This is my first ever medical statistics/epidemiology questions, so please be patient if I come across as naive, I normally focus on drug and protein chemistry.
I have a huge medical dataset. From this set I have divided up the population by certain characteristics, a particular disease, an age group, and gender. I have four disease to consider over seven age groups of gender, thus 56 sub populations. Per sub-population I determine the prevalence within that population of reported comorbidities. This will leave me with 56 lists of "disease (a name) - prevalence ((0-1])". That is a lot of data. Putting aside a particular hypothesis or particular descriptive question, how would you go about displaying this amount of data in a report or a publication?
I think based on your description of there being 14 groups that you are not presenting comorbidities (i.e. disease A plus at least one of diseases B, C, or D), but rather prevalence of 4 separate diseases. If this is correct, then I would suggest listing each disease as a column, and each of your 14 groups as rows since I find longer tables easier to work with. Table 3 of this paper (https://link.springer.com/content/pdf/10.1007%2Fs00127-016-1250-3.pdf) will give you a sense of what this would look like to have the long by wide layout (replacing relative risks with prevalence rates, and showing how you could include confidence intervals with the rates) You might want to have a column for any illness 1st, and then the breakdowns for each of the 4 illnesses separately in their own columns. Depending on your question, you could alternatively show any illness, followed by the number of illnesses if you don't need to specify the prevalence by illness.
If you need to show all of the comorbidities, this gets a lot more complicated and you'd want a good rationale for needing to show all of the combinations. All of this to say, if you can provide a bit more in the way of specifics, it could be possible to get more helpful responses, but in general, 14 rows x 4-8 or 9 columns seems more reasonable than the other way around.
Hopefully this helps.
I'm not sure your description sounds the depth of your question. Anyhow, each time you face a disease whose freqs in 7x2 age-by-sex classes are to be illustrated, a graphic(al) presentation may be this:
Take a square, mark 0 % through 100% horizontally and vertically (unless you decide to use transformed frequencies), mark one axis Male and the other Female; then draw a "snake," i.e. seven points successively connected by arrows (six arrows needed), starting with the first point having coords. (freq in boys, freq in girls) and ending at the last point (freq in oldest men, freq in oldest women). Depending on legibility (clutter), two or more snakes may be drawn in one graph - to display similarities or discrepancies, e.g. If convenient, the marked freqs may be made relative ones by division by the overall (preferably, non-sex-specific!) prevalence; I'm mostly thinking of rare diseases here. Log transformation etc. of the axes is possible but probably not so instructive.
As regards depicting cooccurrence of diseases A and B across 7x2 classes, this can now be reduced to 7x2 indices of association, such as
P(AB)/(P(A)P(B)) or P(neither A nor B)/(P(not A)P(not B)), or their product, or any other measure of association in a two-by-two table. So again we have 14 numbers to present, possible as an "association snake."
- Badges
- Science topic
- Similar topics
- Medicine
- Epidemiology
- Age Groups