factors have been identified to include:

• Transitions within the Medicaid program from the traditional fee for service structure to a managed care environment.
• Changes in federal and state statutes, regulations and guidelines, as well as court decisions.
• Funding constraints of the Commission's legislative appropriation for OIE and MFADS.
• Full time equivalent staffing caps, staff turnover and attrition difficulty in recruiting qualified investigative staff, and inability to find qualified contracted staff to perform investigative work.

The following chart, which is part of the MFADS Project Management Development Plan, reflects progress on performance measures as of the second quarter of state fiscal year 2001.

Number of investigations initiated against Medicaid providers represents the number of suspects referred for investigation or recoupment action.

Recoveries are reported when the dollars are actually recovered by HHSC-OIE, rather than when dollars are identified or there is an agreement by the provider to refund the program. In the case of duplicate dental claims and the performing provider recovery project, recovery action will be undertaken by the claims administrator and reported to HHSC-OIE on a routine basis.

"Total Identified for Recovery" represents the dollar amount of claims inappropriately paid as identified by target detection queries, ad hoc reports, neural models, or other MFADS tools. It does not represent the actual dollars that may be recoverable.

back to top

VI.  RISK MITIGATION

The Medicaid Fraud and Abuse Detection System (MFADS) is an integral part of reducing Medicaid fraud, abuse, or waste in Texas. Before the award of the MFADS contract, HHSC executed a contract with an independent consulting firm to evaluate the risks associated with the development and implementation of MFADS. The evaluation focused on the management and organization of the project and, besides identifying risks associated with the project, it presented suggestions for risk mitigation.

The fragmentation of the claims processing system across agencies was ranked as a high risk factor, and it has always been one of the most difficult areas within MFADS. The ability to have a single platform from which to conduct fraud, abuse, and waste detection and analysis is one of the key factors of MFADS. Before the implementation of MFADS, the state did not have the ability to conduct an analysis of a provider's participation in multiple programs or across operating agency lines to develop a comprehensive case investigation. MFADS was expected to integrate both historical and current data stored in the various processing systems into a single data repository. MFADS has encountered multiple obstacles in this objective since it depends on input from the operating agencies, which, at times, is not available due to technical problems within the data processing systems of the agencies or projects that delay the receipt of the data.

The independent consultant rated data conversion and cleansing as a very high-risk area. The consultant expressed concerns as to the quality of the data, as well as the vendor's ability to carry out the effort of converting, cleansing, and loading in accordance with the project schedule. The independent consultant identified four major entities, the Texas Department of Health (TDH), the Texas Department of Human Services (DHS), the Texas Department of Mental Health and Mental Retardation, and the Medicaid Claims Administrator that would supply data to MFADS from, at least, six different major automation systems. The consultant felt "a tremendous amount" depended on the data from these systems being available, clean, and accurate.

This concern has proven to be a very valid one. HHSC-OIE and the vendor have encountered numerous problems in this area, and there continue to be concerns specific to various data types including managed care encounter data and long-term care data. Several phases of the project were modified to accommodate the lack of data. Other phases of the project were refocused for the same reason.

Both the primary vendor and the subcontractor have shown willingness to adapt to these circumstances. The creative efforts of the OIE project management team, combined with the vendor's willingness to adapt to external factors outside the control of OIE, have allowed the MFADS project to move forward and achieve significant results.

back to OIE webpage | back to top