| To Register SMDM Homepage | |
|
|
|
||||
Purpose: Aggressive cancer screening has increased the number of patients diagnosed with localized tumors. Many are elderly and have pre-existing comorbidities. Yet, treatment guidelines are surprisingly vague on how competing risks should factor into treatment choice. Traditionally, concerns about overtreatment in cancer have focused on patients diagnosed with late-stage tumors. The increase in early detection raises the issue in a different guise. The purpose of this study is to examine the degree to which prognosis (i.e. the risk of death from competing causes) influences treatment of patients diagnosed with localized prostate tumors.
Methods: SEER-Medicare data were used for the study. The analysis has three steps. 1) A Weibull model was used to estimate the impact of age and comorbidities on survival time among non-cancer Medicare enrollees in 1992 (to allow for 10 years of follow-up; N = 44,880). 2) Parameter estimates from the model were used to predict life expectancy in the absence of cancer for prostate cancer patients diagnosed in 1997-1999 with localized tumors (N = 32,798). This imputed life expectancy measure captures “competing risks”. 3) The association between the imputed life expectancy measure and treatment (prostatectomy versus other) was assessed via logit regression.
Results: Prognosis is a strong predictor of treatment type. A one year increase from the mean value of life expectancy for a patient with a moderately differentiated tumor is associated with a one percentage point increase in the likelihood of surgery (p<0.01). The relationship is highly non-linear, and it varies in the expected direction by tumor grade. The strength of the relationship does not vary by race or SES.
Conclusion: Findings increase confidence that physicians consider prognosis when prescribing treatment and that aggressive screening is not leading to substantial over-treatment.
See more of Oral Concurrent Session B - Health Services Research
See more of The 26th Annual Meeting of the Society for Medical Decision Making (October 17-20, 2004)