Sample Regression Analysis Using SEER*Stat Data

This example is an analysis of trends in colorectal cancer incidence rates from 1993-2013 in the SEER cancer registries.

Using Joinpoint

There are four basic steps involved in generating any Joinpoint trend analysis. Review the description of the process:

  1. Creating an Input Data File for Joinpoint (or for this exercise you can use the files we created below)
  2. Setting Parameters in the Joinpoint Program
  3. Executing the Joinpoint Regression Program
  4. Viewing the Joinpoint Results

Then walk through the example, Running a Sample Regression Analysis in Joinpoint, using files created for this purpose.


Creating an Input Data File for Joinpoint

The Joinpoint input file must be an ASCII text file or Excel spreadsheet. Refer to the Joinpoint help system for details concerning the format of this file. You may use SAS, SPSS, Excel, Word, or any software package to create the text file. For the example, we have used the SEER*Stat system to create a text input data file. SEER*Stat provides a convenient mechanism for generating the statistics, exporting the statistics to a text file, and providing the file format information required by Joinpoint.

The input data file for this example was created by exporting SEER*Stat results. SEER*Stat created two files when exporting data:

  • Sample.Dictionary.dic is the SEER*Stat export dictionary used in this exercise. It contains the information describing the layout of the export data file. You will use this file to set up your session specifications.
  • Sample.Data.txt is the data exported from SEER*Stat. You will use this file as the Input Data File for this exercise.

Setting Parameters in the Joinpoint Program

The Joinpoint Regression Program requires that you specify parameters that are organized on three tabs: the Input File tab, the Method and Parameters tab, and the Advanced Analysis Tools tab.

Input File Tab

The Input File tab specifies the file format of the input data file and some additional settings for the model. For this exercise, the default settings were used for the By Variables, Dependent Variable, Independent Variable, Log Transformation, and Heteroscedastic Errors Option. Joinpoint is able to determine the settings for these based on the information in the SEER*Stat dictionary. If your input data file did not come from SEER*Stat, you would need to set the appropriate values for these controls.

Method and Parameters Tab

The Method and Parameters tab specifies the modeling method, constraints on the location(s) of the joinpoints, number of joinpoints, autocorrelated errors options, model selection method, AAPC confidence intervals, and include/exclude select cohorts. For this exercise, the default settings on the Method and Parameters tab tab were used.

Advanced Analysis Tools Tab

The Advanced Analysis Tools tab can be used to set up a parallel or coincident pairwise comparison or a Jump Model/Comparability Ratio Model analysis. The pairwise comparison is only relevant if you have one or more By Variables defined on the Input File tab. There will also be enhancements to this tab in the future that include multi-group clustering.

Executing the Joinpoint Regression Program

Once an input data file has been created and loaded into the session, and parameters have been selected, the program can be executed by clicking the green execute button on the Joinpoint toolbar. A progress meter will be shown on the screen while the Joinpoint calculation engine processes the data and generates the output. Note, it can take a few minutes to execute, depending on the size of the input data file and the options selected. After execution has completed, Joinpoint opens an output window to display the results.

View the Joinpoint Results

For standard Joinpoint analyses, the output window displays the results on five tabs: Graph, Data, Model Estimates, Trends, and Model Selection. When using the Pairwise Comparison or Jump Model options, an additional tab labeled with the appropriate option will be shown. There is a cohort tree located to the left of the tabs to traverse the cohorts and various joinpoints in your analysis.

The results of your session are not automatically saved. If you close the output window without saving your results, you will need to re-run the analysis. The results can be saved to a Joinpoint output file (i.e. Sample.Session.jpo) by selecting Save or Save As from the File menu, or by clicking on the Save button on the toolbar. The results can be sent to a printer, PDF, Word, or Excel by clicking the Print button from the Joinpoint toolbar, and there are options to customize which elements of the output to include. The results can also be exported as text files and image files (for the graphs) by clicking the Export button on the toolbar making your selections.

Running a Sample Regression Analysis in Joinpoint

  1. Download the following files to use in this exercise by right clicking on the link and using Save Link As... to save the file on your computer:
    • Sample.Dictionary.dic is the SEER*Stat export dictionary used in this exercise. It contains the information describing the layout of the export data file. You will use this file to set up your session specifications.
    • Sample.Data.txt is the data exported from SEER*Stat. You will use this file as the Input Data File for this exercise.

    The input data file for this example contains age-adjusted incidence rates and standard errors by year of diagnosis and sex. We used the SEER*Stat software to generate these rates for colorectal cancer in the 9 SEER registries, 1993-2013.

    If you have the SEER*Stat software, you may open or download the SEER*Stat matrix file Sample.Matrix.sim.

    Or you may view the SEER*Stat results in Sample.SEERStat.Results.pdf (PDF, 293 KB) using Adobe Acrobat. The rates and confidence intervals were exported to a text file (Sample.Data.txt) using the SEER*Stat export feature. View the SEER*Stat export dictionary (Sample.Dictionary.dic) for more information regarding the contents of the sample input file.
  2. Open Joinpoint and click the New button from the toolbar. The Open file dialog will open.
  3. Browse to the folder where you saved the files for the exercise and open the Sample.Dictionary.dic file.
  4. The Joinpoint Session will open to the Input File tab with the following fields filled in:
    • Input Data File - Sample.Data.txt
    • The first 20 records from the Sample.Data.txt file will be displayed
    • Dependent Variable - Run Type – Provided in Data File
    • Type of Variable - Age-Adjusted Rate
    • Age-Adjusted Rate - Age-Adjusted Rate
    • Standard Error - Standard Error
    • Independent Variable - Year of diagnosis - 1993-2013
    • Shift data points by - 0
    • Heteroscedastic Errors Option - Standard Error (Provided)
    • By Variables - Sex
    • Log Transformation - Yes {ln(y) = xb}
    Below is a picture of the window you should see:

    Sample Joinpoint session set up window
  5. Click on the Execute button on the toolbar to execute the session. The progress dialog opens showing the progress of the Joinpoint software. When it is complete, the Output Results opens with the Graph tab displayed.
  6. To save the results so that they can be opened up later in Joinpoint without re-running the analysis, click the Save button and choose an appropriate file name and location. Please note that when you save the results, you also save the session parameters with it. Once a results file is open, you can retrieve the session that was used to produce it by using the Retrieve Session button on the Joinpoint toolbar. There is no need to save BOTH the session and output results to file.
  7. To export the results as text files, click the Export button on the Joinpoint toolbar. The Export dialog will open.

The following graph is the result of the session in the example for a maximum of 3 joinpoints. It is a scatter plot of the age-adjusted colorectal cancer incidence from SEER, for males and females from 1993 to 2013. A Joinpoint Model is also drawn on this graph. In this case, the model is four line segments joined at the joinpoints of 1995, 1998, and 2008. The model decreases until 1995, it increases again until 1998, decreases until 2008 and then decreases more until 2013. The APCs (Annual Percent Change) indicate the magnitude of the trend for each segment or time period.

This example is an analysis of trends in colorectal cancer incidence rates from 1993-2013 in the SEER cancer registries. The input data file used contains age-adjusted incidence rates and standard errors by year of diagnosis and sex. We used the SEER*Stat software to generate these rates for colorectal cancer in the 9 SEER registries, 1993-2013.

Sample Joinpoint output window
Last Updated: 19 Apr, 2023