grlurton / drc_hrh Goto Github PK

number marginally different from http://staff.washington.edu/grlurton/CongoAnalysis.html

check where this comes from

Donnee pas coherente

• could you update heatmap on « proportion des agents qui reçoivent chaque revenu ». It was in the previous ppt but needs updating. Exact same cadres, same everything, just add "informal revenues" and take out "autres revenus". Feasible?
• could i also have that same heatmap, but with numbers and no colors – i.e. a table :-) to use in the narrative? thanks

marginal problem (8 people)

covariables : province, cadre, rural / urbain, age, poste

pas urgent

@grlurton I am updating the graphs in the ppt and have some questions/issues

• Penurie (staffing_heatmap): there's something weird in the HGR. Maybe when updating the cadre gruping for the pharmacists, you forgot the slightly different categories of personnel for HGR (autre tech and non qualif disappeared...). Could you check?
  The rest looks fine but of course I am now panicking that there may a whole lot of such little mistakes that I am missing ... :-|
• Pletore for HGR: similar issue? autre tech and non-qual disappeared, and there are weird data on autre (should be "non qual") and infirmier -- the normes line is not correct and there's no one plethorique, while there should be. For the time being, I am keeping the old graph, which looked better (the only issue there is with Pharmacists -- now we want to separate pharmacists from preposés, if possible)

RecodePost <- function(PostInit){
PostInit[PostInit %in% c('administrateur' , 'administrateur_gestionnaire')] <- 'administrateur'

PostInit[PostInit %in% c('autre' , 'Autre')] <- 'autre'

PostInit[PostInit %in% c("membre_du_personnel" , "membre_personnel")] <- 'autre'

PostInit[PostInit %in% c("technicien_labo" , "technicien_laboratoire")] <- 'labo'

PostInit[PostInit %in% c('infirmier' , 'infirmier_superviseur' ,
'infirmier_titulaire' , 'directeur_nursing')] <- 'infirmier'

PostInit[PostInit %in% c("medecin" , "medecin_chef_staff" , "medecin_directeur" ,
"medecin_generaliste_specialiste")] <- 'medecin'
PostInit
}

add range

@grlurton
see issue #25 -- could you do me a table with the same info of the graph (median USD amount for each component by cadre) and also % of each component. it's for the narrative...

size of bar = median
composition = mean %ages for the category

produire 6 profils 'type' qui serviront d'illustration dans le rapportr

pas urgent

by poste/authority (done for some),

external support (done for some – see below), rural/urban (to be done), level of education (feasible?), age or

years in health sector (I added this, as I think it may be relevant), anything else?

• could I have a table of the MCZ incomes by province, so I can calculate % for each income to use in the narrative. Thanks.
• hey, you used averages, and not medians!! Why? (just for my understanding..)

prepare and send table to Rishma:

find mock tables in Drive / output / tables for joint report_WB

It's just a quick draft. Makes sense? is that what we had agreed on with her? -- stuff highlighted in yellow is for you/to be discussed... I don't think it should take long to do, but Rishma may have questions/ask for other stuff.

by province / cadre

@grlurton Analysis of mecanises and immatricules should be the "heat map status" graphs, right?
You mentioned yesterday that there may be some mistakes (you have no pharmacists in Bandundu...), and the mecanisé stuff looks different from what we had before (linked to issue #16 ??)
Also, remember to do it using the cadre grouping that we agreed on: #18
Let me know once these are done and which folder/title the tables will have so I can paste them in the ppt.

At this point, don't worry about issues #2 and #3 ...

check systematiquement

@grlurton not everything is done as I had wished, but i can't think anymore.. off to bed.
Could you please:

• have a look at the new issues and ideally close #36 #37 #38 #39 #40 before you go to bed tonight? once i have that, i'll add to the ppt and finalize
• have a look at the ppt and the abstract (sending them by email but also in Drive/output): they are not finished, but almost ...
  ** ppt: still need to do the last slides: ideas welcome. Could you also re-read the ppt for French mistakes and to see if I explained the stats bits correctly? Thanks.
  ** abstract: particularly uninspired, results are super watered down, descriptive and so what?... again, ideas welcome! and could you add a sentence on the analysis?

I am aiming to finish both ppt and abstract my tomorrow night (my friday night, your friday midday) and send to Hadia etc.
I am working on something else tomorrow morning, but back on this in the afternoon -- it would be great if you have 30 minutes to chat on your FRIDAY MORNING AS SOON AS YOU WAKE UP so we finalize all -- feasible?
Let me know! thanks,
m

table avec percentage de personne dans chaque type d'activite, by cadre and province and rurban + agrege

mettre des triangles pour les CSR

Police plus grande pour les legendes et les labels

that there is much difference/inequalities between what each one receives and that the difference seems to be driven by factors that are not salary+pdr (almost all receive them and seem quite uniform). ?

• check if there is an outlier or something fishy in informal income for MCZ…. We get weird data (very high informal income for MCZ) in the overall table/graph, but then this seems not to be there in the MCZ specific graph. mmmmmmmm....

une table avec % de mecanise et immatricule par province seulement

where? (for these two I just need the overall % that I add in the narrative

see IGA outside health sectore

• what is the recall period for per diems and for p.partenaire in the frequency tables?
• I think you are right, the results for the model are ugly.. what else could be done about it? Maybe it be better if we use the logistics model instead of the “counting” one?
  What is fac_number_appui.csv file I saw in the output? useful? what about the stuff you had done on appui in the analysis online page? -- NOT URGENT -- just trying to get some ideas on how to improve this bit in the report, as all will be looking at it!

• mecanisation: overall taux by province to use in the narrative
• immatriculation : overall taux national and by prov to use in the narrative

faire juste du descriptif
3 tables differentes
province, cadre, type of facility

Screwcork is in the kitchen's drawer

• prime de partenariat (et prime PBF only) vs. partage des recettes (hypothèse: les agents qui reçoivent une prime de partenariat reçoivent moins des patients)
• Salaire+prime de risque vs. partage des recettes
• Rémunération totale dans le secteur de la santé vs. rémunération à partir des activités hors santé

EXTRA

pour guider Maria in question #27

@grlurton could you select 4 nice/interesting pictures of health centres, maybe with people in them, which i can use in the policy brief? If you don't have time to select them, you can either send 10 random ones, or tell me where i find them all.

I would carry out the analysis of determinants as for the each income component, but

there may be other factors that we are interested to explore: ownership of facility, size of facility, others?? Or

not? (to be discussed)

• Could you redo the individual stacked bars with total income -- update incomes (now we have informal) and do it on same scale
  If you can do on one sheet FoSa staff (6 cadres) and on another ECZ (3 cadres).

thanks.

Averages ?

juste premiere difference par province, au niveau national, type de cente de sante

caracteristique des bancarises, avec cadres, provinces, type of facility

Les medecins apparaissent deux fois...

Prendre la variable qui contient appui financier

Could you have a look at the report (it's in the drive/output/HRH DRC_final report_20012015). There are a couple of comments for you -- things to check and to update the data analysis bit. Feel free to correct, edit, etc. in track change.
Are you ok with the model results I pasted in Annex?

thanks!