Table 2.
Quality measures used in Initiative to Improve Mammogram Interpretation
| Quality measure | Goal | Benchmark |
|---|---|---|
| Proportion of cancers detected at stage 0 or 1 | 80% in 1998 85% by 2003 |
80%1 |
| Sensitivity | 80% | 73%1 |
| Cancers diagnosed per 1000 mammograms | >6 | 62 |
| Diagnosis of new, probably benign lesion | 7% in 1998 4% by 2003 |
5%3 |
| Recall rate for screening mammograms | ≤7%a | 8.3%1 |
| Positive predictive value | 25–40%a | 23–534–6 |
| Annual number of mammograms read per radiologist | >4000 | 4807 36008 |
| Cost per mammogram per relative value unit | ≤ Medicare rate | Medicare b 120–160% c |
| Radiologist satisfaction | >90% | 89.25%d |
a
Measures of individual variation are applied. Generally, interventions are conducted when variation exceeds 2 SD.
b
The 2003 Medicare reimbursement rate, applying the KP Colorado geographic practice cost index (GCPI), is $36.42.
c
Prevailing community commercial reimbursement rate, per KP Colorado External Medical Services Department.
d
As an average of the four indicators of overall physician satisfaction assessed in the Colorado Permanente Medical Group physician satisfaction survey.