ments of posttreatment complications have been found to greatly underestimate the rate reported by patients themselves (Litwin et al., 1998). This finding may suggest that a potentially important area for quality assessment is differences across providers in the patient counseling process.
A specific recommendation from the American Urological Association's (AUA's) clinical guidelines on the management of clinically localized prostate cancer is that all alternative treatment modalities (radical prostatectomy, radiation therapy external beam, interstitial treatment—and expectant management) should be presented to every patient (Middleton et al., 1995). Thus, a potential quality indicator could include whether these recommendations are followed by urologists.
Estimates of complications resulting from primary treatment of prostate cancer vary widely across facilities even when stratifying by treatment modality: surgery, external beam radiation, or brachytherapy (interstitial radiation treatment or seed implants) (Middleton et al., 1995). After radical prostatectomy, rates of stress incontinence range from less than 10 to 50 percent and impotence rates range from 25 to 100 percent across series reports. Complications following external beam radiation included proctitis, with rates ranging from less than 10 percent to more than 50 percent; cystitis, ranging from 0 to 80 percent; and impotence, ranging from less than 10 percent to nearly 40 percent. Similarly, complication rates reported for brachytherapy range from 0 to 75 percent for proctitis, less than 10 percent to 90 percent for cystitis, and less than 10 percent to 75 percent for impotence.
While these widely varying complication rates may reflect differences in quality of care, it is difficult to draw conclusions based on this type of information (Wasson et al., 1993). First, there may be differences in the way the data were collected, which could account for variations in rates of complications. Second, there may be systematic differences in patient case mix (disease severity and comorbidities) across facilities, and these differences must be accounted for before comparing outcomes across institutions. Even if series reports could be adjusted for case mix, there is usually little information available to link differences in results to differences in the technical process of care. Finally, these series report data from only a small number of providers, often large academic clinics. The results for such providers may not represent those of other institutions or clinics.
An early American College of Radiology Patterns of Care study examined the association between types of radiation equipment (a structure measure) and localized prostate cancer treatment outcomes (Hanks et al., 1985). Facilities that used cobalt units were found to have higher stage-adjusted rates of disease recurrence than facilities that used linear accelerators or betatrons. The use of cobalt equipment was also correlated with other structural indicators: these facilities had lower percentages of patients who were staged, had lower staff-patient ratios, and were more