Similar case control designs employing matched groups of Caucasians versus African Americans from the United States Model Systems database have also been employed to examine race effects on rehabilitation outcomes. Putzke et al. (2002) matched race groups according to age, education, gender, occupational status, impairment level, etiology, primary sponsor of care and geographic region whereas Meade et al. (2004) matched according to level of injury, AIS, age and primary sponsor of care. By controlling for all these variables, these authors were able to establish that race acts more as a proxy variable than a predictor of outcomes (Putzke et al. 2002). For example, differences did exist in a wide variety of demographic, rehabilitation outcomes and medical complications for African Americans versus Caucasians but these were generally accounted for by socio-demographic and etiological differences associated with these groups (Meade et al. 2004; Putzke et al. 2002). For example, African Americans were significantly more likely to be injured as the result of violence and have 11th grade education or less while Caucasians were more likely injured as a result of motor vehicle crashes and had high school education or more (Meade et al. 2004; Putzke et al. 2002). It is likely that these etiological and socio-demographic variations have far more to do with differences seen in rehabilitation outcomes than race. In support of this, Pollard and Apple (2003) found that neurological recovery was not affected by race.
Krause et al. (2006) observed that, post-discharge, African Americans in a Southeastern United States SCI population reported a greater number of poor health days, more hospitalizations, and a greater number of days hospitalized. However, by conducting an analysis of the effect of the potential mediating variables of education and income it was found that these had substantially more impact on these findings than did the effect of race. In contrast, Eastwood and colleagues (1999) found that African Americans experienced shorter rehabilitation LOS than Caucasians. Although, this difference was not further explored in the study, it may be attributable to variation in injury severity.
There is level 3 (from two case control studies and three case series: Eastwood et al. 1999; Krause et al. 2006; Meade et al. 2004; Pollard & Apple 2003; Putzke et al. 2002) that there is no difference with respect to race (Caucasians versus African-American) on rehabilitation LOS and neurological or functional outcomes associated with rehabilitation that are not otherwise explained by socio-demographic or etiological differences.