. Author manuscript; available in PMC: 2018 Oct 21.

Published in final edited form as: Am J Gastroenterol. 2016 Aug 16;111(10):1488–1496. doi: 10.1038/ajg.2016.316

Table 4.

Logistic regression analyses to assess the effect of center on the association between race and alcohol and tobacco among patients with CP

Outcome	OR (95% CI)—Black vs. White CP patients
Outcome	All centers	Top 3 centers
Physician-defined alcohol etiology (yes)	4.31 (2.92–6.35)	4.35 (2.57–7.36)
Physician-identified tobacco as risk factor (yes)	2.54 (1.75–3.67)	2.97 (1.79–4.92)
Self-reported heavy–very heavy drinker (yes)	2.62 (1.82–3.78)	2.50 (1.53–4.09)
Self-reported ever smoking (yes)	2.97 (1.80–4.90)	4.16 (1.98–8.77)

CI, confidence interval; CP, chronic pancreatitis; OR, odds ratio. All models were controlled for age, sex, and center. Joint test of interaction terms between center and race was not statistically significant (range of P values: 0.054–0.979) for any of the models.

Top 3 centers—University of Pittsburgh, Virginia Commonwealth University, and University of Alabama.