1

I have a question on the best way to analyse a particular set of clinical data. The background to the study is assessing the effect of treatment delay on the time to resolution of a particular disease. If we delay treatment, does this lead to an increase in disease duration?

The independent variable is time to treatment from patient admission, which I am unsure of whether should be considered continuous, discrete (due to the fact medical treatment would be classed as day 1 after admission, day 2, day 3 etc ) or even if I should consider two arbitrary groups (for instance using a cut-off point of 72 hours and having before/after 72 hour groups).

The outcome variable is the time to resolution of the specified disease. I imagine this should be considered as a discrete variable? For instance, a resolution of disease could appear on day 4 of treatment, or day 7 etc etc. It is not a continuous scale, although I would be very interested to hear opinions on this.

Thanks for your time and apologies if any of the above is not clear.

mkt
  • 18,245
  • 11
  • 73
  • 172
user117006
  • 31
  • 2
  • 2
    I would have thought that a major problem here is going to be some form of confounding by indication as presumably delay in starting treatment cannot be considered as independent of patient characteristics which might also affect treatment success. Whether your variables are discrete or continuous pales into insignificance compared with that issue. – mdewey Jan 31 '17 at 14:11
  • In the case in question, the differences in starting treatment are not based on patient characteristics but on time it takes for patient samples to be processed and the results given to the physician. If I am interpreting what you have written correctly, you assume that patients who are more ill will be treated earlier which will invariably lead to differences in the outcome. This is not an issue in this scenario. It is more dependent on differences in how quickly patients are assessed and the appropriate investigations being present. – user117006 Jan 31 '17 at 16:29

0 Answers0