Parameter | Base Case | Distribution | Range for one-way sensitivity analysis |
Natural history | |||
Annual background mortality rate for non-wasted children 1-5y in Mali23 31 | 1.7% | Beta: α=114, β=6556 | 1.4–2.0% |
DALY lost due to death from MAM or SAM32 | 27.8 | Fixed | 19–60 |
Proportion of malnutrition cases with SAM15 | 13.1% | Beta: α=3, β=20* | 2.9–29% |
Probability of developing SAM among children with MAM33 | 9.3% | Beta: α=4, β=39* | 3.0–19.5% |
HR of mortality among children with untreated MAM3 | 3.4 | Log-normal: μ=1.2, σ=0.09 | 2.8–4.0% |
HR of mortality among children with untreated SAM3 | 11.6 | Log-normal: μ=2.45, σ=0.09 | 9.7–13.8% |
Duration of untreated SAM episode (weeks)34 | 20.2 | Log-normal: μ=2.98, σ=0.22 | 12.8–30.3 |
Duration of untreated MAM episode (weeks)34 | 11.6 | Log-normal: μ=2.45, σ=0.08 | 10.0–13.4 |
HR of mortality among children post-recovery35 | 1.2 | Log-normal: μ=0.18, σ=0.18*,† | 1–1.6 |
SAM treatment | |||
Duration of SAM treatment (weeks)36 | 6.3 | Log-normal: 1.8, 0.3 | 3.0–11.0 |
Probability of defaulting from SAM treatment programme37 | 8.0% | Beta:α=4, β=46* | 2.3–16.9% |
Weight for calculating average of the duration of SAM and MAM among defaulters (higher weight assumes defaulters are more like recovered children; lower weight assumes defaulters are more like untreated children) | 50% | Beta: α=1, β=1* | 2.5–97.5% |
MAM treatment arm in parent trial | |||
RUSF Treatment | |||
Probability of recovering from MAM after RUSF treatment | 69.9% | Beta: α=234, β=101 | 64.8–74.5% |
Probability of defaulting from RUSF treatment | 6.6% | Beta: α=22, β=313 | 4.2–9.5% |
Average Weeks to recovery | 4.3 | Weibull: shape=1.3, scale=4.6 | 0.3–12.9 |
Average Weeks to default | 5.7 | Weibull: shape=2.2, scale=6.5 | 1.2–12.0 |
CSB++Treatment | |||
Probability of recovering from MAM after CSB++treatment | 61.1% | Beta: α=209, β=133 | 55.9–66.2% |
Probability of defaulting from CSB++treatment | 4.1% | Beta: α=14, β=328 | 2.3–6.4% |
Average Weeks to recovery | 4.2 | Weibull: shape=1.2, scale=4.4 | 0.2–13.1 |
Average Weeks to default | 5.0 | Weibull: shape=1.8, scale=5.6 | 0.7–11.9 |
MI Treatment | |||
Probability of recovering from MAM after MI treatment | 57.2% | Beta: α=175, β=131 | 51.6–62.6% |
Probability of defaulting from MI treatment | 7.8% | Beta: α=24, β=282 | 5.1–11.1% |
Average Weeks to recovery | 4.7 | Weibull: shape=1.2, scale=5.0 | 0.3–14.6 |
Average Weeks to default | 4.0 | Weibull: shape=1.4, scale=4.4 | 0.3–11.3 |
LMF Treatment | |||
Probability of recovering from MAM after LMF treatment | 57.7% | Beta: α=162, β=119 | 51.8–63.4% |
Probability of defaulting from LMF treatment | 1.1% | Beta: α=3, β=281 | 0.2–2.9% |
Average Weeks to recovery | 4.8 | Weibull: shape=1.2, scale=5.1 | 0.2–15.6 |
Average Weeks to default | 9.4 | Weibull: shape=7.84, scale=9.9 | 6.3–11.7 |
Adverse Events | |||
Probability of incident SAM/hospitalisation during MAM treatment | 0.4% | Beta: α=5, β=1264 | 0.1–0.8% |
Average weeks to hospitalisation or SAM | 6.3 | Weibull: shape 1.2, sigma 6.6 | 0.3–19.7 |
*Distribution set so that SD is 50% of the mean such that the 95% interval will be approximately ±100% of the mean.
†Hazard ratios were set to equal the maximum of one or the random draw from the log-normal distribution such that the post-recovery probability of death would not be less than the background mortality.
CSB, corn–soy blend; DALY, disability-adjusted life year; LMF, locally milled flour; MAM, moderate acute malnutrition; MI, Misola; RUSF, ready-to-use supplementary food; SAM, severe acute malnutrition.