Table 2.

Effect of age, risk group, Charlson comorbidity index, year of diagnosis, and race on observation in men with localized prostate cancer^†

Characteristic	OR	95% CI	P-value*
Year of Diagnosis‡
2004 (ref)	—	—
2005	1.42	1.15–1.76
2006	1.39	1.12–1.71	<.001
2007	1.47	1.19–1.81
2008	1.69	1.38–2.05
2009	2.12	1.73–2.59
Age Group¥
66–69 (ref)	—	—
70–79	1.73	1.08–2.00	<.001
80–89	6.02	4.88–7.43
Risk Group€
Low	5.37	4.25–6.77
Intermediate	1.72	1.35–2.18	<.001
High (ref)	—	—
CCI Group£
0 (ref)	—	—
1–2	0.94	0.81–1.09	<.001
≥3	1.38	1.10–1.73
Race§
White (ref)	—	—
Black	1.28	1.03–1.61	<.001
Other	1.32	0.98–1.79

^†Multivariable logistic regression model with a-priori selected exposures of interest: age, risk group, CCI, year of diagnosis, and race and adjusted for median income of census tract, marital status, geographic region as well as all two-level interactions between exposures of interest. The odds ratios are shown for each level of main exposure variables, while other variables are held constant

^‡Year = 2009;

^¥age group =60–69;

^€risk group = low;

^£CCI group = 0;

^§race = white).

^*P-values for main effects test whether the main variable or any interaction term containing the main variable equals 0.

Abbreviations: CI, confidence interval; OR, odds ratio; CCI=Charlson Comorbidity Index