Param Revamp - ALRI

[mmsuarezcosta]

Files

Deleted

• Calculations.csv, Lazzerini CFR.csv, and Pathogen_specific.csv were not in use and have been deleted

Parameters

Note: There is discrepancy between the write-up and the parameter values on the CSV file (parameters names, list, values, ...)

Dummy/Assumed

I am documenting the following parameters in the CSV file that have been either marked as dummy or assumed : prob_lung_abscess_in_pulmonary_complicated_pneumonia, prob_pneumothorax_in_pulmonary_complicated_pneumonia, prob_cough_in_other_alri, prob_difficult_breathing_in_other_alri, prob_chest_indrawing_in_other_alri, prob_danger_signs_in_sepsis, prob_iCCM_severe_pneumonia_treated_as_fast_breathing_pneumonia, prob_IMCI_severe_pneumonia_treated_as_non_severe_pneumonia

Local v. Universal

• Any proportion has been defined as local, and any probability (parameter with prefix with prob) has been defined as universal. A recommended exercise will be to revisit to prob parameters and determine if any are local-specific
• All treatment failure parameters (prefixed with tf) are assumed to be based on the true effect of the drug treatment, not on any location-specific factors (eg drug adherence) and thus are deemed to be universal.

[mmsuarezcosta]

@marghe-molaro ready for review

[marghe-molaro]

Hi @mmsuarezcosta,

Many thanks for this!

I believe all relevant hardcoded parameters have been removed (the few standing ones are more "by definition" anyway, so I'm flagging it as such in the overview file.

For the parameters flagged as "undetermined"/"calibration", was this established via the write-up?

I am not sure I am in a position to judge whether all those universal parameters can be labelled as such, I think this is @tbhallett's call!

[mmsuarezcosta]

Hi @mmsuarezcosta,

Many thanks for this!

I believe all relevant hardcoded parameters have been removed (the few standing ones are more "by definition" anyway, so I'm flagging it as such in the overview file.

For the parameters flagged as "undetermined"/"calibration", was this established via the write-up?

I am not sure I am in a position to judge whether all those universal parameters can be labelled as such, I think this is @tbhallett's call!

Hi @marghe-molaro

• Hardcoded params: OK
• Undetermined/Calibration: These parameters (rr_all_strains_Strep_pneum_ALRI_with_PCV13, overall_progression_to_severe_ALRI,scaler_on_risk_of_death) were noted as 'for calibration' in the CSV file itself. The ALRI write-up does not discuss calibration.
• Dummy/Assumed: Given that there are some parameters that were labeled as calibration and some that had a different label of dummy/assumed, I did not take the assumed ones to be calibration. There is definitely an open question as to the best way to label them. Here has been my logic:
  • Local/Universal: It serves more as a 'placeholder' value rather than something that was calibrated to any data numbers. This may be assumed from conversations or general knowledge. They may have chosen this number and left as-is, and calibrated other parameters to match the data appropriate country data. For the labeling, I use the logic of " if it was backed by a reference, what it would be". (I have chosen this logic/classification)
  • Undetermined, Calibration: This MUST be indicated by the developer. It is clear that the value was adjusted based on specific data.

[marghe-molaro]

Hi @mmsuarezcosta thanks for the response!

Another thing I wanted to check - I see you have deleted some of the files inside ResourceFile_Alri/ directory which weren't being used (will double check whether this is the protocol we actually want to follow vs storing elsewhere)

Just wandering why GBD_Malawi_estimates.csv and McAllister_2019.csv were kept? I don't see them being called by the module itself

[mmsuarezcosta]

Hi @mmsuarezcosta thanks for the response!

Another thing I wanted to check - I see you have deleted some of the files inside ResourceFile_Alri/ directory which weren't being used (will double check whether this is the protocol we actually want to follow vs storing elsewhere)

Just wandering why GBD_Malawi_estimates.csv and McAllister_2019.csv were kept? I don't see them being called by the module itself

@marghe-molaro thanks for the review.

Regarding deleted files:
A) Initially, I deleted any file that was not referenced ANYWHERE in the codebase (e.g. GBD_Malawi_estimates is referenced in analysis files, so it has been kept. Pathogen_specific is not referenced anywhere, so it has been deleteD)
B) After your input, I stopped deleting any files regardless of whether they were referenced or not. This would primarily apply to Malaria, TB, and HIV, and a couple that I may have revisited