Analysis Of SEER-Medicare Data Research Articles

Visit and treatment patterns over time among elderly patients (pts) with M1 prostate cancer (PC): An analysis using SEER-Medicare

5170 Background: Treatment options for the management of pts diagnosed with M1 PC have evolved over time. It is unknown whether visits to medical oncologists/hematologist oncologists (MOH) and treatment patterns are also changing over time. The objective of this study is to examine changes in visit and treatment patterns over time. Methods: A retrospective analysis of SEER Medicare data included pts diagnosed with M1 PC from 1994–2002 (age > 65 years) and residing in SEER registries that were present for the entire period. The study included pts with a post-diagnosis visit to a urologist; pts who saw a MOH prior to the urologist visit were excluded. Pts were grouped as 1) no MOH visit, 2) MOH visit w/in 3 mos of a urologist visit, 3) MOH visit => 3 mos after a urologist visit. Treatment with hormone therapy or chemotherapy was defined as 1) none received; 2) timely (i.e. within 6 mos of diagnosis); and 3) delayed, i.e. => 6 mos following the diagnosis. Time periods were defined as early (1994–1996), middle (1997–1999) and late (2000–2002). Results: 3,269 pts (mean age 77, 81% white) were available for analysis. Ninety-three percent of pts received treatment. Thirty-eight percent of pts saw a MOH during the study period; over the study period (early; middle; late) 13% (10%; 13%; 17%) of pts had a timely visit to the MOH and 25% (24%; 28%; 25%) had a delayed visit to the MOH. The proportion of patients seeing a MOH increased (34%; 41%; 42%, p < 0.001) and the proportion of treated pts increased (93%; 93%; 95%, p = 0.03) over the early, middle, and late periods. Conclusions: Approximately one-third of patients with M1 disease and a post-diagnosis urologist visit also see a medical oncologist. The vast majority of these patients receive treatment. Over time, a larger proportion of patients are seeing medical oncologists. [Table: see text]

Total societal costs attributable to the prevention of recurrent HER2+ breast cancer.

Abstract Abstract #6101 Introduction: Adjuvant therapies are available to prevent breast cancer recurrence that may reduce morbidity and mortality. However, the associated savings to society have not been quantified, particularly for HER2+ patients. Methods: We reviewed the literature (PubMed) to estimate the costs attributable to breast cancer recurrence including direct (medical and non-medical) and indirect (loss income, productivity, informal care) costs. The following sources were used to estimate the recurrence cost. The 10-year average direct medical costs of recurrence were obtained from an analysis of SEER-Medicare data (Stokes 2008). This study compared women with recurrence to similar women without recurrence and accounts for differential costs associated with survival (ie, it did not estimate only the cost of treating recurrent cancer). In the initial year after cancer diagnosis, women required an additional 66 hours of personal time to obtain medical services (Yabroff 2007). After recurrence, an excess of 9% of patients would leave the work force if they were aged 50-64 (Lidgren 2007). Women were found to take an additional 8.5 months away from work during the 3 years following a recurrence compared with those without recurrence (Drolet 2005). The intensity of informal care was similar (0.8 hour/week) in the first year after recurrence as it was for women without recurrence (Lidgren 2007). The valuation of utilization was conducted from a societal perspective and costs reported as 2008 US dollars. Because certain costs were relevant to specific ages, we estimated costs in 3 age cohorts (30-49, 50-69, and ≥70). Lost income from early retirement was calculated as the number of years retired from work before the age 65 times the annual average income derived using the national average wage and benefit data. Patient time required for receiving care, time absent from work, and caregiver time were also valued in the same way. We combined this information into a simulation used to estimate the number of HER2+ women diagnosed in one calendar year in the United States, and whose recurrence could be prevented with trastuzumab. The model accounted for variability of clinical and economic inputs by sampling from distributions using 5,000 replications. The mean costs per person and the middle 95% of the distribution were reported. Results: For ages 50-69, early retirement costs were $39,600. Costs due to work absences were $41,600 (age 30-49) and were $31,200 (age 50-69). The total societal costs attributable to a single recurrence were $60,400 ($32,000 to $129,000) for ages 30-49, $89,600 ($45,000 to $203,000) for ages 50-69, and $18,900 ($13,000 to $25,000) for ages ≥70. Overall, the savings from preventing recurrence with trastuzumab were estimated at over $167 million ($70 million to $385 million) per year of diagnosed cases of HER2+ tumors. Conclusion: Therapies that can prevent breast cancer recurrence can lead to substantial savings to society and represent a considerable opportunity cost. Citation Information: Cancer Res 2009;69(2 Suppl):Abstract nr 6101.

Sensitivity of the hazard ratio to nonignorable treatment assignment in an observational study

In non-randomized studies, estimation of treatment effects generally requires adjustment for imbalances in observed covariates. One such method, based on the propensity score, is useful in many applications but may be biased when the assumption of strongly ignorable treatment assignment is violated. Because it is not possible to evaluate this assumption from the data, it is advisable to assess the sensitivity of conclusions to violations of strong ignorability. Lin et al. (Biomet. 1998; 54:948-963) have implemented this idea by investigating how an unmeasured covariate may affect the conclusions of an observational study. We extend their method to assess sensitivity of the treatment hazard ratio to hidden bias under a range of covariate distributions. We derive simple formulas for approximating the true from the apparent treatment hazard ratio estimated under a specific survival model, and assess the validity of these formulas in simulation studies. We demonstrate the method in an analysis of SEER-Medicare data on the effects of chemotherapy in elderly colon cancer patients.