Patient Race May Be a Proxy for Other Factors in Predictive Models for UTI Risk in Children

Although previous prediction models have shown an association between patient race and urinary tract infection (UTI) risk, the use of a modified model that replaced race with UTI history and fever duration resulted in similar predictive accuracy, according to results of a study published in JAMA Pediatrics.

Although previous prediction models established an association between patient race and UTI risk, researchers conducted a systematic review and meta-analysis to determine whether race is a proxy for other factors in models that predict UTI risk among children. The primary outcome was the odds for UTI among children who were Black vs those who were of non-Black race/ethnicity. In addition, the researchers assessed the risk of bias among the included studies, and meta-regression was used to explore potential heterogeneity.

A total of 16 studies representing 17,845 children were included in the initial analysis, and the final and secondary analyses comprised 11 and 5 studies, respectively. Results of the primary analysis showed that the pooled odds ratio (OR) for UTI was significantly increased among children who were nonBlack (odds ratio, 2.44; 95% CI, 1.87-3.20) vs those who were Black. However, significant heterogeneity resulting from selection bias was noted. After the analysis was limited to studies with a low or very low risk of bias, the pooled OR for UTI among children who were non-Black was increased further (OR, 4.84; 95% CI, 3.16-7.41).

Similar results were noted in the secondary analysis. Results showed that the pooled OR for UTI among children who were non-Black (OR, 1.77; 95% CI, 1.08-2.90) was increased vs those who were Black. Significant heterogeneity among these studies also was noted, and 42% of the patients included were Hispanic. In addition, limiting the secondary analysis to the 1 study that had a low risk of selection bias resulted in an additional increase in the odds for UTI.

Compared with the predictive model used among the included studies, the researchers found that using a modified model that replaced race with UTI history (positive vs negative) and fever duration (<48 vs ≥48 hours) had similar sensitivity (95% vs 96%). However, in models that replaced race with either UTI history alone or fever duration alone, the sensitivity decreased to 90% (95% CI, 79-100). In a model that had no replacement for the variable of race, the sensitivity decreased further (82%; 95% CI, 68-96).

This study was limited by the lack of data on patients’ family history of UTI.

Despite these findings, the researchers noted that “the association between race and UTIs observed in many previous studies still requires explanation as clinical and translation science [shifts] toward race-conscious medicine.”

Reference

Shaikh N, Lee M, Stoke LR, et al. Reassessment of the role of race in calculating the risk for urinary tract infection: A systematic review and meta-analysis. JAMA Peds. Published online April 18, 2022. doi:10.1001/jamapediatrics.2022.0700