The reliability of a newborn foot length measurement tool used by community volunteers to identify low birth weight or premature babies born at home in southern Tanzania
1 Faculty of Infectious and Tropical Disease, London School of Hygiene and Tropical Medicine, Keppel St, London WC1E 7HT, UK
2 MARCH (Maternal, Adolescent, Reproductive and Child Health Centre), London School of Hygiene and Tropical Medicine, Keppel Street, London WC1E 7HT, UK
3 Ifakara Health Institute, Dar es Salaam P.O. BOX 78373, Tanzania
BMC Public Health 2014, 14:859 doi:10.1186/1471-2458-14-859Published: 20 August 2014
Low birthweight babies need extra care, and families need to know whether their newborn is low birthweight in settings where many births are at home and weighing scales are largely absent. In the context of a trial to improve newborn health in southern Tanzania, a counselling card was developed that incorporated a newborn foot length measurement tool to screen newborns for low birth weight and prematurity. This was used by community volunteers at home visits and shows a scale picture of a newborn foot with markers for a ‘short foot’ (<8 cm). The tool built on previous hospital based research that found newborn foot length <8 cm to have sensitivity and specificity to identify low birthweight (<2500 g) of 87% and 60% respectively.
Reliability of the tool used by community volunteers to identify newborns with short feet was tested. Between July-December 2010 a researcher accompanied volunteers to the homes of babies younger than seven days and conducted paired measures of newborn foot length using the counselling card tool and using a plastic ruler. Intra-method reliability of foot length measures was assessed using kappa scores, and differences between measurers were analysed using Bland and Altman plots.
142 paired measures were conducted. The kappa statistic for the foot length tool to classify newborns as having small feet indicated that it was moderately reliable when applied by volunteers, with a kappa score of 0.53 (95% confidence interval 0.40 – 0.66) . Examination of differences revealed that community volunteers systematically underestimated the length of newborn feet compared to the researcher (mean difference −0.26 cm (95% confidence interval −0.31—0.22), thus overestimating the number of newborns needing extra care.
The newborn foot length tool used by community volunteers to identify small babies born at home was moderately reliable in southern Tanzania where a large number of births occur at home and scales are not available. Newborn foot length is not the best anthropometric proxy for birthweight but was simple to implement at home in the first days of life when the risk of newborn death is highest.