Initiative to Improve Mammogram Interpretation.

Introduction Mammography quality is a significant issue of national concern. The Mammography Quality Standards Act (MQSA) extends regulation to an unprecedented level of detail in the practice of medicine; however, the Act pertains to the technical quality of mammography and is largely silent on the critical issue of the radiologist’s proficiency in interpreting examination results. The reason that no measure of radiologist proficiency is required may be partly because radiologists often do not know whether a patient whose mammogram they interpreted received a diagnosis of cancer months or years later or lived a long, cancer-free life. Kaiser Permanente (KP), with its well-established databases of patient information, is unique in its ability to monitor and track patient outcome. Of 370,000 members in the KP Colorado Region, approximately 101,000 are women who are eligible for mammography. For these women, breast cancer is a leading cause of cancer-related deaths. During the past five years, KP Colorado has averaged more than 80% penetration for screening mammography by Health Employer Data Information Set (HEDIS) criteria. However, internal quality audits in late 1995 indicated that breast cancer detectable on mammograms was sometimes being missed. In 1996, KP Colorado began to implement a multifaceted initiative to reduce variation and improve accuracy in the interpretation of mammograms. The initiative was conceived and sustained by the radiology leadership team, including staff from Health Plan and medical groups, with extensive sponsorship from Kaiser Foundation Health Plan and Operations. The integrity of vision among top management and the radiology department informed an organizational team spirit that fueled this initiative from its inception. The initiative team members are listed in Table 1. The project consisted of organizational redesign, quality improvement, and performance management and reflected many innovations in health care delivery, patient safety, continuous quality improvement, and development of subspecialty practice in radiology. The objective of this initiative was to maximize the number of cancerous lesions detected at an early, curable stage by achieving industry-leading performance in mammographic diagnosis of breast cancer. To achieve this objective, we investigated three issues: reasons for differing levels of performance among radiologists interpreting mammograms; the potential for improvement and barriers to realizing this potential; and innovations that result in sustained improvement in performance.