Skip to content
Quick Start for:

Section I. Medicaid Fee-for-Service (FFS) Study

Overview of the Texas Medicaid Program
Medicaid Fraud and Abuse Detection Activities
Medicaid FFS Study
Medicaid FFS Study Recommendations

Medicaid Fee-for-Service Study

Purpose

The purpose of this study is to measure the incidence of potential overpayments, which could include occurrences of fraud and abuse in the Texas Medicaid program. Section 403.028 of the Texas Government Code requires the Comptroller's office to perform this study biennially in consultation with the State Auditor's Office.

An overpayment is a payment for a service that is not in accordance with the policies of the Medicaid program.

The Medicaid definitions of fraud and abuse are: [39]

  • fraud is an intentional deception or misrepresentation made by a person with the knowledge that the deception could result in some unauthorized benefit to him or herself or some other person. It includes any act that constitutes fraud under applicable federal or state law; and
  • abuse includes provider practices that are inconsistent with sound fiscal, business or medical practices and result in an unnecessary cost to the Medicaid program, or in reimbursement for services that are not medically necessary or that fail to meet professionally recognized standards for health care. It also includes beneficiary practices that result in unnecessary costs to the Medicaid program.

Past studies indicate that overpayments are usually not due to fraud, but to abuse through incomplete documentation of a service, inappropriate coding and clerical
errors.

Most health care service providers strive to provide the best medical care possible for Medicaid patients. Also, providers identified with overpaid claims frequently did not intentionally set out to abuse or commit fraud in the Medicaid program. The few providers who commit fraud or abuse for financial gain are receiving money that could and should be available to the Medicaid program to expand benefits, serve more uninsured Texans and improve reimbursements to compliant providers.

This study, through a statistically structured sampling and review methodology, helps identify specific service areas, and in some cases providers, with a high incidence of overpayments.

This section describes the sampling technique, review methodologies and study findings for the Medicaid FFS study.

Texas Medicaid Fee-for-Service (FFS) Sample

Sample Stratified Categories

The study uses stratified health care categories, such as home health and physicians, that consist of services performed by providers with similar practice types and specialties designated within the Medicaid claim system. Medicaid assigns the provider type a numerical value to represent the area of service performed, such as laboratory services or medical supplies. The provider specialty numerical value represents more specific information about the provider, such as an individual or group practice, the number of beds in a hospital or a specific medical specialty like pediatrics. Appendix A.8 lists the health care categories with the provider types and specialties for each category.

Table I-2 lists the number of services for the study categories:

Table I-2: March 2005 Study Sample Services
Category Number of Clients Number of Services
Ancillary/Outpatient 100 313
Dental 100 488
Home Health 100 143
Hospitals 100 432
Mental Health 100 139
Other 100 214
Physicians 100 258
Supplies/Durable Medical Equipment 100 215
Total 800 2,202

Sample Selection

The Medicaid FFS study used the random sampling methodology created for the January 2001 study and used in the March 2003 study. The sample selection was performed using special programming written by National Heritage Insurance Company (NHIC) system staff. (See Appendix A-9 for the detailed summary of the selection program.) The sample consists of 800 patient sample days randomly selected from detailed dates of service for the specified categories of medical health care services. Hospital services without detailed services, such as those paid by Diagnosis Related Grouping (DRG), the patient sample days were selected from the date ranges of the admission date through the discharge date. All the sample days were selected from paid claims for services provided from September 1 to November 30, 2002.

DRGs consist of a given set of diseases, like diabetes, and disorders, such as a bone fractures, compiled into clinically cohesive groups that use similar hospital resources and length-of-stay patterns. Each DRG represents the average resources needed to treat patients grouped to that DRG relative to the national average of resources used to treat all Medicare patients. [40] The Texas Medicaid program uses the DRG payment system to reimburse all enrolled hospitals except children's hospitals and some psychiatric facilities.

A patient sample day was used as a starting point to provide the study with a core date to sample all health care services provided to a unique Medicaid client within each stratified category. Using this core date, the sampling program identified all paid services billed for the sample client on the sample day, regardless of the provider type. These services were labeled "associated sample services." Both the core sample service and the associated sample services were tagged with a study identification number built from the core sample. For example, the first client core sample service in the Physician category was assigned a study identification value of PH1 and the associated services were assigned study ID values of PH1A1, PH1A2, and so on.

Since providers bill their services on claims, the sample services were also tracked by the claim number assigned for the sample services within the Medicaid claim processing system. This claim number was required for all research on the system and in communicating with the providers. It was also used to compare some of the sample data.

In the Texas Medicaid program, providers may bill multiple services on a single claim performed on the same day, or a range of multiple services performed daily or weekly. This study reviewed only the core sample service and all associated sample services on a claim with multiple services.

The following information was collected on each core and associated service from the claim processing system:

  • Client Medicaid number
  • Client number assigned by the Texas Department of Human Services in their System for Application, Verification, Eligibility, Reports, and Referrals
  • Client Social Security Number
  • Client name
  • Client date of birth
  • Client address
  • Claim number
  • Service detail number (location on claim as billed to NHIC)
  • Service detail date of service (sample day)
  • Program designation code for core sample service or associated sample service
  • Current Procedural Terminology code for professional services or Diagnosis Related
  • Grouping code for inpatient services
  • Procedure code or DRG code description
  • Diagnosis code
  • Diagnosis code description
  • Provider Medicaid number
  • Provider name
  • Provider type and specialty for MMIS
  • Provider address
  • Provider telephone number
  • Dollar amount paid for sample service

Since the associated services were in some instances performed by providers that belonged to one of the other study categories, the selected sample was resorted by the provider type and specialty before the final overpayment rate was calculated. For example, following the re-sorting, all services performed by physicians were grouped together, regardless of the category in which they were selected and assigned in the primary sample. The individual services had to be grouped by appropriate health care service areas to identify the sources of potential overpayments within a specified health care category.

Table I-3 shows the distribution by number of sample services before and after resorting the sample by the provider type and specialty for each category:

Table I-3: Medicaid FFS Study Distribution of Sample Services Before and After Provider Re-sort
Sample Categories Number of Sample Services
Before Provider Type
and Specialty Re-sort
Number of Sample Services
After Provider Type
and Specialty Re-sort
Ancillary/Outpatient 313 252
Dental 488 490
Home Health 143 113
Hospitals 432 405
Mental Health 139 132
Other 214 115
Physician 258 479
Supplies/Durable Medical Equipment 215 216
Total 2,202 2,202

Sample Exclusions

Before selecting the sample, some eligible client groups and provider types were excluded from the sample of paid claims. Medicaid clients receiving services exclusively from HMOs and the Long Term Care program were excluded.

The Texas Commission for Alcohol and Drug Abuse clinics were excluded because HHSC performs a 100 percent review on all the services they provide. The other provider exclusions were due to the client eligible groups they served, such as the military hospitals and providers who bill only for Medicare clients. Also, providers under current investigation by the Texas OIG or Attorney General's Office were excluded.

Medicaid FFS Methodology

Only two review methods were used in the 2005 study, which differs from the previous studies: contextual data analysis and medical record reviews. The client telephone interview method was deleted from the FFS methodology after the results from the January 2001 and March 2003 studies were deemed insufficient to measure program overpayments. The tools used in each of these methods are included in Appendices A.10 through A.13. The sampling, medical record requests and study reviews were performed prior to HHSC's reorganization. The OIE performed these duties along with the now-defunct Texas Department of Health. Following the reorganization, the OIG performed the interagency reviews and will be responsible for follow-up activities. The main duties of the PAM and future PERM projects are the responsibility of the OES.

Contextual Data Analysis Review Method

Research analysts conducted contextual data analyses, also called online reviews, of all services on paid claims from the sample clients using the Medicaid claim processing system. They reviewed the services in each client's claim in relation to other claim services paid from two months before to one month after the sample day. In some
samples, all claims billed by a provider for a specific procedure were reviewed to determine if a pattern of potential abuse existed.

The objective of this analysis was to note:

  • any unusual provider billing or client utilization trends that might not be observed when reviewing only the sample claim's services;
  • any services that did not appear related to the diagnoses or procedure billed on the claim; and
  • any services billed and paid that were not in accordance with the policies or procedures stated in the 2002 Texas Medicaid Provider Procedure Manual or the 2002 Texas Medicaid Bulletins.

The 2002 version was used because all the sample services occurred in 2002 and the providers were required to follow the policy and procedures in this manual and any Medicaid Bulletin received prior to the sample dates of service.

Analysts used a number of reference materials to ensure consistency during this review. The main sources were:

  • 2002 Texas Medicaid Providers Procedure Manual;
  • 2002 Texas Medicaid Bulletins;
  • American Medical Associations' Physician Cumulative Procedure Manual;
  • International Classification of Diseases, 9th Version;
  • Healthcare Common Procedure Coding System Manual;
  • Lab book; and
  • Medicaid Policy Interpretation (online procedure manual).

Each analyst signed a confidentiality statement before beginning the review. Analysts conducted the medical record reviews at the Comptroller's office or at HHSC's Office of Inspector General, and the team kept the review documents in locked cabinets.

Medicaid FFS Medical Record Review Method

The second method used in this study was a review of medical records pertaining to the sample services. This review consisted of sending letters to providers to request medical records and reviewing the records using specific guidelines. Each step in this process is discussed below.

Medicaid FFS Medical Record Request Letter

To obtain the medical records, HHSC's OIE department sent letters to providers. The letter informed the provider of HHSC's authority to obtain the medical records as well as the provider's responsibility to deliver the medical records to OIE under federal and state regulations and the provider's Medicaid enrollment agreement with TDH. OIE used a three-phase process to send these letters.

OIE staff sent an initial certified letter to each provider with an attachment identifying the client by name and date of birth, the sample date of service, the client's Medicaid identification number, the social security number and the number of the claim as it appeared on their Medicaid payment statement. If there were multiple claims for different clients for the same provider, OIE listed all of the claims on one attachment. OIE also included an affidavit form for the provider to affirm that the record sent was the complete documentation for the client and date of service requested. An example of the letter with the request sheet and affidavit are in Appendix A-10.

If the provider failed to return the appropriate documentation, OIE contacted the provider to verify the address, explaining what had been requested and sent a second certified letter. If there was no response, OIE sent a third and final certified letter, followed by a final telephone call to explain the consequences of noncompliance.

All three of the letters had specified deadlines ranging from seven to 14 days, a list of the information requested and an explanation of the consequences for noncompliance. According to OIE, non-compliance is total recoupment of the claim because Medicaid regulations require providers to comply with HHSC's requests for medical records.

Medical Record Review Criteria

Professional nurses reviewed all the medical records received to verify that clients received the sample services. The reviewers used standardized worksheets that included the required documentation guidelines implemented by the Medicaid program after the last study. A quality review was performed by nurse consultant on all the record reviews to ensure the consistency and accuracy of the error assignments.

In addition to the general required documentation guidelines, the Medicaid program also implemented program-specific documentation guidelines for private duty nurses and services provided by psychiatrists and counselors, including psychologists, licensed master social worker-advanced clinical practitioners and licensed professional counselors. A copy of all these guidelines is in Exhibits 1 through 4 of Appendix A.11.

After the project team reviews, professional staff from HHSC/OIG reviewed the medical records to verify the error assignments. A physician or dentist reviewed all records with questions about the necessity of medical or dental care. All the reviewers used standardized review worksheets (Appendix A.12.) The project team entered data from the worksheets into a database.

The project team classified the sample services into distinct types of potential errors based on the policies and procedures in effect for the sample dates of service, including the general documentation guidelines. These errors are listed in the Medicaid Medical Record Review Discrepancy Error Table (Appendix A.13). Each type of error has a specific plan of action recommendation for the HHSC/OIG staff, including what money to count as potential overpayments. These error discrepancies are based on
basic principles required of all providers billing claims to the Medicaid program for reimbursement.

All services must be:

  • documented as required in the provider's manual;
  • billed for the date that the services occurred, for the client that received them and by the provider performing them; and
  • billed as stipulated by the policy or procedure guidelines published by the Texas Medicaid program.

Medicaid FFS Medical Record Review Findings

The medical record review continues to be the best source for identifying potential overpayments. Of the 2,202 sample services reviewed, 390 had errors. This was 12 percent fewer than the 442 overpayments out of 2,122 sample services from the March 2003 study.

Before finalizing the results of the medical record review, the professional utilization review nursing staff at HHSC OIG reviewed the medical records with potential overpayment errors. These nurses have experience in all the health care areas related to the study's category areas. Both the medical reviewers and the interagency reviewers used the 2002 Texas Medicaid policies and procedures in place when the sample services were performed. The OIG staff was given an opportunity to agree or disagree with the project review nurses' findings. A total of 22 sample services were reviewed by a dentist for appropriate coding and billing compliance and 45 sample services were reviewed by a physician for medical necessity and potential fraud or abuse.

Types of Errors

The main types of errors found in the study's medical record review were: documentation, clerical, unbundling of laboratory services, upcoding and potential fraud or abuse. Unbundling is billing services separately that are required to be billed a single item. Upcoding is billing a procedure code for a higher level of service than was provided to a patient.

Table I-4 illustrates the distribution of these discrepancy errors by category for all the sample services in the 2005 study. The incomplete documentation and clerical/billing error numbers are shaded because these errors are addressed with educational letters and are not counted as potential overpayments.

Table I-4: Type of Medical Record Review Errors by Study Categories
  STUDY CATEGORIES  
TYPE OF ERRORS Ancillary/
Outpatient
Dental Home
Health
Hospitals Mental
Health
Other Physicians Supplies/
DME
TOTAL
ERRORS
No Medical Record
Received
6 0 4 7 37 4 25 7 90
No Document for the Sample Day Received 2 0 6 26 1 5 10 16 66
No Document of the
Sample Service
7 17 2 3 22 6 62 20 139
Incomplete Document (Multiple Items) 117 141 38 65 123 68 280 33 865*
Clerical/Billing Errors 15 2 1 5 4 2 6 8 43*
Lab/ Procedure
Unbundled
26 0 0 0 0 0 0 0 26
Upcoding 1 5 2 4 2 4 19 0 37
Potential Fraud or Abuse 0 0 4 3 4 4 10 3 28
Claim System
Process Error
0 0 0 0 0 1 0 0 1
Total Potential
Overpayments
42 22 18 43 66 24 126 46 387
*Not counted as overpayments.

Documentation Errors

The documentation errors are separated into four types based on the problems found:

  • medical records not sent after receipt of certified request letters;
  • medical records received without any documentation for the sample date of service requested;
  • medical records received for the sample date of service requested without documentation of one of the services billed and paid; and
  • medical records received for the sample date of service with incomplete information based on the Medicaid requirements, such as the client's name.

As in the January 2001 and March 2003 studies, some providers disregarded the request for complete records. Some of the providers sent records unrelated to the dates requested and a lot of providers sent records with multiple omissions, such as client names, allergy information or their own signatures.

Services for which no medical record documentation was sent for review from providers were all considered errors. This error occurred in 90 of the sample services provided. This was 37 percent fewer than the March 2003 results of 144. Two hundred and five of the sample services were not documented in the medical records received by the providers. Sixty-six of these services were from providers whose medical records did not contain documentation for the sample date of the service; the other 139 services were from medical records for the sample date of service that did not have documentation of the services billed on the claim. This is a decrease from 219 in the March 2003 study.

The last documentation overpayment error (no documentation of service in a medical record for the sample date-of-service) is of particular concern because it implies some providers are billing for what are sometimes called "phantom" services that were never provided. The concern is that the only way to identify this abuse is to request medical records on all suspected providers, which is costly and labor intensive. From this random selection, it is difficult to determine if these phantom service discrepancies are due to a provider's poor coding and billing practices or if they were intentional. If the problem is due to unintentional coding or billing errors, there are ways to help providers avoid overpayment errors. For example, some providers use assistants to transcribe visit or service information as the provider performs the service. Assistants can also transcribe the information using software. The state can only identify intentional billing of phantom services through detailed investigative techniques. HHSC's OIG staff will investigate each provider and pursue the providers who appear to bill for phantom services consistently.

Nearly the same amount of errors made due to a lack of documentation were found in 2005 as in 2003. As in the 2003 study, several providers notified the HHSC/OES staff that they had made an error in billing the service and would send an adjustment immediately. Several providers sent refund checks. These examples count as missing documentation errors.

General guidelines for all providers list 10 documentation items required for all services provided to Medicaid clients, such as legible handwriting, a full date and allergy information. Some of the items are not applicable for all provider types. For example, independent laboratories do not usually schedule follow-up visits. One of the items, the client's Medicaid number, may not be available for all providers to document at the time the service is performed, such as in the case of a newborn who may not get a number for a few weeks depending on the mother's promptness in notifying a caseworker of the birth. This item was the most common omission found in the medical records. Medical records frequently had multiple omissions per document. For example, many of the records lacked the client's Medicaid number, allergy information and the provider's signature or initials. These findings were the same as noted in the March 2003 study.

In most cases, these discrepancies were not counted as overpayments because the sample service was confirmed as provided to the client. The HHSC/OIG staff will send educational letters to these providers. If critical items, such as the client's name or the full date, were omitted, preventing confirmation of the service, the project team designated the service with either the "no documentation for the sample date" or "no documentation of the service for the sample date" error. If the medical record lacked documentation of a laboratory or x-ray result, this omission counted as an overpayment
error.

Medicaid program guidelines state that omitting some of the program-specific documentation guidelines can result in recoupment of the amount paid for those paid services. The guidelines also state that providers must submit items critical to determining if the service billed was actually provided. For example, psychological services documentation requirements specify a narrative description of the counseling session along with behavioral observations and descriptions of the assessment, treatment plan and any recommendations made. Fifty-six of the 130 mental health services in the sample lacked these items, and the team counted them as overpayments. Fifty-four of these overpayments were found in the March 2003 study. Table I-5 provides a detailed breakdown of the documentation discrepancy findings.

Table I-5: Summary of Incomplete Documentation Discrepancy Errors
Study Categories Illegible
Handwriting
Full Date
of Service
Provider
Signature
(Initials Accepted)
Client Name Client Medicaid
ID Number
Allergy
Information
Presenting
Complaint
Follow-up Visit Unresolved
Problems
Missing Lab/X-ray Results Private Duty Nursing Counseling/
Psychiatric
Totals
Ancillary/Outpatient 3 0 3 1 84 19 0 6 1 0 0 0 117
Dental 0 0 36 9 63 32 0 1 0 0 0 0 141
Home Health 0 0 0 0 20 8 0 0 0 0 10 0 38
Hospitals 0 0 1 0 44 6 0 0 14 0 0 0 65
Mental Health 0 0 3 0 69 0 0 0 0 0 0 49 121
Other 0 0 12 0 46 9 0 1 0 0 0 0 68
Physicians 7 1 22 1 189 54 0 2 1 0 0 3 280
Supplies/Durable Medical Equipment 0 1 8 0 20 0 0 0 0 0 0 4 33
Totals 10 2 85 11 535 128 0 10 16 0 10 56 863

Clerical

The team found clerical or unintentional billing errors on 43 sample services. The most common examples were when a provider billed for a medical supply on an accounting posting date rather than the actual invoice or delivery date. None of these errors were counted toward the overpayment calculation because no adjustments were necessary to the claim payment. HHSC/OIG staff members will send these providers educational letters addressing the discrepancies. Twice as many of these errors were found in 2005 as in 2003.

Laboratory Tests/Procedure Unbundling

The survey identified 26 samples that were unbundled, meaning the provider billed them separately when Medicaid requires them to be billed as a single item. These services were lab tests billed individually instead of in panels as directed by the 2002 Texas Medicaid Provider Procedures Manual. The HHSC/OIG staff will adjust these errors and let the providers know how to bill them properly.

Upcoding

In 37 instances, providers coded office visits for a more complex level of care than documented in the medical records. This is also referred to as "upcoding" and is considered abusive because it is primarily done to maximize the reimbursement to the provider. The majority of these occurrences (19) were for physician office visits billed for codes that indicate the physician spent 25 to 40 minutes with the patient performing a detailed comprehensive examination and complex medical decision. The reviewers found these codes used by the physicians for visits with established patients with recurring medical diagnoses and minimal documentation on the level of care provided. These complex office visit codes were also used by physicians who identified medical problems during a routine Texas Health Steps (THSteps) check-up. Billing the additional office visit is appropriate for the Medicaid program, however, the detailed examination and comprehensive history is included in the THSteps check-up reimbursement of $75. This upcoding accounts for $7 to $30 in overpayments for each occurrence. A professional physician performed a medical necessity evaluation on all these records prior to confirming them as errors.

Also found were bills from dentists for a full mouth debridement, a procedure that involves the removal of plaque and calculus below the gum level that is obstructive and prevents a dentist from performing an oral evaluation, when the actual procedure performed on each documentation was a routine prophylactic cleaning. This billing practice costs an overpayment of $47 for each occurrence. These services were reviewed by a dentist prior to designating them as errors. The number of this type of error found in 2005 was slightly lower than in March 2003.

Potential Fraud or Abuse

The reviewers noted potentially fraudulent or abusive billings with 28 sample services billed by and paid to 19 providers. Some examples were:

  • a physical therapist billed physical therapy for a phone conversation with a client to see how "he was feeling about (his) electric wheel chair at this time";
  • a professional counselor billed for individual (face-to-face) counseling of a patient that documents showed was a phone call while the patient was in a hospital;
  • an optometrist billed for a vision exam and test on a two-month-old baby with no documentation of an eye or vision problem; and
  • two instances of providers billing for services performed by other providers.

HHSC/OIG will investigate all of these services with potential recoupment or sanctions.

System Processing Error

The reviewers found one system processing error. The 2002 Texas Medicaid Provider Procedures Manual states that a refraction test (an eye exam for glasses) on a patient with dual Medicare/Medicaid benefits is only payable when the client has aphakia (disease or injury to an eye such as a missing lens) or an ocular disease. The only diagnosis billed on the claim was presbyopia, (unable to focus on close objects) which does not meet the payment criteria. Research of this procedure on the claim payment system indicates it is set-up to pay incorrectly. HHSC/OIG staff members will notify TMHP to correct the system's payment for this procedure code.

Which Provider Types Made The Most Errors?

The project team re-sorted the sample by provider type and specialties to place the appropriate providers into the correct category of service. For example, following the provider re-sorting, all services performed by laboratory and x-ray facilities were grouped together in the Ancillary/Outpatient category, regardless of the category in which they were selected and assigned in the primary sample. Grouping the individual services by appropriate health care service areas per the provider type and specialty allowed the team to identify the amount of potential overpayments.

Table I-6 illustrates which categories of services had the greatest number of errors in the study.

Table I-6: Number of Services with Errors by Rank per Re-sorted Provider Types in Each Stratified Category
  Documentation Errors    
Study Categories by Rank No Medical Record No Document for
Sample Date
No Document for
Sample Serviced
Incomplete Document* Clerical/Billing Lab/Procedure
Unbundling
Upcoding Potential Fraud or Abuse System Process Error Total Rank
Physicians 25 10 62 280 6 0 19 10 0 126 1
Mental Health 37 1 22 123 4 0 2 4 0 66 2
Supply/ DME 7 16 20 33 8 0 0 3 0 46 3
Hospital 7 26 3 65 5 0 4 3 0 43 4
Ancillary/ Outpatient 6 2 7 117 15 26 1 0 0 42 5
Other 4 5 6 68 2 0 4 4 1 24 6
Dental 0 0 17 141 2 0 5 0 0 22 7
Home Health 4 6 2 38 1 0 2 4 0 18 8
Totals 94 65 139 865 43 26 37 28 1 387  
* Documentation errors not counted as overpayment errors.

The majority of the errors, 126 (32%) of 390 found in this study, were for sample services billed in the physician category. Most of these errors were due to lack of documentation of the sample services on the sample date. Specifically, physicians who did not send any record after three request letters, who did not include documentation for any of the services on the sample date or who did not include documentation on the sample date-of-service for specific services billed. Fifty percent (62 out of 126 errors) of the lack of documentation errors were for services billed that were not provided. The most significant example of this error was physicians who billed for incomplete well-child check-ups, known as Texas Health Steps exams. This occurred in 42 percent of the 62 services billed, with no documentation that the exams were provided. The 2002 Provider Procedures Manual lists seven services that must be provided and documented to bill for this well-child check-up procedure code. These services are health history, unclothed physical exam, appropriate immunizations, specific laboratory tests, health education, age-appropriate vision and hearing tests, and referral for dental exams after 12 months of age. The manual includes a detailed schedule that identifies the components of each service required for the child's age to assist the provider. These comprehensive exams are an important opportunity for the provider to identity significant health problems that might go undetected until a child is seriously ill.

Upcoding errors were identified in 19 (15%) of the 126 errors. This error is when a physician bills procedure codes beyond the level of care documented in the medical record. Many of the physicians' records lacked signatures and patients' allergy information.

The second most common error was the mental health category (psychologists, licensed master social worker-advanced clinical practitioners and licensed professional counselors). Most of the errors found in this group were for providers who did not send the requested records and billing for counseling services without the required documentation. Documentation omissions included the session times and narrative descriptions of the session or behavioral assessments. Two providers billed for family counseling when the documentation indicated only the client was present.

Providing invoices signed by the client continues to be the main source of errors in the supply/durable medical equipment category. The 2002 Texas Medicaid Provider Procedures Manual requires these providers to maintain delivery documentation that proves the client received the supply or equipment. This documentation must be a signed invoice or delivery slip or a dated carrier tracking document with the shipping date attached to the invoice. [41]

The majority of errors in the hospital category were from a lack of documentation. Sixty percent (26 out of 43 errors) of the records with errors from the hospital providers did not have documentation of the sample date of service requested. Four records included diagnoses that were not treated during the inpatient stay. Three records were identified with potential fraud or abuse. One hospital billed for a hip x-ray on a patient treated for an allergic skin disorder. Two providers performed lab tests deemed medically unnecessary for the patient's medical condition by the reviewing physician.

The most common error in the ancillary/outpatient category was the unbundling of lab services. This occurs when a provider bills separately for services that are part of a group of services with one billing code. Two independent laboratory providers billed 26 blood tests individually. They should have been part of an obstetric lab panel. The other major error in this category was a lack of documentation confirming that the services were provided. One provider billed a higher outpatient surgical procedure code than the procedure documented in the medical record.

The "other" category consists of health care professionals such as chiropractors, physical and occupational therapists, nurses, and speech therapists. The most common error made within this category was a lack of documentation for a sample service. Of the 11 records lacking documentation of the sample service, eight were for health services such as speech therapy provided in public schools. Four of the services were coded at higher levels of care than was actually provided. For example, an optician billed for
bifocals when the documentation shows single vision lens were provided. The study found potential fraud or abuse with four sample services.

Of the dental errors found, 77 percent (17 of 22) were for instances when providers billed for services that were not provided. The other five errors were for services billed at a higher level of care than that provided. These providers billed for detailed scaling and cleaning when only routine teeth cleanings were performed. This is a significant error that has been identified in all past studies. The March 2003 study identified an annual risk of $332,601 in erroneous payments due to this billing practice.

Home health providers had the fewest errors, with 18 out of the total of 390. The most common error was providers who sent records without documentation of the sample date or service. Home health providers had the fewest number of errors in both the January 2001 and March 2003 studies. Four of the 18 errors found were potential fraud or abuse. One provider billed for therapeutic services that another provider had also billed and was paid to provide. The HHSC OIG staff will investigate to determine which provider actually performed the services. Another provider billed for twice the amount of time for a service than the medical record documented.

Potential Overpayment Measurement

In order to calculate the amount of overpayments, the cost of each type of overpayment must be determined.

In this study, the amount for each overpayment was the entire dollar amount of the service, an adjusted dollar amount or zero. For example, if the service was classified as an error because the service was not documented in the medical record, the entire dollar amount of the service was counted as an error. However, if a service was documented in the medical records but was not coded correctly, as in upcoding errors, the amount of the error would be the difference in the amount paid and the amount that should have been paid. The dollar amount of the error was zero if the service was documented in the medical record and the error was an incorrect date or some other clerical error.

Of $187,708 paid for medical services, the study found that $25,869 were potential overpayments.

The State Auditor's Office (SAO) performed the statistical calculation of the overpayment percentage measurement. Below is a simplified sequential summary of SAO's calculation of the percentage of potential overpayments using the dental category as an example.

The Total Sample Paid Dollars for all the sample services in the health care categories was summed:

Total Sample Paid Dollars = $187,708

  • The Total Universe Paid Dollars for all the paid claims in the sample universe was summed:

Total Universe Paid Dollars = $135,214,047

  • The Sample Proportion of Dollars was calculated for each health care category by dividing the number of paid dollars in a category by the Total Sample Paid Dollars:

Dental Category Sample Proportion of Dollars Example:
$16,617
$187,708
= 8.9 percent
The Universe Proportion of Dollars for each health care category was calculated by dividing the number of dollars in a category by the Total Universe Paid Dollars:

Dental Category Universe Proportion of Dollars Example:
$1,500,975
$135,214,04
= 1.1 percent
Each sample category was then weighted by the ratio of the Universe Proportion Dollars divided by the Sample Proportion Dollars:

Dental Category Weighted Proportion Example:
1.1
8.9
= 0.13 percent
The Total Overpaid Dollars was calculated and summed for all eight study categories:

Total Overpaid Dollars = $25,870
An Error Rate was calculated for each category by dividing the number of category dollars overpaid by the category's sample paid dollars:

Dental Category Universe Error Rate Example:
$880.24
$16,617
= 5.30 percent
The weighted Error Rate was calculated by multiplying the Error Rate and the Weighted Proportion:

Dental Category Weighted Error Rate Example:
5.30 X 0.13 = 0.66%

*Definitions:
Universe: is the three month study period of September 1, 2002 through November 30, 2002. The Universe dollars do not include the dollars for the provider types excluded from the study.
Weighted: a process to indicate the relative importance of each quantity's contribution to the average. This process was used in the study to indicate each health care category's representation of the provider types in the study universe of paid claims.

See Table I-7 below for a summary of the calculations and the tables in Appendix A.14 for all the calculations. Appendix A.15 is a detailed description of the steps SAO used the overpayment rate.

This series of calculations determines the weighted average overpayment measurement of 13.71 percent. The margin of error is plus or minus 0.2 percent.

Table I-7: FFS Summary Data for the Overpayment Calculation
Study Categories Total Paid for
Sample Services
Total Paid
for Claims
Proportion of
Category to Total
Total Overpaid
for Errors
Overpayment
Error Rate
Weighted
Overpayment
Error Rate
Ancillary/Outpatient $11,912 $6,813,566 0.05 $1,678 14.09% 11.19%
Dental $16,617 $1,500,975 0.13 $880 5.30% 0.66%
Home Health $20,259 $41,933,496 2.87 $2,857 14.10% 40.53%
Hospitals $71,250 $7,345,140 0.14 $6,660 9.35% 1.34%
Mental Health $18,750 $3,470,667 0.26 $2,992 15.96% 4.10%
Other $4,928 $15,517,545 4.37 $728 14.78% 64.61%
Physicians $21,754 $30,956,011 1.98 $6,625 30.45% 60.16%
Supplies/DME $22,238 $27,676,646 1.73 $3,312 14.90% 25.73%
Totals $187,708 $135,214,046 1.00 $25,732* 13.71% 13.71%
*Due to rounding this total is off by $1.90.

Dollars at Risk

The overpayment measurement can be applied to the annual Medicaid expenditures to determine the amount that could be overpaid, or the "dollars at risk" in the Medicaid program. The Comptroller's office uses the term "dollars at risk" because it would be impossible, given the size of the Texas Medicaid program, to identify and recover all overpayments unless each questionable service was identified through a complete medical record review of all services submitted for payment. However, HHSC can use the findings in this study to target specific areas of investigation as they did in both the January 2001 and March 2003 studies.

This study identified $544 million dollars at risk. This figure was computed by multiplying the overpayment measurement rate of 13.71 percent by the total Medicaid FFS expenditures for 2002-$3.9 billion.



ENDNOTES

[39] U.S. Department of Health and Human Services, Health Care Financing Administration, Guidelines for Addressing Fraud and Abuse in Medicaid Managed Care: A Product of the National Medicaid Fraud & Abuse Initiative (August 2000), p. 12, http://cms.hhs.gov/states/letters/smd80700.asp. (Last visited January 14, 2005).

[40] St. Anthony's Publishing, DRG Guidebook (Reston, Virginia: St. Anthony's Publishing, Inc., 2001), p. Intro-3.

[41] Texas Department of Health, National Heritage Insurance Company, 2002 Texas Medicaid Provider Procedures Manual (Austin, Texas), pp. 23-13.