Family and friends have become increasingly responsible for providing informal care for stroke survivors. Studies have noted race/ethnic variation in caregiving of older adults and those with dementia or cancer, but few have described social networks of stroke survivors and examined network characteristics by race/ethnicity. As part of the Discharge Educational Strategies for Reduction of Vascular Events (DESERVE) trial, stroke/TIA patients were asked to identify up to 5 persons (“alters”) with whom they discuss important matters and provide information about each relationship. Network characteristics included median number of alters; having an alter in the same household, someone who talks with the participant every day, or with whom they discuss health concerns; relationship “closeness”; and having a family-only (vs. family and friend) network. Household composition was also recorded. Network characteristics were described overall and by race/ethnicity using frequencies and Wilcoxon rank-sum and Chi-square tests. Of the 500 study participants, most identified at least one alter (86%), 61% had an alter in their household, 88% talked to an alter every day, 82% were extremely/very close with all alters identified, 96% discuss health concerns with an alter, and 74% had a family-only network. Less than 27% lived alone, 21% with a spouse only, 15% with a non-spouse, and 38% with >1 person. NHBs and Hispanics were more likely than NHWs to identify no alters (19% and 16% vs. 8%, p=0.006) but talk to an alter every day (90% and 92% vs. 79%, p=0.007). NHWs were more likely to live in the same household as an alter than NHBs and Hispanics (67% vs. 53% and 59%, p=0.006). Hispanics had greater prevalence of having a family-only network than NHWs and NHBs (83% vs. 71% and 68%, p=0.014). NHWs were more likely to live with a spouse, while NHBs and Hispanics were more likely to live with a non-spouse or with >1 person (p<0.001). Given the importance of social networks in the aftermath of stroke, understanding the structure of different types of networks will help inform targeted interventions to promote vascular risk factor management and reduce race/ethnic disparities in stroke outcomes.