[使用JoinPoint对流行病学观察数据趋势进行描述性分析：一个看似用户友好的工具]
[Descriptive analysis of trend of epidemiological observational data using JoinPoint: a user-friendly tool, seemingly].
作者信息
Colonna M
机构信息
Registre du cancer de l'Isère, CHU de Grenoble, Grenoble, France.
出版信息
Rev Epidemiol Sante Publique. 2011 Apr;59(2):123-33. doi: 10.1016/j.respe.2010.11.002. Epub 2011 Mar 15.
BACKGROUND
Analysis of the temporal trend of epidemiological data is becoming increasingly popular. For the purpose, the JoinPoint is a widely used software. We present important elements involved when using some of the options of this software.
METHODS
In order to identify potential breakpoints in the trends and estimate average rates of change, JoinPoint applies piecewise regression to model the expectation of a variable that fluctuates over time.
RESULTS
Using breast cancer incidence data from Isère, an administrative district in France, during the period 1979-2007, we show the effects of user choices concerning the potential number of breakpoints, the research method used by the final model, the length of the period, and the weighting coefficients.
CONCLUSION
JoinPoint is useful to describe changing trends. Nevertheless, user choices have an impact on output and must be clearly identified. Moreover, the JoinPoint approach cannot replace other possible approaches for temporal analysis of observational data.
背景
对流行病学数据的时间趋势进行分析正变得越来越普遍。为此，JoinPoint是一款广泛使用的软件。我们介绍了使用该软件某些选项时涉及的重要因素。
方法
为了识别趋势中的潜在断点并估计平均变化率，JoinPoint应用分段回归来对随时间波动的变量的期望值进行建模。
结果
利用法国一个行政区伊泽尔省1979年至2007年期间的乳腺癌发病率数据，我们展示了用户在潜在断点数量、最终模型使用的研究方法、时间段长度和加权系数方面的选择所产生的影响。
结论
JoinPoint有助于描述变化趋势。然而，用户的选择会对输出结果产生影响，必须明确识别。此外，JoinPoint方法不能替代用于观察性数据时间分析的其他可能方法。
Zotero 插件