Estimation, CohortMethod, Results: Improvements #114
Comments
|
The KM issue has been resolved previously and the power table now has a download option (3rd and last bullets). |
|
Nate - can you check which of these issues remain? I think most are resolved and this could be closed. In future, it may be easier to post each issue separately as that makes it easier to keep track. |
Sure, I will go through the list and check. I also think we have addressed most of this, and those which we haven't are probably no longer needed or a priority. I will close it if this indeed turns out to be the case. |
Standardizing power table outputs to adhere to OSM standard theme (reactable via resultTableViewer and resultTableServer). Also split the tables into 2 sub-tabs for easier readability
* Fixing Power tables re: #114 Standardizing power table outputs to adhere to OSM standard theme (reactable via resultTableViewer and resultTableServer). Also split the tables into 2 sub-tabs for easier readability * Updating propensity score table Adding conditional formatting of colors to values to aid in interpretation, and also adding an absolute value of beta column. * Updates to propensity score plot Overlaying equipoise statistic onto the plot, and truncating cohort names in legend * Updating covariate balance plot Adding maxSDM to covariate balance plot * Updating captions and systematic error Adding to caption descriptions in figures and overlaying EASE statistic on systematic error plot * Updating systematic error Adding more descriptive filenames for plot downloads * Develop (#340) * Fixing #301 (#326) Adding alphabetical sort to DB input options * Fixing #296 (#327) Standardizing pickerInput type for database selection where multiple options are possible to improve UX. Also fixed some colDefs in Exposed Cases * Fixing #302 and Filtered Data Downloads (#332) * Fixing #302 and Filtered Data Downloads All filtered data downloads across all modules should work now, with a button styled the same as the full download. It needs to be a CSV handler though due to the reactable statte * Updating the fix Also fixed download buttons in Cohorts module, addressing #298 * Update R_CMD_check_Hades.yaml * Update R_CMD_check_Hades.yaml * Update R_CMD_check_Hades.yaml * fixing R checks fixing R checks * Update cohort-diagnostics-databaseInformation.R (#333) Fixing the issue reported in #162 * fixing issue issue_330 (#334) fixing time plot x-axis * fixing issue 167 (#335) added code to get long database names on multiple lines * Update cohort-diagnostics-timeDistributions.R (#336) adding fix for issue 168 * Update patient-level-prediction-modelSummary.R (#337) --------- Co-authored-by: Nathan Hall <[email protected]> * fixing testing errors --------- Co-authored-by: jreps <[email protected]>
|
This is all now fixed/incorporated here: #339 |
Create section for options to select before generate: Target, Comparator, Outcome, Database, Analysis
Show /hide columns - default: only show Cal HR, Cal LB, Cal UB, Cal p
Results/Power:
o Need to download results, is it two tables/csv for Table 1a and Table 1b? (desirable to be able to download across databases/analyses, not only in drilldown to 1 DB/1 analysis)
Results/Attrition:
o Some text overlaps box outline
Population characteristics:
o Show raw table (not just ‘pretty’) with target/comparator std diff before and after PS adjustment
Propensity model:
o Sort beta by value (current bug: doesn’t sort as number correctly), need to cast as sortable value
o Consider column with abs(beta) to sort on covariate with largest coefficient
o Add column for covariate prevalence for T and C pre-adjustment (so that you can determine feature importance) ….same values in Population Characteristics
Propensity score:
o overlay equipoise statistic
o Red/blue legend; need to truncate cohort name to allow for both categories to be shown
Covariate balance:
o overlay maxSDM
Systematic error:
o overlay EASE
o – description of attrition from negative control list provided to negative controls with estimates
o When downloading graph, has some identifier for database/analysis (could be title on the plot or name of the image itself)
ASE003 doesn’t show Kaplan-Meier plots (was on ASE001)
o Aesthetics to improve in KM plot:
Long cohort name label in Life table distorts the graph
The text was updated successfully, but these errors were encountered: