Background info

Background Program Evaluation of Mental Health Services

The program evaluation is looking at patient-level outcomes to determine what impact our services have on patients with specific mental illnesses. The contractor has proposed a telephone survey of selected veterans to supplement the data found in VA’s administrative databases and patient charts. Two groups of veterans will be surveyed- veterans who use VA services to see how our services have impacted them, and veterans who do not use VA for their health care to find out why they don’t. We are specifically looking a vets with five diagnoses: schizophrenia, bipolar, major depressive disorder, PTSD and substance abuse.

The following notes are taken from the contractor’s proposal.

• Client Survey Sample. In order to build upon existing VA data, we will use a random sample of SMI veterans by stratified by VISN and within VISN. The sample will be stratified further by a combination of client characteristics, including some, but not all of the following data: demographics (age, gender and race), socio-economic status, service-connected status, comorbidities, and housing status.

• There will be 150 veterans per VISN, per diagnosis, or a total sample of 3,150 veterans per diagnosis and 12,600 total for the client survey. We expect to have a response rate of 50%, yielding 75 completed surveys per diagnosis and VISN, for a total of 6,300.

• The advantages of linking the client survey with the medical record review are that data elements from both sources could be analyzed together to provide a more comprehensive assessment of whether the needs and requirements of veterans are being met.

• We will survey an additional non-VA comparison sample using the same client survey, which will allow us to compare VA and non-VA samples on important outcomes such as satisfaction, functioning and quality of life. This survey will be augmented by questions about current and past treatment utilization. We will also ask about current family income and family/social support and why they are not using VA care. We plan to sample 7,200 veterans not receiving VA care.

• We will look at several measures of access, including contact with providers, volume of services, and continuity of care. We will also look at barriers to access, including geographic accessibility, individual preferences, family resources and past experiences with VA and non-VA care.

• We will conduct the entire data collection through a telephone survey of about 30 minutes in length.