Section 1: How much do countries spend on health? (per capita & as a % of GDP)

[jm-rivera]

An issue to track our discussion on what should be included in this section, methodology, etc.

[nupur-parikh]

ideally the time frame for analysis for this section and the rest of the sections would be 2000/2001-2020. I think we should exclude 2021 given only ~20 countries, most of which are high income European countries, have reported data to WHO.

[jm-rivera]

Thanks @nupur-parikh, very helpful.

For the per capita figures, should we do constant 2020 prices (since that would be the latest year considered). That would allow you to compare changes over time.

If we wanted to do country-level data we could also provide the data in $USD. Thoughts?

[jm-rivera]

Second question @nupur-parikh.
When we say "spend on health" do we mean the che - total current health expenditure indicator?

And for context/info, in line with other research, we will use GDP and GDP estimates from the IMF World Economic Outlook

[nupur-parikh]

Yes let's do constant 2020 prices for per capita figures.

For country level data, yes let's do $USD instead of national currency for better comparability across countries where possible.

[nupur-parikh]

Second question @nupur-parikh. When we say "spend on health" do we mean the che - total current health expenditure indicator?

And for context/info, in line with other research, we will use GDP and GDP estimates from the IMF World Economic Outlook

Is Total CHE the same as "ghed_current_health_expenditure" in the list below? If so then yes that's what we mean when we say "spend on health":

Sounds good to me, this makes sense re: GDP/GDP estimates!

[jm-rivera]

@nupur-parikh the GHED data is provided in (among other) LCU, USD, per capita USD, in both current and 2020 constant prices.

They calculate their figures using GDP data from the World Economic Outlook (though the version is fixed to April 2022 (and so the October 2022 version is not taken into account, nor future updates).

They calculate their USD figures using exchange rates from the WB.

Finally, their base year is 2020.

All of the above is fine. But it means we depend on them to update GDP figures, exchange rate information, and to produce constant prices figures. I don't think that is ideal.

We can just as easily create shares of GDP, USD figures, per capita and constant prices. We can use the same data sources as them. However, that would mean that our figures (when not current prices) wouldn't be exactly the same as theirs. They would not be very different, but they wouldn't be exactly the same.

What do you think? tagging @lpicci96 as well for thoughts.

[lpicci96]

If we are following their process (as explained in their methodology) with updated data from WEO, we are adding more value rather than copying their data only. As long as we document this and state it anywhere we show the data it should be fine. And it would need an extra step of vetting to make sure the calculated values are correct, check if they deviate significantly from GHED values etc etc. Fairly standard to do this, so why not

[nupur-parikh]

Agreed with Luca. We would also not be the only ones doing something similar to the GHED data, IHME calculates GDP data for their GHED analysis from a few other sources (WB, IMF, UN, Penn World Tables). So this process makes sense to me, I leave to you what the best source to use for GDP is though as I'm not sure

[jm-rivera]

Thank you for the quick reaction @lpicci96 and @nupur-parikh! For consistency with other products (and with GHED) we should go for:

• UN population numbers (for per capita)
• IMF WEO for GDP/GDP estimates
• WB WDI for exchange rates

and work from their NCU/LCU numbers to produce the others.

[lpicci96]

@jm-rivera do we want to keep the original GHED data in the MongoDB collection and use this repository to do the transformation or have the calculated values in the collection in place of the original?

[jm-rivera]

@jm-rivera do we want to keep the original GHED data in the MongoDB collection and use this repository to do the transformation or have the calculated values in the collection in place of the original?

Thanks @lpicci96, let's leave the MongoDB version untouched. But we could consider adding our version as a separate collection when we are done. That way both versions are accessible as needed.

[jm-rivera]

@nupur-parikh a related question of this section was:
Did COVID-19 impact country spending in 2020?

How would you like to assess that? Possible options:

• We compare spending levels (in real terms) with the average of the previous 2-3 years (if so how many years)
• We compare with other notable changes in spending levels (other crises like Ebola?)
• We index to some year (like 2015 to link to the SDGs) and we see how spending has evolved since then (so 2015 levels = 100), to assess if 2020 was indeed out of the norm
• other ideas...?

What do you think?

[nupur-parikh]

Visualization idea: bar or line chart, x-axis years: 2000-2020, y-axis: health spending per capita in USD, health spending as a percent of GDP. with drop down to show the following groupings (where possible): all countries, income groups, Africa (all), and individual countries with available data.

[nupur-parikh]

Another option would be to show someething like this, but by income groups and maybe also include Africa and the US for comparison sake?

[jm-rivera]

Something like this by income:

Show the "world total" statistic next to the viz.

The main point of the viz is that:

• health spending has increased
• But very unequally (income level)

Add an option to see the country level data on demand like this

Two charts:

• Money (in constant USD)
  -% of GDP or Per capita

[jm-rivera]

Per our discussion, #41, will remove total health spending as a share of government spending from chart 1.

[nupur-parikh]

Per our discussion, #41, will remove total health spending as a share of government spending from chart 1.

Great, thanks! Have updated the text for the chart and tagged you in the google doc.