Wasting: leftovers for next PR

[EvaJanouskova]

Leaving some TODOs for next PR, following current #543:

important

• the death from untreated SAM: if SAM determined as clinical acute malnutrition state, do also determine whether the individual will die due to SAM (make sure these individuals are not on treatment or are not already scheduled for death) in xx days since then, i.e., assuming an average number of days to death. Do we think the death from SAM should be determined every time the WHZ changes and the individual has SAM (after the change), or only if the individual didn't have SAM before the change? I would vote for the former, as it can happen in many combinations. If they had SAM but WHZ changed for sure, MUAC and oedema may have changed, it could be that one symptom gets better but another worsens, so it might be time to check if they are supposed to die again. If they are enrolled in the tx before they die, then death should be prevented (at least after some time on the tx).
• ? natural recovery from severe wasting: when progress to severe wasting initiated, and hence the WHZ changes, all the clinical signs are updated—MUAC, oedema, clinical state, complications/death. It will be checked if the person is scheduled to die, and if not, the recovery to moderate wasting will be scheduled after the duration of the severe wasting. When they recover from severe wasting to moderate wasting, they should be scheduled again after the duration of moderate wasting to either progress back to severe wasting or recover to no wasting.
• ? progression from uncomplicated to complicated SAM, or even gradual progression: well -> MAM -> uncomplicated SAM -> complicated SAM
• stop progression/recovery event from running if treatment issued:
  • return in the beginning of apply()
  • set the empty appt footprint so the resources are not used
• additional risks for moderate wasting incidence: HIV, exclu_breastfeeding, diarrhoea, WASH, TB, other underlying infections (see guidelines for CMAM)
• parametrise module properly (more detailed calibration? --- diagnosis, treatment, mortality etc.)
• involve an expert in malnutrition e.g. Carlos or Andy Seal in IGH...
• description of linear models to be included in the wasting module's write-up
• re-consider full recovery from SAM (the guideline seems to me pointing to SFP always following the SAM treatment, so either full recovery is not expected, or SFP provided as precaution)

non-essential

• see probability_of_severe -> Wouldn't it be more efficient if the wasted categories are generated by the choice fnc at once for all wasted individuals, and also the properties (in the following lines) are set at once?
• ? apply treatment coverage and cure rates at initiation
• looking up of consumables in one function and then access them from there as done in the other modules like RTI and COPD
• update GBD data to GBD 2021
• HIV-positive determined for SFP, should be send to OTP
• children < 6 months with SAM should go always for ITC
• issuing the treatment each visit, to allow possibility of not finishing the tx, or re-enrolment on tx due to missed two following tx supplies
• allow progression in the beginning of tx (it would require to add, referencing to diff. tx if progressed)
• if a tx not available, it is not provided - but if lower level tx is available, should be provided at least that (but with higher prob of death, and not being cured in the average time--prolonged tx or non-cure)
• relative risk for becoming wasted (WHZ < -2) being ever wasted/stunted before (rr 13.68/1.67, respectively, see Wright et al. 2021)
• schedule initial wasting cases for progression or natural recovery & their last onset of wasting set to random date between today and duration_of_untreated_mod_wasting in past
• commodities reducing prob to recover https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0238311#sec020
• monthly clearance event for 5y+ (all to recover)