A study of predictive validity of physician assistant students' reported practice site intent

With the downward trend in numbers of primary care physicians, it is anticipated that mid-level providers will increasingly fill the resulting void in rural health care. Recruitment of health care providers into rural areas has been historically difficult, and West Virginia has been at the forefront of developing programs to address rural recruitment. One of these programs is the West Virginia Rural Health Education Partnership, which provides rural clinical rotation opportunities for students from multiple disciplines and schools in West Virginia. This study assessed the predictive validity of physician assistant (PA) students' prediction, after completion of rural rotations, of a rural or an urban practice site and also assessed for factors associated with subsequent rural practice.

The Institutional Review Board of West Virginia University approved this study. The West Virginia Rural Health Education Partnership's electronic database, which has collected students' postrotation evaluation information since 2001, was coupled with West Virginia licensing information that included information on PA practice site(s). Data on 168 practicing West Virginia PAs who had completed rural rotation questionnaires during their clinical education were examined. Designation of a rural or an urban practice site was done using rural-urban commuting area codes.

Student prediction of rural practice was significantly associated with subsequent rural practice (P < 0.04). Variables most predictive of rural practice were student reporting of high school hometown as rural and of West Virginia practice intent.

This study suggests moderate predictive validity of PA student reporting on rural practice and on West Virginia practice intent; such methods may have potential in prediction of the future rural PA workforce.