Table 2. Prevalence Ratios From Univariable Bilevel Log-Binomial Regression Modelsa for Cervical Cancer Screening for Women in Selected Oregon and California OCHIN-Affiliated Community Health Centers (N = 11,560), 2008–2010.
| Demographic Characteristic | N | Women Receiving Pap Test, n (% ) | Prevalence Ratio (95% Confidence Interval)b |
|---|---|---|---|
| Total | 11,560 | 7,346 (63.5) | NA |
| Insurance coverage | |||
| Continuous coverage | 5,062 | 3,237 (63.9) | 1 [Reference] |
| Partial coverage | 3,856 | 2,322 (60.2) | 0.98 (0.95–1.01) |
| No coverage | 2,642 | 1,787 (67.6) | 1.08 (1.04–1.12) |
| Age, y | |||
| 21–39 | 5,324 | 3,531 (66.3) | 1 [Reference] |
| 40–64 | 6,236 | 3,815 (61.2) | 0.93 (0.91–0.96) |
| Race/Ethnicity | |||
| Non-Hispanic white | 5,426 | 2,899 (53.4) | 1 [Reference] |
| Hispanic | 4,424 | 3,337 (75.4) | 1.39 (1.34–1.44) |
| Non-Hispanic other | 1,710 | 1,110 (64.9) | 1.16 (1.11–1.22) |
| Household income | |||
| ≥ 100% of FPL | 3,551 | 2,335 (65.8) | 1 [Reference] |
| 0-99% of FPL | 8,009 | 5,011 (62.6) | 0.95 (0.92–0.98) |
Abbreviations: FPL, federal poverty level.
a
Regression models included patient-level factors as fixed effects at level 1 and a clinic-level random intercept at level 2.
b
Prevalence ratios are significant at the α = .05 level if the 95% confidence interval does not contain 1.00.