This is the accessible text file for GAO report number GAO-03-402R entitled 'Military Treatment Facilities: Eligibility Follow-up at Wilford Hall Air Force Medical Center' which was released on April 21, 2003. This text file was formatted by the U.S. General Accounting Office (GAO) to be accessible to users with visual impairments, as part of a longer term project to improve GAO products’ accessibility. Every attempt has been made to maintain the structural and data integrity of the original printed product. Accessibility features, such as text descriptions of tables, consecutively numbered footnotes placed at the end of the file, and the text of agency comment letters, are provided but may not exactly duplicate the presentation or format of the printed version. The portable document format (PDF) file is an exact electronic replica of the printed version. We welcome your feedback. Please E-mail your comments regarding the contents or accessibility features of this document to Webmaster@gao.gov. April 4, 2003: The Honorable Dennis J. Kucinich: Ranking Minority Member: Subcommittee on National Security, Emerging Threats: and International Relations: Committee on Government Reform: House of Representatives: The Honorable Edolphus Towns: Ranking Minority Member: Subcommittee on Government Efficiency and Financial Management: Committee on Government Reform: House of Representatives: The Honorable Janice D. Schakowsky: House of Representatives: Subject: Military Treatment Facilities: Eligibility Follow-up at Wilford Hall Air Force Medical Center: In October 2002, we reported to you on the results of our audit of selected internal control activities at three military treatment facilities: Eisenhower Army Medical Center, Augusta, Georgia; Naval Medical Center-Portsmouth, Portsmouth, Virginia; and Wilford Hall Air Force Medical Center, San Antonio, Texas.[Footnote 1] As part of our work for that report, we requested data files of all patients who had been admitted, treated as outpatients, or received pharmaceutical benefits during fiscal year 2001. Despite considerable effort by the three facilities, only Wilford Hall Air Force Medical Center was able to provide a file of beneficiaries who received pharmaceuticals during the year. We compared this file to data in the Social Security Administration’s (SSA) Death Master File as a technique to identify instances of potential fraud or abuse.[Footnote 2] For Wilford Hall, we identified 41 cases in which a prescription was ordered for an individual after the date of his or her death as recorded in the SSA Death Master File. You requested that we determine whether individuals fraudulently obtained pharmaceuticals or other health benefits by assuming the identity of a dead person, and, if so, to identify the specific breakdowns in internal controls that allowed such fraud to occur. As agreed to with your staffs, we confined our investigation to the 41 cases described above. Results in Brief: We did not find indications of individuals fraudulently obtaining health care benefits in our examination of the 41 cases we identified of people receiving treatment after they were listed in SSA’s Death Master File. In 40 of the 41 cases, data entry errors and/or internal control weaknesses at either SSA or at the military treatment facilities created the impression that a deceased person had received prescriptions. Of the 40 cases, * 10 were instances in which SSA’s Death Master File had incorrectly listed as deceased the individual to whom a prescription was dispensed and: * 30 resulted from Department of Defense (DOD) data entry errors. In the 10 cases involving inaccurate SSA death records, most of the individuals concerned found out about the erroneous report of their deaths when they were notified that their SSA benefits had ended. The individuals had their benefits restored, and most did not experience significant problems as a result of the errors; however, some had other problems, including temporary suspension of their retired military payments and difficulty in getting reimbursed for a prescription filled at a retail pharmacy. Inaccurate information in the SSA database has caused DOD to expend resources researching inaccurate death information for living individuals, not only at Wilford Hall, but also for the eligibility system DOD-wide. Thirty of the 40 cases were data entry errors that occurred during the process of entering a prescription into DOD’s health care database. For 14 of these 30 cases, the pharmacy dispensed a prescription to the intended eligible individual but inadvertently recorded the prescription under a deceased person’s Social Security number (SSN). In 10 of these 14 cases, the deceased person was either the spouse or another eligible relative of the individual receiving the prescription. In 2 of the 30 cases, we could not determine who received the prescriptions, but they totaled only 3 prescriptions of small value. In 14 of the 30 cases, the prescriptions were not dispensed. Pharmacy records show that 8 were canceled before they were filled and 6 were never picked up. The 30 cases of data entry errors at DOD were the result of human error as well as the result of DOD not having adequate controls over the data entry process. Specifically, DOD does not have a preventive control in its data entry process that prohibits entering new clinical data such as prescriptions into a deceased person’s record in the DOD automated health care database. When this happens, prescriptions are not entered into the correct individual’s file and the potentially significant patient safety issue of hazardous drug interactions may not be addressed. The remaining case involved an elderly former spouse of a retired service member who continued to receive prescriptions valued at about $350 after she became ineligible when they divorced. We concluded that this situation existed because the retired service member may not have reported the divorce as required by DOD policy. Therefore, DOD’s eligibility system continued to show the former spouse as eligible. The ineligible former spouse told us she was not aware she was ineligible, and that she thought she could continue to get prescriptions until her identification card expired, about 3 years after her divorce. Although this case resulted in inappropriately provided health care benefits, i.e., improper payments, our investigation did not conclude that the payments were fraudulently obtained. Rather, they most likely resulted from a lack of information about eligibility criteria. This letter includes a recommendation to the Secretary of Defense to develop and implement a preventive control for data entry errors involving a deceased person’s clinical record. In a written response to a draft of this report, SSA agreed that the Death Master File has some problems with accuracy and discussed the improvement efforts it has underway. In oral comments, DOD agreed with our findings but disagreed with our recommendation and said that our report overstates the extent of the problem because of the small number of data entry errors. We disagree with DOD. In our work, we focused only on the 41 cases and did not attempt to determine the overall extent of the problem. We believe that a preventive control can effectively avoid the data entry problems we identified and thereby help ensue that patient safety issues are addressed. Background: Wilford Hall is the Air Force’s largest medical facility. It provides a wide range of medical services, including pharmacy-dispensed prescription drugs to active and retired military personnel and their dependents. Wilford Hall reports that it fills approximately 2.6 million prescriptions annually for about 100,000 people. Wilford Hall’s clinical records are contained in DOD’s Composite Health Care System (CHCS). CHCS is DOD’s primary medical information system, which medical treatment facilities use to support their various activities, including registering patients, documenting inpatient activity, and tracking pharmacy prescriptions. Since 1997, DOD has had a project underway to replace CHCS with a new system, CHCS II, that DOD envisions as a state-of-the-art automated medical information system. Part of DOD’s goal for CHCS II is to assist clinicians in making health care decisions.[Footnote 3] In our October 2002 report based on work at three military treatment facilities,[Footnote 4] we reported that erroneous eligibility information contained in DOD information systems, including CHCS, precluded the military treatment facilities from providing reasonable assurance that medical care was provided only to eligible persons. We found that unreliable and inaccurate data, system inadequacies, complicated processes, and a lack of adherence to policies and procedures contributed to internal control weaknesses. Military treatment facilities such as Wilford Hall are required to verify a person’s eligibility for DOD health care benefits before providing treatment, except in emergencies. The facilities use a two- step process to verify eligibility. One step is for a staff person to physically review the person’s military identification card, which includes a picture of the person, and visually verify the identity of the person requesting health care. The military identification card is issued at over 900 DOD locations and is used DOD-wide to access a variety of DOD services in addition to health care. Sponsors--the military active duty persons or retirees upon whom their dependents’ eligibility is based--are responsible for reporting any changes in status for themselves and their dependents. The other step is for the facility’s staff to access the person’s clinical record in CHCS, which verifies the person’s eligibility status by interfacing with the Defense Enrollment Eligibility Reporting System (DEERS). DEERS is a DOD-wide system that contains eligibility information on active, reserve, and retired military and their dependents. It is used by DOD facilities such as commissaries and base exchanges as well as military treatment facilities to determine eligibility for various types of DOD benefits. DEERS regularly receives updated data from SSA regarding deaths reported to it. SSA’s Death Master File is the agency’s repository of death information and is available for use by both public and private sector organizations. The Death Master File is a national file listing the SSNs of individuals whose deaths have been reported to SSA. Data sources include friends and relatives of deceased individuals, funeral directors, financial institutions, postal authorities, and other federal and state agencies. Scope and Methodology: To determine if any ineligible persons were using the identity of a deceased person to obtain health care benefits, we compared a data file from Wilford Hall Medical Center of patients who had received a prescription to data from SSA’s Death Master File. The patient data file was extracted by Wilford Hall staff from CHCS and identified about 100,000 individuals in Wilford Hall’s database who had a pharmacy prescription during fiscal year 2001. These files included prescriptions recorded at Brooke Army Medical Center and Randolph Air Force Base Clinic as well as Wilford Hall because the facilities share computer services for health care matters. As of April 2002, the Death Master File contained about 70 million records of persons with SSNs who, according to SSA, have been reported as deceased. We first matched only on SSN and identified 266 matches. However, most matched only on SSN but not on other critical data such as name and date of birth. Because the military treatment facilities’ eligibility verification process is to match both the sponsor’s SSN and the patient’s name, we selected for further analysis and investigation only the 41 cases in which the SSN matched in both files and other identifying information, such as the same name and date of birth, raised questions about how the deceased person in the SSA database could have received care after his or her reported death. For all 41 people, we also obtained from Wilford Hall a list of prescriptions ordered after the date of death recorded in the SSA Death Master File. We also obtained eligibility information from DOD’s automated eligibility systems. To obtain an explanation of the facts of each case and to identify indications of fraud, our investigators reviewed other records such as death certificates and divorce decrees as needed. For the 10 cases of inaccurate reports of death, our investigators interviewed patients, family members, and others, as needed. We conducted our work from November 2002 though January 2003 in accordance with U.S. generally accepted government auditing standards, and we performed our investigative work in accordance with standards prescribed by the President’s Council on Integrity and Efficiency, as adapted for GAO’s work. We provided a draft of this letter to DOD and SSA for comment. DOD provided oral comments, which are discussed in the “Agency Comments and Our Evaluation” section, and SSA provided written comments, which are reprinted as an enclosure. Benefits Provided to Eligible Individuals but Data Entry Errors Raise Concerns: In 40 of the 41 cases we investigated, a data entry error and/or internal control weaknesses either at SSA or at the military treatment facility caused these cases to appear to have had a prescription ordered for a deceased person. We did not find indications of potential fraud in any of these 40 cases. A data entry error at SSA caused 10 of the errors. The remaining 30 cases stemmed from data entry errors made at Wilford Hall. They occurred in part because DOD has not built a control into CHCS’ data entry process to prevent entering new clinical data into a deceased person’s record rather than the correct record. Table 1 summarizes our analysis. The remaining case is discussed in the next section of this letter. Table 1: Results of Analysis of 40 People for Whom a Prescription Was Ordered after Their Reported Date of Death: [See PDF for image] Source: DOD and SSA data. Note: GAO analysis of DOD and SSA data. [End of table] Individuals Incorrectly Listed as Deceased by SSA: Ten of the 40 cases involved individuals who were incorrectly listed as deceased in SSA’s Death Master File. These individuals were not only alive, but they were also eligible for health care benefits. Our interviews with the individuals or their family members disclosed that the erroneous entry typically occurred when the individual reported the death of a spouse. The SSA official receiving the report of death appears to have recorded not only the death of the actual deceased person but also the individual reporting the death. In each case, the individual who was incorrectly recorded as deceased told us that he or she notified SSA of its error and benefits were restored. However, these individuals continued to be listed in the SSA Death Master File. These inaccuracies in SSA’s database had generally persisted for years. For example, 5 of the 10 had been listed as deceased for over 10 years. Incorrect recordings of death are not isolated incidents. SSA’s Inspector General has reported that erroneous dates of death continue to exist in the Death Master File database.[Footnote 5] These erroneous dates stayed in the database because SSA’s payments and Death Master File systems were not fully integrated. Although SSA restarted payments, changes in the payment system database to restart the payments did not trigger subsequent changes in the Death Master File. According to the Inspector General report, these erroneous dates of death have caused other agencies to expend resources researching death information for living individuals. In our work, a DOD official told us that DEERS officials have to reverify that individuals were alive and eligible for health care not only at Wilford Hall but also throughout the DOD-wide eligibility system. In a January 2003 report on SSA’s efforts to improve its Death Master File, the SSA Inspector General reported that as of September 2002, SSA had implemented an automated process to (1) identify inaccurate death data and (2) generate a quarterly report that lists names and SSNs requiring investigation.[Footnote 6] In addition to causing agencies additional work, erroneous reports of death in the Death Master File can result in living individuals’ SSN and other personal information becoming public information because SSA makes the Death Master File information available to the public upon request. The SSA Inspector General reported that as a result, at least some erroneously reported deceased individuals had experienced various continuing difficulties, such as obtaining credit.[Footnote 7] In one case we investigated, for example, the individual, whose SSN had been listed in the Death Master File since 1991, reported experiencing periodic problems ever since her reported death. She told us she had been denied a cell phone and had difficulty getting reimbursement for a prescription filled at a retail pharmacy. In two other cases, the individuals said that their retired military and/or Social Security payments were temporarily suspended when the problem first occurred in the 1990s, but their benefits were restored within a couple of months. They said they had not experienced additional problems caused by the inaccurate death file. In the remaining cases we investigated, the individuals reported that they had not experienced significant problems because of these errors. They had found out about the erroneous reports of their deaths when they received a notification that their Social Security or other government benefits had ended. However, they reported the error to SSA and had not experienced subsequent difficulties, although the Death Master File continued to show them as deceased. Prescriptions Dispensed to an Eligible Individual but Recorded under a Deceased Person’s SSN: For 14 of the cases, prescription drugs were dispensed to an eligible individual but were recorded under a deceased person’s SSN. We concluded that these situations were data entry errors made by physician or pharmacy staff when they entered a prescription into the CHCS database. Usually, only one or two prescriptions were dispensed under the incorrect SSN for the 14 cases, and the errors were one-time events limited to a single day. To record a prescription in the patient’s CHCS clinical record, physician or pharmacy staff must access the patient’s record in the CHCS database, which also includes records of deceased patients. The staff is to use the first letter of the last name and the last four digits of the SSN of the individual’s sponsor to search for and select the appropriate record. In these 14 cases, the person who entered the prescription into the CHCS database selected the wrong individual’s record. In 5 of the 14 cases, they chose the patients’ deceased sponsor’s record. In 5 other cases, they chose another related individual’ s record. In the remaining 4 cases, they appear to have chosen the record of an individual unrelated to the patient. We identified the likely recipients of the prescriptions by examining relevant data such as the prescription history and physician appointments of the deceased person’s family members and others with similar names. For example, one case involved a deceased sponsor whose widow’s first name was very similar to his. The widow had a history of taking the pain medication that showed up in her deceased sponsor’s CHCS record, and she also had a doctor’s visit on the same day that the prescription was entered into her sponsor’s CHCS record. In another example, an individual with a similar last name and the same last four digits of the SSN as our case had a history of using the same ophthalmic medication that showed up in our case’s CHCS record. Even though our work indicated that the intended individuals received the prescriptions, we believe these cases raise a clinical issue because the prescriptions were not entered into the correct individuals’ records, leaving those records incomplete. When they are incomplete, patient safety issues such as potentially dangerous drug interactions for those individuals may not surface and be addressed. Prescription Dispensed to Unknown Individual and Recorded under a Deceased Person’s SSN: For two cases, a prescription was dispensed and recorded under a deceased individual’s SSN, but we could not determine who received the prescription. A total of three prescriptions were dispensed. In one case, a single prescription was dispensed for the generic equivalent of the sleeping aid Ambien. The other case was for two prescriptions for four pills each of the inexpensive antibiotic Amoxicillin. Although we were not able to determine who received these prescriptions, the limited number and small value of the prescriptions dispensed led us to conclude that these two cases were probably not indications of fraudulent or abusive activity. Rather, we concluded that these cases were caused by the same type data entry errors as just discussed. Prescriptions Not Dispensed: For 14 cases, Wilford Hall’s records show that the prescriptions did not leave the pharmacy and were canceled. We concluded that these cases involved data entry errors similar to the ones discussed in the previous two sections except that in these cases the prescriptions were not dispensed, according to the clinical records. For 8 of these 14 cases, the physician or pharmacy staff identified the data entry errors and canceled the prescriptions in the CHCS database before they were filled. In most of these cases, they caught and corrected their own error within minutes. In the remaining 6 cases, the prescription was filled but was not picked up. At Wilford Hall, the pharmacy’s practice is to return medications to inventory if they have not been picked up after 7 days. A prescription is canceled in the individual’s CHCS record when the medication is returned to inventory. CHCS Missing Important Data Entry Control: Thirty of these errors were caused by Wilford Hall staff accessing the wrong person’s CHCS record to enter a prescription. DOD’s process for entering clinical data into an individual’s CHCS record does not include a preventive edit or control to prohibit entering new data into a deceased person’s record. While such data entry errors would not necessarily be unexpected given the workload, they should be anticipated and mitigated. These types of data entry errors can create a risk that a prescription does not get into the correct person’s clinical record, which can result in a potential patient safety issue not being addressed since the clinical record is incomplete. Neither CHCS nor its planned successor system, CHCS II, have edits or controls built into them to prevent new data from being entered into a deceased person’s clinical record, according to DOD officials responsible for the successor system. Both CHCS and CHCS II have an alert/reminder feature that can notify clinicians of potentially dangerous drug interactions based on comparing the prescriptions a patient is currently taking to a new one that is prescribed. However, this alert feature cannot work effectively when prescription information is entered into the wrong individual’s record. Various edits and controls to help ensure the integrity of data entered into clinical records are possible, such as making the records of deceased persons “read-only” so that new data cannot be entered. Another possibility includes programming CHCS so that when a deceased individual’s clinical record is accessed, a warning message appears saying that the individual is deceased and asking if new data should be entered. Prescriptions Dispensed to an Ineligible Individual: The last of the 41 cases involved prescriptions dispensed to an ineligible individual. However, based on our investigation and analysis of the circumstances of this case, we did not identify health care benefits that we could conclude were fraudulently obtained. In this case, an elderly retired military member’s second wife was listed under her name as eligible in DOD’s DEERS eligibility system but was incorrectly assigned the member’s deceased first wife’s SSN. Therefore, when we compared the SSNs in the Wilford Hall file to the SSNs in the Death Master File, she was identified as having prescriptions ordered after the date of her death. According to Wilford Hall records, the divorced second wife received 39 original prescriptions and refills that Wilford Hall valued at less than $500 from 1997 through 2001. However, she became ineligible for DOD health care benefits upon her divorce from the retired service member in March 1998. We determined that 31 of these prescriptions, valued in Wilford Hall’s records at about $350, were for prescriptions after she became ineligible. DOD’s policy is that sponsors are to report any change in dependent status, which enables DOD facilities to determine when a divorced spouse or other dependents are no longer eligible for benefits. In this case, we were unable to determine if the sponsor had reported his divorce to DOD because the sponsor’s very poor health at the time of our investigation precluded our contacting him on this matter. The second wife explained that when her husband established her eligibility, he used his deceased first wife’s SSN. The second wife said she did not correct the error because she was provided benefits under her sponsor husband’s SSN, which the military treatment facility uses to access the clinical care records. She was issued an identification card before she was divorced from the sponsor that was valid until September 2001, 3 years after her divorce. Absent a record of the divorce, DEERS--DOD’s eligibility system--showed her eligible for benefits. As of January 2003, the last recorded prescription in Wilford Hall’s database for the patient was in August 2001, the month before the expiration date on her identification card. According to this individual, no one told her that she became ineligible when she was divorced. She said she stopped using the military treatment facility when her identification card expired. We have provided our documentation on this case to DOD to correct its eligibility records. Cases similar to this one do not appear to be unusual, and may, in fact, be quite commonplace. In a January 2000 report on DEERS,[Footnote 8] the DOD Inspector General reported that in 30 of the 81 cases it analyzed in which individuals were ineligible for benefits, the sponsor had not reported a divorce to DEERS, as required by DOD policy. Fifteen of the divorces had been final for at least a year, and of those, 9 had been final from 4 to 26 years. In these 9 cases, the identification cards had been renewed at least one time after the divorces became final. Some cards were renewed with the sponsor’s signature on the application and some with the sponsor’s divorced spouse’s signature. In the latter cases, the former spouses used their expiring identification cards as the basis for obtaining new cards. Based on the Inspector General’s recommendations, DOD established a 30-day time limit for sponsors to report a change in their dependents’ eligibility status. Conclusion: We did not find evidence of fraudulently obtained health benefits in the 41 cases we investigated. However, our follow-up work suggests that the process for entering data into patients’ clinical records at DOD’s military treatment facilities has a key flaw. While the 10 cases related to errors in the SSA Death Master File are beyond DOD’s control, the other 30 are not. They are the result of human data entry errors that, while not unexpected in a busy environment such as the one at Wilford Hall, can result in incomplete medical records and significant patient safety issues such as potentially hazardous drug interactions not being identified. These errors could reasonably be addressed by adding preventive data entry controls. Recommendation for Executive Action: To strengthen controls over data entry into the DOD clinical records database and to help ensure that patient safety issues are identified, we recommend that the Secretary of Defense direct the Assistant Secretary of Defense for Health Affairs, in conjunction with the military services’ Surgeons General, to institute a standardized preventive control procedure or procedures to prevent inadvertent entry of new clinical data into a deceased person’s record clinical record in CHCS and CHCS II. Agency Comments and Our Evaluation: We provided a draft of this report to both SSA and DOD for their review. SSA, in its written comments reprinted as an enclosure, agreed that some issues of accuracy exist about information contained in the Death Master File. SSA explained why these inaccuracies exist and the efforts it has underway to improve file accuracy. In DOD’s oral comments, the Assistant Secretary of Defense for Health Affairs concurred with the findings of the report but did not concur with the recommendation. DOD’s position is that the report overstated the extent of the problem and that the small number of data entry errors compared to the number of prescriptions written annually does not warrant a global change to its processes for entering data into its clinical database. DOD said that its current data entry procedures and oversight controls are adequate to prevent errors in medical care or the delivery of significant levels of inappropriate health care, and it believed a continuing emphasis on ongoing pharmacy training programs to ensure correct data entry was a more feasible approach. DOD also said that the results of our work verify that DOD’s health care eligibility system works extremely well. With regard to DOD’s health care eligibility system, we do not agree with DOD. Our work was narrowly focused on investigating the 41 cases for potential fraud. In our work, we did not attempt to measure the full extent of the problem of data entry errors, and we neither evaluated nor do we comment on the effectiveness of controls over DOD’s health care eligibility system. In the course of investigating the 41 cases, instead of identifying fraud, we determined that DOD clerical errors in 30 of the 41cases had created the appearance that individuals had received a prescription drug after their death. : With regard to the best approach to avoiding clerical data entry errors, we continue to believe that the practical solution to these clerical errors is for DOD to implement our recommendation to develop a preventive control over the process for entering data into the clinical database. The problems we discuss in the report are a matter of entering prescription information into the wrong individual’s medical file, which can raise patient safety concerns. When a prescription is not entered into the file for the individual who is to receive the prescription, CHCS’ ability to compare the prescription to others the individual may be taking and identify potentially hazardous drug interactions is jeopardized. The problems we identified were caused by human error in the data entry process. While we understand that human errors will always occur to some extent and that training is very valuable, we do not believe that additional training alone is the best approach to preventing these types of errors. We believe they can be even more effectively avoided by adding a systemic preventive control to the data entry process. For example, CHCS II could be programmed to present a “flag” to the data entry person when a deceased person’s record is accessed that presents a message such as the following on the screen. “This person is deceased. Are you sure you want to enter new clinical data?” The system could also be programmed to not allow further data entry until the question is answered. When patient safety is at stake, we believe that DOD should take all reasonable safeguard measures, particularly during the development stage of a new system when changes are comparatively less costly. We believe DOD will miss a significant opportunity to improve its control over data entry and help ensure the safety of its patients if it does not address this weakness in the data entry process, especially during the development of the CHCS II pharmacy module. Unless you publicly announce its contents earlier, we will not distribute this letter until 15 days from its date. At that time, we will send copies of this report to the Chairmen of the Subcommittee on National Security, Emerging Threats and International Relations and the Subcommittee on Government Efficiency and Financial Management of the House Committee on Government Reform as well as other congressional committees. We are also sending copies to the Secretary of Defense; the Assistant Secretary of Defense for Health Affairs; the Surgeons General of the military services; the Secretary of the Air Force; and the Commanders of Brooke Army Medical Center, Randolph Air Force Base Clinic, and Wilford Hall Medical Center. Copies will be made available to others upon request. In addition, the letter will also be available at no charge on the GAO Web site at http://www.gao.gov. Please contact Greg Kutz at (202) 512-9095 or by e-mail at kutzg@gao.gov or Linda Garrison, Assistant Director at (404) 679-1902 or by e-mail at garrisonl@gao.gov if you or your staffs have any questions concerning this report. Major contributors to this correspondence were Mario Artesiano, Ray Bush, Carl Higginbotham, Ken Hill, Sue Piyapongroj, John Ryan, and Lisa Warde. Gregory D. Kutz: Director, Financial Management and Assurance: Robert J. Cramer: Managing Director: Office of Special Investigations: Signed by Gregory D. Kutz and Robert J. Cramer: Enclosure: Enclosure: Comments from The Social Security Administration: SOCIAL SECURITY ADMINISTRATION: The Commissioner: March 7, 2003: Mr. Gregory D. Kutz: Director, Financial Management and Assurance U.S. General Accounting Office: Washington, D.C. 20548: Dear Mr. Kutz: Thank you for the opportunity to review and comment on the draft report “Military Treatment Facilities: Eligibility Follow-up at Wilford Hall Air Force Medical Center” (GAO-03-402R). We are pleased to know that our Death Master File (DMF) assisted you in conducting this review. Also, while some issues exist with respect to the accuracy of information contained in the DMF, I want to take this opportunity to note that prior audits of the DMF found that the file is over 95 percent accurate. With respect to the 10 instances where SSA’s DMF had the individual incorrectly listed as deceased, I offer the following reasons why an individual who is alive may be shown as deceased on the DMF. * Erroneous Termination Cases - Prior to 2000, two actions in different venues were required to return a person to payment status when erroneously terminated: one to correct the payment record and one to correct the Numident/DMF. A review of those processes found that often times the DMF was not corrected. In November 2000, we modified our Death Alert, Control and Update System to recognize reinstatement cases and correct the DMF automatically. * Returned Payment Policy - We also found that many erroneous death terminations were due to returned payments marked “deceased” from the postal authority and financial institutions. Under previous procedures, these death notices were processed without further verification until after the termination action occurred. However, in May 2002, we changed our policy and now verify these payments for title 11 beneficiaries marked deceased before terminating benefits, not after. As demonstrated by the actions described above, we are committed to working to improve the accuracy of the DMF. In 1999, we entered into contracts with the National Association for Public Health Statistics and Information Systems and with the individual States to fund the Electronic Death Registration (EDR). EDR is a State system that provides us with a verified death report and is reported within 5 days of the person’s death. When EDR is fully implemented, most of the death data we process will be a State report with a verified SSN for beneficiaries and for non-beneficiaries. We expect full implementation of EDR will produce a nearly 100 percent accuracy rate for death records reported via EDR. If you have any questions, please have your staff contact Laura Bell at (410) 965-2636. Sincerely, Jo Anne B. Barnhart: Signed by Jo Anne B. Barnhart: (192081): FOOTNOTES [1] U.S. General Accounting Office, Military Treatment Facilities: Internal Control Activities Need Improvement, GAO-03-168 (Washington, D.C.: Oct. 25, 2002). [2] We used a database of pharmacy prescriptions recorded in fiscal year 2001 provided to us by Wilford Hall that included prescriptions recorded for about 100,000 individuals at Brooke Army Medical Center and Randolph Air Force Base Clinic, which share health-care-related computer files with Wilford Hall. In this report, we refer to them collectively as Wilford Hall. [3] U.S. General Accounting Office, Information Technology: Greater Use of Best Practices Can Reduce Risks in Acquiring Defense Health Care System, GAO-02-345 (Washington, D.C.: Sept. 26, 2002). [4] GAO-03-168. [5] Social Security Administration, Office of the Inspector General, The Social Security Administration’s Procedures to Identify Representative Payees Who Are Deceased, A-01-98-61009 (Baltimore, Md.: September 1999) and Disclosure of Personal Beneficiary Information to the Public, A-01-01-01018 (Baltimore, Md.: December 2001). [6] Social Security Administration, Office of the Inspector General, The Social Security Administration’s Efforts to Process Death Reports and Improve its Death Master File, A-09-03-23067 (Baltimore, Md.: January 2003). [7] SSA, Office of the Inspector General, A-01-01-01018. [8] Department of Defense, Office of the Inspector General, Evaluation of The Criminal Investigative Environment In Which The Defense Enrollment Eligibility Reporting System Operates, CIPO2000S001 (Washington, D.C.: Jan. 7, 2000).