Skip to content
Quick Start for:

Medicaid Vendor Drug Program Pilot Study

The Medicaid Vendor Drug study was performed as a pilot study to develop a sampling and review methodology for a reliable and comparable performance measurement of the Medicaid Vendor Drug Program (VDP) prescription overpayments. The Comptroller’s office looked at the VDP program in both the December 1998 and January 2001 studies but did not conduct a study for an error rate due to time and data factors.

The remainder of this section describes the sampling technique, review methodologies and study findings for the Medicaid Vendor Drug Program pilot study.

Medicaid Vendor Drug Pilot Sample Selection
The Medicaid VDP pilot study sample was a simple random selection of 225 prescription claims from a universe of paid claims from October 1, 2001 through December 31, 2001. This time period differs from the Medicaid Fee-for-Service (FFS) to capture claims at the end of the calendar year to compare to a full calendar year of medical health care claim history. Since this is the first overpayment measurement study, this timeframe will serve as the baseline for future studies.

The sample was selected by National Heritage Insurance Company (NHIC) using special programming written by their system staff. This special programming required more time than originally expected causing a delay in running and receiving the final sample for this pilot study. Also, several test selections were necessary to obtain a sample size that statistically represented the universe of paid prescription claims. The estimated run time was February/March 2002, however, the final sample was run in April 2002. The NHIC sampling program for this study is in Appendix B.2. The sample was selected from the Texas Medicaid Vendor Drug Program’s (VDP) paid history file housed in the NHIC’s data warehouse system and validated by the State Auditor’s Office with statistical measurement tests, such as the Chi-squared test and the Z-test of the mean of the total amount paid.

Following the random selection of the VDP sample claims, NHIC identified all paid prescription claims for each sample recipient within the sample time frame. The total number of sample prescription claims for this study was 1,070 with a total paid amount of $46,936.76.

Once the sample was selected from the VDP history file, NHIC system staff cross-reference the recipient and prescribing provider data to the Medicaid claim processing Compass 21 system to extract demographic information not carried in the VDP paid claim history file. This step allowed NHIC staff to send the recipient notification letters and to conduct the Medicaid online medical services review. Additionally, the project team manually identified the pharmacy provider’s addresses as this data is not stored in the VDP paid history file. Consequently this sample was not available to conduct the study reviews until May 2002.

The following information was collected on each recipient and prescription claim:

1. Recipient Medicaid number

2. Recipient number assigned by the Texas Department of Human Services (TDHS) in their System for Application, Verification, Eligibility, Reports, and Referrals (SAVERR)

3. Recipient Social Security Number (SSN)

4. Recipient name

5. Recipient date of birth

6. Recipient address

7. Guardian name (if applicable)

8. Prescription number

9. Date prescription written

10. Date prescription claim paid

11. Pharmacy provider number – National Council of Prescription Drug Programs (NCPDP)

12. Brand name of drug (based on National Drug Code Directory)

13. National Drug Code

14. Refill number

15. Drug cost amount (claim billed amount)

16. Prescription paid amount

17. Basis of cost determination (Cost Basis Code)

18. Basis of cost description

19. Pharmacy provider name

20. Pharmacy address

21. Pharmacy telephone number

22. Prescribing physician name

23. Prescribing physician Medicaid number

24. Prescribing physician address

25. Prescribing physician telephone number.

Denied claims were excluded from this sample.

Medicaid VDP Pilot Study Methodology
The Medicaid FFS review methodology was repeated for this pilot study as illustrated in Table II-1. Each review method is discussed in the sections following the table.

Table II-1: Medicaid VDP Pilot Study Review Methods:

Method: Description:
Recipient Telephone Survey Telephone interviews with the sample recipients.
Online Medicaid Claim History Review A review of all the sample recipientŐs Medicaid medical claims for the calendar year 2001.
Drug Prescription Review A review of each sample prescription.

Recipient Telephone Survey
The project team patterned the telephone interview after the FFS study interview tool with the exception of using only two of the questionnaire sections, the demographic section and an enhanced prescription drug section. The enhancements focused on each prescription to determine if the recipient actually received the prescription from the prescribing physician listed on the pharmacy claim and filled the prescription at the pharmacy billing the claim. Similar to the FFS study, there is a version for an adult recipient and a child recipient.

All individuals administering the interview were experienced interviewers, and each was given specific instructions by the telephone survey contractor for this prescription study before its initiation. The project team emphasized the importance of confidentiality to the interviewers during their orientation, and each interviewer signed a statement of confidentiality. Information sheets were provided to interviewers by the telephone survey contractor when they were scheduled to conduct the interviews, and returned to a supervisor at the close of the interview session. The supervisor monitored and tallied all information sheets and filed them in binders. Interviewers made up to 10 attempts to contact each recipient. The interviewers conducted the interviews in English unless the recipients only communicated in another language. For this study the only other language needed for the recipients was Spanish. Interviewers recorded responses into an electronic database, which made compiling and analyzing the results easier. The English version of the adult recipient prescription interview tool is in Appendix B.3.

Recipients were not required to participate in the interviews and they were allowed to end the interview at any point. If a recipient responded that they could not remember a prescription, the interview was continued and the interviewer skipped to questions for the next sample prescription for the sample recipient.

The interviewers used the recipient’s Social Security Number (SSN) to verify the recipient’s identity. The interview tool provided three opportunities for the recipient (or parent/guardian) to remember receiving a prescription:

1. the recipients were asked if they filled a prescription in the sample month;

2. the recipients were asked if they filled the prescription on the sample date; and

3. the recipients were asked if they filled the prescription at the sample pharmacy on the sample date.

The same two issues experienced in the FFS study also occurred with this pilot study. These issues were that the staff experienced difficulties locating the recipient’s telephone number and too much time elapsed from filling the prescription to conducting the interview, especially since the interviews did not begin until June 2002, six to eight months after a prescription was filled.

The project team took the same actions in this pilot study to identify correct telephone numbers to increase the number of completed telephone interviews.

• TDHS asked their Medicaid caseworkers assigned to the recipients to provide the most current number from their files;

• Texas Health and Human Services Commission mailed informational postcards (Appendix B.4) with a reply form to sampled recipients in May 2002; and

• the project team culled telephone numbers from the pharmacy records received for the sampled recipients from June through August 2002.

Overall these actions did not improve the final number of completed interviews. Out of the 225 recipients, telephone numbers were provided for 156 recipients. Out of these 156 recipients, 79 interviews were successfully completed, only 35 percent of the sample. There were two refusals and 66 recipients had wrong or disconnected numbers. The survey contractor was unable to contact (no answer) nine of the 156 recipients with telephone numbers.

The survey had nine recipients, 6 percent of those contacted, who stated that they did not receive the sample prescriptions billed and paid to the sample pharmacies for a total of 17 prescriptions. This information was referred to the Vendor Drug Program’s Department of Utilization Review (DUR) for further investigation. Forty recipients, 26 percent of those contacted, were unsure or could not remember if they had received the sample prescription.
Table II-2 shows the results of the telephone interviews for the sample. The refusals listed refer to recipients who declined to conduct an interview.

Table II-2: Results of the VDP Recipient Telephone Survey

Interview Results Totals
Completed 79
Deceased 0
Refusals 2
Disconnected/wrong numbers 66
Unable to contact/no answer 9
Totals for Telephone Numbers 156
No phone/no phone number 69
Totals for Sample 225

The project team halted phone calls in August 2002 in order to compile the results of the study.

Online Medicaid Medical Claims History Review
An online review was performed using the NHIC Compass 21 claim processing system to locate the VDP sample recipients’ related medical claims. Using the prescription filled dates, review nurses researched all medical claims billed for calendar year 2001 for each sample recipient. This review was performed concurrently with the drug prescription review.

Three objectives were sought by the project team from this review:

  1. to determine if the sample recipient had visited a physician or outpatient clinic (including emergency room) at or near the date of the sample prescription;
  2. to compare the recipient’s diagnoses from these medical claims with the drug’s actions and indications for prescribing them; and
  3. to identify any follow-up visits following the prescription sample date.

Besides identifying related medical claims, the initial intent for this review was to identify claims that the team would request medical records from the prescribing physicians and perform a medical necessity review. This action was not completed due to lack of project time. This did not impact this study because of all the prescriptions with related medical claims only one case was found to have prescription drugs totally unrelated to the sample recipients prescribed medications. This case involved a one-year-old child with prescription drugs to lower blood cholesterol. This prescription may be a pharmacy billing/entry error. This error can be avoided by using a system edit to compare recipients’ ages to the medication and suspend payment if the drug is not appropriate for the recipient’s age. The prescription claim information was referred to the VDP Utilization Review department for further investigation.

Medicaid VDP Drug Prescription Review
The third review method in this pilot study was a review of a hard copy of the original prescription written by the prescribing physician. HHSC staff requested copies of the original prescriptions from the pharmacies through letters similar to the one used in the Medicaid FFS study. Once the pharmacy’s documents were received, professional nurses reviewed the prescriptions using specific prescription review guidelines and designated potential overpayments. In addition to the nurse review, a Vendor Drug Program auditor reviewed all the sample cases identified with errors. Each step in this process is discussed below.

Medicaid VDP Prescription Request Letter
HHSC and the project team adapted the Medicaid FFS request letter to request sample prescription copies by changing the description of documents needed in the body of the letter and changing the request attachment to list all the prescriptions from a specific pharmacy provider with the appropriate recipient information. HHSC requested these prescription copies from the sample pharmacy providers.

Due to the sample delay, only two letters were sent to the pharmacies, the initial letter and a second request letter. Since these pharmacy providers did not receive a third request letter, they were not penalized for not complying with the request with either a discrepancy error designation or a recoupment of paid monies as done in the FFS study. A copy of the initial request letter is in Appendix B.5.

Prescription Drug Review Criteria
Professional nurses reviewed all the prescription records received to confirm that they met both the federal and state prescription requirements. These prescription requirements were included in the VDP Discrepancy Error Code Table (Appendix B.6) for the nurses to designate overpayments.

Texas State Board of Pharmacy requires the following information on each prescription: [13]

• patient’s name and address;

• name, strength and quantity of the drug to be dispensed;

• directions for use;

• the intended use of the drug, if appropriate;

• the name, address and telephone number of the physician;

• the name, address, telephone number and identification number of the advanced practice nurse or physician assistant completing the prescription drug order;

• the date; and

• the number of refills permitted.

Additionally, prescription drug orders must be manually signed by the physician in the same manner that they would sign a check or legal document. Currently, electronically produced or rubber stamped signatures can not be used.[14] Prescriptions for controlled substances (narcotics, sedatives, etc.) must be dated “as of” and signed on the day issued and have the physician’s Drug Enforcement Agency (DEA) registration number.[15]

Each prescription form must have the practitioner’s name, address, telephone number and required identification number. The form must also contain two signature lines, side by side, at the bottom of the form with the following words under each line
respectively: “product selection permitted” and “dispense as written.”[16] The term “product selection permitted” allows the pharmacist to use a generic brand drug to fill the prescription.

Besides the prescription documentation requirements, other discrepancies that counted as overpayments were:

• no documentation (original hard copy) of the written, fax or telephone drug prescription in the pharmacy records received;

• medication prescribed not the same as the drug given to the recipient;

• prescription medication contraindicated with the recipient’s other known medications;

• refilling prescription orders for long term/rehabilitation residential recipients without confirmation from the prescribing provider;

• prescription claim for a drug not appropriate for the patient’s age; and

• prescriptions with no corresponding medical visit for the recipient, excluding Medicare and MCO recipients.

Similar to the FFS study, the Discrepancy Error Code Table has an action plan recommendation for the HHSC/OIE staff.

The reviewers used standardized review worksheets (Appendix B.7) to record the review findings. The reviewers used the Nursing 2001 Drug Handbook[17] and specific drug manufacturer’s Internet sites as resources to identify the drug actions, indications, dosage and clinical uses. The information from the worksheets was entered into a database to compile the results for the overpayment calculation.


[13] Texas State Board of Pharmacy, Texas Pharmacy Laws and Regulations Code §291.34 Records (b), (Austin, Texas, August 31, 2000), p. 107.

[14] Texas State Board of Pharmacy, Texas Pharmacy Laws and Regulations Code §291.34 Records (2), (Austin, Texas, August 31, 2000), p. 127.

[15] United State Department of Justice, Drug Enforcement Administration, “Pharmacist’s Manual – An Information Outline of the Controlled Substances Act of 1970,” (Austin, Texas, January 31, 2000), p. 396.

[16] Texas State Board of Pharmacy, Texas Pharmacy Laws and Regulations, Texas Pharmacy Act, Chapter 563, Subchapter A, “Prescription Requirements for Practitioners,” Sections 563.001 and 563.002, (Austin, Texas, January 31, 2000), p. 40.

[17] Naina Chohan, Rita M. Doyle and Patricia Nale, Nursing 2001 Drug Handbook (Pennsylvania: Springhouse Corporation, 2001).