Param Revamp - Pregnancy Supervisor

[mmsuarezcosta]

Parameters

Assumed

Variables that have been defined as 'assumed' in the write-up because no reliable data source identified for the parameter:

• prob_ectopic_pregnancy
• prob_mod_sev_aph
• prob_glycaemic_control_diet_exercise

General

• @joehcollins , I noticed the column Primary Source (or calibration data). Can you please elaborate on what data is documented in this column?
  • For example, for the parameter prob_ectopic_pregnancy, the value column is populated with [0.01, 0.01] and the Primary Source (or calibration data) value is populated with [0.004937, 0.00366]. It seems that the values from Primary Source (or calibration data) are not used in the module?

Local v. Universal

• Most parameters with prefix prob labeled as local, and those with prefix rr, odds, and aor labeled as local
• Prob/odds/aor related to ANC visits all documented as local (source is MS National Statistical Office (NSO))
• treatment effects: Those related to case management and care defined as local (treatment_effect_ectopic_pregnancy_treatment, treatment_effect_post_abortion_care, prob_glycaemic_control_diet_exercise, treatment_effect_gdm_case_management, treatment_effect_still_birth_food_sups), the rest (related to supplementation and medication) labeled as universal . Currently have treatment_effect_still_birth_food_sups labeled as local as I believe the type of food supplement may be location dependent.
• min_age_reproductive & max_age_reproductive: Defined as universal
• death_delay_days_after_abortion & death_delay_days_after_ectopic_rupture: both set to local because has to do with duration of time that treatment/case management can take into effect to prevent death

Scenario

• Currently have all scenario parameters and their corresponding parameters labeled as scenario (e.g. alternative_anc_coverage and anc_availability_odds which is the target that is set when alternative_anc_coverage is TRUE)

Undetermined, Design Decision

• hsi_event_window_days: this parameter is used across multiple HSI events as it is it is the 'buffer' that is used for scheduling. To be discussed if helpful to have as this widely applicable parameter, or if each HSI event should have its own parameter (e.g. hsi_event_first_anc_appt_window_days). I have opted to have them all fall under this singular parameter for readability + simplicity

Undetermined, Calibration

• prob_still_birth_per_month: Is 'scaled at initiation' the same as calibrated?

[joehcollins]

Great work Mariana on what is, as always, a very annoying module to make change to. Please see my comments, i think we could work on the names of some of the new parameters being more explict but happy to chat on that.

[joehcollins]

On you 'general' point above - we can delete any values which arent in the 'value' column. we can keep the references to specific studies however. Just going to look at your labelling of the parameters now

[mmsuarezcosta]

@joehcollins thanks so much for taking the time to review this modifications. Super helpful, and really do agree with your recommendations and the push to rethink some of these parameter names to add clarity to the code. Going to go through your comments/recommendations in more detail.

[joehcollins]

@joehcollins thanks so much for taking the time to review this modifications. Super helpful, and really do agree with your recommendations and the push to rethink some of these parameter names to add clarity to the code. Going to go through your comments/recommendations in more detail.

no worries! Could you share the criteria for labelling the parameters? I wanted to use that to review but couldnt find it - will do next week

[mmsuarezcosta]

@joehcollins thanks so much for taking the time to review this modifications. Super helpful, and really do agree with your recommendations and the push to rethink some of these parameter names to add clarity to the code. Going to go through your comments/recommendations in more detail.

no worries! Could you share the criteria for labelling the parameters? I wanted to use that to review but couldnt find it - will do next week

Yes, for the latest guidance, you can reference the tab 'Guide for Modelers' in this document HERE. Also, I'll be sure to tag you in the other labour modules that I've worked through that would also be helpful to start getting your eyes on. thanks!

[marghe-molaro]

Hi @mmsuarezcosta @joehcollins, just checking what the status is on this - has Joe signed off on all changes as module lead, or does this require some further thought? Thanks in advance!

[joehcollins]

hi @marghe-molaro - I need to come back this to check the parameter labels etc.

Hoping to get to this next week. is that ok??

[marghe-molaro]

Of course!! Thanks @joehcollins .

[joehcollins]

I've made some updates to the parameter labels and names of new paramters which have been created along with simplfying some of the associated logic. Tests permitting I'm happy!