search

CRSU-Apps / MetaInsight

An interactive web-based tool for analyzing, interrogating, and visualizing network meta-analyses using R-shiny
https://crsu.shinyapps.io/MetaInsight/
GNU General Public License v3.0
8 stars 6 forks source link

MR-38 remaining covariate tabs #133

Closed JanionNevill closed 6 months ago

JanionNevill commented 7 months ago

Added result details and deviance report pages. More tabs to follow, but this task is blocking Tom, so I'm truncating it here.

Deconfliction will be done after review to not pollute the review.

BalancedmonkeY commented 7 months ago

One thing I noted is that for the standard Bayesian analysis, the result details will change according to what the user sets as the reference treatment. However, this isn't the case for the covariate analysis tab. Maybe have a look to see if it can be (if your struggling I can try help). Furthermore, the covariate 'result details' says it is taking the covariate value at the center, which I don't think can be changed. It might be worth adding a note (maybe below, but above the gelman plots) to mention this (i.e. the covariate value selected above it not fed into here).

I don't think this point has been looked at?

JanionNevill commented 6 months ago

One thing I noted is that for the standard Bayesian analysis, the result details will change according to what the user sets as the reference treatment. However, this isn't the case for the covariate analysis tab. Maybe have a look to see if it can be (if your struggling I can try help). Furthermore, the covariate 'result details' says it is taking the covariate value at the center, which I don't think can be changed. It might be worth adding a note (maybe below, but above the gelman plots) to mention this (i.e. the covariate value selected above it not fed into here).

I don't think this point has been looked at?

I had a look at these and I think that it's a limitation of gemtc. I've raised a couple of GitHub issues on the gemtc repo:

BalancedmonkeY commented 6 months ago

One thing I noted is that for the standard Bayesian analysis, the result details will change according to what the user sets as the reference treatment. However, this isn't the case for the covariate analysis tab. Maybe have a look to see if it can be (if your struggling I can try help). Furthermore, the covariate 'result details' says it is taking the covariate value at the center, which I don't think can be changed. It might be worth adding a note (maybe below, but above the gelman plots) to mention this (i.e. the covariate value selected above it not fed into here).

I don't think this point has been looked at?

I had a look at these and I think that it's a limitation of gemtc. I've raised a couple of GitHub issues on the gemtc repo:

Great - thanks!

Could I ask for the following sentence to be added to the app in the meantime? "Please note: the printout of result details may not reflect the chosen reference treatment and chosen covariate values. This is a limitation that we are exploring with the {gemtc} developers." I reckon it would fit nicely between the title "Result details for all studies" and the printout.

JanionNevill commented 6 months ago

Could I ask for the following sentence to be added to the app in the meantime? "Please note: the printout of result details may not reflect the chosen reference treatment and chosen covariate values. This is a limitation that we are exploring with the {gemtc} developers." I reckon it would fit nicely between the title "Result details for all studies" and the printout.

I've added it at the top of the page because it was too awkward to put elsewhere

tommorris168 commented 6 months ago

One thing I noted is that for the standard Bayesian analysis, the result details will change according to what the user sets as the reference treatment. However, this isn't the case for the covariate analysis tab. Maybe have a look to see if it can be (if your struggling I can try help). Furthermore, the covariate 'result details' says it is taking the covariate value at the center, which I don't think can be changed. It might be worth adding a note (maybe below, but above the gelman plots) to mention this (i.e. the covariate value selected above it not fed into here).

I don't think this point has been looked at?

I had a look at these and I think that it's a limitation of gemtc. I've raised a couple of GitHub issues on the gemtc repo:

Great - thanks!

Could I ask for the following sentence to be added to the app in the meantime? "Please note: the printout of result details may not reflect the chosen reference treatment and chosen covariate values. This is a limitation that we are exploring with the {gemtc} developers." I reckon it would fit nicely between the title "Result details for all studies" and the printout.

This is a problem with MetaInsight, not with gemtc. The cause is the same as why the model fit statistics table disappeared - mixing up summary(model) with summary(relative_effects) in gemtc_analysis.R. It may sort itself out if the changes to gemtc_analysis in mr-41 are merged.

BalancedmonkeY commented 6 months ago

One thing I noted is that for the standard Bayesian analysis, the result details will change according to what the user sets as the reference treatment. However, this isn't the case for the covariate analysis tab. Maybe have a look to see if it can be (if your struggling I can try help). Furthermore, the covariate 'result details' says it is taking the covariate value at the center, which I don't think can be changed. It might be worth adding a note (maybe below, but above the gelman plots) to mention this (i.e. the covariate value selected above it not fed into here).

I don't think this point has been looked at?

I had a look at these and I think that it's a limitation of gemtc. I've raised a couple of GitHub issues on the gemtc repo:

Great - thanks! Could I ask for the following sentence to be added to the app in the meantime? "Please note: the printout of result details may not reflect the chosen reference treatment and chosen covariate values. This is a limitation that we are exploring with the {gemtc} developers." I reckon it would fit nicely between the title "Result details for all studies" and the printout.

This is a problem with MetaInsight, not with gemtc. The cause is the same as why the model fit statistics table disappeared - mixing up summary(model) with summary(relative_effects) in gemtc_analysis.R. It may sort itself out if the changes to gemtc_analysis in mr-41 are merged.

Right you are Tom! Thanks! (The output was so similar, I wasn't expecting them to be)

JanionNevill commented 6 months ago

This is a problem with MetaInsight, not with gemtc. The cause is the same as why the model fit statistics table disappeared - mixing up summary(model) with summary(relative_effects) in gemtc_analysis.R. It may sort itself out if the changes to gemtc_analysis in mr-41 are merged.

Thanks Tom! I was just building a minimum reproducible example to send you. I've closed those issues on the gemtc repo with an apology.