# Shallow Report on Shigella

**Summary**

Considering the expected benefits of eliminating shigella (i.e. improved health and greater economic output), as well as the tractability of a shigella vaccine advance market commitment (AMC), I find that the marginal expected value of a vaccine AMC for shigella to be **192 DALYs per USD 100,000**, which is around 30% as cost-effective as giving to a GiveWell top charity (CEA).

Key Points

Importance: This is a moderately important cause, with

**1.43 * 10**^{7}**DALYs**at stake from now to the indefinite future. Around 73% of the burden is health related, while 27% is economic in nature. A big reason why this issue isn’t as critical as it would otherwise be is that the shigella disease burden has drastically declined over the past few decades, and is reasonably expected to continue doing so as economic development brings improved sanitation/nutrition/access to healthcare, and as education improves hygiene habits.Neglectedness:

**8 vaccines are currently under development**, with one in Phase 3 clinical trials, two in Phase 2, and five in Phase 1. If any are successful, they would probably be deployed – albeit at a slower pace and to a lesser extent – even without an AMC. GAVI’s active role in the area of vaccines is significant, and makes EA involvement probably less counterfactually valuable than it otherwise would be.Tractability: A

**moderately tractable solution**is available, in the form of an advanced market commitment for a shigella vaccine, to ensure that manufacturers have the commercial incentive to scale up production after a vaccine is successfully developed.

Caveats

This report was produced with only one week of research, and critically, only desktop research was used, without experts consulted due to the lack of time. More research – at the intermediate stage and subsequently deep stage – will be needed before we can have high confidence in these findings.

The headline cost-effectiveness will almost certainly fall if this cause area is subjected to deeper research: (a) this is empirically the case, from past experience; and (b) theoretically, we suffer from optimizer’s curse (where causes appear better than the mean partly because they are genuinely more cost-effective but also partly because of random error favouring them, and when deeper research fixes the latter, the estimated cost-effectiveness falls). As it happens,

__CEARCH does not intend to perform deeper research in this area__, given that the headline cost-effectiveness does not meet our threshold of 10x that of a GiveWell top charity.

Further Discussion

While we have high certainty that the disease burden of shigella will continue falling – given that it has been doing so thus far, and given that expert modelling agrees on a future decline – there is real uncertainty as to the exact rate of decline, even as this is what drives the results of our analysis.

I am somewhat sceptical of the economic burden of shigella being as low as it is, and there is in my view a real chance that this analysis underestimates the cost of stunting.

It is uncertain as to how antimicrobial resistance in the future will make treatment less effective and hence shift the shigella disease burden upwards relative to trend; if that happens, the vaccine AMC will of course be more valuable.

**Expected Benefit: Improved Health from Eliminating Shigella**

The primary expected benefit from eliminating shigella is improved health, in terms of fewer deaths as well as less disability and suffering. Overall, around **1.04 * 10 ^{7} DALYs** are at stake here, with this benefit modelled in the following way.

**Moral Weights & Scale**: The global disease burden of shigella in DALY terms for our baseline year of 2024 is around **7.05 * 10 ^{6} DALYs**. This is calculated by using the latest available 2019 figures and projecting forward using a model that will be discussed in greater detail subsequently.

**Persistence**: The problem of shigella is likely to persist, albeit in a declining fashion, and eliminating it will bring benefits not just for one year but across multiple years. In terms of how this multi-year benefit is calculated:

Firstly, I discount for the extent to which the solution does not persist (i.e. the treatment effect being reversed). The solution here is taken to be a shigella vaccine AMC, the choice of which will be discussed at length later.

On the one hand, solution reversal can be at the vaccine-level, with the effectiveness of a vaccine waning over time. I use the effectiveness of a rotavirus vaccine over time up till 5 years after injection in high mortality countries to estimate this as a block discount, under the assumption that the vast majority of the risk from shigella is posed to under 5s, and that the majority of the DALY burden is in high mortality countries. Overall, this translates to a block discount of 43.7% – this is the proportion of the total potential benefit of treating shigella that is nonetheless lost due to the vaccine waning in effectiveness over time.

On the other hand, solution reversal can also occur at the population level, with the treatment effect of the vaccine being reversed as vaccinated individuals die off and are replaced by new, unvaccinated individuals. This population turnover rate is approximated as the inverse of life expectancy, to produce a discount of 1.4% per annum.

Secondly, I account for the proportion of disease burden remaining after being counterfactually solved (i.e. shigella declining due to the intervention of other actors or else due to structural changes).

On the one hand, there are all the non-vaccine considerations to account for. To model this, I use the simple theoretical model that takes total DALYs lost to shigella to be a function of DALYs lost per capita and population size.

Note that whatever efforts that agents (i.e. governments, nonprofits and businesses) are making to solve the problem (e.g. expanding diagnosis and treatment at public hospitals/private non-profit hospitals/private for-profit clinics), and whatever impact that non-population structural trends are having (e.g. economic growth improving sanitation/nutrition/access to healthcare; or improving education causing better hygiene habits) all this will impact either prevalence of shigella or the disease burden per illness, and hence DALYs lost per capita; in short, the variable of DALYs lost per capita accounts for all these agentic and structural factors. The only exception, of course, is population size (where a larger population mechanically increases the disease burden for a given disease burden per capita), with that handled separately in this model.

Overall, to model how the problem is expected to evolve over the years, I (a) project future DALYs lost per capita to shigella by estimating the year-on-year change via a linear regression of past DALYs lost per capita on discrete time; (b) use UN estimates of projected future population growth; and (c) multiply each future year’s DALYs lost per capita and population size to obtain the expected total DALY burden for each year. I cap the projected fall in future DALYs lost per capita to the 2019 rich world levels, on the assumption that there is a limit to what agentic efforts and structural effects can achieve. I also limit the population extrapolation to 2100, as the high-confidence UN estimate ends there, and assume constant population thereafter.

Note that the projected future decline in DALYs lost to shigella per capita is in line with expert estimates and hence justified, with Anderson et al projecting mortality and morbidity estimates from 2015 to 2034, by (a) estimating annual rates of decline for non-rotavirus under-five diarrheal mortality, on the basis that diarrheal mortality rates have declined over time; and (b) assuming morbidity declined at a rate of 0.45% per year, as calculated from YLDs from diarrheal disease from 1990 to 2010 from the GBD. To further ensure that these projections are reasonably accurate and not overly optimistic, I compare against Anderson et al’s projections of the disease burden of shigella in children under 5 in 79 countries for 2025 to 2034 (discounted) and find that CEARCH’s mainline projections are moderately similar (~15% difference) and indeed more conservative (i.e. CEARCH projects a higher disease burden and hence a slower decrease in DALY burden per capita). Note further that the counterfactual effect of vaccines is separately modelled, as a vaccine has not been deployed and its effects would not already be captured by the historical fall in DALY burden per capita.

All this helps produce a per annum discount rate, and allows us to project the growth of DALYS lost due to shigella up to 2100 after taking into account non-vaccine counterfactual solution. The results are shown in **Diagram 1**.

**Diagram 1**: Remaining DALYs lost to shigella after non-vaccine counterfactual solution

And this brings us to second consideration with respect to the problem being counterfactually solved – vaccines, which are already under development and which if successfully developed, would presumably be deployed regardless (albeit more slowly or to a less comprehensive extent). Such counterfactual introduction (i.e. the mere speeding up effect) is calculated using Kremer, Levin & Snyder’s approach of comparing the pneumococcal vaccine AMC to the rotavirus vaccine non-AMC introduction. In modelling, I make the assumption that most of the value of the vaccine comes from the first decade or so. Overall, this translates to a block discount of 37.3% – this is the proportion of the problem that would be counterfactually solved by vaccine deployment even in the absence of an AMC.

Thirdly, I discount for the probability of the world being destroyed anyway (i.e. general existential risk discount) – around 0.07% per annum. This takes into account the probability of extinction, since the benefits of saving people from shigella in one year is nullified if they would die in an extinction event anyway. For how this risk is calculated, refer to CEARCH’s shallow research on nuclear war.

Fourthly, I apply a broad uncertainty discount of 0.1% per annum to take into account the fact that there is a non-zero chance that in the future, the benefits or costs do not persist for factors we do not and cannot identify in the present (e.g. actors directing resources to solve the problem when none are currently doing so).

Overall, by taking the remaining DALYs lost to shigella after non-vaccine counterfactual solution, and discounting each year’s DALY burden using the other per annum discounts (i.e. solution reversal due to population effects, existential risk, uncertainty), the total amount of DALYs lost to shigella that is available for a vaccine AMC to counterfactually avert without block discounts applied is shown in **Diagram 2**.

**Diagram 2**: DALYs available for vaccine AMC to counterfactually avert, without block discounts applied

Finally, by summing the discounted per annum relative values for 2024-2100, and then using a perpetual value formula for 2101 to infinity, all the while factoring in the block discounts, we see that the benefit of improved health from eliminating shigella will last for the equivalent of **1.5 baseline years**.

**Value of Outcome**: Overall, the raw value of improved health from eliminating shigella is **1.04 * 10 ^{7} DALYs**.

**Probability of Occurrence**: Unlike longtermist problems, there is no uncertainty that this is an actual problem. Recent estimates attribute Shigella as causing 125 million diarrhoeal episodes annually, leading to around 160,000 deaths, with a third of these associated with young children. The problem has hardly faded away since; hence, the probability that shigella is a problem can be assigned ~1.

**Expected Value**: Hence, the expected value of improved health from eliminating shigella is **1.04 * 10 ^{7} DALYs**.

**Expected Benefit: Increased Economic Output**

Beyond the health benefits, there are also economic benefits to eliminating shigella. Around **3.9 * 10 ^{6} DALYs** are at stake here, as calculated in the following manner.

**Moral Weights**: I take the value of doubling consumption for one person for one year in DALY terms to be **0.21**. This is calculated as a function of (a) the value of consumption relative to life from GiveWell’s IDinsight survey of the community perspective, as adjusted for social desirability bias, and (b) CEARCH’s estimate of the value of a full, healthy life in DALY terms. For more details, refer to CEARCH’s evaluative framework.

**Scale**: I start by calculating the economic burden of shigella relative to annual income per shigella sufferer, using three separate estimates to do so. The first estimate is Anderson et al’s; I take this data on the total direct treatment costs of shigella for under 5s in 79 low and lower-middle income countries for 2025-2034, divide by number of cases, and then divide by average low and lower-middle income GDP per capita. The expected cost of stunting (calculated by taking stunting to case ratio, as pulled from Anderson et al, and multiplying by the cost of stunting in terms of GDP per capita (as pulled from Galasso et al) is then added to this. The second estimate is Scharff’s; I take this data on the per case costs (both direct treatment and indirect productivity costs) of shigella in the United States, and then divide by US GDP per capita, before adding the stunting costs as previously calculated. The third estimate is Kawai et al’s; I take this data on the per case costs (both direct medical and non-medical or indirect costs) of rotavirus across various Asian countries, divide by each country’s GDP per capita, and then calculate a weighted average. In producing a weighted average of the three estimates, I penalize (a) Anderson et al for not taking into account indirect productivity costs; (b) Scharff for being a one-country estimate; and (c) Kawai et al for being an extrapolation from rotavirus costs. Overall, therefore, equal weightage is used, producing an economic burden of shigella relative to annual income per shigella sufferer of 0.05 – which in turn yields the degree of consumption doubling per shigella sufferer if their shigella is eliminated.

At the same time, the total number of shigella sufferers in the baseline year of 2024 (using cases as a proxy) is 270 million.

Multiplying the degree of consumption doubling per shigella sufferer if their shigella is eliminated, by the total number of shigella sufferers as of 2024, this yields the total number of consumption doublings achievable by eliminating shigella in the baseline year of 2024 – around **12.6 million**.

**Persistence**: The same block discounts, same per annum discounts, and same projections of the disease burden (and hence economic burden) over time, as discussed in the previous section, are used here as well, such the benefit of increased economic output will similarly last for the equivalent of **1.5 baseline years**.

**Value of Outcome**: Overall, the raw value of increased economic output is **3.9 * 10 ^{6} DALYs**.

**Probability of Occurrence**: Same probability as before is applied.

**Expected Value**: All in all, the expected value of increased economic output is **3.9 * 10 ^{6} DALYs**.

**Tractability**

To summarize our tractability findings: we can solve 21% of the problem with a USD 1.53 billion investment into a vaccine AMC (factoring in both the cost of the AMC as well as subsequent deployment costs as borne by developing country governments or GAVI), which means the proportion of the problem solved per additional USD 100,000 spent is around **0.00001**.

In terms of potential solutions, there appear to be a couple of candidates: (a) vaccine development and deployment, (b) generally improving sanitation, and (c) diagnosis and treatment. Both (a) and (b) can be classified as prevention while (c) can be considered treatment, and there are strong theoretical arguments in favour of the former over the latter: (a) the health burden will be higher under treatment vs prevention, given that treatment may not start simultaneously with the onset of the disease and given that treatment may not fully eliminate the disease burden; and (b) the economic burden will be correspondingly higher as well. On the empirical front, shigella is difficult to diagnose and antimicrobial resistance is growing, so even taking into account the difficulty of vaccine development and deployment, (a) vaccines seem a better bet than relying on (c) diagnosis and treatment.

Meanwhile, in terms of vaccines vs general efforts to improve sanitation – using both very rough GiveWell estimates as well as academic estimates of cost-effectiveness that do not factor in critical considerations like long-term effects or the probability of success, vaccine development and deployment for shigella does seem more cost-effective than general WASH interventions, and hence we examine vaccines here as the candidate solution. Note that the precise intervention is an advanced market commitment to purchase a shigella vaccine, with the goal being both to incentivize new vaccine development and also to get vaccine manufacturers to scale up manufacturing.

As for whether a shigella vaccine is even possible – there are currently 8 shigella vaccines under development, including the Beijing Zhifei Lvzhu Biopharmaceuticals vaccine currently in Phase 3 trials, as well as the Eveliqure vaccine and the LimmaTech vaccine in Phase 2 trials. Hence, not only do scientists judge this theoretically possible – as evidenced by attempts to develop vaccines in the first place – certain candidates have passed Phase 1 (i.e. small groups of people receiving the trial vaccine) and Phase 2 (i.e. even larger groups receiving the trial vaccine, with the people having characteristics like age and health similar to the eventual target audience); hence, a vaccine is clearly possible.

In terms of our theory of change:

Step 1: An effective shigella vaccine is successfully developed

Step 2: The AMC boosts shigella vaccine deployment in countries that need it

Step 3: Adoption of the vaccine in such countries reduces the global burden of disease of shigella.

Step 1: To estimate the probability of an effective shigella vaccine being successfully developed, conditional on this even being possible, I take an outside view, and consult three reference classes in doing so. The first reference class is the probability that any vaccine candidate from the relevant clinical trial phase reaches market entry. This is calculated by taking the 8 shigella vaccines under development and then consulting the base probability that any vaccine from any particular clinical trial phase reaches market entry. The assumption is that there will not be further vaccine development beyond these 8, since second-generation suppliers will have little incentive – they will require more time to enter the market and demand will likely be met by their competitors by then. The second reference class is the probability that any infectious disease drug from the relevant clinical trial phase reaches market entry; the same approach as above is used, albeit with a different reference class. Finally, the third reference class is the probability that any new drug from the relevant clinical trial phase is ultimately approved by the FDA); once more, the same approach from above is used, just with a different reference class. In creating a weighted average, I weighed the reference classes of vaccines and of infectious disease drugs more, given greater relevance – yielding a probability of 92% that a shigella vaccine is successfully developed (from the 8 attempts). An inside view is not used, given the lack of deep technical knowledge necessary to perform useful inferences.

Step 2: To estimate the expected extent of shigella vaccine deployment in countries that need it, I take both an outside and inside view.

For the outside view, I consult the three following reference classes. First, there is the adoption rate of the rotavirus vaccine amongst GAVI-eligible countries). Rotavirus is relevant as another diarrhoea-causing disease (indeed, it is the leading disease in causing DALY loss from diarrhoea; shigella is 2nd) that is vaccine-preventable. Meanwhile, GAVI-eligibility is used as these are countries that are presumably poorest, suffering the highest disease burdens, and correspondingly will be ones we need to actually deploy the vaccine. For our second reference class, there is the adoption rate of the pneumococcal vaccine amongst GAVI-eligible countries, with pneumococcus being relevant as another infectious bacterial disease that is vaccine-preventable. And third, there is the adoption rate of the measles vaccine amongst GAVI-eligible countries, with measles being relevant as another infectious disease that is vaccine-preventable. In producing the weighted average, I use a weightage that reflects the degree to which the reference class diseases are similar to shigella (i.e. diarrhoea-related vs not; bacteria-based vs not), given that such differences may impact uptake rates (e.g. policymakers may be more likely to think that diarrhoea can be combated purely through improved sanitation without needing a costly vaccination campaign; or bacteria-based illnesses taken less seriously insofar as antibiotics are available). This yields an estimated deployment extent of 87%.

For the inside view, I reason as follows. A shigella vaccine will save a lot of lives, so it is not improbable that poor countries adopt it (i.e. not <10%). However, the adoption of a new shigella vaccine will compete with other vaccine-preventable diseases in a crowded immunization schedule, and with shigella’s disease burden being comparatively lower (e.g. compared to malaria or HPV or rotavirus), countries may well deprioritize shigella; hence, I rate adoption as being <50%. That said, Shigella may have higher perceived value from it addressing an important cause of antibiotic use and preventing long term consequences (i.e. stunting), which is something policymakers will take note of; hence, and I assign about a 33% chance that any individual country deploys the vaccine (which equivalently translates to the extent of global deployment).

Taking the outside view and adjusting it by the inside view, I penalize the latter due to inferential uncertainties, yielding an estimate of 82% for shigella vaccine deployment in countries that need it.

Step 3: Finally, for the degree to which adoption of the vaccine in the GAVI countries reduces the global burden of disease of shigella, I utilize the outside view once more, with three attendant reference classes – the vaccination rate and effectiveness of the rotavirus vaccine, of the pneumococcal vaccine and of the measles vaccine respectively. The idea is that the degree to which adoption of the vaccine in the GAVI countries reduces the global burden of disease of shigella is a function of (a) country level vaccination rate, (b) vaccine coverage with respect to different shigella strains, (c) vaccine effectiveness, and (d) the proportional disease burden in GAVI-eligible poor countries. The reasons these three diseases are used as reference classes are as previously discussed, while effectiveness draws from high mortality countries given greater relevance. I do not take into account vaccine coverage with respect to different shigella strains, on the assumption that this is accounted for by estimated vaccine efficiency. Low-income countries per GBD classification is used as a proxy for GAVI-eligibility for simplicity. As before, weightage reflects the degree to which the reference class diseases are similar to shigella (i.e. diarrhoea-related vs not; bacteria-based vs not), given that such differences may impact the effectiveness of the vaccine rates, thus producing an estimate of 27% for the degree to which adoption of the vaccine in the GAVI countries reduces the global burden of disease of shigella. An inside view is not used, given the lack of deep technical knowledge necessary to perform useful inferences.

Overall, the proportion of disease reduction from an advance market commitment for a shigella vaccine –as a function of (a) the probability of effective shigella vaccine being successfully developed; (b) the extent of shigella vaccine deployment in countries that need it; and (c) the degree to which adoption of the vaccine in such countries reduces the global burden of disease of shigella – is ultimately 21%.

Meanwhile, for the cost of the AMC – I calculate total cost to be a function of (1) cost per vaccinated individual, and (2) number of vaccinated individuals. (1) is a function of cost per dose, including not just vaccine cost but also supply chain cost and service delivery cost (pulling from Portnoy et al), AMC adjustment above marginal cost of production to incentivize manufacture (pulling from the original pneumococcal AMC design of 2x marginal cost for the first 20% procured), and number of doses required (averaging and rounding the number of doses required for the rotavirus vaccine). (2) is roughly estimated using the global population in low income countries (on the assumption the whole populations will eventually be vaccinated through vaccinating children and process of time thereafter) and discounting for vaccination rates (using rotavirus figures). Costs – at the per dose component level—are discounted on the basis of the lower counterfactual cost of the average poor government (or equivalent GAVI spending on poor countries) relative to EA funding going to top GiveWell charities or similar, as a function of the top GiveWell health charity’s cost-effectiveness relative to just giving cash to poor people, correcting for GiveWell’s undervaluation of life vs income. Additionally, the probability of a successful vaccine being developed (and hence purchases even occurring) as well as the extent of deployment (and hence the extent of expenditure), are discounted for.

Consequently, the proportion of the problem solved per additional USD 100,000 spent is around **0.00001**.

**Marginal Expected Value of Vaccine AMC for Shigella**

All in all, the marginal expected value of a vaccine AMC for shigella is **192 DALYs per USD 100,000 spent**, making this around 30% as cost-effective as a GiveWell top charity.

Thanks for the research and write up, Joel. I will be participating in a Shigella vaccine challenge study at the beginning of February (I learned of this particular study through 1Day Sooner). If anyone has questions, let me know.

Sounds great! Hope the side effects aren’t too bad, and (for what it’s worth), I think it’s really commendable to volunteer for challenge trials.