The diagnosis numbers are data of the past ~5 months of the project scaled up to 6 months (×208/154 (clinic-weeks)). REO records data on all diagnoses (by week) and the provided are the sums.
For the QALY estimates, I used my best judgment, also based on the potential decrease of the Health-related Quality of Life (HRQoL) multiplied by the years lived with the (consequences of the) condition (equivalent Years Lost due to Disability) and on the potential Years of Live Lost. For example, for the mild condition treatments, deworming or vitamin A supplementation can have great long-term quality of life effect. However, treating other minor conditions may have only relatively small short-term effects. Thus, the wide ranges in estimates.
In terms of the QALY improvement, this will depend on the facility of treatment of the condition. For example, treating severe malaria, severe acute malnutrition, averting maternal deaths due to antenatal care, or deworming may be relatively easier than treating cancer, cardiovascular disease, or hypertension. This should be already factored in the ranges—for example, for life-threatening conditions, the upper QALY estimate is that the clinics save all lives and the lower estimate is that perhaps only ~1/3 of lives is saved or that all interventions save a life but its quality is decreased by 67 percentage points, or a combination of the two that results in the same increase in overall quality×quantity of life.
I’m sure there are many giving opportunities in global health that are better than the GiveWell top charities, and I’m pleased to see promising small or medium-sized projects like this being brought to the attention of EAs.
However, I think you should try to get better estimates of QALYs gained (or DALYs averted)—especially if you’re going to feature the cost-effectiveness ratio so prominently in your write-up. This should be possible by referring to the relevant literature. The current estimates don’t seem all that plausible to me, e.g. an episode of “simple malaria” (by which you presumably mean there are no other complications like anaemia) tends to last a few weeks or less, so even if it could be immediately cured at the beginning, it wouldn’t reach your lower estimate of 0.1 QALYs, let alone the upper of 5 QALYs. For life-threatening conditions, I don’t think you should have the theoretical maximum of “save all lives” as the upper estimate, as that wouldn’t happen in any context, and certainly not this one. If you must rely on your intuitive guesstimates, perhaps you should use 90% or 95% credible intervals.
I found the dataset that I thought I saw before: the Institute for Health Metrics and Evaluation (IHME) Global Burden of Disease Study 2017 (GBD 2017) Disability Weights. Disability weights are the changes of Health-related Quality of Life (HRQoL) due to a condition. I re-ran the calculations and found the cost-effectiveness of the mobile clinics project as 26.63 USD/QALY, with a low estimate of 184.14 USD/QALY and high estimate of 6.33 USD/QALY. I used the same data to estimate the cost-effectiveness of AMF and found 56.07 USD/QALY (low 112.14 and high 11.21). The Business Insider AMF number is about 49.76 USD/QALY. Thus, these updated calculations may be more accurate. Still, the calculations do not take into account the preventive care outcomes, deaths averted due to the Ebola outbreak response, and economic benefits (e. g. of deworming) that may lead to further health improvements, leave alone the positive long-term virtuous cycle of improved health and wealth—but that may apply to other health-related programs too.
Hello. I apologize for the late reply. I was moving over the weekend. I am looking at the IHME DALY by cause data (my calculations here) but these do not seem to take into account the long-term effects of the diseases. For example, deworming and vitamin A supplementation may have positive long-term effects in terms of schooling and economic gains that may far outweigh the direct short-term QALY losses. From there the upper estimate of 5. Simple malaria I would presume one that does not require immediate medical attention but one that still may result in severe condition if untreated (CDC). For the life-threatening conditions, my rationale was also that children treated with severe acute malnutrition are younger than average-age patients and that persons who survive 5 years live on average longer than life expectancy.
Also, the QALY estimates are not taking into account the effects of preventive measures—e. g. almost 90,000 persons informed on STIs and the response to a cholera outbreak (training and material provided) - before the intervention, 5 persons died, after no other deaths occurred.
On that note, I would actually appreciate if anyone could provide more credible estimates, taking into account the effectiveness and long-term consequences of the treatment. I am sure that REO would welcome such cooperation, also for capacity building reasons.
The diagnosis numbers are data of the past ~5 months of the project scaled up to 6 months (×208/154 (clinic-weeks)). REO records data on all diagnoses (by week) and the provided are the sums.
For the QALY estimates, I used my best judgment, also based on the potential decrease of the Health-related Quality of Life (HRQoL) multiplied by the years lived with the (consequences of the) condition (equivalent Years Lost due to Disability) and on the potential Years of Live Lost. For example, for the mild condition treatments, deworming or vitamin A supplementation can have great long-term quality of life effect. However, treating other minor conditions may have only relatively small short-term effects. Thus, the wide ranges in estimates.
In terms of the QALY improvement, this will depend on the facility of treatment of the condition. For example, treating severe malaria, severe acute malnutrition, averting maternal deaths due to antenatal care, or deworming may be relatively easier than treating cancer, cardiovascular disease, or hypertension. This should be already factored in the ranges—for example, for life-threatening conditions, the upper QALY estimate is that the clinics save all lives and the lower estimate is that perhaps only ~1/3 of lives is saved or that all interventions save a life but its quality is decreased by 67 percentage points, or a combination of the two that results in the same increase in overall quality×quantity of life.
I’m sure there are many giving opportunities in global health that are better than the GiveWell top charities, and I’m pleased to see promising small or medium-sized projects like this being brought to the attention of EAs.
However, I think you should try to get better estimates of QALYs gained (or DALYs averted)—especially if you’re going to feature the cost-effectiveness ratio so prominently in your write-up. This should be possible by referring to the relevant literature. The current estimates don’t seem all that plausible to me, e.g. an episode of “simple malaria” (by which you presumably mean there are no other complications like anaemia) tends to last a few weeks or less, so even if it could be immediately cured at the beginning, it wouldn’t reach your lower estimate of 0.1 QALYs, let alone the upper of 5 QALYs. For life-threatening conditions, I don’t think you should have the theoretical maximum of “save all lives” as the upper estimate, as that wouldn’t happen in any context, and certainly not this one. If you must rely on your intuitive guesstimates, perhaps you should use 90% or 95% credible intervals.
Good luck with the project!
Hello!
I found the dataset that I thought I saw before: the Institute for Health Metrics and Evaluation (IHME) Global Burden of Disease Study 2017 (GBD 2017) Disability Weights. Disability weights are the changes of Health-related Quality of Life (HRQoL) due to a condition. I re-ran the calculations and found the cost-effectiveness of the mobile clinics project as 26.63 USD/QALY, with a low estimate of 184.14 USD/QALY and high estimate of 6.33 USD/QALY. I used the same data to estimate the cost-effectiveness of AMF and found 56.07 USD/QALY (low 112.14 and high 11.21). The Business Insider AMF number is about 49.76 USD/QALY. Thus, these updated calculations may be more accurate. Still, the calculations do not take into account the preventive care outcomes, deaths averted due to the Ebola outbreak response, and economic benefits (e. g. of deworming) that may lead to further health improvements, leave alone the positive long-term virtuous cycle of improved health and wealth—but that may apply to other health-related programs too.
Hello. I apologize for the late reply. I was moving over the weekend. I am looking at the IHME DALY by cause data (my calculations here) but these do not seem to take into account the long-term effects of the diseases. For example, deworming and vitamin A supplementation may have positive long-term effects in terms of schooling and economic gains that may far outweigh the direct short-term QALY losses. From there the upper estimate of 5. Simple malaria I would presume one that does not require immediate medical attention but one that still may result in severe condition if untreated (CDC). For the life-threatening conditions, my rationale was also that children treated with severe acute malnutrition are younger than average-age patients and that persons who survive 5 years live on average longer than life expectancy.
Also, the QALY estimates are not taking into account the effects of preventive measures—e. g. almost 90,000 persons informed on STIs and the response to a cholera outbreak (training and material provided) - before the intervention, 5 persons died, after no other deaths occurred.
On that note, I would actually appreciate if anyone could provide more credible estimates, taking into account the effectiveness and long-term consequences of the treatment. I am sure that REO would welcome such cooperation, also for capacity building reasons.