Table 4.

Polytomous logistic regression model assessing factors associated with the aggressiveness of primary care physicians’ lung cancer screening recommendations^a (n=962)

	Recommends for 1–3 Vignettes vs Does not recommend for any Vignette	Recommends for 4–5 Vignettes vs Does not recommend for any Vignette
	OR 95% CI	OR 95% CI
Physician Characteristics
Board certified
Yes	1.0	1.0
No	1.0 (0.5, 2.0)	2.4 (1.3, 4.3)
International medical graduate
No	1.0	1.0
Yes	2.1 (1.2, 3.7)	2.2 (1.3, 3.6)
Practice Setting Characteristics
Primary practice arrangement
Employee of practice/other/missing	1.0	1.0
Full/part owner of practice	1.1 (0.7, 1.9)	1.6 (1.0, 2.5)
Physician Beliefs and Practice Style
Believes any expert group recommends lung cancer screening
No	1.0	1.0
Yes	2.1 (1.1, 3.8)	5.0 (2.9, 8.6)
Believes any screening test2 is very effective in reducing lung cancer mortality for never smokers
No	1.0	1.0
Yes	0.4 (0.1, 1.4)	1.2 (0.4, 4.1)
Believes any screening test2 is very effective in reducing lung cancer mortality for former smokers
No	1.0	1.0
Yes	4.5 (1.6, 12.5)	5.5 (1.9, 15.5)
Believes any screening testb is very effective in reducing lung cancer mortality for current smokers
No	1.0	1.0
Yes	2.6 (1.3, 5.5)	3.7 (1.5, 8.8)
Intensity of colorectal cancer screening recommendations
Low/moderate	1.0	1.0
High	2.2 (1.4, 3.6)	3.0 (1.9, 4.7)
Unknown	2.7 (1.0, 7.1)	1.1 (0.3, 3.6)

Data source: National Survey of Primary Care Physicians’ Recommendations and Practices for Breast, Cervical, Colorectal, and Lung Cancer Screening

^aOther model covariates included physician gender, race/ethnicity, specialty, and medical school affiliation; practice size, geographic location, Census region, type of medical record system used, % of patients who are uninsured, % of patients with Medicaid, weekly patient volume, and # patients who have asked about lung cancer screening in the past 12 months. These covariates were not significant in the multivariate model.

^bChest × ray, sputum cytology, or low–radiation dose spiral CT