P4-10-04: Automated Breast Cancer Risk Assessment: Identifying High Risk Women in the Primary Care Setting

Background: Despite the availability of risk assessment tools and risk-reducing interventions, high risk women are not routinely identified in clinical practice, and the few that are, rarely choose interventions — often due to misperception of risk. Identification of patients at risk needs to begin in the primary care setting, rather than the breast health specialty world. To accomplish this, we developed a web-based tool that provides automated risk assessment and personalized decision support designed for collaborative use between patients and clinicians in the primary care setting. We assessed the feasibility and efficacy of using this tool in a primary care clinic at an academic hospital to identify women at high.

Methods: Women aged 40 to 65 years of age were recruited from a schedule of patients attending annual physicals at a primary care clinic in an academic hospital. Patients with a history of breast cancer, genetic testing, or chemoprevention education were excluded. Information used to assess breast cancer risk was gathered from medical record review and phone interviews when necessary. Risk assessment was performed on all patients using four risk assessment models (Gail, Claus, BRCAPRO, and BCSC Density). Patients were randomized to view the decision aid either before their appointment or with their PCP during their appointment. Prior to each visit, providers received a risk report that summarized patient risk and recommended referrals. Outcomes were gathered from surveys administered to patients before and after appointments, and to providers after appointments.

Results: Over 4-months, 98 women were approached to reach the enrollment goal of 60 women (61%). 24/60 (40%) patients were identified to be high risk for breast cancer using standard high risk thresholds. 9/60 (15%) of patients fit criteria for referral to genetic counseling, while 15/60 (25%) fit the criteria for referral to a BC specialist. Out of the 24 patients who fit the referral criteria, 17 (71%) were referred to a high risk clinic by their PCPs. 9/17 (53%) patients followed through in scheduling the appointment within 4 months of the referral date. The PCPs’ perceptions of these patients’ risk was in line with the calculated risk for 21 (88%) of the patients. A discussion regarding breast cancer risk reduction occurred with 22/24 (92%) of these patients during the visit, while the PCPs chose to use the decision aid with only 13/24 (54%) of them. PCP chose to view the decision aid during the visit with more high risk patients than average risk patients (p=0.04). Use of the DA during the appointment did not alter provider satisfaction with visit. A majority of patients thought the DA was helpful in making a decision and would recommend it to women like them.

Conclusions: Performing personalized risk assessment and use of the decision aid in the primary care setting was feasible and acceptable. These results suggest risk assessment alone may be enough to encourage a discussion about breast cancer risk reduction for some providers. This method of risk assessment and decision support holds promise in the effort to reduce the incidence and burden of breast cancer.

Citation Information: Cancer Res 2011;71(24 Suppl):Abstract nr P4-10-04.

PD06-08: Projecting the Impact of Adopting Trial Results

Background: There is often a substantial delay between the presentation of clinical trial results and change in practice. Reasons for slow adoption include the disruption of practice routines, doubt about the validity of results, financial disincentive, and resistance to change. We propose a decision-making process to inform adoption that considers three factors: likelihood the trial's conclusions will not change with longer follow up, and consequences of early and late adoption. We apply this framework to the published 4-year results of the international TARGIT-A trial, which compares intraoperative radiation therapy (IORT) to external beam radiation therapy (EBRT).

Methods: To find whether the trial results will remain robust, we reviewed the TARGIT-A trial's annual hazards for local recurrence (LR). We then reviewed 5–10 year LR rates from recent clinical trials with similar patient populations undergoing RT, no RT or partial breast RT. To assess the impact of an early change in practice, a Markov model was used to evaluate life expectancy, quality adjusted life years (QALYs), and cost. Sensitivity analysis estimated the impact of varying the LR rate 1–20x the TARGIT-A 4 year results. To estimate the impact of late adoption, we generated the expected number of N0, grades 1 & 2, ductal cancers in postmenopausal women (>50 yrs) from SEER and US Census Bureau data. Using Medicare rates, costs of EBRT vs. IORT were compared and potential savings to the health care system calculated.

Results: The TARGIT-A peak hazard for LR for IORT and EBRT occurred at 3 years. In the START and ATAC trials, the peak local recurrence Kaplan Meier estimates are between 2 and 3 years. In other trials (ATAC, and in the 2005 RT Lancet overview) there is no second peak of recurrence, making it unlikely that longer follow up will change the conclusion of the TARGIT-A results. The LRR in similar trials have dropped steadily over time for post menopausal, stage I patients with or without RT. Impact of early adoption: If treatment were adopted early, the impact on life expectancy and QALYs would be minimal unless IORT LR rate exceeds 20% over a 10 year period of time. Impact of late adoption: In the US, 45% of new breast cancer cases fit the described population. EBRT costs $6,400 more than IORT per patient on average. If adoption were delayed 5 years to allow the trial results to mature, 70,136 patients per year are expected to receive EBRT resulting in a societal burden of >$2.2 billion barring the capital investment required for new technology. This also assumes that patients have equal utility for a single dose of IORT as they do for 3–6 weeks of postoperative radiation therapy. If IORT utility is higher than that of EBRT, then the IORT strategy is both more effective and less costly.

Conclusion: The process of modeling the impact of both early and late adoption when considering the stability of trial results can serve as a tool to evaluate whether to change practice. This analysis was tested on the results of the TARGIT-A trial, and demonstrated that prompt adoption of the IORT intervention would cause minimal harm, provide an improved quality of life, and offer significant societal savings.

Citation Information: Cancer Res 2011;71(24 Suppl):Abstract nr PD06-08.

Decision analysis and Alzheimer disease: three case studies

Decision analysis may be useful to people facing Alzheimer disease (AD) decisions. The use of decision analysis in three such cases is reported. The first case involved a middle-aged person worried about early-onset AD and deciding whether to seek genetic testing. The analysis let the participant reject testing and consider innovative care options. The second case involved a middle-aged person concerned about later-onset AD. The analysis for her was more complex, and led to the assignment of some limited value on genetic testing for her. The third case revolved around a caregiver's treatment decisions for a patient with severe AD. It led her to recognize the importance of factors she had not previously considered. In each of the three cases, the intensive process of decision analysis appears to have improved the subject's decision.

Case management applications in Australia

Case management approaches in Australia were first introduced in the fields of intellectual disability and vocational rehabilitation in the 1970s and 1980s by major system reformers very much influenced by international trends in these particular fields. A broader diffusion occurred in the 1980s and 1990s with federal and state governments beginning to incorporate case management approaches into key policy documents, particularly in the fields of aging, mental health, and employment. The models of case management introduced have not been particularly sophisticated in terms of financial control nor have many models been specifically evaluated. They have, however, been introduced into a system in which they are underpinned by a basic safety net of universal health coverage and a noncontributory pension system funded out of general revenue. There is now fairly widespread diffusion of the concept and practice of case management across a broad range of client populations and an interest in more sophisticated implementation models.