Abstract：
BACKGROUND: Medical record abstraction (MRA) and self-report questionnaires are two methods frequently used to ascertain cancer treatment information. Prior studies have shown excellent agreement between MRA and self-report, but it is unknown how a recall window longer than 3 years may affect this agreement.
METHODS: The Women\'s Environmental Cancer and Radiation Epidemiology (WECARE) Study is a multicenter, population-based case-control study of controls with unilateral breast cancer individually matched to cases with contralateral breast cancer. Participants who were diagnosed with a first primary breast cancer from 1985 to 2008 before the age of 55 years completed a questionnaire that included questions on treatment. First primary breast cancer treatment information was abstracted from the medical record from radiation oncology clinic notes for radiation treatment and from systemic adjuvant treatment reports for hormone therapy and chemotherapy. Agreement between MRA and self-reported treatment was assessed with the kappa statistic and corresponding 95% confidence intervals (CIs).
RESULTS: A total of 2808 participants with MRA and self-reported chemotherapy treatment information, 2733 participants with MRA and self-reported hormone therapy information, and 2905 participants with MRA and self-reported radiation treatment information were identified. The median recall window was 12.5 years (range, 2.8-22.2 years). MRA and self-reported treatment agreement was excellent across treatment modalities (kappachemo, 98.5; 95% CI, 97.9-99.2; kappahorm, 87.7; 95% CI, 85.9-89.5; kapparad, 97.9; 95% CI, 97.0-98.7). There was no heterogeneity across recall windows (pchemo = .46; phorm = .40; prad = .61).
CONCLUSIONS: Agreement between self-reported and MRA primary breast cancer treatment modality information was excellent for young women diagnosed with breast cancer and was maintained even among women whose recall window was more than 20 years after diagnosis.
METHODS: The Women\'s Environmental Cancer and Radiation Epidemiology (WECARE) Study is a multicenter, population-based case-control study of controls with unilateral breast cancer individually matched to cases with contralateral breast cancer. Participants who were diagnosed with a first primary breast cancer from 1985 to 2008 before the age of 55 years completed a questionnaire that included questions on treatment. First primary breast cancer treatment information was abstracted from the medical record from radiation oncology clinic notes for radiation treatment and from systemic adjuvant treatment reports for hormone therapy and chemotherapy. Agreement between MRA and self-reported treatment was assessed with the kappa statistic and corresponding 95% confidence intervals (CIs).
RESULTS: A total of 2808 participants with MRA and self-reported chemotherapy treatment information, 2733 participants with MRA and self-reported hormone therapy information, and 2905 participants with MRA and self-reported radiation treatment information were identified. The median recall window was 12.5 years (range, 2.8-22.2 years). MRA and self-reported treatment agreement was excellent across treatment modalities (kappachemo, 98.5; 95% CI, 97.9-99.2; kappahorm, 87.7; 95% CI, 85.9-89.5; kapparad, 97.9; 95% CI, 97.0-98.7). There was no heterogeneity across recall windows (pchemo = .46; phorm = .40; prad = .61).
CONCLUSIONS: Agreement between self-reported and MRA primary breast cancer treatment modality information was excellent for young women diagnosed with breast cancer and was maintained even among women whose recall window was more than 20 years after diagnosis.
摘要：
背景：医疗记录摘要（MRA）和自我报告问卷是确定癌症治疗信息的两种常用方法。先前的研究表明，MRA和自我报告之间有很好的一致性，但尚不清楚超过3年的召回窗口会如何影响该协议。
方法：妇女环境癌症和辐射流行病学（WECARE）研究是一个多中心，基于人群的病例对照研究,对单侧乳腺癌对照与对侧乳腺癌病例进行单独匹配.从1985年到2008年，在55岁之前被诊断出患有首次原发性乳腺癌的参与者填写了一份问卷，其中包括有关治疗的问题。首次原发性乳腺癌治疗信息是从放射治疗的放射肿瘤学临床记录以及激素治疗和化疗的全身辅助治疗报告中提取的。使用kappa统计量和相应的95％置信区间（CI）评估MRA与自我报告治疗之间的一致性。
结果：共有2808名患者接受MRA和自我报告的化疗治疗信息，2733名具有MRA和自我报告激素治疗信息的参与者，并确定了2,905名具有MRA和自我报告放射治疗信息的参与者.召回窗口中位数为12.5年(范围，2.8-22.2年）。MRA和自我报告的治疗协议在不同的治疗方式上表现优异(kappachemo，98.5；95%CI，97.9-99.2；卡帕霍姆，87.7；95%CI，85.9-89.5；kapparad，97.9；95%CI，97.0-98.7）。在召回窗口中没有异质性(chemo=.46；phorm=.40；prad=.61)。
结论：对于诊断为乳腺癌的年轻女性，自我报告和MRA原发性乳腺癌治疗方式信息之间的一致性非常好，甚至在诊断后20年以上的召回窗口的女性中也能保持。
方法：妇女环境癌症和辐射流行病学（WECARE）研究是一个多中心，基于人群的病例对照研究,对单侧乳腺癌对照与对侧乳腺癌病例进行单独匹配.从1985年到2008年，在55岁之前被诊断出患有首次原发性乳腺癌的参与者填写了一份问卷，其中包括有关治疗的问题。首次原发性乳腺癌治疗信息是从放射治疗的放射肿瘤学临床记录以及激素治疗和化疗的全身辅助治疗报告中提取的。使用kappa统计量和相应的95％置信区间（CI）评估MRA与自我报告治疗之间的一致性。
结果：共有2808名患者接受MRA和自我报告的化疗治疗信息，2733名具有MRA和自我报告激素治疗信息的参与者，并确定了2,905名具有MRA和自我报告放射治疗信息的参与者.召回窗口中位数为12.5年(范围，2.8-22.2年）。MRA和自我报告的治疗协议在不同的治疗方式上表现优异(kappachemo，98.5；95%CI，97.9-99.2；卡帕霍姆，87.7；95%CI，85.9-89.5；kapparad，97.9；95%CI，97.0-98.7）。在召回窗口中没有异质性(chemo=.46；phorm=.40；prad=.61)。
结论：对于诊断为乳腺癌的年轻女性，自我报告和MRA原发性乳腺癌治疗方式信息之间的一致性非常好，甚至在诊断后20年以上的召回窗口的女性中也能保持。
