Param Revamp - Care of Women During Pregnancy

[mmsuarezcosta]

To Note

• Within the ANC visits, there are many checks based on gestational age. These checks seemed to be based on the age cut-offs of the ANC visits (e.g. if someone wen to ANC 4 with the gestational age of ANC 7, then ANC 7 interventions applied). Thus, I used the parameter that I created (e.g. anc_gestational_age_weeks_anc7 in this case) within the conditions of the other ANC appointments.
• Calibration information described in detail in the write-up. My review of write-up was not exhaustive, recommend reviewing to ensure proper documentation of calibrated parameters.

Parameters to review

• prob_seek_anc3: Had [0.77, 0.77] in an undefined column. I've moved it to be the prior_max value
• prob_delivery_modes_ec, prob_delivery_modes_spe: Assume to be 'local'
• gestational_age_weeks_upper_bound: I created this parameter. Set this as 'undetermined' with the description 'structural'. It is used as an upper bound of max weeks of pregnancy (not defined by local or universal literature)
• treatment_effect_blood_transfusion_anaemia: marked as 'universal'

Params that I have left hard-coded (list not exhaustive, aimed to document the main ones)

• All med consumable dosages
• Date
• The DateOffsets windows for tlclose (e.g. 1, 3, 7 days). I believe it will add confusion to add too many of these variables. See example:

          self.sim.modules['HealthSystem'].schedule_hsi_event(visit,
                                                                priority=0,
                                                                topen=visit_date,
                                                                tclose=visit_date + DateOffset(days=7))

• 7 weeks: Seemed to be different conditions with 7 week pregnancies, and did not believe they could be collapsed into single parameter

        # Here we block the event from running for previously scheduled ANC1 HSIs for women who have lost a pregnancy
        # and become pregnant again
        if (
            (date_difference > pd.to_timedelta(7, unit='D')) or
            (df.at[individual_id, 'ac_total_anc_visits_current_pregnancy'] > 0) or
            (df.at[individual_id, 'ps_gestational_age_in_weeks'] < 7)
             ):
            return False

• Delivery weeks for grouped parameters, I have left hard-coded, see below:

	# Women with abruption, praevia or chorioamnionitis prior to 28 weeks will not be delivered until they have
	# reached that gestation
	if ((mother.ps_placental_abruption or
	    ((mother.ps_placenta_praevia and (mother.ps_antepartum_haemorrhage == 'severe')) or
	     mother.ps_chorioamnionitis)) and
	    (mother.ps_gestational_age_in_weeks < 28)):
	    beddays = int((28 * 7) - (mother.ps_gestational_age_in_weeks * 7))
	
	# Similarly more mild bleeding or PROM without infection occuring prior to 37 weeks will not be delivered until
	# they have reached that gestation
	elif ((mother.ps_placenta_praevia and (mother.ps_antepartum_haemorrhage == 'mild_moderate')) or
	      (mother.ps_premature_rupture_of_membranes and not mother.ps_chorioamnionitis)) and \
	     (mother.ps_gestational_age_in_weeks < 37):
	    beddays = int((37 * 7) - (mother.ps_gestational_age_in_weeks * 7))

Initially I considered creating unique parameters for them, but they are really groupings of other parameters. Namely,

• 37 weeks is for the parameters:
  delivery_weeks_placenta_praevia_mild,
  delivery_weeks_prom_no_infection

• And the 28 weeks is for the parameters:
  delivery_weeks_placental_abruption
delivery_weeks_placenta_praevia_severe
delivery_weeks_chorioamnionitis

Options moving forward are to:
a) Separate out each condition and apply individual statement
b) Believe that universally these will always be the same, so do a "min" or "max" of these in the statement
c) Create a new parameter (e.g. delivery_weeks_placenta_praevia_mild_no_infection) that encompasses all of them

[mmsuarezcosta]

@marghe-molaro ready for review