Challenges of estimating the annual caseload of severe acute malnutrition: The case of Niger

TY - JOUR

T1 - Challenges of estimating the annual caseload of severe acute malnutrition

T2 - The case of Niger

AU - Deconinck, Hedwig

AU - Pesonen, Anaïs

AU - Hallarou, Mahaman

AU - Gérard, Jean Christophe

AU - Briend, André

AU - Donnen, Philippe

AU - Macq, Jean

N1 - CURIS 2016 NEXS 399

PY - 2016

Y1 - 2016

N2 - Introduction: Reliable prospective estimates of annual severe acute malnutrition (SAM) caseloads for treatment are needed for policy decisions and planning of quality services in the context of competing public health priorities and limited resources. This paper compares the reliability of SAM caseloads of children 6-59 months of age in Niger estimated from prevalence at the start of the year and counted from incidence at the end of the year. Methods: Secondary data from two health districts for 2012 and the country overall for 2013 were used to calculate annual caseload of SAM. Prevalence and coverage were extracted from survey reports, and incidence from weekly surveillance systems. Results: The prospective caseload estimate derived from prevalence and duration of illness underestimated the true burden. Similar incidence was derived from two weekly surveillance systems, but differed from that obtained from the monthly system. Incidence conversion factors were two to five times higher than recommended. Discussion: Obtaining reliable prospective caseloads was challenging because prevalence is unsuitable for estimating incidence of SAM. Different SAM indicators identified different SAM populations, and duration of illness, expected contact coverage and population figures were inaccurate. The quality of primary data measurement, recording and reporting affected incidence numbers from surveillance. Coverage estimated in population surveys was rarely available, and coverage obtained by comparing admissions with prospective caseload estimates was unrealistic or impractical. Conclusions: Caseload estimates derived from prevalence are unreliable and should be used with caution. Policy and service decisions that depend on these numbers may weaken performance of service delivery. Niger may improve SAM surveillance by simplifying and improving primary data collection and methods using innovative information technologies for single data entry at the first contact with the health system. Lessons may be relevant for countries with a high burden of SAM, including for targeted emergency responses.

AB - Introduction: Reliable prospective estimates of annual severe acute malnutrition (SAM) caseloads for treatment are needed for policy decisions and planning of quality services in the context of competing public health priorities and limited resources. This paper compares the reliability of SAM caseloads of children 6-59 months of age in Niger estimated from prevalence at the start of the year and counted from incidence at the end of the year. Methods: Secondary data from two health districts for 2012 and the country overall for 2013 were used to calculate annual caseload of SAM. Prevalence and coverage were extracted from survey reports, and incidence from weekly surveillance systems. Results: The prospective caseload estimate derived from prevalence and duration of illness underestimated the true burden. Similar incidence was derived from two weekly surveillance systems, but differed from that obtained from the monthly system. Incidence conversion factors were two to five times higher than recommended. Discussion: Obtaining reliable prospective caseloads was challenging because prevalence is unsuitable for estimating incidence of SAM. Different SAM indicators identified different SAM populations, and duration of illness, expected contact coverage and population figures were inaccurate. The quality of primary data measurement, recording and reporting affected incidence numbers from surveillance. Coverage estimated in population surveys was rarely available, and coverage obtained by comparing admissions with prospective caseload estimates was unrealistic or impractical. Conclusions: Caseload estimates derived from prevalence are unreliable and should be used with caution. Policy and service decisions that depend on these numbers may weaken performance of service delivery. Niger may improve SAM surveillance by simplifying and improving primary data collection and methods using innovative information technologies for single data entry at the first contact with the health system. Lessons may be relevant for countries with a high burden of SAM, including for targeted emergency responses.

U2 - 10.1371/journal.pone.0162534

DO - 10.1371/journal.pone.0162534

M3 - Journal article

C2 - 27606677

AN - SCOPUS:84990913748

SN - 1932-6203

VL - 11

JO - PLoS Computational Biology

JF - PLoS Computational Biology

IS - 9

M1 - e0162534

ER -