Examining the Structure of the Course of Illness

To find out more about how the two dimensions were related to the four diagnosis categories and the four time points, a supplementary variable Time/diagnosis interaction was created by a cross-classification of the four categories of Patient diagnosis and the four categories of Time of interview. Thus, Time/diagnosis interaction has 16 categories, where the first category indicates the anorexia nervosa patients at their first visit. The fifth category indicates the anorexia nervosa patients at time point 2, and so on, with the sixteenth category indicating the atypical eating disorder patients at time point 4. The use of the supplementary variable Time/diagnosis interaction allows for the study of the courses of illness for the different groups over time. The variable was given a multiple nominal scaling level, and the category points are displayed in the following figure.

Figure 1. Category points for time/diagnosis interaction
Category points for time/diagnosis interaction

Some of the structure is apparent from this plot: the diagnosis categories at time point 1 clearly separate anorexia nervosa and atypical eating disorder from anorexia nervosa with bulimia nervosa and bulimia nervosa after anorexia nervosa in the second dimension. After that, it’s a little more difficult to see the patterns.

However, you can make the patterns more easily visible by creating a scatterplot based on the quantifications. To do this, from the menus choose:

Graphs > Chart Builder...

  1. Select the Scatter/Dot gallery and choose Grouped Scatter.
    Figure 2. Chart Builder
    Chart Builder
  2. Select Time/diagnosis interaction Quantification dimension 2 as the y-axis variable and Time/diagnosis interaction Quantification dimension 1 as the x-axis variable.
  3. Choose to set color by Patient Diagnosis.
  4. Click OK.
  5. Then, to connect the points, double-click on the graph, and click the Add interpolation line tool in the Chart Editor.
  6. Close the Chart Editor.
Figure 3. Structures of the courses of illness
Mulitple line chart with separate lines for each diagnosis

By connecting the category points for each diagnostic category across time, the patterns immediately suggest that the first dimension is related to time and the second, to diagnosis, as you previously determined from the object scores plots.

However, this plot further shows that, over time, the illnesses tend to become more alike. Moreover, for all groups, the progress is greatest between time points 1 and 2; the anorectic patients show some more progress from 2 to 3, but the other groups show little progress.

Next