traumar - traumar 1.2.1

• Within the trauma_performance() function, renamed the variable predicted_prob_death to scale_factor which is commensurate with the source literature.

• updated comments in trauma_performance() for z_method method of the Z_score to reflect the right text

• In trauma_performance(), completed the comment where the scale_factor is created so that it is complete and clear

• Corrected a test error at CRAN from using bootstrap CI process in testing with 100,000 observations and 100 bootstrap samples to make sure rmm() and rm_bin_summary() ran in under 60 sec. That test now does not use the bootstrap process so the core function can be tested and will always run in under a minute with 100,000 observations.

• Cleaned up other tests within for relative_mortality.R that were checking for correct error / warning handling where multiple lines of output were sent to the console. Built a custom function to deal with those scenarios and correctly perform those unit tests.

- R
Published by bemts-hhs 8 months ago

traumar - traumar 1.2.0

• This minor release introduces functionality to demonstrate how the Iowa trauma system currently engages in the quality improvement process using the System Evaluation and Quality Improvement Committee (SEQIC) Indicators.
• Added a Contributor Code of Conduct and support information. _ Additionally, a convenience function is_it_normal() provides the ability for users of traumar to get descriptive statistics on one or more numeric variables, with optional normality tests, and diagnostic plots (for one variable only). Grouping functionality is also supported in is_it_normal() to conduct exploratory data analysis of one or more variables within zero or more groups.
• Added the following functions:
  • seqic_indicator_1()
  • seqic_indicator_2()
  • seqic_indicator_3()
  • seqic_indicator_4()
  • seqic_indicator_5()
  • seqic_indicator_6()
  • seqic_indicator_7()
  • seqic_indicator_8()
  • seqic_indicator_9()
  • seqic_indicator_10()
  • seqic_indicator_11()
  • seqic_indicator_12()
  • seqic_indicator_13()
  • is_it_normal()
• A fix was applied to nonlinear_bins() to make the percent column calculate correctly when groups were not introduced.
• Removed hard-coded rounding from most calculations within the package where possible.
• Improved examples for the the package's README, probability_of_survival(), nonlinear_bins(), rmm(), and rm_bin_summary() using more helpful data.
• Improved error messages coming from nonlinear_bins() when the argument Ps_col does not follow the expected distribution of the calculated probability of survival, and/or a sample size too small to calculate bins is passed to the function, including when passed to rmm() and rm_bin_summary().
• Code formatting changed to the air package through the RStudio IDE.
• Updated data validation for trauma_case_mix(), trauma_performance(), nonlinear_bins(), rmm(), and rm_bin_summary() to provide improved messaging related to missings in Ps_col and outcome_col .
• Across functions using the probability of survival calculation, it is expected that Ps values have a range of [0, 1]. Functions will no longer handle values in percentage format (e.g. 10, 50, 98).
• The outcome argument was removed from trauma_performance() to remove ambiguity in the nature of the outcome_col values. Only values of TRUE/FALSE and 1/0 are accepted.
• The diagnostics argument was removed from trauma_performance() to make the user interface smoother. Instead of providing guidance via outputs to the console, users are encouraged to seek assistance with interpreting results via the source academic literature and the package documentation.
• trauma_performance() will no longer provide a pivoted output as a default. Users can elect to pivot the outputs as needed in their workflows.
• rmm() and rm_bin_summary() now have a new argument bootstrap_ci that allows a user to elect to use the bootstrap CIs, or not. bootstrap_ci defaults to TRUE in order to better support backward compatibility.

- R
Published by bemts-hhs 9 months ago

traumar - traumar 1.1.0

New Features

• Added optional grouping functionality to nonlinear_bins(), rmm(), and rm_bin_summary().
  
  • Setting group_vars = NULL applies the functions to the entire dataset without subgrouping.
    
  • For pivoting the rmm() outputs longer, setting pivot = TRUE will work when group_vars is invoked by pivoting longer with the grouping context.

Enhancements

• Improved NA handling in rmm() and rm_bin_summary().
  
• Ensured RMM calculations remain within the expected range of [-1 to 1], including their 95% confidence intervals.
  
• Optimized nonlinear_bins() by replacing its internal for loop with dplyr functions, enhancing accuracy and efficiency without introducing breaking changes.
• Improved command line messaging and documentation within rmm() and rm_bin_summary() regarding probability of survival values Ps_col < 0 and Ps_col > 1. Now, these functions will throw an error if probability of survival values are Ps_col < 0 or Ps_col > 1.
• The nonlinear_bins() function has improved data validation for the Ps_col variable.

- R
Published by bemts-hhs 11 months ago

traumar - traumar 1.0.0

- R
Published by bemts-hhs 12 months ago