Abstract
There is a need for easily administered, low‐cost measures to assess child development in large field studies. Many researchers evaluate the age of attainment of motor milestones, but there is little information on their validity. A large longitudinal study (MINIMat) was conducted in a poor rural area of Bangladesh and we assessed the age of attainment of motor milestones in a subsample of over 2000 children. We examined their association with scores on the Bayley psychomotor development index (PDI) and mental development index (MDI) at 18 months and with scores on the Movement Assessment Battery for Children and with intelligence quotient (IQ) on the Wechsler Preschool and Primary Scale of Intelligence at 64 months. A field worker visited the children's homes monthly from 3 to 12 months of age and then at 15 months and examined the children. Mothers recorded the date of attainment of the milestones. Age of attainment of walking and standing alone was moderately correlated with the PDI and had significant but low associations with later motor development. They were as good as the PDI in predicting later motor development and could be used in field studies for that purpose. Milestone age of attainment had significant but low correlations with MDI and later IQ. Height for age at 15 months was related to milestones and later IQ and motor development and accounted for some of the association between milestones and IQ. Milestone age of attainment may not be sensitive enough to be used as an indicator of later IQ.
Keywords: motor milestone, motor development, cognitive development, children, Bangladesh, MINIMat
Introduction
It is estimated that over 200 million children under 5 years of age in developing countries are failing to reach their developmental potential (Grantham‐McGregor et al. 2007) and poor nutrition and stimulation in the home are the main causes (Walker et al. 2011). There has been a recent increase in early childhood programmes; however, their impact on children's development is usually not well evaluated. One of the main reasons for lack of evaluation is that there are no readily available easy methods of assessing children's development.
Full developmental assessments with infant tests such as the Bayley Scales of Infant Development are the ‘gold standard’ but they are difficult to use at scale because they take a long time, require a quiet room and call for trained psychologists to conduct them. There is a great need for indicators of child development that can be used when evaluating large‐scale nutrition and early child development programmes so that measurements can be conducted in the field by relatively low‐skilled personnel.
Many investigators have used motor milestones such as age of crawling or walking (e.g. Faber et al. 2005; Olney et al. 2006; Adu‐Afarwuah et al. 2007) to evaluate the effects of programmes on children's development. Unfortunately, there is little data from developing countries demonstrating the relationship between motor milestones and current or future levels of child development. A large study in Finland used health records from welfare clinics and found that children at 12 months of age who had reached infant developmental milestones (walking and standing alone or saying one or two or more words) sooner were more likely to achieve better teacher ratings of achievement at 16 years of age and higher educational levels at 31 years of age than children who attained them later (Taanila et al. 2005), although the effect was small. However, a recent study in the Netherlands found very limited predictive ability of later intelligence quotient (IQ) from motor assessments in the first year in a healthy population (Roze et al. 2010). How poor nutrition in early childhood affects the predictive ability of age of attainment of milestones is unknown.
We examined the age of achievement of a series of motor milestones from age 3 to 15 months in over 2000 children. These children were then given full developmental assessments at age 18 months and tests of motor function and intelligence at 64 months. These tests were conducted as part of a large longitudinal study, MINIMat, which was run by the International Centre for Diarrhoeal Disease Research, Bangladesh (icddr,b) in a poor rural area of Bangladesh.
We hypothesised that motor milestone attainment in the first year of life would predict motor development and to a lesser extent mental development at 18 and 64 months of age. The aim was to examine the relationship between age of milestone attainment and internationally recognised tests of motor and mental development at 18 and 64 months of age. A second aim was to examine the relation of age of attainment of milestones with nutritional status and socio‐economic background. A final aim was to assess the relationship of the Bayley test scores at 18 months with the above tests at 64 months.
Key messages
Mothers' reports of age of attainment of motor milestones were reasonably accurate.
The ages of attainment of walking and standing alone had moderate correlations with the Bayley PDI at 18 months but low correlations with MDI.
Ages of attainment of motor milestones had significant but very low associations with IQ and motor development at 64 months.
Height for age in this undernourished population accounted for some of the association of age of attainment of milestones with later IQ.
The use of combined indicators of nutritional status and milestones and words used should be explored.
Materials and methods
Study area
The study was conducted in a poor rural area located 53 km southeast of Dhaka.
The study was part of a pregnancy supplementation trial, in which the effect of food and micronutrient supplementation on fetal growth and size at birth was evaluated.
Sample
All pregnancies in the study area were identified during monthly home visits. Pregnant women were randomised to either an early start of a 600 kcal daily food supplement (at week 8) or the routine supplementation (later start, around week 17). They were also randomised to either daily 60 mg iron/folate (routine); 30 mg iron/folate; or multiple micronutrient supplementation (UNICEF tablets, 15 micronutrients; UNICEF/UNU/WHO 1999).
After delivery, a subsample (n = 2853), comprising all singleton infants born between May 2002 and December 2003, was selected for long‐term follow‐up of the children's development. Supplements in women had small benefits only to the development of infants of mothers with low body mass index at 7 months (Tofail et al. 2008), but showed no effect at 18 months (Hamadani et al. 2011).
Measurements
Family
The following measurements were made in the main study at enrolment and were used in this study; parental education, wealth index (computed by factor analysis of ratings of quality of housing, household possessions and land ownership; Saha et al. 2008) and income/expenditure deficit. In addition, stimulation in the home was measured at 18 months using the infant and toddler version of the home observations for measurement of the environment (HOME) inventory (Caldwell 1967).
Children
Children's gestational age at birth was measured using ultrasound and weight, length and head circumference were measured at birth. Weight, length and head circumference were also measured at 15 months of age by trained field workers (FWs) according to the World Health Organization (WHO) guidelines (WHO 1983; Frisancho 1990) in the homes. All measurements were converted to standard scores of the WHO growth references (WHO 2006a).
At 18 months, we assessed the children's development on the psychomotor development index (PDI) and mental development index (MDI) of the revised version of Bayley Scales of Infant Development (Bayley 1993) and the interobserver reliabilities of each of the testers with the trainer in 10 children (ages from 13 to 30 months) ranged from r = 0.88 to 0.99 [intraclass correlations (ICCs)] for both MDI and PDI.
At 64 months, we assessed their motor development using the Movement Assessment Battery for Children (MABC; Henderson & Sugden 1992) and IQ on Wechsler Preschool and Primary Scale of Intelligence (WPPSI; Wechsler 2002). The WPPSI was modified for Bangladesh to ensure that the wording and pictures were culturally appropriate; however, the constructs underpinning the test items were not changed. For example the analogy ‘dog and kennel’ was substituted with ‘bird and cage’ as they do not have kennels in Bangladesh. The scores appeared to have good validity as they were significantly associated with the child's nutritional status, socio‐economic conditions and home stimulation measures. They were also significantly related to the Bayley MDI scores assessed at 18 months of age.
The test‐retest reliability by four testers after 7–10 days for 60 children was good for full scale IQ (ICC r = 0.98, P < 0.001, n = 53) and moderate for MABC (r = 0.68, P < 0.001, n = 51). Ongoing interobserver reliabilities were conducted during the tests and the ICCs ranged from 0.77 to 1.0 for the five testers (n = 85).
The MABC was piloted in the same population of children. We used three subscales of the MABC viz. threading the beads, catching the bean bag and jumping over cord, after they were slightly modified for use in Bangladeshi children. The other subscales were found either too easy or too difficult for Bangladeshi children of this age and were excluded. The children received scores on each subscale based on their performance on a scale of 0 to 5. Higher item scores indicated failure to perform the task and lower scores indicated greater success. The scores of the three subscales were summed to calculate a total score on the MABC ranging from 0 to 15.
Motor milestone assessment
We report seven motor milestones, which are shown in Table 2, six of them were taken from the WHO Multicentre Growth Reference Study (WHO‐MGRS; WHO 2006b). We added one milestone, ‘pull to stand’ (taking the definition from the same item in the Bayley scales) that was very easy to identify and we thought might be important.
Table 2.
Median, quartiles and 10th and 90th percentiles of age of attainment of each milestone (months) by Mothers' and field worker's (FW) report by sex and total
| Milestone | Median | 25th and 75th percentiles | 10th and 90th percentiles | Mann–Whitney U P‐value for sex differences |
|---|---|---|---|---|
| Sits alone according to mother (n = 2203) | 5.97 | 5.50, 6.60 | 5.10, 7.17 | 0.009 |
| Boys (n = 1141) | 6.0 | 5.50, 6.63 | 5.13, 7.23 | |
| Girls (n = 1062) | 5.90 | 5.43, 6.53 | 5.07, 7.13 | |
| Sits alone according to FW (n = 2255) | 6.83 | 5.97, 7.10 | 5.83, 8.13 | 0.549 |
| Boys (n = 1167) | 6.83 | 5.97, 7.10 | 5.87, 8.10 | |
| Girls (n = 1088) | 6.80 | 5.97, 7.10 | 5.83, 8.13 | |
| Pulls self to stand according to mother (n = 2214) | 8.30 | 7.60, 9.10 | 6.87, 10.20 | 0.759 |
| Boys (n = 1146) | 8.28 | 7.60, 9.08 | 6.93, 10.31 | |
| Girls (n = 1068) | 8.33 | 7.57, 9.13 | 6.83, 10.13 | |
| Pulls self to stand according to FW (n = 2257) | 8.97 | 8.00, 9.97 | 7.13, 11.07 | 0.455 |
| Boys (n = 1167) | 8.93 | 8.00, 9.93 | 7.13, 11.04 | |
| Girls (n = 1090) | 8.97 | 8.00, 9.97 | 7.10, 11.07 | |
| Stands assisted according to mother (n = 1923) | 8.67 | 7.83, 9.57 | 7.37, 10.50 | 0.754 |
| Boys (n = 1020) | 8.70 | 7.83, 9.57 | 7.33, 10.53 | |
| Girls (n = 903) | 8.67 | 7.83, 9.57 | 7.38, 10.50 | |
| Stands assisted according to FW (n = 2253) | 8.97 | 8.00, 9.97 | 7.17, 10.97 | 0.577 |
| Boys (n = 1165) | 8.97 | 8.00, 9.97 | 7.17, 10.97 | |
| Girls (n = 1088) | 8.97 | 8.00, 9.97 | 7.16, 10.97 | |
| Walks assisted according to mother (n = 2171) | 10.00 | 9.07, 11.20 | 8.37, 12.26 | 0.265 |
| Boys (n = 1121) | 10.03 | 9.08, 11.23 | 8.40, 12.29 | |
| Girls (n = 1050) | 10.00 | 9.03, 11.17 | 8.30, 12.13 | |
| Walks assisted according to FW (n = 2242) | 10.33 | 9.30, 11.90 | 8.90, 14.90 | 0.439 |
| Boys (n = 1158) | 10.77 | 9.80, 11.90 | 8.90, 14.90 | |
| Girls (n = 1084) | 10.25 | 9.27, 11.90 | 8.90, 14.88 | |
| Stands alone according to mother (n = 1984) | 11.60 | 10.5, 13.63 | 9.45, 14.63 | <0.001 |
| Boys (n = 1020) | 11.72 | 10.67, 13.7 | 9.60, 14.73 | |
| Girls (n = 964) | 11.47 | 10.2, 13.47 | 9.22, 14.6 | |
| Stands alone according to FW (n = 2143) | 12.00 | 11.00, 15.00 | 9.97, 15.20 | <0.001 |
| Boys (n = 1094) | 12.10 | 11.03, 15.03 | 10.03, 15.27 | |
| Girls (n = 1049) | 11.97 | 10.93, 14.97 | 9.93, 15.17 | |
| Walks alone according to mother (n = 1987) | 13.90 | 12.27, 14.77 | 11.30, Not achieved | <0.001 |
| Boys (n = 1022) | 13.97 | 12.50, 14.83 | 11.51, Not achieved | |
| Girls (n = 965) | 13.73 | 12.0, 14.73 | 11.12, Not achieved | |
| Walks alone according to FW (n = 2054) | 15.00 | 14.80, 15.20 | 11.90, Not achieved | 0.002 |
| Boys (n = 1052) | 15.00 | 14.83, 15.23 | 11.93, Not achieved | |
| Girls (n = 1002) | 14.97 | 12.23, 15.17 | 11.83, Not achieved |
We also added a second type of ‘creep’ because some children according to the WHO study never creep on hands and knees (WHO 2006b). The definition of the additional creep was ‘Moves forward or backward easily by any means other than hand and knee crawl (using stomach & arms, buttock & hand/leg, hands & feet, etc. – at least three movements in a row)’.
Trained FWs visited the homes monthly from birth and assessed the age of attainment of the milestones from 3 months of age until the children were 12 months of age; thereafter, they visited the children at 15 months. When the child was 2 months old, mothers were given a chart with pictures of each milestone arranged in order of difficulty and asked to keep a note on the chart when the child attained any of the first three milestones. They were also shown how to record on the chart when the child attained a milestone. During the next monthly visit, the FWs checked the recording then assessed the child for the first three milestones. Mother's reports and the FW's findings were recorded on a questionnaire at each visit. Then the FW ticked the attained milestones on the chart and showed the mother the next three milestones she should look for until the next visit. If the mother was illiterate, she was told to ask another literate family member or neighbour to note down the date of attainment on the milestone chart for her. At every visit, the FWs assessed the child beginning from the last attained milestone marked on the chart. The whole examination took around 10 min.
The FWs were instructed to ensure that the children were happy during the observations and if necessary to take sufficient time to calm them before performing the assessment. If a child was so sick that it interfered with the assessment, the FWs revisited the child when he/she was better.
Statistical analysis
We first calculated the median and percentiles for age of attainment of each milestone for both the mother's record and the FW's observation. We then examined concordance between Mothers' records and FWs' observations for each milestone with correlations. We also calculated the percentage of times that the mother recorded a milestone attained when the FW recorded it as not attained at the next visit and the number of times the mother failed to record the milestone attained when the FW recorded it attained.
As children were examined at different time points, we selected those children who were tested at 64 months and only included those in the analysis. Some children had not yet achieved the last milestones when the study finished at 15 months of age. Rather than omit this important group of children, we included them in all the analyses by giving the children who had not achieved the milestone an arbitrary age of 17 months. We also tried using 20 months instead of 17 and it made little difference to the findings except that the standard deviations for mean age of attainment of standing and walking alone were larger when 20 months were used instead of 17 months.
We then examined the association between the attainment of each milestone and scores obtained on tests at 18 and 64 months of age by computing correlations controlling for precise age at testing. We also computed correlations between the age of attainment of the milestones with family and child characteristics. We then conducted a series of multiple regressions of the developmental outcomes (Bayley Scores, WPPSI and MABC) to identify the effect of the selected milestones on development, controlling for child and parental characteristics and home background variables. We conducted additional regressions for walk and stand alone using quartiles for the age of attainment instead of age as a continuum to allow for children who had not attained the milestones at the end of the study. The earliest quartile was contrasted with the later three quartiles. Unlike the previous regressions, we only controlled for age and sex in these regressions to simplify the analyses. In addition, if milestones were to be used in the field as an indicator of future development, it would not be feasible to control for other variables. We also examined the role of height for age on the outcomes and whether it explained some of the relation between age of milestone attainment and IQ by running the regressions with and without height for age.
Finally, we computed similar regressions predicting the 64‐month tests using the Bayley MDI and PDI at 18 months instead of the milestones in order to determine the predictive validity of the Bayley in this sample.
Ethics
Written informed consent was obtained from mothers at their enrolment in the study and again when their child was being tested at 64 months of age. The project was approved by the institutional review board of icddr,b.
Results
The characteristics of the 2260 children who were tested at 64 months are shown in Table 1. In general, the children came from poor homes and had poor nutritional status.
Table 1.
Characteristics of the sample (n = 2260)
| Variable | Mean ± SD or % (n) |
|---|---|
| Family characteristics | |
| Fathers' education (<5th grade) | 44% (2241) |
| Mothers' education (<5th grade) | 45% (2255) |
| Poor housing | 21% (2254) |
| Child's characteristics | |
| Male | 52% (2260) |
| Birthweight (g) | 2688 ± 394 (2127) |
| Low birthweight (%) | 31% (2127) |
| Gestational age at birth (weeks) | 39.2 ± 1.7 (2137) |
| 15 months measures | |
| HAZ | −2.0 ± 1.1 (2005) |
| WAZ | −1.5 ± 1.1 (2005) |
| WHZ | −0.8 ± 1.1 (2006) |
| 18 months measures | |
| HAZ | −2.0 ± 1.1 (2058) |
| WAZ | −1.6 ± 1.0 (2059) |
| WHZ | −0.9 ± 1.0 (2060) |
| Age at Bayley test (months) | 18.6 ± 1.2 (1886) |
| MDI | 78.9 ± 12.3 (1885) |
| PDI | 93.8 ± 13.3 (1885) |
| HOME | 83.5 ± 6.96 (2116) |
| 64 months measures | |
| HAZ | −1.6 ± 0.9 (2260) |
| WAZ | −1.8 ± 0.9 (2260) |
| Age at test (months) | 64.5 ± 1.9 (2260) |
| WPPSI IQ | 75.1 ± 9.4 (2260) |
| MABC | 6.2 ± 2.8 (2260) |
| HOME | 8.86 ± 4.98 (2187) |
HAZ, height for age z‐score; HOME, home observation for measurement of environment; MABC, Movement Assessment Battery for Children; MDI, mental development index; PDI, psychomotor development index; WAZ, weight for age z‐score; WHZ, weight for height z‐score; WPPSI IQ, Wechsler Preschool and Primary Scale of Intelligence full scale intelligence quotient.
Age of attainment
The median time of attainment of each milestone along with the 10th, 25th, 75th and 90th percentiles according to mother and examiner are shown in Table 2. We omitted ‘creep’ because of concerns over the testing. We had added ‘move forward in any way’ to the data collection but the field workers appeared to have been confused by this and some neglected to also check ‘creep on hands and knees’ if it occurred after ‘move forward in any way’.
There was a significant sex difference with girls attaining several milestones earlier than boys. However, the differences between the sexes in median age of attaining milestones were small. The difference was 8 days in the age for standing alone and 7 days in the age of walking. At the end of the study, 24 (0.7%) of the children could not sit alone, 23 (1%) could not pull self to stand, 7 (0.4%) could not stand with assistance, 54 (2.5%) could not walk with assistance, 195 (6%) had not attained ‘standing alone’ and 554 (18.6%) had not attained walking alone. All the milestones were attained in the expected hierarchical order. The difference in months between the 10th and 90th percentiles generally increased for milestones with increasing age of attainment. The numbers varied slightly from milestone to milestone; this is due to the children not being at home during the visit. As our concern here is the validity of the measurements themselves and not the representative of the district or the initial sample, we did not control for loss.
Concurrence between the mothers report and FW report
The FW consistently recorded an older age of attainment than the mothers and the differences between the medians were less than 1 month for all milestones except for ‘age of walking’ (1.1 month). The larger difference at 12 months was probably due to monthly visiting being stopped then. The concurrence between the mother's report and FW's report was good with correlations ranging from r = 0.90 for ‘pull to stand’ to 0.95 for ‘stands with help’ and ‘walking alone’ (Table 3). The percentage of mothers reporting attainment of a milestone occurring before the last FW report of ‘not attained’ was small (maximum 8.9% for ‘stand alone’) and is given in Table 3. Mothers were extremely unlikely to record a date of attainment later than the FWs' report.
Table 3.
Correlations between field worker (FW)'s and Mothers' reports and the percent of mother's reports that occurred outside the possible time according to the FW (between oldest negative and youngest positive recording)
| Milestones | % of Mothers' report of age of attainment younger than FW's highest age of not achieving | % of Mothers' report of age of attainment older than FW's age of attainment | Correlation between FW and mother's report |
|---|---|---|---|
| Sits alone | 1.6% | 0 | 0.93 |
| Crawl | Dropped | ||
| Pulls self to stand | 7.8% | 0 | 0.90 |
| Stands assisted | 2.8% | 0 | 0.95 |
| Walks assisted | 6.5% | 0 | 0.94 |
| Stands alone | 8.9% | 0.1% | 0.93 |
| Walks alone | 1.9% | 1 case | 0.95 |
Correlations of milestones with the developmental test scores
We examined the correlations between the age of attainment of the milestones (reported by the mothers and FWs) and the developmental measures at 18 and 64 months (Table 4). The milestones attained at an older age were more highly correlated with the Bayley scores at 18 months and WPPSI IQ and MABC at 64 months than those attained earlier.
Table 4.
Correlation of mother's and field worker (FW)'s report of attained age of each milestone with Bayley tests (MDI and PDI) at 18 months and with WPPSI IQ and Movement Assessment Battery for Children (MABC) at 64 months controlling for age at test as well as correlation of Bayley tests with WPPSI IQ and MABC
| Milestones | MDI † | PDI † | WPPSI IQ † | MABC †‡ | ||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| FW | n | Mother | n | FW | n | Mother | n | FW | n | Mother | n | FW | n | Mother | n | |
| Sits alone | −0.08* | 1882 | −0.13** | 1859 | −0.13** | 1882 | −0.19** | 1859 | −0.07* | 2252 | −0.14** | 2225 | 0.07* | 2252 | 0.09** | 2225 |
| Pulls self to stand | −0.07* | 1882 | −0.10** | 1851 | −0.26** | 1882 | −0.30** | 1851 | −0.11** | 2254 | −0.14** | 2211 | 0.13** | 2254 | 0.15** | 2211 |
| Stands assisted | −0.11** | 1881 | −0.14** | 1844 | −0.29** | 1881 | −0.31** | 1844 | −0.16** | 2250 | −0.18** | 2201 | 0.13** | 2250 | 0.15** | 2201 |
| Walks assisted | −0.13** | 1873 | −0.15** | 1851 | −0.34** | 1873 | −0.37** | 1851 | −0.19** | 2239 | −0.21** | 2204 | 0.16** | 2239 | 0.16** | 2204 |
| Stands alone | −0.13** | 1799 | −0.18** | 1787 | −0.31** | 1799 | −0.37** | 1787 | −0.16** | 2140 | −0.21** | 2127 | 0.17** | 2140 | 0.19** | 2127 |
| Walks alone | −0.20** | 1730 | −0.23** | 1718 | −0.34** | 1730 | −0.38** | 1718 | −0.21* | 2051 | −0.25** | 2038 | 0.19** | 2051 | 0.23** | 2038 |
| Bayley | Psychologist | n | Psychologist | n | ||||||||||||
| MDI | 0.36** | 1882 | −0.16** | 1882 | ||||||||||||
| PDI | 0.27** | 1882 | −0.18** | 1882 | ||||||||||||
MDI, mental development index; PDI, psychomotor development index; WPPSI IQ, Wechsler Preschool and Primary Scale of Intelligence full scale intelligence quotient. *P < 0.05, **P < 0.01. †controlling for age at test. ‡higher scores represent poorer achievement.
18 months
As hypothesised, the age of attainment of milestones correlated more strongly with the motor scale (PDI) than with the mental scale (MDI). The highest correlations with the MDI reached r = 0.23, (P < 0.001) with ‘walks alone’, whereas correlations with PDI reached r = 0.38 (P < 0.001), also with ‘walks alone’.
64 months
All correlations between the age of attainment of milestones and the 64‐month IQ and MABC were significant but low. Surprisingly, those with the IQ were little different from those with the MABC. The milestones with the highest correlations were stands alone (r = −0.21, for IQ and r = 0.19, for MABC, P < 0.001 for both) and walks alone (r = −0.25, for IQ and r = 0.23 for MABC, P < 0.001 for both).
We also examined the correlations between the Bayley scores and the 64‐month tests. The correlation between MDI and IQ was moderate (r = 0.36, P < 0.001) and higher than with any milestone, whereas the correlation between MDI and MABC was low (r = −0.16, P < 0.001). The correlation between the 18‐month PDI and the motor scores at 64 months (MABC) was actually slightly lower (r = −0.18, P < 0.001) than those of ‘stands alone’ and ‘walks alone’ milestones with MABC [r = 0.19 and r = 0.23, respectively (P < 0.001 for both)]. However, the PDI correlation with IQ was slightly higher (r = 0.27, P < 0.001) than those with any milestone.
In general, the correlations with the Mothers' report of age of attainment were slightly higher than those with the FW's report, except in case of walks assisted where they were the same size. The FWs recorded their observations only once a month, whereas the mother observed the child every day so it should be more accurate. Furthermore, the FW–mother concurrence was good; therefore, we have used the Mothers' report in all further analyses.
Correlations of milestone age with family and child characteristics
The milestones were related to many parent and home background variables (Table 5) but most of the correlations were very small. Correlations with child characteristics were slightly higher; correlations with birthweight and gestational age at birth were between r = 0.18–0.28 (P < 0.001 for all) The nutritional status variables ‘height for age’ and ‘weight for age at 15 months’ were the most highly correlated with ‘walk alone’ and ‘stand alone’ (Height for age: r = −0.27 and r = −0.25, weight for age: r = −0.25 and r = −0.23, respectively (P < 0.001 for both).
Table 5.
Significant correlations between age of attainment of milestones (Mothers' report) and parental characteristics, home background and child characteristics
| Milestones | Sits alone | Pulls self to stand | Stands assisted | Walks assisted | Stands alone | Walks alone |
|---|---|---|---|---|---|---|
| Family characteristics | ||||||
| Assets | −0.07** | −0.05* | −0.043* | −0.07** | −0.12** | −0.13** |
| Housing | −0.04* | −0.02 | −0.08** | −0.10** | −0.09** | −0.13** |
| Total HOME at 18 months | −0.12** | −0.06* | −0.13** | −0.15** | −0.15** | −0.19** |
| Mother's education | −0.08** | −0.07* | −0.09** | −0.12** | −0.15** | −0.19** |
| Father's education | −0.07** | −0.07** | −0.10** | −0.13** | −0.13** | −0.14** |
| Child characteristics | ||||||
| Birthweight | −0.22** | −0.20** | −0.22** | −0.21** | −0.19** | −0.20** |
| Gestational age at birth | −0.28** | −0.22** | −0.22** | −0.20** | −0.18** | −0.21** |
| HAZ at 15 months | −0.18** | −0.15** | −0.18** | −0.22** | −0.25** | −0.27** |
| WAZ at 15 months | −0.24** | −0.17** | −0.19** | −0.23** | −0.23** | −0.25** |
| WHZ at 15 months | −0.19** | −0.11** | −0.12** | −0.14** | −0.13** | −0.15** |
HAZ, height for age z‐score; HOME, home observation for measurement of environment; WAZ, weight for age z‐score; WHZ, weight for height z‐score. *P < 0.05, **P < 0.01.
Many of these variables were also related to the child development outcomes. It is therefore likely that these socio‐economic and child variables accounted for some of the relation between milestone attainment age and later child development.
Multiple regressions
We then examined the effect of age of attainment of milestones on the developmental test scores, adjusting for possible confounding effects of the above variables. We computed a series of multiple regressions predicting the four outcome measures (IQ, MABC, MDI and PDI) from age of attainment of milestones controlling for child and family covariates. We used ‘walking alone’ and ‘standing alone’, the two milestones that were attained last and were the most highly associated with the developmental outcomes. We also used ‘sitting alone’, the first one attained. We controlled for sex, age at test, gestational age at birth, birthweight, height‐for‐age z‐score (HAZ) at 15 months, parent's education, housing index, assets index, birth order and HOME at 18 months. A large number of covariates were significant predictors of IQ (Table 6). HOME and HAZ were the variables most consistently associated with IQ and MDI scores. Fewer variables were related to motor development.
Table 6.
Results from a series of multiple regressions using milestones (age of walking, standing and sitting) to predict developmental outcomes at 18 and 64 months, giving standardised regression coefficients (β) and 95% confidence interval (CI) for the effect of milestones
| Variables | MABC‐64 months | WPPSI IQ‐64 months | MDI‐18 months | PDI‐18 months |
|---|---|---|---|---|
| β (95% CI) | β (95% CI) | β (95% CI) | β (95% CI) | |
| n = 1767 | n = 1767 | n = 1535 | n = 1535 | |
| Age of walking alone (months) | 0.21 (0.23, 0.37)** | −0.12 (−0.72, −0.34)** | −0.13 (−2.7, −1.7)** | −0.32 (−2.5, −1.8)** |
| Other significant variables | Age at developmental test, sex, HAZ | Age at developmental test, parent's education, housing, birth order, HOME, HAZ | Age at developmental test, asset index, birth order, HOME, HAZ | Age at developmental test, gestational age at birth, HOME, HAZ |
| R 2 | 0.10 | 0.28 | 0.18 | 0.21 |
| n = 1764 | n = 1764 | n = 1534 | n = 1534 | |
| Age of standing alone (months) | 0.18 (0.17, 0.29)** | −0.11 (−0.60, −0.26)** | −0.10 (−0.86, −0.32)** | −0.32 (−2.2, −1.6)** |
| Other significant variables | Age of test, sex, HAZ | Age of test, parent's education, HOME, housing, birth order, HAZ | Age of test, HOME, asset index, birth order, HAZ | Age of test, gestational age at birth, HOME, HAZ |
| R 2 | 0.09 | 0.28 | 0.17 | 0.21 |
| n = 1760 | n = 1760 | n = 1530 | n = 1530 | |
| Age of sitting alone (months) | 0.10 (0.16, 0.47)** | −0.06 (−1.0, −0.15)* | −0.06 (−1.6, −0.14)* | −0.13 (−2.9, −1.3)** |
| Other significant variables | Age of test, sex, HAZ | Age of test, parent's education, HOME, housing, birth order, HAZ | Age of test, HOME, birth order, asset index, HAZ | Age of test, gestational age at birth, HOME, HAZ |
| R 2 | 0.07 | 0.28 | 0.16 | 0.13 |
HAZ, height for age z‐score; HOME, home observation for measurement of environment; MABC, Movement Assessment Battery for Children; MDI, mental development index; PDI, psychomotor development index; WPPSI IQ, WPPSI full scale intelligence quotient.
Model: Entered: Step 1: age at specific developmental test, sex, age of milestone; Step 2: gestational age at birth, birthweight, height for age z‐score at 15 months; Step 3: maternal and paternal education, housing index, assets, birth order, HOME at 18 months. *P < 0.05, **P < 0.001.
All the regression coefficients for the age of attainment of the milestones predicting the four development outcomes were significant (P < 0.001 for all except ‘age of sitting alone’ when P < 0.05 for regressions on IQ and MDI; Table 6).
Association of milestones with Bayley score at 18 months
The effect of age of attainment of each milestone on the Bayley scores adjusting for covariates was higher on PDI than on MDI. One standard deviation (SD) score of the age of walking (2.0 months) and standing (2.2 months) had an effect of −0.32 SD on PDI (both P < 0.001), whereas one SD of the age of sitting (0.96 month) had an effect of 0.13 SD on PDI (P < 0.001). The effect of one SD of milestone age of attainment on MDI was −0.13 SD and −0.10 SD (both P < 0.001) for age of walking and standing, respectively, and −0.06 SD for the age of sitting (P = 0.020).
Association of milestones with IQ and MABC at 64 months
After adjusting for covariates, the effect of the milestones on IQ at 64 months was significant but small. A difference of one SD in the age of walking or standing alone was associated with an effect of 0.12 and 0.11 SD of IQ (P < 0.001), respectively. The effect of one SD of the age of sitting alone was associated with 0.06 SD (P = 0.009) of IQ.
The effect of the age of attainment of milestones on motor development (MABC) at 64 months was also significant and slightly larger. One SD of the age of walking had an effect on the MABC score of 0.21 SD (P < 0.001) and one SD of the age of standing alone and sitting alone had effects of 0.18 and 0.10 SD (P < 0.001), respectively, on MABC.
Predictive ability of the Bayley scores
We then examined the independent effect of the Bayley scores at 18 months on the IQ and MABC, using similar regressions (Table 7). The adjusted effect of a standard score of MDI was associated with 0.19 SD of IQ (P < 0.001) and slightly less (−0.12 SD, P < 0.001) of the MABC. The predictive ability of PDI was similar; one SD of PDI was associated with 0.14 SDs (P < 0.001) of IQ and with −0.16 SD (P < 0.001) of MABC.
Table 7.
Results from a series of multiple regressions predicting the WIPPSI IQ and Movement Assessment Battery for Children (MABC) scores at 64 months from the Bayley MDI and PDI at 18 months controlling for child and family covariates, giving standardised regression coefficients (β) and 95% confidence interval (CI) for the effect of Bayley scores
| Variables | MABC‐64 months | WPPSI IQ‐64 months |
|---|---|---|
| β (95% CI) | β (95% CI) | |
| n = 1551 | n = 1551 | |
| MDI 18 months | −0.12 (−0.04, −0.02)* | 0.20 (0.11, 0.18)* |
| Other significant variables | Age, sex, HAZ | Age of test, parent's education, HOME, housing, birth order, HAZ |
| R 2 | 0.07 | 0.30 |
| n = 1551 | n = 1551 | |
| PDI 18 months | −0.16 (−0.04, −0.02)* | 0.14 (0.06, 0.12)* |
| Other significant variables | Age, sex, HAZ | Age of test, parent's education, HOME, housing, birth order, HAZ |
| R 2 | 0.08 | 0.29 |
HAZ, height for age z‐score; HOME, home observation for measurement of environment; MDI, mental development index; PDI, psychomotor development index; WPPSI IQ, Wechsler Preschool and Primary Scale of Intelligence full scale intelligence quotient. Model: Entered: age at specific developmental test, sex, MDI/PDI, gestational age at birth, birthweight, height for age z‐score at 15 months, maternal and paternal education, housing index, assets, birth order, HOME at 18 months. *P < 0.001.
The use of quartiles and the role of nutritional status
Because of the number of children who did not attain walking or standing alone by 15 months, we ran another set of regressions using quartiles of age of walking and standing alone (Table 8). In addition, we explored whether nutritional status (HAZ) had an independent effect on the outcomes and explained any of the relationship between milestones and IQ. To simplify the regressions, we only controlled for age and sex in a first step and then added height for age to the model in a second step.
Table 8.
Results from a series of multiple regressions of developmental outcomes at 18 and 64 months on quartiles of age of walking and standing, giving standardised regression coefficients (β) and 95% confidence interval (CI) for the effect of quartiles
| Variables | MABC 64 months | WPPSI IQ 64 months | MDI 18 months | PDI‐18 months |
|---|---|---|---|---|
| β (95% CI) | β (95% CI) | β (95% CI) | β (95% CI) | |
| Age of walking | n = 1849 | n = 1849 | n = 1778 | n = 1778 |
| Step 1 | ||||
| Quartile 2 | 0.06 (0.02, 0.71) P = 0.039 | −0.05 (−0.205, 0.21) P = 0.109 | −0.03 (−2.43, 0.65) P = 0.255 | −0.11 (−4.83, −1.62) P < 0.001 |
| Quartile 3 | 0.11 (0.34, 1.03) P < 0.001 | −0.18 (−4.94, −2.67) P < 0.001 | −0.13 (−5.23, −2.09) P < 0.001 | −0.16 (−6.48, −3.21) P < 0.001 |
| Quartile 4 | 0.28 (1.50, 2.20) P < 0.001 | −0.27 (−6.80, −4.49) P < 0.001 | −0.25 (−8.98, −5.81) P < 0.001 | −0.44 (−15.56, −12.25) P < 0.001 |
| Step 2 | ||||
| Quartile 2 | 0.04 (−0.06, 0.63) P = 0.110 | −0.01 (−1.37, 0.82) P = 0.623 | −0.01 (−1.88, 1.17) P = 0.650 | −0.09 (−4.24, −1.06) P = 0.001 |
| Quartile 3 | 0.09 (0.20, 0.90) P = 0.002 | −0.13 (−3.85, −1.63) P < 0.001 | −0.10 (−4.40, −1.28) P < 0.001 | −0.13 (−5.61, −2.34) P < 0.001 |
| Quartile 4 | 0.25 (1.26, 1.98) P < 0.001 | −0.18 (−5.00, −2.70) P < 0.001 | −0.20 (−7.46, −4.23) P < 0.001 | −0.39 (−13.96, −10.58) P < 0.001 |
| HAZ‐15 months | −0.29 (−0.40, −0.17) P < 0.001 | 0.27 (1.88, 2.63) P < 0.001 | 0.17 (1.41, 2.48) P < 0.001 | 0.16 (1.50, 2.62) P < 0.001 |
| Age of standing | n = 1733 | n = 1733 | n = 1665 | n = 1665 |
| Step 1 | ||||
| Quartile 2 | 0.06 (0.04, 0.77) P = 0.031 | −0.05 (−2.32, 0.09) P = 0.070 | −0.08 (−3.87, −0.56) P = 0.009 | −0.10 (−4.55, −1.33) P < 0.001 |
| Quartile 3 | 0.09 (0.18, 0.90) P = 0.003 | −0.10 (−3.29, −0.92) P < 0.001 | −0.08 (−3.89, −0.67) P = 0.006 | −0.14 (−5.40, −2.26) P < 0.001 |
| Quartile 4 | 0.20 (0.89, 1.61) P < 0.001 | −0.17 (−4.50, −2.16) P < 0.001 | −0.19 (−6.89, −3.66) P < 0.001 | −0.30 (−10.06, −6.91) P < 0.001 |
| Step 2 | ||||
| Quartile 2 | 0.05 (−0.02, 0.71) P = 0.062 | −0.03 (−1.86, 0.46) P = 0.237 | −0.07 (−3.54, −0.28) P = 0.022 | −0.09 (−4.25, −1.07) P = 0.001 |
| Quartile 3 | 0.07 (0.08, 0.80) P = 0.018 | −0.07 (−2.48, −0.20) P = 0.021 | −0.06 (−3.28, −0.09) P = 0.038 | −0.12 (−4.84, −1.73) P < 0.001 |
| Quartile 4 | 0.17 (0.73, 1.44) P < 0.001 | −0.11 (−3.25, −0.98) P < 0.001 | −0.15 (−5.87, −2.63) P < 0.001 | −0.27 (−9.11, −5.97) P < 0.001 |
| HAZ‐15 months | −0.12 (−0.45, −0.21) P < 0.001 | 0.29 (2.04, 2.82) P < 0.001 | 0.17 (1.51, 2.63) P < 0.001 | 0.16 (1.37, 2.46) P < 0.001 |
Model: Step 1: age at each specific developmental test, sex, quartiles of age of walking/standing based on mother's report; Step 2: height for age z‐score at 15 months.
Quartiles of age of attaining milestones and Bayley scores at 18 months
When adjusted for age and sex only, the effect of being in the latest quartile for age of attainment of walking and standing alone on the Bayley PDI was substantial (0.44 and 0.3 SD (both P < 0.001), respectively) but the effect on MDI was smaller [0.25 and 0.19 SD (P < 0.001), respectively]. HAZ had a significant effect (P < 0.001) on both Bayley scores. One SD HAZ was associated with 0.17 SD of MDI and 0.16 SD of PDI. The effect of being in the slowest quartile of age of walking on MDI was reduced by 25% to 0.2 SD when HAZ was entered and the effect on PDI was reduced by 11.3% to 0.39 SD.
Quartiles of age of attaining milestones and later IQ
When adjusted for age and sex only, there was no significant difference between the earliest two quartiles of age of walking or standing alone in their effect on later IQ. Being in the third quartile (walking between 13.9 and 14.8 months, or standing between 11.6 and 13.6 months) had a significant but small effect on IQ compared with those in the earliest quartile (walking by 12.3 months, standing by 10.5 months). Children in the third quartile of age of walking and standing alone had a lower IQ of 0.18 SD and 0.1 SD (both P < 0.001) of IQ, respectively. Being in the fourth or latest quartile (not walking at 14.8 months, or standing at 13.6 months) was associated with lower IQ of 0.27 and 0.17 of a SD (P < 0.001), respectively.
In the second step, one SD of HAZ had an effect of 0.27 SD (P < 0.001) on IQ and the loss associated with being in the latest quartile for age of walking alone was reduced by 33.3% to 0.18 SD of IQ. Similarly, the effect of being in the latest quartile for age of standing was reduced by 35.3% to 0.11 SD when HAZ entered the regression. A combination of a reduction of one SD of HAZ and being in the latest quartile of age of walking had a substantial effect of 0.45 SD on IQ.
Quartiles of age of attaining milestones and MABC
Being in the latest quartile for age of walking and standing alone had a similar effect on the MABC [0.28 and 0.2 SD (both P < 0.001), respectively] to that of their effect on IQ. One SD of HAZ had an effect of 0.29 SD (P < 0.001) on the MABC. A combination of HAZ and being in the latest quartile of age of walking had a combined effect of 0.54 SD on MABC.
Discussion
We have used a large, longitudinal, population‐based study to examine the association between age of attainment of motor milestones and other internationally recognised measures of motor and mental development at 18 and 64 months. We found that Mothers' reports of milestones were reasonably accurate and probably gave more accurate data than a trained FW visiting once a month. However, it should be noted that the mothers were given a chart to fill in and were taught how to recognise the milestones and were visited once a month. This should be distinguished from mothers recall on one occasion that is sometimes used to evaluate interventions and may not be as valid.
Associations of milestones with later motor development
We found that ages of attainment of the milestones were moderately well correlated with a full assessment of motor development (PDI) with the Bayley Scales at 18 months. The later milestones (walking and standing alone) that were attained nearest in time to when the Bayley PDI was assessed were more highly associated than those attained earlier. The milestones had lower but significant correlations with motor development at 64 months of age. When adjusted for covariates, the effects of milestones on motor development at 64 months were slightly larger than for mental development, but generally small. However, even the Bayley mental and motor scores, (which were assessed 3 months later than the last milestone assessment) had only low associations with IQ and motor development at 64 months. When adjusted for covariates, the effect of PDI on MABC was small and almost similar to those of the later milestones. This suggests that if motor development is the focus of interest, age of attainment of motor milestones are as good an indicator as the Bayley PDI, albeit both have only weak associations with later motor development. However milestones are easier to assess and can be recorded at home by lower skilled field workers.
Associations of milestones with later mental development
The correlations between ages of attainment of the milestones and mental development at 18 and 64 months were low but significant. After adjusting for covariates, the effect of one SD in age of attainment of every milestone was 0.12 SD or less of IQ at 64 months.
The correlation between the Bayley MDI at 18 months and IQ at 64 months was moderate in size (r = 0.36) and higher than for any milestone. When adjusted for covariates, the effect of MDI was small (one SD of MDI was associated with 0.20 SD of later IQ), although better than any milestones. Thus, even a full developmental assessment had limited ability to predict future IQ. It has been recognised previously that it is difficult to predict future IQ in the first 18 months (Colombo 1993) and motor milestones may not be sensitive enough of later IQ to be useful as an indicator of later development.
Comparisons with language tests
We previously developed a short language inventory for rural Bangladeshi children based on Mothers' report (Hamadani et al. 2010). It had a comprehension and an expression scale and contained 60 words. The inventory was developed from the concepts of the MacArthur Communicative Development inventory (Fenson et al. 1994). The test was given at home to mothers of the children from this study at 18 months of age and results have been reported (Hamadani et al. 2010). Not surprisingly, the correlations between the language scores and the Bayley MDI scores at 18 months were higher (expression r = 0.41 and comprehension r = 0.32, n = 2110) than any of the motor milestones. More importantly, the correlations with IQ at 64 months were also higher (expression r = 0.37, comprehension r = 0.41, n = 2114) and as good as those of the Bayley MDI with IQ (r = 0.36). Furthermore, giving the language inventory took no more than 15 min and was administered by relatively low‐skilled FWs with a very short training. These results suggest that if the interest is in predicting mental development from an easy‐to‐give, low‐cost measure, the use of a language inventory based on Mothers' report justifies further research. The language inventory given at 18 months of age was more predictive of future IQ than motor milestones assessed up to 15 months. However, it is possible that age is critical when assessing the number of words spoken by Mothers' report and 18 months may be an optimal age. A vocabulary test would not be appropriate for children younger than 12 months. In a large Finnish study (Taanila et al. 2005), the number of words spoken at 12 months was associated with later educational attainment but the association was not stronger than with age of walking or standing.
Nutritional status
These children had very poor nutritional status and HAZ at 15 months was significantly associated with both age of attainment of milestones and later IQ and motor development at 64 months. HAZ had an independent effect on IQ and MABC when age of attaining a milestone was in the regression. Furthermore, HAZ accounted for approximately a third of the effect of age of walking or standing on later IQ. Therefore, when HAZ is not controlled, the effect of age of attainment of milestones on later development is increased. It is likely that in populations where stunting is not as prevalent, the unadjusted associations between motor milestones and later IQ would be smaller. However, a combination of both HAZ and age of attainment of milestones increased the strength of association with both later IQ and motor development and may be useful in some studies as an indicator of later development.
Some social background variables, especially maternal education and HOME, were also associated with both milestones and IQ and would have explained some of the effects of age of attaining milestones on IQ. However, they tended not to be associated with later motor development.
Limitations of study
The main limitation of the study is that due to resource constraints, after 12 months, we only collected data again at 15 months, when some children had not attained the last two milestones. However, we ran the regressions of these two milestones using quartiles for age of attainment and the effect sizes were compatible with those in analyses using the continuous measure for age of attainment.
A further limitation is that the Bayley scales, the WPPSI and MABC tests were all standardised in other countries and like most developing countries, there are no well standardised tests available. However, even in developed countries, where the tests are standardised, the predictive validity of infant tests in the first 18 months is limited. A review of studies found a median correlation of r = 0.34 for the association between infant tests between 13 to 18 months of age and child IQ between 5 and 7 years (Colombo 1993). Furthermore, a recent much smaller longitudinal study in the Netherlands found a correlation of r = 0.19 between the MDI at 18 months and IQ at 73 months (Roze et al. 2010). The size of the correlations between WPPSI IQ and the MDI (r = 0.37) in this study compares well with those reported globally and provides confidence in the validity of the tests. Also, both the MDI and WPPSI correlated with many socio‐economic and nutritional variables in a theoretically expected way, providing more evidence that the tests are valid for this population. Although the tests can accurately identify variability in development within the population, the mean MDI, PDI and IQ scores were generally low compared with the standardised population in the United States. The lack of standardisation for the country and our modifications suggests that it is better to use internally standardised scores.
Conclusions
In conclusion, the ability to predict later IQ or motor development from early development is limited even with full developmental assessments. Mothers' reports of age of attainment of motor milestones were reasonably accurate. Ages of attaining motor milestones were a moderately good indicator of a full assessment of motor development at 18 months and had low but significant associations with longer term motor development. However, milestone age of attainment was as good as the Bayley PDI in predicting long‐term motor development and could be used to predict motor development in large field studies.
The association of age of attainment of milestones with MDI and later IQ was significant but weak, and may be too weak to be useful. The association of MDI with IQ was only moderate but MDI was a better predictor of IQ than age of attainment of milestones.
Height for age was related to milestone age of attainment and later IQ and in this undernourished population accounted for some of the association of age of attainment of milestones with later IQ. A combination of HAZ and age of attainment of milestones improved the prediction of future IQ and motor development. It is probable that in the second year, Mothers' report of words used combined with HAZ and motor milestones would be more useful than age of attainment of motor milestones alone. These approaches justify further research.
Source of funding
The MINIMat study was funded by United Nations Children's Fund (UNICEF), UK Medical Research Council, Swedish Research Council, Department for International Development (DFID), International Center for Diarrhoeal Disease Research, Bangladesh (ICDDR,B), Global Health Research Fund‐Japan, Child Health and Nutrition Research Initiative (CHNRI), Uppsala University, and United States Agency for International Development (USAID). This work was also supported by the EU project Public Health Impact of long‐term, low‐level Mixed Element Exposure in susceptible population strata (PHIME), sponsored by the EC (FOOD‐CT‐2006‐016253; the European Community is not liable for any use that may be made of the information contained therein), the Swedish Research Council, the Swedish International Development Cooperation Agency (SIDA), SIDA's Department for Research Cooperation (SAREC) and Karolinska Institutet. The analysis of the data was supported by Global Alliance for Improved Nutrition.
Conflicts of interest
The author does not have any conflict of interest.
Acknowledgements
We thank all the mothers and children who participated in the study. We are also grateful to all the field workers, interviewers, testers and their supervisors.
References
- Adu‐Afarwuah S., Lartey A., Brown K.H., Zlotkin S., Briend A. & Dewey K.G. (2007) Randomized comparison of 3 types of micronutrient supplements for home fortification of complementary foods in Ghana: effects on growth and motor development. American Journal of Clinical Nutrition 86, 412–420. [DOI] [PubMed] [Google Scholar]
- Bayley N. (1993) Bayley Scales of Infant Development , 2nd edn The Psychological Corporation: San Antonio, TX. [Google Scholar]
- Caldwell B.M. (1967) Descriptive evaluations of child development and of developmental settings. Pediatrics 40, 46–54. [PubMed] [Google Scholar]
- Colombo J. (1993) Infant Cognition: Predicting Later Intellectual Functioning. Individual Differences and Development Series, Vol. 5 Sage Publications, Inc: Newbury Park, CA. [Google Scholar]
- Faber M., Kvalsvig J.D., Lombard C.J. & Benadé A.J.S. (2005) Effect of a fortified maize‐meal porridge on anemia, micronutrient status, and motor development of infants. American Journal of Clinical Nutrition 82, 1032–1039. [DOI] [PubMed] [Google Scholar]
- Fenson L., Dale P.S., Reznick J.S., Bates E., Thal D.J. & Pethick S.J. (1994) Variability in early communicative development. Monographs of the Society for Research in Child Development 59, 1–173. [PubMed] [Google Scholar]
- Frisancho A.R. (1990) Anthropometric Standards for the Assessment of Growth and Nutritional Status. University of Michigan Press: Ann Arbor, MI. [Google Scholar]
- Grantham‐McGregor S., Cheung Y.B., Cueto S., Glewwe P., Richter L., Strupp B. et al (2007) Developmental potential in the first 5 years for children in developing countries. Lancet 369, 60–70. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Hamadani J.D., Baker‐Henningham H., Tofail F., Mehrin F., Huda S.N. & Grantham‐McGregor S.M. (2010) The validity and reliability of Mothers' report of language development in 1‐year‐old children in a large‐scale survey in Bangladesh. Food & Nutrition Bulletin 31 (2 Suppl.), S198–S206. [DOI] [PubMed] [Google Scholar]
- Hamadani J.D., Tofail F., Nermell B., Gardner R., Shiraji S., Bottai M. et al (2011) Critical windows of exposure for arsenic‐associated impairment of cognitive function in pre‐school girls and boys: a population‐based cohort study. International Journal of Epidemiology 40, 1593–1604. [DOI] [PubMed] [Google Scholar]
- Henderson S.E. & Sugden D.A. (1992) Movement Assessment Battery for Children. The Psychological Corporation: London. [Google Scholar]
- Olney D.K., Pollitt E., Kariger P.K., Khalfan S.S., Ali N.S., Tielsch J.M. et al (2006) Combined iron and folic acid supplementation with or without zinc reduces time to walking unassisted among Zanzibari infants 5‐ to 11‐mo old. Journal of Nutrition 136, 2427–2434. [DOI] [PubMed] [Google Scholar]
- Roze E., Meijer L., Van Braeckel K.N., Ruiter S.A., Bruggink J.L. & Bos A.F. (2010) Developmental trajectories from birth to school age in healthy term‐born children. Pediatrics 126, e1134–e1142. [Epub 2010 Oct 4]. [DOI] [PubMed] [Google Scholar]
- Saha K.K., Frongillo E.A., Alam D.S., Arifeen S.E., Persson L. & Rasmussen K.M. (2008) Appropriate infant feeding practices result in better growth of infants and young children in rural Bangladesh. American Journal of Clinical Nutrition 87, 1852–1859. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Taanila A., Murray G.K., Jokelainen J., Isohanni M. & Rantakallio P. (2005) Infant developmental milestones: a 31‐year follow‐up. Developmental Medicine and Child Neurology 47, 581–586. [PubMed] [Google Scholar]
- Tofail F., Persson L., Arifeen S.E., Hamadani J.D., Mehrin F., Ridout D. et al (2008) Effects of prenatal food and micronutrient supplementation on infant development: a randomized trial from Maternal and Infant Nutrition Interventions, Matlab (MINIMat) study. American Journal of Clinical Nutrition 87, 704–711. [DOI] [PubMed] [Google Scholar]
- UNICEF/UNU/WHO (1999) Composition of Multiple Micronutrient Supplement to be Used in Pilot Programmes among Pregnant Women in Developing Countries. UNICEF: New York, NY. [Google Scholar]
- Walker S.P., Wachs T.D., Grantham‐McGregor S., Black M.M., Nelson C.A., Huffman S.L. et al (2011) Inequality in early childhood: risk and protective factors for early child development. Lancet 378, 1325–1338. [DOI] [PubMed] [Google Scholar]
- Wechsler D. (2002) Wechsler Preschool and Primary Scale of Intelligence, 3rd edn Harcourt Assessment, Inc., San Antonio, TX. [Google Scholar]
- World Health Organization (WHO) (1983) Measuring Change in Nutritional Status. Guidelines for Assessing the Nutritional Impact of Supplementary Feeding Programmes for Vulnerable Groups. World Health Organization: Geneva. [Google Scholar]
- WHO (2006a) Global Database on Child Growth and Malnutrition. World Health Organization: Geneva. [Google Scholar]
- WHO Multicentre Growth Reference Study Group (2006b) WHO Motor Development Study: windows of achievement for six gross motor development milestones. Acta Paediatrica Supplement 450, 86–95. [DOI] [PubMed] [Google Scholar]