Skip to main content

Table 3 Patient characteristics and their relationship with receiving mammography screening in the 12 months after visiting a study physician

From: Do physician communication skills influence screening mammography utilization?

  Mammography screening   
Patient Characteristics Yes No Odds Ratioa(95% CI) P- Value
N=32 689 N=64 019
Demographic mean (SD) mean (SD)   
Age, per 10 year increase 56.9 (5.6) 57.4 (5.9) 0.99 (0.98-0.99) <0.001
Percent high school drop-outs, per 20% increase 31.3 (12.9) 32.1 (13.1) 0.94 (0.90-0.98) 0.001
Family income, in dollars, per $25,000 increaseb 52,302 (23,302) 50,431 (22,754) 1.08 (1.06-1.11) <0.001
Patient region No. (%) No. (%)   
 Urban 28 263 (86.5) 57 035 (89.1) reference  
 Rural region 4 426 (13.5) 6 984 (10.9) 1.44 (1.23-1.69) <0.001
Health Status No. (%) No. (%)   
Charlson Comorbidity Index     
 0 25 595 (78.3) 47 912 (74.8) reference  
 1-2 6 448 (19.7) 14 448 (22.6) 0.88 (0.84-0.91) <0.001
 3-4 416 (1.3) 1 114 (1.7) 0.71 (0.63-0.82) <0.001
 >=5 230 (0.7) 545 (0.9) 0.84 (0.72-0.98) 0.02
Hospitalization in past year     
 No hospitalization 27 642 (84.6) 52 237 (81.6) reference  
 ≥ 1 hospitalization 5 047 (15.4) 11 782 (18.4) 0.97 (0.93-1.02) 0.21
Breast Related Health Care No. (%) No. (%)   
Breast biopsy in past year     
 No biopsy 32 404 (99.1) 63 514 (99.2) reference  
 Biopsy 285 (0.9) 505 (0.8) 1.38 (1.15-1.65) <0.001
Mammogram in past year     
 Yes 4 329 (13.2) 22 290 (34.8) reference  
 No (due for mammogram) 28 360 (86.8) 41 729 (65.2) 3.94 (3.47-4.48) <0.001
Self-referral screening available     
 No 17 945 (54.9) 31 922 (49.9) reference  
 Yes 14 744 (45.1) 32 097 (50.1) 0.86 (0.82-0.89) <0.001
  1. aEstimated using the generalized estimating equation (GEE) extension of multivariate logistic regression, adjusting for physician sex, specialty, medical school within Canada, years of practice, MCCQE1 exam score, MCCQE2 communication sub-score, MCCQE2 data acquisition sub-score, patient age, patient percent with high school diploma, patient income, patient region urban versus rural, patient comorbidity, patient prior hospitalization, patient breast biopsy or mammogram in previous year, and self referral for screening available. 499 patients had missing values for percent with a high school diploma and family income, and were not included in the regression model.
  2. bin Canadian Dollars.
\