significantly less experience with the traditional doctor-patient relationship (Inui et al., 1984). These data clearly indicate that veterans receiving care within VHA are more medically and socioeconomically vulnerable than either nonveterans or veterans who do not use VHA.
Second, published VHA studies rely on information available in the Patient Treatment File (PTF), a national administrative database that contains a record for each admission to a VA hospital since the 1970s (Lamoreaux, 1996). PTF was developed to provide information about health-care use to VHA policy makers and managers. It does not, however, contain clinical data. Without valid clinical data, any relationship between race and health must be seen as theoretical. The nature of administrative data sets limits investigators’ ability to reach cause-and-effect conclusions about observed racial variations. On the other hand, primary data are needed for prospective studies designed to better understand whether observed racial variations can be accounted for by factors such as patients’ clinical characteristics, health/functional status, preferences for treatment options, or racism.
Given the above limitations, why should we nevertheless pursue studies examining the relationship between race and health among veterans receiving care within VHA? First, from a policy point of view, Kenneth W.Kizer, VHA Undersecretary for Health, has stated that one goal of VHA is to document that the quality of care it provides is equal to, or better than, care available in the private sector (Kizer, 1998). The Undersecretary specifically addressed actions needed to increase the accessibility of VA services. Second, VHA provides a functional safety net for a group for whom investigations about race and health may be most important—the most socioeconomically vulnerable veterans (Iglehart, 1996; Wilson and Kizer, 1997). Third, from the standpoint of research, because access to health care for eligible veterans is not limited by income, well-designed prospective studies can separate the effects of race and income. This factor is unique to studies conducted in VHA. Fourth, the veteran population contains a greater proportion of underrepresented minorities than exists in the country overall. The percentage of Blacks among all veterans is growing: they comprise 6.1 percent of World War II veterans, 8.3 percent Korean Conflict veterans, 8.6 percent of Vietnam Era veterans, and 17.1 percent of post-Vietnam veterans. The number of Hispanic veterans has also grown since World War II, from 3.4 percent to 5.7 percent (Heltman and Bonczar, 1990). Finally, research based on data from the PTF database is limited to vital statistics, as is research based on data from most administrative databases. The limitations of these data give emphasis to the critical need to supplement PTF with primary (clinical) data collection.