Open robin-wang opened 2 years ago
@a24680bjiang Sorry for my delayed response. Here is the assignment for you, Pakistan2017 in DHS RecodeVII. You can find the working branch and raw data linked in the issue content above; if any of this is unclear at this point by writing alone, let me know : )
there are some questions that I put a "?" there like: whether lady health visitors equals to ready health workers; whether basic health unit( bhu) can be defined as formal delivery institutions; for Measles, data value of h9a does not equal to h9; and what should I do if there are any vaccination else for child?
there are some questions that I put a "?" there like: whether lady health visitors equals to ready health workers; whether basic health unit( bhu) can be defined as formal delivery institutions; for Measles, data value of h9a does not equal to h9; and what should I do if there are any vaccination else for child?
Noted! Let me look into it. Thanks much for looking into such level of details. Will be back with resolutions asap
Skilled ANC: The survey report, page 156 tells us 86% received ANC from skilled providers. When you look at the coded output, what is the rate for this variable with and without lady health visitors?
Per table 9.7, note 1, skilled provider includes doctor, nurse, midwife, lady ehalth visitor, and community midwife
For facility delivery, page 171, report 9.6 shows 65.9%. Could you please benchmark?
on whether m15 is a filter question, you can cross-tabulate m15 against m17, and m15 against m61. When you do tab m61 m15, it becomes clear that time spent at place of delivery is not recorded when giving birth at home.
Other Vaccinations: great to have the information recorded. Let's only recognise their presence.
Measles: Good catch. if respondent reported having received measles vaccination on either h9 or h9a, let's treat them as having vaccination.
question about do file 8-c_diarrhea_pro, report wrote as "The survey data show that the prevalence of diarrhoea among children under age 5 is 19%. Overall, 71% of children with diarrhoea were taken to a health provider for advice or treatment ", and the mean of c_diarrhea_pro is 0.61. When I add the var to the !regexm, the mean of c_diarrhea_pro does not change. I believe there isn't the story of double-counting.
question about do file 8-c_diarrhea_pro, report wrote as "The survey data show that the prevalence of diarrhoea among children under age 5 is 19%. Overall, 71% of children with diarrhoea were taken to a health provider for advice or treatment ", and the mean of c_diarrhea_pro is 0.61. When I add the var to the !regexm, the mean of c_diarrhea_pro does not change. I believe there isn't the story of double-counting.
Interesting observation here. I suspect maybe one of the categories that contributed to c_diarrhea_pro is making up for the 10 percentage point difference. Do you have a sense of which categories are constituents of the c_diarrhea_pro variable?
question about do file 8-c_diarrhea_pro, report wrote as "The survey data show that the prevalence of diarrhoea among children under age 5 is 19%. Overall, 71% of children with diarrhoea were taken to a health provider for advice or treatment ", and the mean of c_diarrhea_pro is 0.61. When I add the var to the !regexm, the mean of c_diarrhea_pro does not change. I believe there isn't the story of double-counting.
Interesting observation here. I suspect maybe one of the categories that contributed to c_diarrhea_pro is making up for the 10 percentage point difference. Do you have a sense of which categories are constituents of the c_diarrhea_pro variable?
I suppose some categories, but these var also does not appear the difference. Like h12c-bhu is the var I think is not included, but sum h12c still shows observations.
Let me look into this, let's move on with the rest of the recoding : )
I had a quick look at c_diarrhea_pro, you were right to have confusions. Our coding is correct, despite the differences. Let's move on. Thanks for flagging this out!
Hi, Robin! For delivery care. do file, "*c_sba_eff1: Effective delivery care (baby delivered in facility, by skilled provider, mother and child stay in facility for min. 24h, breastfeeding initiated in first 1h after birth)". What do you mean when we say "facility"? Does any place other than home count as facility, or only the formal facility mentioned before in delivery?
I am circling back on reviewing / responding to your question.
This is a good one, and the judgment on the scope of facilities should ultimately come from the report, in two ways:
Whether the survey report has specifically said anything about which are formal facilities. It might appear in the associated tables, as a footnote, or even as header.
Whether our calculated rate of facility delivery agrees with the reported (if possible) rate of children/babes delivered in formal facilities. Refer to the first point also.
Another angle to understand this, in our coding template:
Please note that in the delivery_care.do file, around row 60, we have defined [c_facdel]. This speaks to facility delivery status. The understanding of facility delivery in the context of c_sba_eff1 should be consistent with c_facdel.
c_sba_eff1 is an aggregate indicator, based on multiple inputs, and these inputs are clues across the delivery_care.do file.
In some countries, however, judgments may be harder to arrive at. For example, when combing through the countries in Central Asia, we noted that most of the delivery was carried out in [Maternity homes], and we had gone to great length to validate the definition with the World Bank team, to make sure their Country Assignment Office shares insights on how to interpret them in light of the health care system in the associated region/country. Some caveats here, as background information for you.
[CHECKLIST] WORKFLOW DOC Tracker for Project Management https://github.com/Data-Whale-DC/HEFPI2022-SC-update-tasks/issues/1
[WORKING BRANCH] https://github.com/wengxyu1030/DHS-Recode-VII/tree/Pakistan2017-NingJiang
[RAW DATA] https://worldbankgroup-my.sharepoint.com/:f:/g/personal/sneelsen_worldbank_org/Eu36TLZJIb5NuKfRQk_SdrYBKpl3yQhq8jkPCtdrsfM5_A
{FOR REVIEWERS}
{FOR CODERS} A. General Procedural Steps
[x] Rebase current branch based on Master, ensure your branch is the latest. Tutorial: Rebasing your project branch onto another branch https://docs.github.com/en/desktop/contributing-and-collaborating-using-github-desktop/syncing-your-branch#rebasing-your-project-branch-onto-another-branch
[x] Run script without error
[x] Pass internal quality checking (the embedded quality check result: successfully produced value for the indicators and check if there're flags that the indicator value produced is too off from the published data)
[x] Document as comments any notable changes/caveats (if any)
[x] When Fully Completed Survey Edits, proceed to record hours worked: https://forms.gle/tLh7fxCMMsLobGHA6
B. Checklist and Guidance
[x] ANC
[x] Delivery Care
[x] PNC
[x] Child Vaccination (search receive/month in birth.dta. if unknown vaccine pops up, search definition in DHS report)
[x] Child Illness
[x] Child Mortality
[x] Sleeping Under Bednet
[x] Child Anthropometrics
[x] Reproductive Health
[x] Female Anthropometrics and Cancer Screening【no data found】
[x] Miscellaneous: Hospitalisation, NCD & HIV 【no data found 】
Hospitalisation, general observations
Hypertension treatment , blood pressure measure in the past or during the survey, for example check hm.dta/ind.dta/men.dta (generate a_bp_treat a_bp_sys a_bp_dial a_bp_meas a_hi_bp140 a_hi_bp140_or_on_med from ind.dta & men.dta by combining them)
Note from WB team for reference: "don't worry if you cannot find men.dta, not all surveys have it. if it is collected from ind.dta and men.dta seperately (some survey have female & male questionnarie, collecting these data seperately), you will need to combine these 2 dataset, generating these indicators and merge them back to hm.dta. If only from ind.dta or men.dta, i.e. only for female or male, don't generate these variables. If a new/survey-specific datafile occurs, it is highly likely that it contains info. for 13.do. "
Diabetes treatment/measure in the past (for example, a_diab_treat)
HIV, general observations
HIV questionnaire might exsit in Section 4 of "HEALTH ISSUES QUESTIONNAIRE". If they don't ask hiv test results, they are not helpful. but please document these observations.
[x] Cross-validate statistics in survey report, if applicable.
[x] Reference the NAME-SURVEY branch when your edits are finalised.
[x] Update micro data in Onedrive
[x] Examine Quality Checking Output Files