Comparing Performance of Simplified Mid-Upper Arm Circumference Devices ("Click-MUACs") to Detect Acute Malnutrition

Overview

The Click-Mid-Upper Arm Circumference study aims to describe and compare the performance of a set of three simplified mid-upper arm circumference classification devices ("Click-MUAC") for use by mothers and caregivers to screen for acute malnutrition in young children. The study will be conducted in Isiolo County, Kenya. The study builds on the recent approach of training mothers to screen for acute malnutrition in their own children (mother MUAC approach). The primary aim of the study is to describe and compare the performance of a set of three prototype Click-MUAC devices against a gold standard of classification, in terms of measures of agreement (sensitivity, specificity, accuracy, Cohen's Kappa, Youden's J) for the classification (diagnosis) of acute malnutrition. A secondary aim of the study is to determine the difference in agreement with the gold standard classification amongst mothers/caregivers using a Click-MUAC prototype versus mothers/caregivers using a regular MUAC tape. The hypothesis of the study is that at least one of the Click-MUAC prototypes is equal to the gold standard of classification in terms of measures of agreement for the classification of acute malnutrition. A secondary hypothesis of the study is that agreement with the gold standard classification is better amongst mothers/caregivers using a Click-MUAC prototype compared to mothers/caregivers using a regular MUAC tape. The outcomes of the study will be used to decide on whether production of a Click-MUAC device is desirable and, if this is the case, to select a design based on the performance of the prototype devices and the stated preferences of intended users.

The study will be a non-randomised, non-blinded, clinical diagnostic trial describing and comparing performance that will be conducted in two phases: (i) calibration of the devices, (ii) testing of Click-MUAC prototypes against "gold standard" classification. The study will take place at designated health facilities within Isiolo County that already have an IMAM programme running. The investigators expect that the health facilities with the highest number of acute malnutrition admissions will be selected as study sites in order to be able to test the prototypes on a large number of attending Severe Acute Malnutrition (SAM)/Moderate Acute Malnutrition (MAM) children. The first phase of work will concentrate on the calibration of Click-MUAC prototypes. The performance of the Click-MUAC prototypes with regard to accuracy of classification will be assessed in children with a measured MUAC at or close to the 115 mm and 125 mm case-defining thresholds. This phase of work is required as prior testing of the Click-MUAC prototypes will only have been performed on semi-rigid testing tubes. The second phase of work will test the Click-MUAC prototypes. Three sets of measurements/classifications will be taken during the Click-MUAC testing trial: (i) Mother/caregiver MUAC classification performed by the mother/primary caregiver on their own child, using each of the three Click-MUAC prototypes and a regular MUAC tape; (ii) IMAM staff MUAC classification performed by the Integrated Management of Acute malnutrition (IMAM) staff on the same child, using each of the three Click-MUAC prototypes and a regular MUAC tape; (iii) Gold standard MUAC measurement performed by the study coordinator and/or data collection team members trained in the measurement of MUAC and selected for their accuracy (i.e. against the study supervisor) and precision of measurement, following a validation exercise using the Habicht method, on the same child. In addition, mothers/caregivers and IMAM staff will also be asked to state their preference of Click-MUAC prototype and asked to give reasons for their preference. The Click-MUAC devices are screening/diagnostic devices. Appropriate measures of agreement between the different measurer groups (mothers/primary caregivers, IMAM staff, gold standard) can be calculated from two-by-two contingency tables. The following appropriate measures of agreement will be calculated from these tables: sensitivity, specificity, accuracy, Cohen' s Kappa, Younden's J. The analysis (i.e. two-by-two contingency table and estimation of appropriate measures of agreement) will be performed for each prototype Click-MUAC device. A discordancy analysis will also be performed. The aim of the discordancy analysis is to describe misclassifications in terms of the range, interquartile range, and median MUAC at each type of discordancy. This analysis provides further calibration information that may inform design changes that avoid/minimize gross misclassifications and to minimize the frequency of errors made at or close to class boundaries.