Table 2.
Multiple logistic regression models assessing characteristics associated with specialty physicians who refer or enroll patients in clinical trials (N = 1533)*
| Characteristic | Medical and radiation oncologists (n = 721) |
Surgeons (n = 794) |
||||
| Adjusted OR† (95% CI) | Adjusted percentage‡ (95% CI) | P§ | Adjusted OR† (95% CI) | Adjusted percentage‡ (95% CI) | P§ | |
| Physicians | ||||||
| Specialty | ||||||
| Medical oncology | 5.2 (2.8 to 9.4) | 87.9 (84.9 to 91.0) | <.001 | — | — | .010 |
| Radiation oncology | 1.00 (referent) | 66.2 (59.1 to 73.3) | — | — | ||
| General surgery | — | — | 1.00 (referent) | 31.9 (27.8 to 36.0) | ||
| Colorectal surgery | — | — | 1.0 (0.5 to 2.0) | 31.3 (19.6 to 43.1) | ||
| Thoracic surgery | — | — | 1.3 (0.7 to 2.2) | 35.9 (27.2 to 44.6) | ||
| Surgical oncology | — | — | 3.9 (1.8 to 8.6) | 58.3 (43.2 to 73.3) | ||
| Other surgical subspecialty | — | — | 2.1 (0.8 to 5.5) | 46.0 (27.8 to 64.2) | ||
| Age, y | ||||||
| <50 | 1.7 (1.0 to 3.1) | 82.7 (78.7 to 86.7) | .203 | 1.2 (0.7 to 1.9) | 36.6 (31.8 to 41.4) | .413 |
| 50–59 | 1.5 (0.8 to 2.8) | 81.4 (76.8 to 86.1) | 0.9 (0.5 to 1.5) | 31.9 (26.7 to 37.2) | ||
| ≥60 | 1.00 (referent) | 76.0 (69.4 to 82.6) | 1.00 (referent) | 33.9 (27.0 to 40.7) | ||
| Sex | ||||||
| Male | 1.00 (referent) | 79.6 (76.2 to 82.9) | .068 | 1.00 (referent) | 34.3 (30.8 to 37.8) | .726 |
| Female | 1.8 (1.0 to 3.3) | 86.0 (80.6 to 91.3) | 1.1 (0.6 to 2.1) | 36.2 (26.0 to 46.4) | ||
| US or Canadian medical school graduate | ||||||
| Yes | 1.00 (referent) | 79.9 (76.7 to 83.2) | .136 | 1.00 (referent) | 34.0 (30.5 to 37.5) | .355 |
| No | 1.6 (0.9 to 3.0) | 85.1 (79.6 to 90.7) | 1.3 (0.7 to 2.4) | 38.9 (28.7 to 49.1) | ||
| Practice setting | ||||||
| Income increases as a result of enrolling patients on clinical trials | ||||||
| Yes | 1.9 (0.7 to 5.1) | 87.0 (78.3 to 95.6) | .214 | 0.9 (0.3 to 2.9) | 33.5 (14.6 to 52.5) | .919 |
| No | 1.00 (referent) | 80.3 (77.3 to 83.4) | 1.00 (referent) | 34.5 (31.2 to 37.9) | ||
| Practice type | ||||||
| Office based, solo | 0.3 (0.1 to 0.6) | 65.5 (53.8 to 77.2) | .018 | 0.8 (0.4 to 1.6) | 32.5 (23.2 to 41.8) | .673 |
| Office based, group‖ | 0.8 (0.4 to 1.4) | 81.0 (76.4 to 85.6) | 0.8 (0.5 to 1.3) | 32.9 (27.0 to 38.7) | ||
| Hospital based | 1.00 (referent) | 83.8 (79.6 to 88.0) | 1.00 (referent) | 36.0 (31.2 to 40.8) | ||
| Practice size, No. of physicians | ||||||
| 1–5 | 1.00 (referent) | 78.4 (74.3 to 82.5) | .290 | 1.00 (referent) | 32.2 (27.6 to 36.9) | .251 |
| 6–10 | 1.6 (0.8 to 3.1) | 84.0 (77.7 to 90.4) | 1.1 (0.6 to 1.8) | 33.0 (25.7 to 40.4) | ||
| 11–20 | 1.7 (0.6 to 4.4) | 84.5 (75.3 to 93.6) | 1.8 (0.9 to 3.5) | 42.6 (31.4 to 53.8) | ||
| >20 | 2.4 (0.8 to 7.3) | 88.0 (79.3 to 96.7) | 1.5 (0.9 to 2.4) | 38.8 (31.1 to 46.4) | ||
| Study site | ||||||
| Five large HMOs¶ | 1.00 (referent) | 83.0 (73.9 to 92.0) | .129 | 1.00 (referent) | 46.9 (34.4 to 59.3) | .004 |
| Northern California counties | 1.0 (0.4 to 2.6) | 82.6 (76.9 to 88.4) | 0.5 (0.3 to 1.0) | 34.3 (28.1 to 40.5) | ||
| Los Angeles County | 0.8 (0.3 to 2.1) | 80.2 (74.4 to 85.9) | 0.4 (0.2 to 0.9) | 30.9 (24.4 to 37.3) | ||
| State of Alabama | 1.1 (0.4 to 3.4) | 84.2 (77.4 to 91.1) | 0.3 (0.1 to 0.7) | 25.0 (17.7 to 32.2) | ||
| State of Iowa | 1.0 (0.3 to 3.1) | 83.0 (74.8 to 91.1) | 1.0 (0.4 to 2.6) | 46.6 (32.2 to 61.1) | ||
| State of North Carolina | 1.0 (0.3 to 3.2) | 83.2 (74.5 to 91.8) | 0.9 (0.4 to 2.1) | 45.6 (36.4 to 54.9) | ||
| Veteran’s Administration | 0.3 (0.1 to 0.8) | 65.9 (55.0 to 76.9) | 0.4 (0.1 to 1.0) | 29.0 (18.2 to 39.9) | ||
| Practice affiliated with a CCOP | ||||||
| Yes | 2.1 (1.2 to 3.7) | 86.1 (81.7 to 90.5) | .016 | 1.4 (1.0 to 2.0) | 38.3 (32.6 to 44.0) | .094 |
| No | 1.00 (referent) | 77.8 (73.9 to 81.7) | 1.00 (referent) | 32.6 (28.6 to 36.6) | ||
| Practice affiliated with an NCI-designated cancer center | ||||||
| Yes | 2.8 (1.2 to 6.4) | 89.7 (83.5 to 95.8) | .023 | 1.7 (1.1 to 2.6) | 41.4 (34.9 to 47.9) | .010 |
| No | 1.00 (referent) | 79.2 (75.9 to 82.5) | 1.00 (referent) | 31.6 (27.7 to 35.5) | ||
| Physician practice style | ||||||
| Percentage of patients in managed care, quartiles | ||||||
| 0–20 | 0.9 (0.4 to 1.9) | 79.6 (73.8 to 85.4) | .759 | 1.7 (0.9 to 2.9) | 37.5 (30.1 to 44.8) | .190 |
| 21–49 | 1.3 (0.5 to 3.3) | 83.8 (77.6 to 90.0) | 1.8 (1.0 to 3.4) | 39.4 (31.6 to 47.1) | ||
| 50–78 | 0.9 (0.4 to 2.0) | 79.9 (73.8 to 86.1) | 1.5 (0.9 to 2.5) | 36.1 (30.0 to 42.3) | ||
| 79–100 | 1.00 (referent) | 80.8 (74.0 to 87.7) | 1.00 (referent) | 29.1 (23.4 to 34.8) | ||
| Teaches medical students and/or residents | ||||||
| Yes | 2.7 (1.5 to 5.0) | 87.4 (83.2 to 91.7) | .003 | 1.6 (1.1 to 2.4) | 38.4 (33.7 to 43.1) | .019 |
| No | 1.00 (referent) | 75.8 (71.2 to 80.4) | 1.00 (referent) | 30.1 (25.2 to 35.0) | ||
| Attends tumor board meetings | ||||||
| Weekly | 1.1 (0.5 to 2.5) | 82.7 (79.3 to 86.2) | .076 | 2.9 (1.8 to 4.6) | 42.3 (36.3 to 48.3) | <.001 |
| Monthly | 0.6 (0.2 to 1.3) | 74.3 (67.6 to 81.1) | 2.1 (1.3 to 3.5) | 36.6 (30.8 to 42.3) | ||
| Quarterly or less frequently | 1.00 (referent) | 81.7 (72.9 to 90.5) | 1.00 (referent) | 23.9 (18.3 to 29.5) | ||
| No. of colorectal or lung cancer patients seen per month | ||||||
| <5 | 1.00 (referent) | 75.2 (66.5 to 83.9) | .400 | 1.00 (referent) | 28.7 (24.7 to 32.8) | <.001 |
| 5–9 | 1.8 (0.9 to 3.8) | 82.6 (76.8 to 88.3) | 2.0 (1.3 to 3.3) | 41.8 (33.5 to 50.1) | ||
| 10–19 | 1.9 (0.8 to 4.3) | 82.9 (77.3 to 88.5) | 2.1 (1.2 to 3.7) | 42.6 (32.7 to 52.5) | ||
| ≥20 | 1.5 (0.7 to 3.2) | 80.4 (75.9 to 84.9) | 5.1 (2.3 to 11.3) | 60.1 (45.2 to 75.1) | ||
| Time spent with a newly diagnosed cancer patient, min | ||||||
| <60 | 1.00 (referent) | 81.5 (76.2 to 86.8) | .792 | 1.00 (referent) | 34.9 (31.2 to 38.7) | .584 |
| ≥60 | 0.9 (0.5 to 1.6) | 80.6 (77.1 to 84.1) | 0.87 (0.5 to 1.4) | 32.7 (25.5 to 39.8) | ||
*
— = not applicable; CCOP = Community Clinical Oncology Program; CI = confidence interval; HMO = health maintenance organization; NCI = National Cancer Institute; OR = odds ratio.
†
Adjusted using multiple logistic regression for all other characteristics in table.
‡
Adjusted using predictive margins for all other characteristics in table.
§
Based on an overall Wald χ2 test for association obtained from the logistic regression models (two-sided).
‖
Includes office-based HMOs, community health centers, and other (a questionnaire item)
¶
Group Health Cooperative of Puget Sound, Harvard Pilgrim Health Care, Henry Ford Health System, Kaiser Permanente Northwest, and Kaiser Permanente Hawaii.