Info

(0.5%)

cancers were minimal; using the more rigorous Martin-Gallager criteria (invasive cancers smaller than 6 mm) (Martin and Gallager, 1971), there were 150 (34.2 percent) minimal cancers.

Table 6.13 summarizes selected mammography screening results separately for each radiologist in the practice demonstrating considerable variability in performance. These data indicate that the most experienced radiologist generated more biopsies, and identified more nonpalpable and early-stage cancers per abnormal screening interpretation.

6.3.3 How to Interpret Audit Results

For the first medical audit of a mammography practice, interpretation of audit results is based primarily on comparison with parallel data from previously published reports. However, there are pitfalls inherent in such an exercise because substantial variations in results can arise due to differences in the patient populations studied and due to differences in the methods and definitions used to compile the data.

By far the most confusing situation occurs when comparing audit results derived from screening and problem-solving examinations. As stated previously, the definitions basic to data analysis may vary widely for these two types of examination. Furthermore, the likelihood of finding breast cancer is much greater for problem-solving examinations, which often involve women having palpable masses. Either or both of these factors may cause substantial differences in observed audit statistics (Dee and Sickles, 2001). Therefore, results of any audit must be interpreted in the context of (1) the percentage of symptomatic women examined and (2) whether (and to what extent) problem-solving examinations are intermixed with screening studies. Problem-solving and screening data should be segregated during auditing. If this is not possible, analysis of combined results should be based on known differences between problem-solving and screening examinations (Sohlich et al., 2002).

There also are pitfalls in the interpretation of audit data when results of seemingly similar types of examinations are compared. For example, the rates of detecting prevalent and incident cancers are higher in the expert screening practice than in large population-based screening studies (Tabar et al., 1984). It might be tempting, albeit misleading, to conclude that the screening practice results indicate superior performance.

Table 6.13—Overall screening results for each radiologist expressed in terms of number of abnormal interpretations.

Radiologist

Abnormal Interpretations

Biopsies Performed3

Cancers Detecteda

Nonpalpable Cancers Detecteda

Stage 0 + 1 Cancers Detecteda

Table 6.13—Overall screening results for each radiologist expressed in terms of number of abnormal interpretations.

Radiologist

Abnormal Interpretations

Biopsies Performed3

Cancers Detecteda

Nonpalpable Cancers Detecteda

Stage 0 + 1 Cancers Detecteda

0 0

Post a comment