C3
Deprivation
The third of four harm channels, and the first where nothing is counted. Nobody observes a death from unsafe water; it is attributed, using the share of a population exposed and a relative risk from the published literature.
This channel exists to answer one instruction — count the people who died of bad water and bad food, and exclude natural deaths — without anybody deciding what "natural" means. A death enters only if it carries an attributable fraction against an exposure that economic activity produces, distributes or withholds. Everything with no such fraction is excluded by arithmetic rather than by judgement, and most of what belongs here is excluded today because the evidence has not been read yet.
What this channel counts
What is in the channel now is diarrhoeal disease attributable to unsafe drinking water, unsafe sanitation and the absence of handwashing. The figures are the World Health Organization’s own attribution, published with an uncertainty range for every country.
Beside them sits a second series this project computes independently, from the coverage figures for water, sanitation and hygiene and relative risks taken from a cited medical paper. It is never added to the first. It exists so the two can be compared, and at the median the two agree to within two per cent — which is the best evidence available that the method is sound where it can be checked.
Enters ONLY with a population attributable fraction against a modifiable exposure. A cause with no PAF against any exposure in the register is excluded by construction.
What is in the ledger today
0
Events in the ledger
0
Deaths recorded by the source
0
Countries
0
Conflicts
–
Years covered
These figures are the publisher’s, republished with the credit its licence requires. Nothing on this page has been attributed to a product, a company or a material.
| Confidence of the death rows | Ledger rows |
|---|
The causes inside it
| Cause | Needs a PAF | Ledger rows |
|---|---|---|
| Acute lower respiratory infection C3-ARI |
yes | not counted here |
| Diarrhoeal disease C3-DIARRHOEA |
yes | not counted here |
| Disease attributable to household air pollution C3-HAP |
yes | not counted here |
| Undernutrition C3-UNDERNUTRITION |
yes | not counted here |
What this channel does not capture
- Undernutrition is not in this channel. It is the largest thing that belongs here and no relative risk for it has been read and cited yet, so the gate refuses it. That is a gap in the work, stated rather than filled with an assumption.
- Household air pollution from cooking fuel is not here either, and for a more awkward reason: the attributed deaths are published, but no attributable fraction is, and a death whose share of its cause cannot be stated does not enter this ledger. The data is fetched and checksummed and deliberately left out.
- The attributable fractions this channel uses do not add together. A child with unsafe water and no sanitation is not two attributable deaths; the exposures combine on the share that survives each one, and adding them would overstate the channel by more than a tenth on two exposures alone.
- The figures cover a single year, because that is what the publisher published. A channel with one year in it cannot show a trend, and nothing here should be read as one.
- A death here may also be a death in another channel: cholera in a besieged city is deprivation harm and conflict harm at once. No channel total is added to another until the overlap between them is measured.
Sources and credit
| Release | Retrieved | SHA-256 |
|---|
Questions
What does “attributable” mean here?
That a share of the deaths from a cause would not have happened without the exposure. It is a statistical share of a population, never a claim about any individual death, and it is computed from how many people are exposed and how much more likely the exposed are to die.
Why is so little in this channel?
Because each exposure needs a published exposure–response relationship, read and cited with its confidence interval, and reading them is slow. The comprehensive source for all of them at once is licensed for non-commercial use only, which this project cannot use. What is missing is listed rather than estimated.
Whose numbers are the headline ones?
The World Health Organization’s. Where the custodian of the data has done the attribution itself, that is better evidence than ours, and our own calculation is kept beside it as a check rather than used as the answer.
Is this channel a lower bound too?
More so than any other. Two of its three largest exposures are missing entirely, it covers one year, and harm that cannot be attributed is not counted at all.