Table 4.
Mixed effects modeling of sensitivity and specificity by each response about malpractice claims and concerns adjusting for patient characteristics*
| Variable | Sensitivity
|
Specificity
|
||||
|---|---|---|---|---|---|---|
| OR | 95% CI | P | OR | 95% CI | P | |
| Medical malpractice insurance | ||||||
| Self pay/other | 1.00 | (referent) | .93 | 1.00 | (referent) | .81 |
| Facility pays | 0.98 | (0.60 to 1.59) | 0.96 | (0.70 to 1.32) | ||
| Ever had a malpractice claim | ||||||
| No claims | 1.00 | (referent) | .28 | 1.00 | (referent) | .25 |
| Non-mammogram related | 0.79 | (0.59 to 1.06) | 1.19 | (0.97 to 1.47) | ||
| Mammogram related | 0.86 | (0.60 to 1.22) | 1.11 | (0.84 to 1.46) | ||
| Concerned about malpractice | ||||||
| Disagree | 1.00 | (referent) | .03 | 1.00 | (referent) | .14 |
| Neutral | 0.77 | (0.43 to 1.35) | 1.11 | (0.73 to 1.69) | ||
| Agree | 1.21 | (0.73 to 2.01) | 0.87 | (0.60 to 1.25) | ||
| Malpractice influences recommendation for ultrasound | ||||||
| Not changed | 1.00 | (referent) | .52 | 1.00 | (referent) | .46 |
| Moderately increased | 1.18 | (0.87 to 1.59) | 0.92 | (0.74 to 1.15) | ||
| Greatly increased | 1.19 | (0.77 to 1.82) | 0.81 | 0.58 to 1.13) | ||
| Interpreting mammograms is tedious | ||||||
| Disagree | 1.00 | (referent) | .56 | 1.00 | (referent) | .70 |
| Neutral | 0.82 | (0.57 to 1.18) | 0.95 | (0.72 to 1.25) | ||
| Agree | 0.94 | (0.70 to 1.28) | 0.91 | (0.73 to 1.13) | ||
| Worry when not sure of a mammogram | ||||||
| Disagree | 1.00 | (referent) | .52 | 1.0 | (referent) | .25 |
| Agree | 0.90 | (0.67 to 1.23) | 0.88 | (0.70 to 1.10) | ||
*
Odds ratio (OR), 95% confidence intervals (CI), and omnibus Wald test P values (two-sided) were calculated using logistic regression, adjusting for patient age, breast density, and prior mammography. Additionally, random effects for radiologist were estimated with a likelihood ratio test.