School of Nursing, Queen's University, Kingston, Ontario, Canada.
L.S. Bloomberg Faculty of Nursing, University of Toronto, Toronto, Ontario, Canada.
Psychooncology. 2018 Nov;27(11):2602-2608. doi: 10.1002/pon.4839. Epub 2018 Aug 9.
Fear of cancer recurrence (FCR) is a common concern among cancer survivors, and the Fear of Cancer Recurrence Inventory (FCRI) is a frequently used measure to assess FCR. Given that the dimensionality of FCR has received recent debate, the overall goal of this secondary analysis was to re-examine the dimensionality of the FCRI using confirmatory factor analyses (CFA) to compare models of FCR, using data from a large sample of cancer survivors.
Three models of FCR (including unidimensional and multidimensional models of the FCRI) were informed by the literature and proposed a priori. Separate CFAs were conducted to test the fit of each model to the data, and models with acceptable fits were compared.
Of all the tested FCR models, a multidimensional first-order model aligned with the originally developed 7-subscale FCRI revealed the best fit to the data (χ = 3359.135, P < .0001, df = 795, RMSEA = 0.057 [0.055, 0.059], CFI = 0.897, TLI = 0.888). When this 7-factor structure was loaded onto a single, second-order factor of overall FCR, the model fit statistics were slightly poorer (χ = 3459.632, P < .0001, df = 807, RMSEA = 0.058 [0.056, 0.060], CFI = 0.893, TLI = 0.886). However, the difference between the models was significant (chi-square difference = 103.142, P < .0001, df = 12) indicating that the first-order model was a better fit to the data.
These results align with empirical and theoretical literature that supports the use of the FCRI as a multidimensional scale. Implications of results are discussed in light of FCR conceptualization and measurement.
癌症复发恐惧（FCR）是癌症幸存者的常见关注点，而癌症复发恐惧量表（FCRI）是评估 FCR 的常用工具。鉴于 FCR 的维度最近受到了争议，本次二次分析的总体目标是使用验证性因子分析（CFA）重新检验 FCRI 的维度，使用来自大量癌症幸存者的数据集。
根据文献和预先提出的假设，提出了 FCR 的三个模型（包括 FCRI 的单维和多维模型）。分别进行 CFA 以检验每个模型对数据的拟合程度，并比较具有可接受拟合度的模型。
在所测试的所有 FCR 模型中，与最初开发的 7 分量 FCRI 一致的多维一阶模型与数据的拟合度最佳（χ²=3359.135，P<.0001，df=795，RMSEA=0.057 [0.055，0.059]，CFI=0.897，TLI=0.888）。当将这种 7 因素结构加载到整体 FCR 的单个二阶因素上时，模型拟合统计数据略有变差（χ²=3459.632，P<.0001，df=807，RMSEA=0.058 [0.056，0.060]，CFI=0.893，TLI=0.886）。但是，模型之间的差异具有统计学意义（卡方差异=103.142，P<.0001，df=12），表明一阶模型对数据的拟合更好。
这些结果与支持将 FCRI 作为多维量表使用的实证和理论文献一致。结果的含义是根据 FCR 的概念化和测量进行讨论的。
