Measuring infants’ length is an important practice initiated soon after birth and continued throughout childhood as a tool for clinical and public health surveillance. In order to globally standardise the length measurements, the WHO Multicentre Growth Reference Study developed the WHO Child Growth Standards, based on length/height, weight and age(1). The curves, used in over 100 countries, describe how children should grow when free of disease and raised following healthy practices, such as adequate breast-feeding and a non-smoking environment from birth to 3 years of age(2). For instance, children with length below 2 sd from the length-for-age curve or with weight above 2 sd from the weight-for-length curve, respectively, present evidence of stunting and overweight(Reference De Onis, Onyango and Borghi3).
In this sense, the length measurement is widely applied as a child development indicator, being also used to detect intra-uterine growth restrictions, inadequate postnatal feeding and chronic malnutrition(Reference Stevens, Finucane and Paciorek4). Hence, performing accurate anthropometric measurements is imperative to assess nutritional status and growth of infants and toddlers(Reference Rifas-Shiman, Rich-Edwards and Scanlon5).
The gold standard equipment to measure length among children younger than 24 months of age is the anthropometer, which measures recumbent length (lying down) on a board made of wood or other sturdy material. The anthropometer is composed of a movable piece that serves as footboard and a fixed part serving as headboard. The instrument should be placed on a flat and stable surface, such as a table. Unless there is a digital counter, a measuring tape should be fixed along the length of the board(6). This method is widely applied and suitable for clinical evaluations.
For population-based studies that conduct household assessments of children’s length, an adequate environment for an accurate measurement is not always feasible. Additionally, the increased size and weight of the anthropometer may be difficult for the transportation of the device by interviewers. One option would be the adoption of portable devices that could be more easily transported.
Although portable devices would facilitate the logistics during field work, it is necessary to ensure that these are accurate devices to measure children’s length, since errors on the measurements could lead to misdiagnosis of overweight and underweight(Reference Rifas-Shiman, Rich-Edwards and Scanlon5). This study was conducted with a subsample of 24-month-old children, members of the 2015 Pelotas (Brazil) birth cohort, who attended to the research clinic during the regular birth cohort follow-up that occurred during 2017. The aim of this study was to assess the validity of a portable anthropometer compared with the gold standard in 24-month-old infants.
The present study was carried out in a subsample of the 2015 Pelotas (Brazil) birth cohort. This birth cohort includes all children who were born alive in the five maternity wards of Pelotas (Brazil) between 1 January and 31 December 2015, and whose mothers resided in the urban area of the municipality (n 4275). Participants were interviewed during perinatal study and followed-up when aged 3, 12 and 24 months. More details about the birth cohort can be obtained elsewhere(Reference Hallal, Bertoldi and Domingues7).
During the 24-month follow-up of the 2015 Pelotas birth cohort – that occurred during 2017, all children and mothers recruited during the perinatal study were invited to attend to a research clinic at the Federal University of Pelotas, Brazil. The mothers were interviewed, and the children were measured by trained personnel. Interviewers informed mothers that two length measures would be performed in different anthropometers (fixed and portable) in order to compare length estimates. The extra length measurements using the portable anthropometer did not modify or interfere on the logistics of interviews.
A fixed Harpenden® infant anthropometer, with amplitude between 30·0 and 110·0 cm and 0·1 cm accuracy, was considered as the gold standard for the measurement of infant’s length due to its greater precision and stability. The portable SANNY® (model ES2000) anthropometer, with amplitude between 20·0 and 105·0 cm and 0·5 cm accuracy, was the instrument to be validated.
To estimate the necessary sample size for our study, we considered Willett’s proposal of using between 100 and 210 individuals for validation studies(Reference Willett8). Still, a sample size calculation was performed assuming an acceptable average length difference between the anthropometers equal to 0·5 cm. We accepted this level of difference as the precision of the portable anthropometer is also 0·5 cm. The estimated sample size required 250 children, who were randomly selected from the children who attended to the research clinic during the 2017 birth cohort follow-up. Data collection occurred between August and September, in Pelotas, Brazil.
Previously to data collection, interviewers and researchers who acted as anthropometrists were trained to carry out length measurements in children of the same age range. A gold standard anthropometrist also participated in the training, being possible to measure the inter-anthropometrist accuracy by comparing length measurements taken by anthropometrists and the gold standard anthropometrist; and intra-anthropometrist reliability by comparing the first and second length measures taken by the same anthropometrist. Four trained researchers performed the length measurements in the portable anthropometer at the research clinic, whereas the measurements on the fixed anthropometer were performed by trained interviewers from the 2015 Pelotas Birth cohort.
The length measurements of both anthropometers followed the given protocol: (a) ask the mother to remove children’s shoes and socks, as well as bulky clothing, particularly diapers and head adorners; (b) the available anthropometers are placed on a flat and firm table; (c) a disposable paper towel is placed over the anthropometers’ platform to prevent the child from being in direct contact with its surface; (d) the interviewer requests the mother to lay the child on her/his back, over the anthropometer, and to remain close to the child’s head to calm him/her down during the measurement, if necessary; (e) during the measurement, the interviewer places a hand on the child’s knee, to force it lightly against the apparatus and, with the other hand, move the mobile platform until it reaches the sole of the feet. The child’s feet should be alongside the platform, and the head should stand against the anthropometer’s fixed platform. If the child is agitated, it is necessary to gently force the knees and ankles down and simultaneously push the mobile platform against child’s feet sole; (f) right after the measurement, the interviewer should discard the paper towel, preferably in front of the mother, and sanitise the equipment with alcohol gel. First, the interviewers did the measurements using the fixed infant anthropometer. Second, a researcher responsible for the study was called to measure the length using the portable infant anthropometer. Thus, researchers had no access to the length information measured by the interviewers. Children who presented any health or physical problem that could interfere on the measurements were not included in the study (n 1).
Parametric tests were used because the length of the sample showed a symmetric distribution. Prevalence and 95 % CI were used to describe the sample according to categorical variables, while mean and standard deviations were used for continuous variables. The sample was described according to children’s and mothers’/families’ characteristics. Children variables were sex (female and male), z-scores of height- and weight-for-age (≤ −2, −2 < z < 2, ≥2)(9,10) and the mean age (in months) at the moment the measurements were taken. The characteristics of the mothers were schooling in years of study (0–4, 5–8, 9–11 and ≥12), self-reported skin colour (black, brown, white and other) and mean age (in years) at the moment the measurements were taken. The monthly family income in minimum wages (≤1, 2–3, 4–6, 7–9 and ≥10) was the only family characteristic considered.
The mean length was calculated considering the first and second length measures of both portable and fixed anthropometers, then compared using paired Student’s t test. The Bland–Altman method was used to characterise the concordance between the length measurements of both anthropometers. The Lin’s Concordance Correlation Coefficient (ρ c) was also calculated to analyse the overall concordance, and the concordance according to sex. All analyses were performed using Stata 15.0 software (Stata Corp.).
The sample included 252 children. Children’s, mothers’ and families’ sociodemographic characteristics are described in Table 1. The mean age of analysed children was 23·5 (sd = 0·6) months, while the mean age of their mothers was 29·3 (sd = 6·3) years. The majority of the children had mothers who completed from 9 to 11 years of study, had white skin colour and were part of a family earning from two to three minimum wages per month. Children from our sample were on average lower compared with the general 2015 Pelotas birth cohort population, which was used as reference population (mean sample height = 86·4 (sd = 3·4) v. mean cohort height = 86·7 (sd = 3·6)).
* Fisher’s exact test comparing sociodemographic variables between the validation study sample and the 2015 Pelotas cohort.
† Totals might not sum 252 due to missing information.
‡ Paired student’s t test to test difference by means.
The mean length of the portable and the fixed anthropometers is described in Table 2. The mean of the differences ranged between 0·2 and 0·3, while the error variance ranged from 1·2 to 1·8 cm. There were no statistically significant differences between the estimates of the fixed and portable anthropometers.
σ 2, variance.
* Student’s t test between the anthropometers.
In Fig. 1, Bland–Altman’s plot shows a mean length difference of 0·2 cm (95 % CI 0·1, 0·4) between the fixed and portable anthropometers, with limits of agreement ranging from −2·2 to 2·6 cm. The pattern of spread of the observations in the graph reveals a high concordance between the length estimates of the fixed and portable anthropometers. Similar results were obtained after analysing according to sex. The mean length difference for boys was 0·2 cm (95 % CI 0·0, 0·5; limits of agreement: −2·4, 2·9), while for girls was 0·2 cm (95 % CI 0·0, 0·39; limits of agreement: −2·0, 2·4) (Fig. 2).
Results in Table 3 support the high concordance between both anthropometers. The overall concordance was 0·94 (95 % CI 0·92, 0·95) and was marginally higher among girls compared with boys.
* Lin’s concordance correlation coefficient.
After following the protocol established by this study, the portable anthropometer was found to be an accurate measure compared with the fixed anthropometer (gold standard), with high concordance between the estimates. To the best of our knowledge, this is the first study that compared length measurements of portable and fixed anthropometers among 24-month-old children. Other validation studies that focused on measuring individuals’ body composition used indirect techniques(Reference Orlandi, Bielemann and Martinez-Mesa11,Reference Bosy-Westphal, Mast and Eichhorn12) .
The fixed anthropometer is the most robust way to accurately measure length of children up to 24 months of age, having a 0·1 cm precision. However, using this device to perform length measurements in home interviews might be difficult due to the large weight and size of the fixed anthropometer. To overcome this limitation, a portable anthropometer could be used, which is lighter and easier to carry compared with the fixed device. Yet, the portable device presents some limitations, such as being less precise (0·5 cm) and less durable since it is composed of a more fragile material. Even with these differences between the devices, our study showed that their measures were comparable.
According to the Brazilian Society of Pediatrics, children older than 24 months of age should preferably have their height measured by stadiometers fixed on a flat wall and without a footer(13). However, the same society recommends that anthropometers positioned on a flat surface are used to measure height for children younger than 24 months of age. Although not recommended, the 2006 Pesquisa Nacional de Demografia e Saúde da Criança e da Mulher already used portable anthropometers to measure the length of children under 24 months of age(14).
A study used portable anthropometers to measure the length of children under 24 months of age, stating that it was a suitable device to measure children’s height. However, the study did not perform any kind of validation(Reference Weidauer, Wey and Slater15). Our manuscript fills a gap in the literature, confirming the high accuracy and validity of the portable anthropometer for the length measurement of 24-month-old children.
We observed heterogeneity in our sample distribution, with differences for income, maternal schooling and maternal skin colour. Most children belonged to families with income between two and three minimum wages, which could lead to differences in length measurements. This happens since income is a strong determinant of children’s nutritional status(Reference Martins, Marinho and De Oliveira16–Reference Leal, de Lira and Oliveira18), so as maternal skin colour and schooling(Reference Martins, Marinho and De Oliveira16). The sample of our study can be considered representative of 24-month-old children. However, our validity results can only be generalised for studies comprising children of the same age group, that use the same portable anthropometer and that follow the same training and measurement procedures described in our protocol.
We did not find any statistically significant difference between the fixed and portable anthropometers, but the measurement error was a bit higher for boys compared with girls. One hypothesis for this finding is that boys were more agitated during the measurement process, being harder to keep still and generating higher measurement error compared with girls. Although we could not empirically test this hypothesis in our study, related researches reported that girls were more focused and quieter compared with boys(Reference Huber and Traxl19,Reference Smith, Ulmer and Wong20) .
Some limitations of this study should be acknowledged. We selected the sample within a 2-month period of the year 2017, while the 24-month follow-up of the 2015 Pelotas (Brazil) birth cohort was taking place. Although our study comprised a limited period, it respected the natural flow of the birth cohort, so the sample used in this study is not characterised as a convenience sample. Furthermore, the total sample of the cohort is heterogeneous independently of the period of the year that it is followed-up. Thus, we believe that the characteristics of the subsample considered in our study would not influence the length measures and that this subsample can be considered representative of all the children who were followed-up during 2017. It is important to note that the 0·5 cm precision of the portable anthropometer could have a higher relative error when measuring the length of shorter and younger children. However, our findings are applicable only to children with 24 months of age. Our study presents as strengths the standardisation of the procedure to assess the length in different anthropometers, ensuring the comparability between the measurements, and the blinding between the independent evaluators to avoid contamination.
It is important to highlight that the measurements were performed in an ideal environment, where both anthropometers were placed in a flat and firm surface. However, this scenario could not be found in many contexts. Even so, according to the anthropometric protocol of length measurement in children up to 24 months of age, examiners should look for the most adequate surface possible to set the equipment (e.g. a table or the floor).
Under the standardised protocol followed in the present study, the portable anthropometer proved to be a valid instrument to measure the length of 24-month-old children. Considering its accuracy and the greater practicality to carry and set up, the portable anthropometer is a good option to be used in field work of large population-based studies. In addition, this study has great practical and clinical relevance since it will enable the use of portable anthropometers among 24-month-old children.
Acknowledgements: None declared. Financial support: This study was financed in part by the Coordenação de Aperfeiçoamento de Pessoal de Nível Superior – Brasil (CAPES) – Finance Code 001. This article is based on data from the study ‘Pelotas Birth Cohort, 2015’ conducted by Postgraduate Program in Epidemiology at Universidade Federal de Pelotas, with the collaboration of the Brazilian Public Health Association (ABRASCO). The 2015 Pelotas (Brazil) Birth Cohort is funded by the Wellcome Trust (095582). Funding for specific follow-up visits was also received from the Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq) and Fundação de Amparo a Pesquisa do Estado do Rio Grande do Sul (FAPERGS) and Children’s Pastorate sponsored follow-up at 24 months. Authorship: T.R.F. idealised the study and participated in all stages from conception, written part, data analysis and discussion of results. A.D.B. coordinated and revised the final version. L.I.C.R., L.R.M. and M.D. participated in data collection, written part and discussion of results. C.B. and R.C.M. analysed the data, wrote the results and revised the final version. M.F.d.S. and G.I.M. revised the final version. I.S.S. revised the final version and oriented the study and analyses. Conflict of interest: The authors have no conflicts of interest to declare. Ethics of human subject participation: This study was conducted according to the guidelines laid down in the Declaration of Helsinki, and all procedures involving study participants were approved by the Ethics Committee in Research of the Escola Superior de Educação Física (ESEF) of the Federal University of Pelotas (522.064). All participants signed a consent term and agreed to participate.