ASA Connect

 View Only

  • 1.  disease rate over time

    Posted 06-20-2018 15:34
    ​Hi,
    Having disease rate over time/years would Wilcoxon test be good to check for year to year differences?
    Thank you,
    Konrad

    ------------------------------
    Konrad

    ------------------------------


  • 2.  RE: disease rate over time

    Posted 06-22-2018 10:40
    Hi Folks:  The disease rate (incidence of disease) over a period of time is an average rate and one  expects smaller periods to have different average rates.  LL

    ------------------------------
    Lawrence Lessner
    Research Scientist and Retired
    Institute of Health and the Environment, SUNY, Albany
    ------------------------------

    Original Message


  • 3.  RE: disease rate over time

    Posted 06-22-2018 22:07
    I am not so sure that your disease rate data are percentage or the number of cases per some populations (e.g., 10,000 people). But, whatever either one you have, I would suggest using the general or generalized mixed model to handle year-to-year comparisons because it is better to take care of temporal autoregressive correlations, especially when you want to control some risk factors.

    ------------------------------
    Lung-Chang Chien
    Assistant Professor
    Epidemiology and Biostatistics
    Department of Environmental and Occupational Health
    University of Nevada, Las Vegas
    ------------------------------

    Original Message


  • 4.  RE: disease rate over time

    Posted 06-25-2018 14:13

    Can you please provide more information on what type of data you have and what hypothesis question you would like answered?

    Very generic suggestions that come to mind are:

    Assuming no trend over time

    -Test for Equal Proportions

    -Goodness of Fit to a Poisson distribution

    Assuming some autoregressive or monotonic trend with time

    -Some type of Regression/ Mixed Model/ ARIMA

    -Somers' D or Cochran-Armitage Test for Trend

    Associating with another variable:

    -Correlated the observed rates with hypothesized explanatory variables

    -Compare the rates over time between two distinct groups using a permutation test



    ------------------------------
    Matthew Robinson
    ------------------------------

    Original Message


  • 5.  RE: disease rate over time

    Posted 07-02-2018 14:25
    The National Cancer Institute uses JOINPOINT in testing for significant changes in the slope over time (on time series of cancer rates).
    https://surveillance.cancer.gov/joinpoint/

    Joinpoint Trend Analysis Software

    Joinpoint is statistical software for the analysis of trends using joinpoint models, that is, models like the figure below where several different lines are connected together at the "joinpoints". Cancer trends reported in NCI publications are calculated using the Joinpoint Regression Program to analyze rates calculated by the SEER*Stat software.

    There are two versions of Joinpoint software available: desktop and command-line.

    1. Desktop version: This version has a graphical user interface and is for interactive use. It is most suitable for an individual to use at their desktop PC.
      Latest Release: Version 4.6.0.0 – April 16, 2018 (What's new in this version?)
    2. Command-Line version: The command-line version is for use in a batch mode or called from other software, such as SAS or R. 
      Latest Release: Version 4.6.0.0 – April 16, 2018 (What's new in this version?)
    Sample Joinpoint Graph


    ------------------------------
    Raid Amin
    Professor
    University of West Florida
    ------------------------------

    Original Message


Related Content