What statistical test should I use?

#1
Hi there,

I'm currently in the process of carrying out a systematic review and I've gathered cost estimates from studies for a specific type of treatment (with two different approaches), I am interested in finding out if these costs have decreased over time.

I have 40 cost estimates for the one treatment approach and 12 cost estimates for the other treatment approach, spanning from 2012 - 2020. When I plot a simple scatter graph (y) Cost vs (x) Year, with a line of best fit I can see that Costs decrease as the Year increases for both approaches.

I was wondering is there any statistical test I can do to show that the costs are decreasing over time, I would like a test with a p-value to 'prove' (or disprove) that there is a decreasing trend in costs over time. I do not want to do linear regression (because I only have 12 estimates for one of the approaches).

The costs for both approaches are significantly positively skewed, so I assume it would need to be a non-parametric test.

Thanks for your help!
 

Attachments

#2
Depending on whether your data are parametric or not, I would say a Pearson test of correlation or a Spearman test.

Since you assume that it would non-parametric, then a Spearman test