Psychomotor development in Argentinean children aged 0 5 years

Transcription

1 Psychomotor development in Argentinean children aged 0 5 years 47 Horacio Lejarraga a, María Cecilia Pascucci b, Sara Krupitzky a, Diana Kelmansky c, Ana Bianco c, Elena Martínez c, Fabián Tibaldi c and Noel Cameron d a Servicio de Crecimiento y Desarrollo, Hospital Garrahan, Buenos Aires, Argentina, b Servicio Nacional de Rehabilitación y Promoción de la Persona con Discapacidad, MSP, c Instituto de Cálculo, Facultad de Ciencias Exactas y Naturales, Universidad de Buenos Aires, Buenos Aires, Argentina, d Department of Human Sciences, University of Loughborough, Loughborough, UK, and Department of Anatomical Sciences, University of the Witswatersrand Medical School, Johannesburg, South Africa Summary Correspondence: Dr Horacio Lejarraga, Service of Growth and Development, Hospital Garrahan, Combate de los Pozos 1881, 1245 Buenos Aires, Argentina. hlejarraga@intramed.net.ar In Argentina, there is no information on ages of attainment of developmental milestones and very few data about environmental factors that influence them. A national survey on the psychomotor development of children under 6 years of age was carried out with the help of 129 paediatricians. Logistic regression was applied to a final sample of 3573 healthy, normal children in order to estimate selected centiles (25th, 50th, 75th and 90th), together with their respective confidence intervals, of the ages of attainment of 78 developmental items belonging to the following areas: personal social (18 items), fine motor (19), language (18) and gross motor (23). The 50th centile obtained for each of the 43 comparable items was compared with those obtained in previously standardised tests: DDST, Denver II, Bayley and Chilean scales. Neither significant nor systematic differences were found between our results and those described in the tests used for comparison. Multiple logistic regressions showed that social class, maternal education and sex (female) were associated with earlier attainment of some selected developmental items, achieved at ages later than 1 year. Selected items achieved before the first year of life were not affected by any of the independent environmental variables studied. The information is useful in helping paediatricians in their daily practice for surveillance of development, as baseline information for epidemiological studies on development in our country and for crosscultural analysis. Introduction The evaluation of psychomotor development in children during the first years of life is an essential part of paediatric care. 1 This evaluation is carried out either at the office, during a general paediatric examination, or during a formal screening procedure, testing the ability of the child to perform a series of developmental milestones. In both cases, the way of drawing significant conclusions from these evaluations is derived from comparison of the age of the child and the number of developmental milestones (usually called items) he/she can perform at a given age. In order to evaluate this, it is necessary to know the individual variation in the age of attaining each developmental item in the normal population. In Argentina, there is no information on the ages of attaining developmental milestones in local children, and there are very few data on the influence of environmental factors on those ages. As there may be population differences in age of attainment of milestones, owing to a wide variety of factors, 2 it is necessary to have that information for local children. Between 1988 and 1995, a nation-wide cross-sectional survey on psychomotor development on 3573 healthy children aged 0 5 years was carried out, with the help of 129 trained paediatricians, who evaluated 229 items related to child development and family health, including the performance of developmental milestones. 3 The sample derived from this survey was considered to be an adequate source of information for studying

2 48 H. Lejarraga et al. individual variation in the age of attainment milestones in a healthy, normal population. From this sample, we estimated centiles of age of attainment of developmental milestones for 78 items, which can be used for clinical purposes. Part of this information has already been published in Spanish in a local journal. 4 Here, we reproduce these data, compare them with information from other communities and study the influence of some environmental variables on selected items. Methods Sampling procedures and main features of the sample The sample on which the centiles were estimated was composed of 3573 boys and girls, aged years. This national sample size was estimated on the criteria of: (1) a minimum desired size age distribution comprising about 400 children of the same sex at each year of age (total of 4800 children, 2400 boys, 2400 girls); (2) the number of paediatricians we decided to invite on the grounds of logistic resources (n = 211) and their time limitations; and (3) an expected failure rate arbitrarily set at 45%. This yielded a raw sample size of 8651 participants. The geographical distribution of children within the sample was calculated from the percentage of children aged <6 years within each of the 24 administrative areas (provinces and federal district) of Argentina according to the National Census of This percentage was applied to the total theoretical figure of children in order to determine the total number of children to be evaluated and the total number of paediatricians (at 52 subjects per paediatrician) undertaking the evaluation in each province. The selection of paediatricians was made from the membership list of the Argentine Society of Paediatrics, on the basis of a complete training period in paediatrics, working in public or private outpatients and a previously demonstrated interest in child growth and development. A total of 121 paediatricians were invited, 181 of them carried out a pilot study, 142 attended a 10-h training workshop (on the administration of the developmental items and the use of an instructive manual) and 129 (61.1% of the initial group) completed the survey. The selection criteria for including children were set before the survey: healthy children (with no chronic diseases or clinical or anthropometric signs of malnutrition, absence of acute disease at the time of evaluation and full-term pregnancy). Children to be examined by the paediatricians were selected at random every day before starting the clinic, by setting the first two (or the number of children that the paediatrician wanted to examine that day) to come to the clinic fulfilling the selection criteria. Paediatricians would each evaluate 20 girls and 21 boys, whose ages were dictated by the number of children required from each of the age subgroups. The final sample of 3573 subjects was 25.6% less than the target sample of 4800, but losses appeared to be evenly distributed between the sexes (25.1% boys and 26.0% girls). There was, however, a clear trend to lose more children in the older age groups, with minimal losses occurring in the first youngest four groups, (<3 months, 3 <6 months, 6 <9 months, 9 <12 months). Five criteria for assessing the external validity of the sample were set: (1) sample size, which represents 0.11% of the total national population of the same age and sex; 5 (2) sex ratio (1.01), compared with 1.02, from the National Census, 5 (3) geographical distribution, which, compared with that of the total national population under 6 years of age, did not differ significantly; (4) social composition of the sample, assessed by the level of maternal education, which showed a marked bias of the sample towards a higher maternal education level, compared with national figures taken on women of the same age range as that of our sample: years. 5 With regard to the source of payment of the appointment, 34.48%, 38.4% and 26.7% of the children studied were from the public, social security and private sectors respectively; (5) the mean (and standard errors) of height and weight of each age group calculated compared with national growth standards for each age were: 6,7 weight of boys 0.06 (0.03), weight of girls 0.05 (0.03), height of boys 0.03 (0.06) and height of girls 0.08 (0.04), none of which indicated a biologically significant difference from the 50th centile of national standards. 6,7 A more detailed description of the sample can be found in a previous paper. 3 Performance of the tests The evaluation of the performance of the different developmental items was carried out in a standardised way by paediatricians in their outpatient clinics. In order to achieve these standards, three instruments were used: (1) selection of paediatricians, on the basis

3 Psychomotor development in Argentinean children 49 of their complete residency in paediatrics and interest in the subject; (2) a cascade-designed training workshop of 10-h duration; (3) an instruction manual with a description of the instruments and techniques with which each developmental item should be tested. Each paediatrician was supplied with a kit with the necessary tools for performing the tests (cubes, tennis ball, figures, etc.) and an instructive manual prepared by us, with the instructions for performing the tests, stating the precise items to be tested at each age range, allowing for fast and slow maturers. The items were selected by a multidisciplinary expert committee from the most widely used items described in the literature at that time The mean number of items to be tested in a single child was 25, with a range of 16 40, depending on his/her age. The age at which the items were performed was estimated only in the items considered developmental ; by this term (referred to an item), we mean that the percentage of children passing the item test increases as the age of the child increases. For instance, if 35% of the children in the age subgroup years pass a given item, such as walk well and if the item was developmental, then a greater percentage of children would pass the item in the age interval years. An even greater percentage would pass the item in the interval years. There is always an age interval at which 100% of the children pass the given item. The item was also considered developmental when the percentage of children passing the item was progressively less as the age increased, such as happens with the Moro reflex. The objective of the task with a cross-sectional sample is to study the relationship between age and a qualitative variable (passing or not passing a given test or item). The estimate of centiles of age of attainment of each item was therefore obtained using a logistic regression model, 19 as has been used by other authors. 8,20 The model was adjusted using only age as the explanatory variable. The coefficients were estimated with the maximum likelihood method. 21 To improve the fit, and hence decrease the errors of estimation for some items, some form of transformation with a monotonic function had to be used, such as, for example, log age. The model estimated for one item might not be valid beyond the observed age range. For instance, if the greater percentage of children passing an item in a given age subgroup is 80%, then it is not possible to estimate the 90th centile without extrapolating the results. As we decided not to carry out any extrapolation, it was not possible for some items to estimate the more important centiles for clinical purposes (25th, 50th, 75th or 90th centile). In these cases, the nearest centile to the selected one estimated without any interpolation was included in the study (e.g. 35th instead of 25th, 85th instead of 90th, etc.). Chronological age was calculated by dividing the number of days elapsed from the birthday up to the date of the survey and dividing by , thus obtaining the exact decimal age. Each estimated curve was subject to goodness-of-fit tests in order to evaluate the adjustment of the model to the data. To this end and in the majority of the cases, the Hosmer Lemeshow test was used. 19 When this test could not be performed, the deviance test was used. These tests measure the discrepancy between the observed and predicted values. When the discrepancy was small, then a conditional test was carried out in order to evaluate whether the age explained the changes in the probability of passing an item. Taking into account the large sample size, the asymptotic significance was estimated from the Chi-square distribution. In the case of goodness-of-fit tests, a large P-value indicates that the estimated curve provides a sufficiently good fit. In this paper, P-values >0.10 were considered acceptable. In the case of the conditional test, a small P-value (P < 0.05) indicates that the age explains the changes in the probability of passing an item (P). If it was not possible to obtain a good fit, or the age did not explain the changes in P, the item was excluded from the study. Centile comparisons The centiles and their corresponding 95% confidence intervals [95% CI] were also calculated. These intervals allowed us to compare our results with those from other studies. In order to do this, the 50th centile of each item from the present study was compared with the same centile of the same item from other studies, in which the items had been performed in exactly the same way and with the same criteria for approval or failure. For instance, the criterion for passing the item jumps on one foot in our survey was the child should jump three times, whereas for the Denver II test, the passing criterion was the child should jump twice. 22 In this case, the item was not subject to comparison. When the 50th centile of a given item was beyond our local 95% CI for the 50th centile, it was classified as advanced or retarded with regard to the study used for comparison.

4 50 H. Lejarraga et al. Environmental variables Together with the tests, in the interviews, the paediatricians collected information about date of birth, gender and environmental variables. Multiple logistic regression was applied in order to evaluate the association between the environmental variables (used as independent variables) and the attainment of a given developmental milestone (used as the dependent variable). To this end, 12 milestones were selected, so that the four main areas of development were represented and a broad age range was covered. In Table 1, we show the categories considered for each environmental variable. A stepwise selection procedure based on the Akaike criterion was used in order to choose the final adjusted model for each item. Chronological age was forced in all equations as an independent variable, and nonsignificant variables were excluded from the final model. The stepwise procedure excludes or includes a given categorical variable with all its categories. Results Out of a total number of 229 items, selected centiles (25th, 50th, 75th and 90th) were estimated on 78 developmental items. Centiles of the remainder of the items were not estimated for any of the following reasons: (1) the age range was insufficient for estimating the selected centiles; (2) there was not a good fit according to the Hosmer Lemeshow test; and (3) there was no significant regression with age. The Appendix table shows the selected centiles of the age of attainment of the 78 items, grouped as usual according to four developmental areas, together with the 95% CI. 4 An important variation in the size of the CI of the different items should be noted; for example, the item response to observer s call (item 11), whose 90th centile is 2.35 years, has a CI between 2.12 and 2.76 years; on the other hand, the item stoop and recover (item 69) whose 90th centile is 1.30 years, has a CI markedly smaller, between 1.23 and 1.41 years. The comparison of the ages of attainment with those from other studies should be carried out only with those items performed in the different studies in exactly the same way. Looking carefully at this condition, we were able to compare the 50th centile of some of the results of our items with those from four other studies: the DDST, 8 the Denver II test 20,21 (a revised version of the former with new data, both from the USA), the Bayley test 9 (the original version from the USA) and the Escala de Evaluación del Desarrollo Psicomotor (EEDP), prepared in Chile. 11 The concept of advanced or retarded refers to the centile (and CIs) of the present paper with regard to the other tests used for comparison. The total number of items subject to comparison was 43. In the Appendix table, a four-sign word marks those items that were advanced, retarded or did not significantly differ from the selected comparison tests. Some items related to the personal social area such as smile responsively (item 2) were advanced with regard to the Bayley 9 and Chilean 11 data, but they were achieved at the same age as the DDST8 and Denver II tests. 20 Within the fine motor area, the delayed items were usually those attained during the first months of life [ regard own hand (item 21), follow to midline (item 19), hands together (item 22)] when compared with DDST 8 and Denver II, 20 and the advanced ones Table 1. The environmental variables used Category Environmental variable Type of appointment Public hospital Social security Private office Maternal age (years) Birth order First Second or more Birthweight (g) <3000 (3000, 4000) 4000 Family size One Two Three 4 Father s occupation Labourer Employee Professional, company owner Father living at home No Yes Child attending day care centre No Yes Sex Male Female Mother s education Primary Secondary Tertiary

5 Psychomotor development in Argentinean children 51 were those attained at later ages, such as pick up raisin (item 24), scribbles (item 28), tower of four cubes (item 29), draw a person with three or six parts (items 33 and 36). With regard to the items in language areas, those attained before the first year of life, such as combine syllables (items 41), dada/mama, nonspecific (item 42), were performed at the same ages as in the Denver II, but they were advanced with regard to the Chilean 11 scale and Bayley s 9 respectively. The items attained at ages later than the first year did not follow a defined pattern. Some of them were attained earlier than the comparison tests [ opposites (item 52), understand prepositions (item 50)], and others did so at later ages [ combine words (item 47)], recognises three colours (item 54)]. The gross motor area contains the greatest number of comparable items (17 out of 23). Some items were delayed [ head up 45 (item 57)], seated, with no support (item 64), pull to stand (item 65), stoop and recover (item 69)]. Others were advanced [ sit, head steady (item 56), seated, helped with own hands (item 61), kick ball forward (item 70), throw ball overhand (item 71), balance each foot, 5 s (item 74), broad jump (item 75), heel-to-toe walk (item 77)], and others were attained at the same age [ walk holding on to furniture (item 66), walk by the hand (item 67), walk well (item 68)]. No definite overall trend could be identified. With regard to DDST, 30 items were compared, of which 16 were advanced and seven retarded. With regard to Denver II, 28 were compared, and 12 and nine were advanced and retarded respectively; only 11 could be compared with BSID, and seven items were advanced and two retarded. Finally, out of 14 items compared, eight were advanced with regard to the Chilean scale (EEDP) and two retarded. The results of the multiple logistic regression of independent environmental variables on the 12 dependent milestone variables are shown in Table 2. The items are ordered in terms of expected median age of attainment. Two variables were not associated with any item: father at home and birthweight, and hence they were not included in the columns of Table 2. In the first column, we report the name of the item preceded by its number and followed by the sample size. These regressions are based only on significant variables. Fractions after some odds ratios represent the categories in which significant differences were found. For example, in the item walk well, the odds ratios (3.31, 5.67) and the fractions (2/1, 3/1) mean that the odds of attaining the item are 3.31 times in favour of category 2 (employee) in relation to category 1 (hand labour) and 5.67 times in favour of category 3 (professional, company owner) in relation to category 1. Fractions in maternal education refer to three categories: (1) primary school; (2) secondary school; and (3) tertiary education. Two items, smile responsively and sit head steady, attained at early ages did not show any association with environmental variables. In general, for items attained at later ages, more variables became significantly related. This result should be treated with caution as the sample size also increases in the same direction. No specific trend of items belonging to a given developmental area (fine motor, etc.) was found to be more linked to significant variables. The variables more often significantly related to items were those associated with social class (type of consultation, father s occupation), maternal education and sex. Interestingly, whenever type of appointment (a variable related to the source of payment of the appointment) became significantly associated with an item, the variable father s occupation did not, and vice versa. Maternal education was the variable most frequently associated with the items. In some cases, it was the secondary educational level over primary level (as happened with the item searches for sound ). In others, it was the tertiary level being the favourable category (as in combine words ) and, in others, both categories, secondary and tertiary educational level, were related to earlier attainment of the item over the primary level, as happened in copy cross or draw a person. Female gender was consistently associated with earlier attainment of five items. Attendance at kindergarten was significantly related to some items in a positive sense ( combine words, draw a person ) and to some others in a negative sense ( recognises three colours ). Discussion The present figures for age of attainment of developmental milestones were derived from a national sample of healthy children attending paediatric practices in the public and private sectors in Argentina. Compared with the country s general population, the sample was biased towards higher maternal education levels and private paediatric offices, which made the sample adequate for the study of development of healthy children.

6 52 H. Lejarraga et al. Table 2. Significant odds ratios [95% confidence interval] from a multiple logistic regression of chronological age and other environmental variables, on the attainment of 12 developmental items Item Type of Father s Mother s Mother s Family Birth Attending (sample size) appointment occupation education age size order Sex day care 2. Smile response (171) NS NS NS NS NS NS NS NS 56. Sit head steady (368) NS NS NS NS NS NS NS NS 39. Searches for sound NS 0.16* 3/1 2.88* 2/1 NS NS NS NS NS (192) [0.03, 0.85] [1.21, 6.85] 25. Neat pincer grasp (189) NS NS NS NS 2.96* 2/1 NS NS NS [1.16, 7.36] 68. Walk well (448) NS 3.31* 2/1 NS NS NS NS NS NS [1.01, 10.82] 5.67** 3/1 [1.56, 20.53] 47. Combine words (987) NS NS 2.07** 3/1 NS NS NS 1.64* 3.61*** [1.24, 3.45] [1.11, 2.38] [1.87, 6.96] 15. Bladder control (1161) NS NS 0.54* 3/1 NS NS NS NS NS [0.33, 0.88] 49. Complete phrase (1704) 1.43* 2/1 NS 1.52* 2/1 NS NS NS 1.69*** NS [1.01, 2.03] [1.07, 2.15] [1.30, 2.22] 1.64* 3/1 2.07*** 3/1 [1.10, 2.44] [1.36, 3.14] 54. Recognises three 3.00**** 2/1 NS 1.92** 3/1 NS NS NS NS 0.43**** colours (1048) [2.07, 4.34] [1.22, 3.01] [0.31, 0.58] 3.96**** 3/1 [2.60, 6.06] 77. Heel-to-toe walk (1182) NS NS NS NS NS 0.68* 1/2 1.42* NS [0.50, 0.92] [1.05, 1.92] 34. Copy cross (996) NS NS 1.62** 2/1 NS NS 1.47* 1/2 1.67** 1.44* [1.11, 2.37] [1.08, 2.02] [1.22, 2.28] [1.04, 2.02] 2.71*** 3/1 [1.74, 4.25] 36. Draw a person NS 1.97* 3/1 1.76*** 2/1 0.57* NS NS 1.20**** 1.86**** in six parts (1455) [1.04, 3.72] [1.27, 2.44] 40/20, 40 [1.56, 2.70] [1.38, 2.51] 1.87** 3/1** [0.34, 0.95] [1.27, 2.77] *P < 0.05; **P < 0.01; ***P < ; ****P < Figures beside the item name are the item number, as in the Appendix table. There had been no previous data of this nature in Argentina. One study was carried out in 1988, based on children attending three hospitals in the Federal District and adjacent locations (greater Buenos Aires area), 23 reported the number of children passing or failing to pass the Chilean scale and DDST. The percentage found was similar to that expected, but no data on the age of attainment of each single milestone were published. The differences found in the 50th centile between our study and the others were not systematic (in the same direction). Some items subject to comparison were attained at earlier and others at later ages than those from other tests. We would have liked to compare our results with those from the Bayley II scales, but this second version does not include the 50th centiles of the items in the publication. 24 It is important to stress that, with these comparisons, we did not consider taking the behaviour of our series of items as a single test, but the individual values for each single item. We would have liked to have values for a greater number of items related to language, accord-

7 Psychomotor development in Argentinean children 53 ing to current trends 20, 25, 26 but, because of either an insufficient age range or lack of adjustment, centiles for those items could not be estimated. The grouping of the items into the four areas of development carried out here may differ to some degree from the grouping made by other authors. In fact, we have found some items included in one area in DDST 8 and in another area in Denver II 20 (made by the same authors) and, in turn, included in yet a further area in another study. 27 For example, the item regard own hand (item 21) may well be located in the personal social area, as in fact it has been in the Denver II test, 20 but it may also be placed in the fine motor area, as we have done in the present study. It is evident that, although the grouping of items seems convenient for operational reasons, authors do not always agree in which area a given item should be located. We think that the performance of a given item includes a very wide scope of complex neurological and psychosocial components, so that the location of an item in a single area is conventional and means that the item is only predominantly, but not exclusively, related to that area. In our study, we have followed these conventions to some extent, but making the changes that we considered suitable. Frankenburg and Dodds 8 and Stott and Ball 28 have stated that a certain group of items that did not seem to be related a priori may, in practice, prove to have much in common and vice versa. This fact suggests caution in the identification of problems in a specific area such as problems in the language area or in the personal social area, when finding failures in a group of items from that area. We strongly agree with this concept; we were dealing with a group of individual items prepared to be used for general screening purposes and not for detecting developmental problems in specific areas. It was not possible to estimate the selected centiles in all the items taken by the paediatricians. In some cases, we had to estimate the nearer centile, e.g. the 85th instead of the 90th. This has also been done in other tests. 8,20 Our centile estimations included the 95% CI. We consider this a very important inclusion because, in the case of some items, the 95% CI may be very large. We have failed to find these CI in the centiles estimated in other papers, 8 10,17 with the exception of a paper by Dick, 29 in which 99% CIs were calculated for only three developmental items selected from the DDST. In this case, the CI was very similar to ours in the three items studied. The centiles estimated here did not differ systematically (in a given direction) from those described in other communities, such as Cardiff, 30 Rennes, 31 Israel, 32 Tokyo 33 and China, 34 although >50% of our items were advanced with regard to the ages reported in those studies. This may be because of the social composition of our sample. The population differences found in the advancement or retardation of a given item may result from a diversity of factors related, for example, to the ways the tests were administered, the training of the observer, cultural differences in child care and rearing habits and values or language differences. In a Kenyan community, it was found that some items from the motor area attained early were those that were best regarded in that community as important and valuable. 35 There are important examples of these cultural influences and interactions in the literature An additional factor that may be playing a role in the differences found in our paper with regard to other studies is that related to the secular trend in development. Many of the items described in the DDST, standardised in 1966, were attained at much earlier ages in the Denver II test, standardised in There are other studies showing secular trends; in Sweden, those children subject to Griffiths tests in 1980 have been shown to be more advanced than those children subject to the same test in The information in our survey was collected in , and the Bayley scale we used for comparison was that published in and the Chilean scale published in The study of associations between environmental variables and attainment of items showed that variables were more likely to be associated with early attainment of milestones when those items were attained at later ages. This is consistent with the expectation that environmental variables become more important as the baby grows older and is more exposed to the external world. However, as we have already mentioned, this should be considered carefully, as it could be confounded by the fact that the sample sizes increase as children s median age of attainment increases. A positive association between upper social level and early attainment of items was found in five cases, with some related to one indicator of social class (type of appointment) and others to father s occupation. In no case were both indicators found to be significant with regard to a single item. This probably means that both indicators are the expression of the same variable.

8 54 H. Lejarraga et al. The association between higher socio-economic status and advancement of age of attainment of developmental milestones has been well proved in toddlers, 8,30,39,41 43 including a recent survey in a province of Argentina. 44 In schoolchildren from La Plata city, an administrative and university city in Argentina, children from lower social levels scored three points lower in IQ (Weschler test) than those from upper levels. 45 However, this association is not that clear in infants. There are some studies showing an advancement in development in infants of low socio-economic levels, 42,46 but those findings were not present in all studies. 8,39,41,42 In Argentina, a survey performed in Cordoba showed, in infants, a positive association between the upper socio-economic levels, maternal education and the age of attainment of the same items described in the present study. 47 We think that the publications in which low social class was associated with early development of infants were those in which this lower social class was in turn related to rearing practices with a strong and positive influence on development, such as breast feeding and close mother infant physical contact. It should also be stated that not all papers exploring relationships between social class and development have used the same indicator of social class. Some studies used father s occupation, 46 others used maternal education, 42 and others used a combination of several social indicators, such as housing, father s occupation and income. 8,39,42,48 In the present study, maternal education (ME) was found to have an important association with seven out of 12 items. A variety of associations with different items may be found in other studies Perhaps the strength of this association is related to the type of sample and to the general level of ME in the sample. In our study, as in many others, this variable was importantly related to the achievement of some items, especially at later ages. We do not think that ME should be used as an indicator of social class, at least in our country. In a survey carried out in La Plata city on 1200 children, it was shown that both ME and social class (as expressed by father s occupation) have quite independent influences on IQ of children aged 4 12 years. 50 Female gender was found to have a positive association with five items combine word, heel-to-toe walk, complete phrase, draw a person and copy cross, in agreement with the international literature. We did not find gender associations with bladder control, although Largo and Stutzle 52 in an important survey found that girls attained this item earlier than boys. In some standardisations of the Denver test 8,32,53 and with the Bayley test, 49 no significant general influence of sex was found on the attainment of items. In these studies, only small associations in favour of girls were found in relation to isolated items, such as heelto-toe walk and some speech items. In a study carried out in Israel, 32 boys attained some gross motor items earlier. In the rest of the revised studies, the significant influences were always in favour of girls in all areas, 45,46,53,54 in speech, 51 in personal social, speech and fine motor. 50 The advantage of girls over boys in psychomotor development has also been found in our region, 11 although not in all countries. In Argentina, there were no important gender differences in 1705 schoolchildren aged 4 12 years. 45 The fact that girls mature faster than boys is a well-known concept with regard to physical maturity. 55 The literature suggests that, at least in some psychomotor items, it is also true, but there are data showing that girls are also at less risk of developmental abnormalities. In England, Butler et al. 56 found a greater prevalence in boys than girls in a study on speech delay. In Argentina, a study on the prevalence of mental retardation carried out in the city of Buenos Aires showed these problems to be more frequent in boys than in girls. 57 Other variables, such as family size, mother s age and attendance at day care centre, showed associations with only one or two items. The influence of birth order on development has been studied by several authors. 30,46,48,53 Bryant et al., 54 in Cardiff, found that first-born children attained milestones earlier than the later born and, in England, Butler et al. 56 found the same association, especially in relation to speech. In our study, this influence was present in only one item. In one survey performed in Argentina, first-born children had some advantage at the age of 24 and 60 months on the Bayley scores over late-born children, 44 but this association was not studied in a multivariate fashion. The variable father living at home was not significant in any of the items studied here. The other variable not having any association with development was birthweight. All these associations should be considered with regard to the general features of our sample. It was composed of healthy, well-nourished children, with a social level biased towards the middle and upper social classes and higher maternal education level, in which children with birthweights <2500 g were not included. This might explain why many

9 Psychomotor development in Argentinean children 55 variables found to be significant in other studies were not so in this study. The results found here can be used as a reference for comparison with other communities, as a guide to be used for screening purposes (after fulfilling certain conditions), 58,59 for other epidemiological studies and for reinforcing the paediatricians instruments of growth and development surveillance. Acknowledgements We are grateful to Dr Marisa Jaitt and Paula Barredo for keeping the data in due order. This paper was partially financed by the Fundación Hospital de Pediatría, Pan-American Sanitary Office, Nestle Argentina S.A., the Embassy of the Republic of South Africa in Argentina and Elvetium Laboratories. References 1 Illingworth R. Development of the Infant and Young Child, Normal and Abnormal, 6th edn. Edinburgh: Livingstone, Lansdown RG, Goldstein H, Shah PM, Orley JH, Di G, Kaul KK, et al. Culturally appropriate measures for monitoring child development at family and community level: a WHO collaborative study. Bulletin of the World Health Organisation 1996; 74: Lejarraga H, Krupitzky S, Giménez E, Diament N, Kelmansky D, Tibaldi F, et al. The organisation of a national survey for evaluating child psychomotor development in Argentina. Paediatric and Perinatal Epidemiology 1997; 11: Lejarraga H, Krupitzky S, Kelmansky D, Martinez E, Bianco A, Pascucci MC, et al. Edad de cumplimiento de pautas de desarrollo en niños argentinos menores de seis años. Archivos Argentinos de Pediatría 1996; 94: INDEC National Census Buenos Aires: Instituto Nacional de Estadisticas y Censos, Publicaciones INDEC, Lejarraga H, Orfila G. Estándares de peso y estatura para niñas y niños argentinos desde el nacimiento hasta la madurez. (Standards for weight and height for boys and girls from birth to maturity). Archivos Argentinos de Pediatría 1987; 85: Lejarraga H, Anigstein C. Desviaciones estándard del peso para la edad de los estándares argentinos desde el nacimiento hasta la madurez. Archivos Argentinos de Pediatría 1992; 90: Frankenburg W, Dodds JB. The Denver Developmental Screening Test. Journal of Pediatrics 1967, 71: Bayley N. Manual for the Bayley Scales of Infant Development. New York: Psychological Corporation, Griffiths R. The Abilities of Babies: A Study on Mental Measurements. New York: McGraw-Hill, Rodríguez S, Arancibia V, Undurraga C. Escala de Evaluación del Desarrollo Psicomotor de 0 a 24 Meses. Santiago de Chile: Galdoc, Martell M, Martínez G, Díaz Rosello JL. Evaluación Promariqa del Crecimiento y Desarrollo Psicomotriz. Publicación Científica Montevideo, Uruguay: CLAP (PAHO/WHO), Brazelton TB. Neonatal Behavioral Assessment Scale. London: Spastic International Medical Publications, Child Development Unit, Child Study Centre. Test for Yale Developmental Examination. New Haven, CT: Yale University, Terman LM, Merrill MA. Stanford Binet Intelligence Scale. Manual for the third revision form L-M. Boston: Houghton Mifflin Co., Illingworth RS. The Development of Infants and Young Children, 2nd edn. Ministry of Health Reports of Public Health and Medical Subjects, Number 102. London: HMSO, Gessel A, Amartruda CS. The evaluation and management of normal and abnormal neuropsychologic development in infancy and early childhood. In: Developmental diagnosis. 3rd edn. Editors: Knoblock H, Pasamanick B. New York: Hoeber, Castro Paez MGC, Molina Brenes MA, Rangel Alfaro A. Escala de evaluación del desarrollo integral dek niño menor de 1 año. Tesis de Licenciatura en Psicología. Ministerio de Salud y Universidad de Costa Rica, Costa Rica, Hosmer DW, Lemeshow S. Applied Logistic Regression. New York: John Wiley, Frankenburg WK, Dodds J, Archer P, Shapiro M, Bresnick B. The Denver II: a major revision and restandaridization of the Denver Developmental Screening Test. Paediatrics 1992; 89: Cox C. Feller s theorem, the likelihood and the delta method. Biometrics 1990; 46: Frankenburg WK, Dodds J, Archer P, Bresnick B, Maschka P, Edelman N, et al. Denver II: Training Manual, 2nd edn. Denver: Denver Developmental Materials, Roy E, Cortiggiani MR, Acosta L, Shapira L. Evaluación del desarrollo psicomotor de 319 niños de 1 a 24 meses de la consulta ambulatoria de los hospitales Ramon Sardá, Pedro de Elizalde, y Luisa de Gandulfo. Archivos Argentinos de Pediatría 1988; 86: Bayley N. Bayley Scales of Infant Development Manual, 2nd edn. New York: The Psychological Corporation, Glascoe FP, Byrne KE, Ashford LG, Johnson KL, Changg B, Strickland B. Accuracy of the Denver II in developmental screening. Paediatrics 1992; 89: Borowitz KC, Glascoe FP. Sensitivity of the Denver Developmental Screening Test in speech and language screening. Pediatrics 1986; 78: República de Colombia Ministerio de Salud. Escala Abreviada de Desarrollo. Manual de Instrucciones. Bogotá: Editora Guadalupe, Stott LH, Ball RS. Infant and Pre-school Mental Test: Review and Evaluation. Monograph of the Society for Research in Child Development. Chicago: University of Chicago Press, 1965; 30:3.

10 56 H. Lejarraga et al. 29 Dick NP. Denver Developmental Screening Test. Developmental Medicine and Child Neurology 1973; 15: Bryant GM, Davies KJ, Newcombe RG. Standardisation of the Denver Developmental Screening test for Cardiff children. Developmental Medicine and Child Neurology 1979; 21: Sénecal J, Bouchard C, Roussey M, Defawe G. Une méthode simple de surveillance du développment moteur et psychologique de l enfant de trois à six ans: le DDST. Annales de Pédiatrie 1982; 29: Shapira Y, Harel S. Standardisation of the Denver Developmental Screening Test for Israeli children. Israel Journal of Medical Sciences 1983; 19: Reiko U. Standardisation of the Denver Developmental Screening Test on Tokyo children. Developmental Medicine and Child Neurology 1978; 20: The Collaborative Study Group of Child Development. Restandardization of DDST from six cities in north China. Chinese Medical Journal 1986; 99: Super CM. Environmental effects on motor development: the case of African infant precocity. Developmental Medicine and Child Neurology 1976; 18: Miller V, Onotera R DDST: cultural variations in Southeast Asia children. Journal of Pediatrics 1984; 104: Fung KP, Lau SP. Denver Developmental Screening Test, cultural variables. Journal of Pediatrics 1985; 106: Lim HC, Chan T, Yoong T. Standardisation and adaptation of the DDST and Denver II for use in Singapore children. Singapore Medical Journal 1994; 35: Caldwell BM. Descriptive evaluations of child development and of developmental settings. Pediatrics 1967; 40: Nordberg L, Rydelius PA, Zetterstrom R. Psychomotor and mental development from birth to age of four years, sex differences and their relation to home environment. Acta Paediatrica Scandinavica Supplement 1991; 378: Bayley N. Comparison of mental and motor test scores for ages 1 15 months by sex, birth order, race, geographical location and education of parents. Child Development 1965; 36: Frankenburg WK, Nathan P, Dick MS, Carland J. Development of pre-school aged children of different social and ethnic groups: implications for developmental screening. Journal of Pediatrics 1975; 87: Frankenburg WK, Dodds J, Archer P, Bresnick B, Maschka P, Ederman N, et al. Denver II: Technical Manual. Denver: Denver Developmental Materials, Carmuega E, O Donell A. Proyecto Tierra del Fuego. Encuesta sobre desarollo infantil. Fundación Jorge Macri, Publicación Cesni, Cusminsky M, Lozano GA, Castro EP, Lejarraga H, Spotti M, Porfiri N, et al. Investigación del desarrollo del niño normal de 4 a 12 años. Estudio Transversal. Comisión de Investigaciones Científicas, Actas del XIV Congreso Internacional de Pediatría, Buenos Aires, Argentina, 1974; 5: Neligan G, Prudham D. Norms for four standard developmental milestones by sex, social class and place in the family. Developmental Medicine and Child Neurology 1969; 11: Sabulsky J, Batrouni L, Agrelo F, Sesa S, Quiroga D, Reyna S, et al. Perfiles Epidemiológicos de Alimentación y Crecimiento y Desarrollo en los Dos Primeros Años de Vida. Publicación no. 2. Córdoba, Argentina: Estudios CLACYD, Bryant GM, Davies KJ. The effect of sex, social class and parity on achievement of DDST item in the first year of life. Developmental Medicine and Child Neurology 1974; 16: Bayley N. Comparison of mental and motor tests scores for ages 1 15 months by sex, birth order, race, geographical location, and education of parents. Child Development 1965; 36: Porfiri H, Spotti M, Petriz G, Lejarraga H, Medina N, Cusminsky M. Effect of age, socio-economic level, maternal education and paternal occupation on intellectual quotient of a representative sample of 900 children aged 4 12 years. Journal of Pediatrics 1976; 89: Yagoob M, Ferngren H, Jalil F, Nazir R, Karlberg J. Early child health in Lahore, Pakistan: XII milestones. Acta Paediatrica Scandinavica Supplement 1993; 390: Largo RH, Stutzle W. Longitudinal study of bowel and bladder. Control by day and at night in the first six years of life. I. Epidemiology and interrelations between bowel and bladder control. Developmental Medicine and Child Neurology 1977; 19: Hindley CB. Growing up in five countries: comparison of data on weaning, elimination training, age of walking and IQ in relation to social class from European longitudinal studies. Developmental Medicine and Child Neurology 1968; 16: Bryant GM, Davies KJ, Newcombe RG. The Denver Developmental Screening Test. Achievement of test items in the first year of life by Denver and Cardiff infants. Developmental Medicine and Child Neurology 1974; 16: Tanner JM. Growth at Adolescence, 2nd edn. London: Blackwell, Butler NR, Peckham C, Sheridan M. Speech defects in children aged 7 years: a national study. British Medical Journal 1973; 1: Fejerman N. Estudio de prevalencia de retardo mental en la población infantil de Buenos Aires. Tesis de doctorado. Facultad de Medicina, Universidad de Buenos Aires, Buenos Aires, Cochrane A, Holland W. Validation of screening procedures. British Medical Bulletin 1969; 27: Rose G. Epidemiology for the uninitiated: screening. British Medical Journal 1978; 2:

11 Psychomotor development in Argentinean children 57 Appendix Selected centiles (in years) [95%confidence intervals] of the age of attainment of developmental milestones, in each area, and results of the comparison with other tests (icons) Centiles Item (sample size) 25th [CI] 50th [CI] 75th [CI] 90th [CI] Personal social 1. Communication with the observer (378) [0.03, 0.05] [0.05, 0.08] [0.10, 0. 15] [0.21, 0.40] 2. Smile responsively (176) [0.05, 0.07] [0.07, 0.10] [0.10, 0.14] [0.14, 0.21] 3. Attitude in the mirror (204) [0.15, 0.30] [0.19, 0.31] [0.36, 0.44] [0.44, 0.65] 4. Resists toy pull (395) /// [0.22, 0.31] [0.35, 0.42] [0.51, 0.62] [0.61, 0.81] 5. Plays peek-a-boo (393) [0.24, 0.33] [0.39, 0.45] [0.52, 0.59] [0.63, 0.75] 6. Search object (385) // [0.43, 0.56] [0.57, 0.66] [0.72, 0.81] [0.85, 1.00] 7. Give an object (466) [0.59, 0.86] [1.00, 1.17] [1.33, 1.74] 8. Symbolic play (1037) [0.98, 1.05] [1.10, 1.21] [1.41, 1.66] 9. Eat alone (272) [0.57, 1.14] [1.23, 1.68] [1.31, 1.99] 10. Help in house (531) [0.94, 1.08] [0.98, 1.11] [1.20, 1.30] [1.42, 1.59] 11. Response to observer s call (743) [0.92, 1.16] [1.48, 1.69] [2.12, 2.76] 12. Imitate activities (536) // [0.94, 1.06] [1.01, 1.12] [1.24, 1.35] [1.51, 1.76] 13. Take off clothes or shoes (701) [1.40, 1.75] [1.56, 1.86] [2.31, 2.57] [2.65, 3.06] 14. Put clothes or shoes on (172) // [1.43, 1.74] [1.87, 2.07] [2.51, 2.79] [2.84, 3.26] 15. Bladder control (day) (1226) [1.79, 1.91] [2.05, 2.15] [2.33, 2.45] [2.62, 2.82] 16. Solve puzzle (919) [1.92, 2.16] [2.20, 2.38] [2.67, 2.81] [3.07, 3.30] 17. Paired colours (929) [2.02, 2.34] [2.82, 3.00] [3.49, 3.78] [3.65, 4.00] 18. Group similar drawings (1237) [3.85, 4.20] [4.63, 4.86] [5.52, 6.09] Fine motor adaptive 19. Follow to mid-line (164) / [0.03, 0.06] [0.07, 0.11] [0.14, 0.25] [0.16, 0.32] 20. Hands semi-open (378) [0.00, 0.06] [0.08, 0.1 2] [0.15, 0.19] [0.21, 0.27] 21. Regard own hand (382) / / [0.13, 0.17] [0.18, 0.22] [0.24, 0.28] [0.31, 0.38] 22. Hands together (378) / // [0.14, 0.19] [0.23, 0.27] [0.32, 0.37] [0.39, 0.47] 23. Pass cube (209) [0.23, 0.29] [0.31, 0.35] [0.37, 0.41] [0.42, 0.50]

12 58 H. Lejarraga et al. Appendix (continued) Centiles Item (sample size) 25th [CI] 50th [CI] 75th [CI] 90th [CI] 24. Pick up raisin (282) / [0.36, 0.41] [0.43, 0.47] [0.49, 0.54) [0.55, 0.62] 25. Neat pincer grasp of raisin (187) / [0.65, 0.80] [0.84, 0.93] [0.93, 1.16] 26. Dump raisin (950) / [0.86, 1.00] [1.22, 1.34] [1.53, 1.71] 27. Put raisin in bottle (453) //+/ [0.82, 0.91] [0.98, 1.05] [1.15, 1.29] [1.35, 1.65] 28. Scribbles (923) [0.87, 0.94] [1.01, 1.08] [1.22, 1.31] [1.51, 1.73] 29. Tower of four cubes (772) // [1.18, 1.27] [1.38, 1.46] [1.61, 1.72] [1.89, 2.09] 30. Tower of eight cubes (1187) // [1.74, 1.88] [2.12, 2.23] [2.54, 2.68] [3.01, 3.26] 31. Corrects tower (1404) [1.50, 1.69] [2.16, 2.33] [3.00, 3.32] [3.59, 4.12] 32. Imitates bridge (932) /// [2.05, 2.25] [2.50, 2.64] [2.99, 3.17] [3.51, 3.87] 33. Draw person three parts (1253) // [2.83, 3.02] [3.38, 3.53] [3.98, 4.18] [4.64, 5.00] 34. Copy cross (1055) // [3.12, 3.33] [3.58, 3.75] [4.13, 4.34] [4.73, 5.20] 35. Folds paper (1038) [2.93, 3.29] [3.61, 3.82] [4.37, 4.63] [4.76, 5.14] 36. Draw person six parts (1544) // [3.48, 3.69] [4.12, 4.27] [4.81, 5.01] [5.56, 5.93] 37. Copy triangle (620) [4.90, 5.20] [5.44, 5.66] [5.73, 6.14] Language 38. Cochleo palpebral reflex (176) Searches for sound with eyes (203) [0.26, 0.33] [0.36, 0.42] [0.44, 0.54] [0.45, 0.57] 40. Answer to no (386) 0.50 // [0.45, 0.54] [0.54, 0.63] [0.73, 1.06] 41. Combine syllables (191) 0.51 /0+/ [0.44, 0.55] [0.57, 0.64] [0.66, 0.80] 42. Dada/MaMa, non-specific (386) / [0.43, 0.53) [0.54, 0.61] [0.66, 0.72] [0.76, 0.85] 43. Phrase word (781) [0.97, 1.05] [1.04, 1.14] [1.33, 1.51] [1.70, 2.25] 44. Point two pictures (696) [0.89, 1.14) [1.28, 1.45] [1.74, 1.90] [2.13, 2.42] 45. Hum in presence of third person (732) [1.06, 1.74] [2.38, 2.88] [2.60, 3.30] 46. Name two pictures (738) [1.13, 1.38] [1.69, 1.85] [2.14, 2.41] [2.25, 2.57] 47. Combine words (1031) // [1.59, 1.69] [1.86, 1.94] [2.11, 2.21] [2.35, 2.49] 48. Gives first and last name (917) [1.94, 2.21] [2.05, 2.30] [2.71, 2.93] [3.41, 3.92]