Joinpoint function signals no error but produces non-sensical results

[chkreis]

Hi Dan,

Thank you for maintaining a package to use NIH Joinpoint Regression Software from R.

We have installed the Command-Line version of Joinpoint for institutional use on our server and updated all the packages listed in imports in https://github.com/DanChaltiel/EDCimport/blob/main/DESCRIPTION .

I have tried to run the R code from the README.md Vignette. The joinpoint function runs for a couple of seconds without producing an error message, but the results from the analysis make no sense, i.e. it finds no joinpoints.

My problem appears similar to closed issues here but these threads did not help me resolve it.

I attach a PDF of the R Console output. Please find the R Code I used below. Your help is highly appreciated.

library(tidyverse)
library(nih.joinpoint)

nih_sample_data %>% group_by(sex) %>% slice(1, 2, n()-1, n())

run_opt = run_options(model="ln", max_joinpoints=3, n_cores=3)
export_opt = export_options()

run_opt
export_opt

reprex::reprex(session_info=TRUE)

jp = joinpoint(nih_sample_data, x=year, y=rate, by=sex, se=se,
               run_opts=run_opt, export_opts=export_opt)

names(jp)

jp_plot(jp) + patchwork::plot_layout(ncol=1)

summary(jp)

str(jp)

test.nih.joinpoint.pdf

[DanChaltiel]

Hi Christian,

As it is written in the text when you opened this issue, I will need the output of sessionInfo():
Please read and follow the guidance text next time.
Also, did you update all your packages as I told you by email?

However, it seems that you are running version 5.0.0.0.1 while this package is meant for v4.9.0.0. There is a significant possibility that backward compatibility was broken when NIH updated their software so that this package doesn't work anymore. For instance, you don't have run_summary in names(jp) as you should. I unfortunately won't have the time to update this package to their new version right now.

[chkreis]

Hi Dan,

Thank you very much for your response. I attach another PDF including the output of sessionInfo(). My apologies if I misread the guidelines. We had downloaded the latest version of all the packages to be imported as you had told me before.

We have indeed installed Windows Command-Line version 5.0. What would you recommend I do, ask NIH if we could use v4.9.0.0?

test.nih.joinpoint2.pdf

[DanChaltiel]

Hi Christian,

I received an email from the NIH team last week saying that they are updating the command line tool and that they will come back to me when they implement a few of my suggestions. However, this could take some time as they seem to be very busy.

If your analysis is in a hurry, you should indeed ask them for v4.9 so you can use my package. Alternatively, you can maybe use the Desktop version.

[chkreis]

Hi Dan,

thank you for your quick response. We do have the Desktop version, but would like to use the Command-line version for simulations. I am going to ask NIH to make v4.9 available to us, which the institute kindly offers to do:
https://surveillance.cancer.gov/help/joinpoint/tech-help/frequently-asked-questions/replicating-previous-results