Comparison of tailored and mainstream primary care delivery for homeless persons: A new, validated patient assessment

Background: Many organizations tailor primary care services for homeless individuals, an initiative recently embraced by the Veterans Health Administration. This study compared homeless patients' assessments of tailored and non-tailored health care settings. Methods: Based on extensive qualitative research, we developed a Primary Care Quality Survey for the Homeless (PCQ-H) to assess patient experiences with 8 aspects of care identified by IoM reports and 3 additional gleaned from the qualitative research (respect, substance abuse/mental illness, homeless-specific needs). The survey was administered to 474 homeless individuals receiving care at a mainstream VA, moderately-tailored VA, highly-tailored VA, or a highly-tailored non-VA program. We compared PCQ-H scores across the sites, adjusting for relevant demographic and health characteristics. As ratings of care were primarily positive, we also compared a binary indicator of offering more than the median number of unfavorable or negative assessments of care across sites. Results: In analyses adjusted for case-mix, the highly-tailored non-VA program exceeded the VA sites for Patient-Clinician Relationship (p<.01), Perceived Cooperation (p<.01), and Homeless-Specific Needs (p<.05). Unfavorable responses were 10-25% more prevalent at the two VA sites compared to both highly-tailored sites (all p<.05). Conclusions: Homeless patients rate their relationship to their provider, cooperation among caregivers, and care for homeless-specific needs more highly when primary care services are highly-tailored. Whether such results generalize to diverse care settings remains to be seen. This new specialized patient assessment survey permits a rigorous examination of the extent to which service tailoring improves the primary care experience of this highly vulnerable population.