Development of a Post-Traumatic Stress Disorder (PTSD) Population Registry for Veterans

This project will be designed to address a range of research questions within the registry itself, as well as in the comparisons with non-registry groups. Specific aims of this project can be divided into the following major research areas.

1. Epidemiology of PTSD Aim 1. To evaluate the natural history of PTSD, including progression and remission of symptoms over time, and the long-term psychosocial, medical and quality of life outcomes associated with the disorder Aim 2. To identify predictors of recovery or remission from PTSD, including treatment utilization, psychiatric and medical comorbidities, and sociodemographic factors Aim 3. To identify clinical characteristics, risk factors and comorbidities of PTSD, including disparities by gender, race/ethnicity, social support, and socioeconomic status, by comparing PTSD patients to a "control" group of veterans without PTSD
2. PTSD Treatment and Outcomes Aim 4. To identify current treatment approaches and outcomes of treatment through time, particularly among service members who are often excluded from clinical trials due to comorbidities and therefore remain unexamined and possibly under-treated Aim 5. To estimate the joint effects of combat exposure and PTSD on long-term psychosocial well-being and quality of life measures, by comparing PTSD patients to a "control" group of veterans who were not deployed to combat/war zones during the same time period
3. Health Services Research Aim 6. To establish the prevalence of PTSD in a comparison group of service men and women who had not received the PTSD diagnosis in their patient record but are high utilizers of the VA medical system, and to identify risk factors for missed PTSD diagnoses Aim 7. To provide data that are immediately useful for determining future health services policy, costs, and needs for upcoming research, manpower and training, by assessing current referral and utilization patterns. Both direct and indirect (e.g., productivity loss, pharmacological burden) costs of treatment will be assessed Aim 8. To develop a large database of servicemen and women with PTSD and network of treatment sites that are potentially available for further observational and interventional studies, as well as concurrent ancillary studies, such as qualitative studies on caregiver/family burden.

The proposed project will be conducted in three phases over three years using both existing data from the VA medical record database and prospective data to be collected by a structured interview in the third phase of the study, as well as periodic data transfers from the medical records. During the first phase (months 1-9), the data de-identification procedures for the study will be pre-tested (months 1-2), the data will be prepared for abstraction (months 2-4), and the study protocol and research interview will be reviewed and finalized (months 6-9). During Phase II (months 4-12), the preliminary PTSD database will be created by data transfer (month 4), query reports/resolutions (month 5). Also during this phase, analyses of existing data extracted from the VA medical record system will be conducted (months 10-12) and target cohorts for follow-up assessment will be identified. Phase III (months 12-36) will include subject recruitment for the PTSD registry and two comparison groups, and the structured telephone interviews will be conducted in all three groups (months 12-24). Data collected in the telephone interview will be coded, de-identified and merged with the overall study database. Analyses will be conducted during this phase (months 24-36) related to the Specific Aims of the study. The PTSD registry database will be updated as new VA electronic medical records for registry participants appear during the course of this project, thereby capturing the trajectory of PTSD patients.

Participants and Sampling Strategy The source of participants will be the VA in/outpatient electronic medical records of all new utilizers of the VA medical system from FY 2002 through 2007. According to data from the DOD Defense Manpower Data Center (DMDC) received by the VA Environmental Epidemiology Service (EES) on April 17, 2007, from FY 2002 through February 2007, there had been a total of 717,196 unique OIF/OEF veterans (excluding 3,303 who died in theater) who had been separated from active duty following deployment, and were therefore eligible for VA benefits. Of these, 252,095 (35%) had obtained VA health care by March 2007. More recent healthcare data through September 2007 indicate that 263,909 individuals have accessed VA health care. The VA health administrative database rigorously records all in-patient and out-patient visits, and includes basic information regarding each visit, such as the patient name, SSN, DOB, gender, and ICD-9 diagnostic codes to describe the purpose of the visit (EES database).

The overarching objective of this project-the creation of a PTSD registry-will be accomplished by identifying male and female OIF/OEF veterans from the entire VA medical records database of new utilizers with a diagnosis of PTSD (as defined by ICD-9 code 309.81) in the electronic medical record. We will sample from the VA database to recruit participants for a structured telephone interview to obtain details on exposures and outcomes of interest that are not captured in the existing database. A sample of 800 PTSD patients with OIF/OEF war zone experience, 400 patients without the PTSD diagnosis but with OIF/OEF experience and high utilization of VA services, and 400 individuals matched by age, gender, race/ethnicity, and year of entry into the VA EES database, who were not sent to the war zones of OIF/OEF will be recruited to participate in the telephone interview. To be representative of all veterans from the OIF/OEF era, the sampling will be stratified by year of entry into the VA database.

Creation of the PTSD Registry Database Patient registries, as defined by the AHRQ and World Health Organization, are organized systems containing information about individual persons,42 and the patient registry database is defined as a file (or files) derived from the registry. A patient-based database is necessary to allow direct access to information on diagnoses of interest and treatment outcomes, as well as to longitudinally track follow-up visits, progress, and treatment courses.

In contrast to these criteria, the current database of utilizers of the VA healthcare system is structured by chronological in-patient and out-patient visits, rather than unique patient identifiers or diagnoses. Thus, with the existing database, there is no readily accessible way to identify unique patients with PTSD or to assess their longitudinal outcomes and utilization of VA healthcare. Therefore, a fundamental objective of this proposal is to establish a registry of patients with PTSD from the existing VA utilization database. Electronic medical records such as those in the VA database are increasingly important sources of data, but data must be extracted, transformed into registry format, and loaded into the registry, where they will reside in the registry database, together with registry-specific data that is imported from other sources. The other sources of data for the PTSD registry will be the OIF/OEF veteran roster (particularly for military specific data, e.g., branch, rank, deployment dates, etc.) and, in Phase III, the proposed telephone interview.

During Phase I of this project, a cumulative roster of unique OIF/OEF veterans who have been separated from active duty following deployment and who have become eligible for VA healthcare will be matched with VA inpatient (PTF) and outpatient (OPC) databases for fiscal year (FY) 2004 through FY 2007. For identification of a potential PTSD patient, any veteran patient with a diagnosis of ICD-9 CM, 308.81 will be selected. The PTSD registry will encompass utilizers of the VA medical system. Each unique patient with a visit record indicating a PTSD diagnosis will be included. The registry creation will involve a multi-step process, including steps to identify all PTSD diagnoses, to identify unique patients within the PTSD diagnoses, and to identify each additional visit made by each PTSD patient. For identification of OIF/OEF veteran non-PTSD patients who are high VA healthcare users (sample #2), we will review in/outpatient records of OIF/OEF veteran patients and identify those who have been hospitalized or treated as outpatients more than 11 times during the post deployment period. A similar step will be taken to identify a non-combat comparison group of patients (sample #3). All veteran patients for three groups will be stratified by year of first entry to VA healthcare data bases (FY 2004-2007) and equal number of veteran patients will be randomly selected for each fiscal year for sample #1 and sample #2. For sample #3, additional matching variables such as age, gender, and ethnicity/race will be used in addition to the year of entry to VA healthcare databases. Ultimately, during Phase II, all existing VA medical utilization data and OIF/OEF veterans' roster data for each veteran patient will be merged and de-identified to ensure privacy and confidentiality.

During Phase III, data obtained from a telephone interview will be merged with the existing data in the PTSD registry database. The PTSD registry will continue to be updated at 6-month intervals during the course of this project as new or updated electronic medical records are entered into the VA in/outpatient healthcare database for registry participants.