Longitudinal Research in Down Syndrome

Longitudinal Research in Down Syndrome JANET CARR DEPARTMENT OF PSYCHOLOGY ST. GEORGE’S HOSPITAL MEDICAL SCHOOL UNIVERSITY OF LONDON LONDON sw17 ORE, ENGLAND

I.

INTRODUCTION

Down syndrome, discoverable at birth and the most common single cause of mental retardation, has been a prime target of longitudinal investigations. These have been of two main kinds: (1) those that aim at charting the patterns of natural development of the children; (2) those that document the influence of intervention programs on the children’s progress. Each is briefly reviewed and succeeded by an outline of the writer’s 21-year prospective study and a more detailed consideration of some of the longitudinal findings from it. The review refers to the following criteria for high-quality longitudinal research, and attempts to assess the extent to which the investigations discussed conform to them: 1. All available subjects in a given geographical area and time span should be included. 2. The research should continue over a matter of years rather than months. 3. All subjects should be examined on every occasion. 4. Subjects who drop out should be accounted for. 5. Measures should be taken on several, rather than on only two, successive occasions.

Where intervention is involved, five further points obtain: 6. The subjects should be carefully assessed at the outset of the experiment. 7. The subjects should be randomly assigned to experimental and contrast groups. INTERNATIONAL REVIEW OF RESEARCH IN MENTAL. RETARDATION. Vul. I X

I97

Copyright 0 1992 by Academic Press. Inc. All rights of reproduction in any lorn reserved.

198

Janet Carr

8. Subsequent assessment should be carried out by researchers who are “blind” to the subjects’ groups. 9. The regime for each group should be clearly described. 10. Follow-up beyond the termination of the intervention should be reported.

These are rigorous requirements and some at least may not always be attainable. Nevertheless, if they are transgressed the reasons for this and any effect it may have on longitudinal findings should be explicitly considered and discussed.

II.

NONINTERVENTION RESEARCH

Early papers provided consistent information on the decline of IQs in young children with Down syndrome, both hospitalized and home reared. Dameron (1963) followed 12 hospitalized infants from 3 to 18 months. Each was tested within 4 days of the birth day but the homogeneity of this sample is attenuated by numbers that fluctuate between I 1 and 5. Mean developmental quotients (DQs) were 59 at 12 months and 46 at 18 months, well below those of home-reared children (described later), findings that were repeated subsequently (Shipe & Shotwell, 1965; Shotwell & Shipe, 1964; Stedman & Eichorn, 1964). In homereared samples mean scores decreased from 70 in the first year, to 60 in the second, 50 in the third, and 45 in the fourth year. (Koch, Share, Webb, & Graliker, 1963; Share, Webb, & Koch, 1961). These means, however, were commonly derived from individual results gained at different ages and then grouped around chronological ages (CAs); the children were clinic attenders whose initial test results happened to be available, with no information given as to why or at what ages they had been tested. Numbers diminished with increasing age, no explanation being given (Fishler, Share, & Koch, 1964). Dicks-Mireaux (1972) appears to have maintained the numbers in her research but this goes only as far as 2 years of age. Melyn and White (1973) amassed findings from 532 children, from 1 to 14 years of age, who had attended an outpatient clinic and been given a wide variety of tests. “Most” are said to have been tested on a longitudinal basis, but after 7 years the number dropped to less than half the original (76), and at 12, 13, and 14 years, only 9, 6, and 4 children are represented. More recently, Berry, Gunn, and Andrews (1984) have provided information on 39 Australian children with Down syndrome, tested between 2$ and 48 months. “Most” children born in one area over a 4- to 5-year span were included. The children were not all assessed at the same intervals on the same number of occasions; each child is estimated to have had an average of nearly five assessments spread over 12 occasions up to 6 years of age. The data are presented only as scattergrams, making it impossible to compare them with any other data.

LONGITUDINAL RESEARCH IN DOWN SYNDROME

199

Thus, of the nonintervention projects reviewed, all continued over a matter of years and contained measures taken on several successive occasions. Most have not examined all subjects on every occasion (Berry et al., 1984; Dameron, 1963; Fishier et al., 1964; Melyn & White, 1973) nor given reasons for omissions. None included all available subjects in a geographical area and time span or accounted for missing subjects. The research gives some consistent information on the early development of children with Down syndrome, but very little is available after about age 7. The majority of this research is so insecurely based that, without access to the original data, few longitudinal conclusions can be drawn.

111.

INTERVENTION STUDIES

Excellent reviews of the remedial effectiveness of intervention projects already exist (Farran, 1990; Gibson & Harris, 1988). Here such projects are considered for their longitudinal validity, assessed on the 10 points outlined above. One of the earliest interventions was that of de Coriat, Theslenco, and Waksman (1967). Details are given of the 12-month stimulation program provided and the 6-monthly assessments made over a total of 5 years. The intervention group initially comprised 30 subjects, rising to 80 at 18 months and diminishing to 9 by 5 years. Numbers in the contrast group ranged from 182 to 15. No information is given on how the children were allocated to their groups nor on the identity of those carrying out the assessment. Subsequently, 16 children between 21 and 69 months of age were divided into two groups matched for age and sex and for mean mental age (Aronson & Fallstrom, 1977). One group received a half-hour of training twice a week for 17 months; the other received the normal preschool regime. All the children were tested on the Griffiths (1954) scale before, during, and after the training period and again 1 year later. Ludlow and Allen (1979), in a potentially valuable investigation, employed an intervention group and two contrast groups, one home and the other institution reared. The stimulation program continued for 2 to 3 years and the children were then followed until age 10. The groups were large, containing respectively 72, 79, and 33 children, but apart from one occasion the full groups are not represented at any age, numbers in the intervention group varying from 25 at 1 year, to 72 at 4 years to 28 at 10 years. Moreover, the children were tested on average only about once in 2 years and “interpolated quotients” (gained by connecting measured points for any child by a straight line and estimating intermediate figures from this) are included. These fictitious figures are not acceptable although the authors state the procedure was preferred by the statisticians consulted. A general description of the intervention regime, but not of the conditions for the other two samples, is given. It is not true, however, as Farran (1990)

200

Janet Carr

avers, that parents of the honie-reared contrast group “had actually refused to participate in the early intervention project” (p. 526); Ludlow and Allen (1979) make it clear that this applied to only four (5%) of these families (p. 30). Piper and Pless (1980) assigned 37 infants with Down syndrome to different groups according to time of entry into the program. Twenty-one infants available from July to December and 16 infants available from March to June were allocated to treatment and no treatment groups, respectively. The former had biweekly center-based 1-hour sessions and their mothers were given instructions to follow between visits. The children were tested “blind” on the Griffiths scale at the outset and at the termination of the project, 6 months later. Cunningham and his associates in Manchester have been engaged in a largescale ongoing study of many facets of intervention, including intensity, frequency, and the age at which the intervention was begun (Cunningham, 1987). The population consisted of about 90% of all births in the area and was judged a representative sample. Means on the Bayley and Binet tests are given from 1 to 5 years, the number of children involved varying from 59 to 112. In one substudy of the timing of visits (Cunningham, 1982) and in another of intensity of training (Cunningham, 1983) 40 and 15% respectively of assessments were done “blind” as to the child’s status, the latter comparison revealing no evidence of bias. The considerable body of data that emerge was subjected to meticulous longitudinal analysis. Children in a recent American program were estimated to constitute about onethird of those in the geographical area and “assumed” to be representative of its population (Pueschel, 1984). A program of motor and sensory stimulation was developed for each child and taught to the parents through demonstration and written guidelines. Between 75 and 88 children were tested biannually on the Bayley scales from 6 months to 3 years, DQs throughout being between 55 and 58 (Schnell, 1984). No controls were used or referred to, attention being given primarily to within-group variables such as cardiac involvement and parental “follow-through” (derived from ratings made at each visit of how well the parents were carrying out the program of Stimulation provided). The policy of enrolling all children with Down syndrome born in Jerusalem in a stimulation program offered an opportunity to study a true population sample (Sharav & Shlomo, 1986). E’ifty-one children were included. A nonspecific description of the intervention program is given, consisting of home visits by an occupational therapist at about 6 weeks, followed after 18 months by a kindergarten milieu with extra speech and physiotherapy. One hundred and eighty-nine assessments were made over 19 occasions from 3 months to 13 years, a maximum of 22 children appearing at 1 and 2 years and 4 or fewer children represented after the age of 6. The findings were compared with those from two previous studies, by Carr (1970) and by Dicks-Mireaux (1972); however, because in Jerusalem infants with severe cardiac and other malformations were

LONGITUDINAL RESEARCH IN DOWN SYNDROME

20 1

excluded, the conclusion that they were superior to Carr’s group, which included such children, is inappropriate. To summarize, all the intervention projects assessed the children at the outset; all except one (Piper & Pless, 1980) continued over some years and included measurements taken on several occasions. All except two (Piper & Pless, 1980; Schnell, 1984) followed up their sample at least one year beyond the termination of the intervention, although the small numbers reported by Sharav and Shlomo (1986) detract from the value of the latter part of their follow-up. In no case were subjects randomly assigned to experimental or contrast conditions, but in one study the groups were compiled geographically (Ludlow & Allen, 1979) and in another, temporally (Piper & Pless. 1980). implying a degree of nonselectivity in their composition. Otherwise matched controls (Aronson & Fallstrom, 1977), noncontemporary controls (Cunningham, 1982; Sharav & Shlomo, 1986), or no controls (Schnell, 1984) were used. Cunningham is, however, unique in employing specific controls for specific purposes. No study gave a clear description of the regime experienced by the controls; all gave some description of that for the intervention children, varying from minimal (Schnell, 1984; Sharav & Shlomo, 1986) to more detailed (Aronson & Fallstrom, 1977; Cunningham, 1987). Finally, in only one study were all (Piper & Pless, 1980) and in another a proportion (Cunningham, 1982, 1983) of the assessments carried out blind. None of the research reviewed fulfills all the criteria. The most satisfactory studies are those of Cunningham (1987) and Aronson and Fallstrom (1977); the most fallible, that of Sharav and Shlomo (1986). A judgment of the methodological strength of any piece of research can be made only on the basis of information supplied; researchers who give clear details of negative, as well as positive, aspects of their work may be penalized, whereas others who fail to provide such details may escape. It is nevertheless apparent that much of both intervention and nonintervention research suffers from serious methodological weaknesses, and that these limit the credence that can be given to their findings.

IV.

THE SURREY STUDY

The Surrey study is of a population-based cohort of Down syndrome subjects. All babies born in one year, December 1963 to November 1964, in part of southeast London (the boroughs of Camberwell and Lewisham, 23 infants) and in the county of Surrey (31 infants) were included. Camberwell and Lewisham are densely populated, city areas, whereas Surrey is predominantly suburban and rural. Because more children came from Surrey than from Camberwell and Lewisham, the project is referred to as the Surrey study. In the first instance the intention was to provide developmental data to complement neurological and chromosomal investigations in the first 10 months of life (Cowie, 1970). Subse-

202

Janet Carr

quently it was decided to extend the developmental study to 4 years, to include nondisabled controls for the home-based children, and to explore, through interviews, the effects on the families of either type of child. Fifty-four babies with Down syndrome were referred, 25 males and 29 females. Careful inquiry over the years has not revealed any missed cases. Fortyfive children were brought up in their own homes, at least until the 11-year assessment, and nine in various foster homes. The children were seen at home at 6 weeks, 6 months, 10 months, 15 months, and 2 , 3, 4, 11, and 21 years. Psychological tests were given on every occasion, and at 15 months, 4 years, 1 I years, and 21 years, the mother (or principal caretaker) was interviewed. Visits were scheduled to occur, at 6 weeks, within 3 days either way of the birth date and, at 6 months to 4 years, within 2 weeks either way. Over 80% were seen within the limits (Carr, 1975, p. 13). At I 1 and 21 years, all subjects were seen within a 3- to 6-month period. All the children were tested on every occasion apart from those who died and 2, 1 , and 3 children who refused parts of the 15month, 3-year, and 4-year tests (Carr, 1975, p. 14). By 21 years, 41 subjects remained in the study: all losses were due to death (5 by 2 years, 2 by 4 years, 1 by 1 I years, and 2 by 21 years) apart from one child whose family refused contact after the first visit (this child was not included in any analysis), one whose family emigrated before she was 2 years old, and one young man whose mother refused contact at 2 1 because of her anxieties about computerized data. Each home-reared child was matched for sex, age, and social class with a nondisabled child; this contrast group was tested, from 6 weeks to 3 years, and the families were interviewed on every occasion as were those with Down syndrome. This contrast group provided contemporary early developmental test data with which to compare those from the children with Down syndrome; in the family study they allowed identification of those problems that arise in rearing a child with Down syndrome as distinct from problems encountered by any family. Because they are not relevant to the topics under discussion in this paper the contrast group will not be referred to further. Formal support for the project was ended at 4 years of age; however, contact was maintained (through Christmas cards) with the Down syndrome families and, at 11 and 21 years, it was possible to follow up both groups. All testing, at every age, for both groups was done by the author. At 1 I years (Carr and Hewett, 1982), Dr. Sheila Hewett interviewed 37 of the 43 Down syndrome families; all other interviews were carried out by the author. The instruments used to assess the children’s abilities were, in the early years, the Bayley scales of Mental and Motor Development (Bayley, 1964); at 11 years, the Merrill-Palmer Scale of Mental Tests ( 1948) and the Reynell Developmental Language Scale (Reynell, 1969); and at 21 years, the Leiter International Performance Scale (Leiter, 1980) and a range of language and academic tests (see Carr, 1988a). The interview schedules used to collect data on the families were based

LONGITUDINAL RESEARCH IN DOWN SYNDROME

203

on, at 15 months, those devised by Newson and Newson (1963) and, at 4 years, those devised by Hewett (1970); at 11 and 21 years, items from the Health, Behavior and Skills (HBS) schedule (Wing, 1980) were included. These are properly devised and commonly used measures, enabling comparisons to be made with other investigations. The strengths of this study are that it concerns a population-based 1-year cohort of children with Down syndrome and their families, with a matched comparison group, who have been followed by a single investigator for a long period, virtually all attrition resulting from death. The weaknesses are, first, that the sample is a relatively small one and, second, that there are periods of several years in which no testing was done. After seven test occasions in the first 4 years, there was a 7-year interval to age 11 and a 10-year interval to age 2 1. These gaps occurred because continuous funding was not available for the project, the two later follow-ups being made possible only through the provision of brief periods of study leave to the author. Nevertheless, it is believed that no other study currently exists that includes both a Down syndrome group and a contrast group, is concerned with both children and families, and has followed these subjects from infancy to adulthood.

A.

Cognitive Aspects

1 . RESULTS

Figure 1 illustrates the results on developmental and intelligence tests for the whole cohort,' showing that mean ratio IQs fell from 80 at 6 months to 45 at 4 years and 37.2 at 11 years. The decline in scores was most rapid in the first 2 years, an average of 17.5 ratio IQ points per annum, whereas the average between 2 and 4 years was 5 points and that between 4 and l l years, 1.1 points per annum. At 21 years mean IQ rose to 41.9, an increase of 4.7 points from the mean at 11 years. Correlations between mental ages from 6 weeks to 4 years were insignificant for the early tests, but from 10 months onward were significant, reaching 0.92 between 3 and 4 years (Carr, 1975). Correlations between IQs at 11 and 21 years with those at all other ages were low until 10 months, between 0.10 and 0.43, confirming the failure of early infant tests to predict later development (Bayley, 1955). At 15 months the correlations rose and, from 2 years onward, were between 0.61 and 0.83. Thus, in this population mean IQs remained very stable over long periods, tests even in early childhood predicting with considerable accuracy scores of the same group in middle childhood and early adulthood, albeit on different tests. 'Figures for the 41 surviving to age 21 are very similar, being two points higher at 6 and 10 months, one point higher at 3 years, and identical at all other ages.

204

Janet Carr

'"1 -

0-0

6 highest at 21 whole cohort 6 lowest at 21

0-0

0-0

90

80-

70-

a

6050 -

40 30 -

20

~

10OJ

'

1.5

6

10

months

15

2

3

4

l U l + 11

21

years

FIG. I . Mean IQs and ranges, 6 months to 21 years.

Between 6 months and 4 years the scores of the non-home-reared children were significantly below those of the home-reared children, but neither at 1 I nor 21 years was any difference apparent. Mean scores of the females were consistently above those of the males, and by 21 years this difference was significant. Not only did females predominate in the upper quartile (Carr,1988a) but Fig. 2 shows that, over the whole range, females more frequently scored at the upper and males at the lower levels. Similar distributions are seen at other ages also (obtainable on request from the author). At no stage has there been a significant difference in IQ between children from the middle- and working-class family groups, although those from the workingclass families have consistently had a slight advantage. Academic skills (reading, writing, and arithmetic) at 21 years were significantly and positively associated with social class and were significantly correlated with IQ from 2 years onward; for example, correlations between 21-year reading level score and IQ at 2, 3, 4, 1 1 , and 21 were, respectively, 0.57, 0.57, 0.62, 0.71, and 0.65. This apparent paradox of significant relationships between academic skills and both IQ and social class when the latter two are not related may be explained by missing data;

205

LONGITUDINAL RESEARCH IN DOWN SYNDROME

60

Males 50

Females

40

0

E

B 30

a

20

10

I 61-70

C

21-30

31-40

41 -50

51-60

IQ FIG. 2. Distribution of 21-year IQs by sex (percentages).

all the subjects were involved in calculating the association between IQ and social class, but when calculating the association between academic skills and social class many subjects (e.g., 10-12 for reading) could not be included as they failed to score on the tests. The relationship of IQ to parental education was also explored, using two measures: (1) years of education reported by each parent and (2) parental educational level, a score of 1 representing education that did not go beyond secondary school, 2 any further training that did not include academic qualifications, and 3 training that included academic qualifications. Parental educational factors were significantly related to social class, but in this population neither showed a clear

206

Janet Carr

association with child ability. This agrees with the findings of Bennett, Sells, and Brand (1979) but conflicts with the positive association between child IQ and parental education reported by Cunningham (1987) and by Fraser and Sadovnick (1976). Multivariate analyses were undertaken to explore the relative importance of the effects of social class, place of rearing, birth weight, illness, health, hospitalization, vision, hearing and weight, parental age at birth, parental education, mother’s scores on the Malaise scale (Rutter, Tizard, & Whitmore, 1970), and “telling,” a score derived from mothers’ descriptions of how well or badly the news their babies’ condition was broken to them. None of these factors had any demonstrable effect on IQ scores except social class at 2 years, the advantage being to the working-class children. When previous IQ was added into the equation, this was shown to be by far the most important predictor of later IQ. The major factor predicting IQ from 15 months onward was previous 1Q and no other factor contributed to any great extent. The scores of two subgroups, those at the extremes at 21 years, were selected for comparison (mean IQs of each of these groups are included in Fig. 1). The six young people who had the lowest scores and were profoundly disabled by 21 years have been described elsewhere (Carr, 1988b), and to balance these an examination has been made of those six (15% of the whole population at 21 years) who gained the highest scores at 2 1 years. (The means for 6 weeks are not given as only two in the profound group and four in the high group were seen at that time.) The two sets of scores run on fairly even lines, the gap between them gradually widening from 6 points at 6 months to 24 points at 3 years. Thereafter, the scores of the profound group continued to decline but those in the high group rose minimally at 4 years, declined slightly to 11 and rose, by 1 I points, to a mean of 60 at 21 years. Means for the high group were consistently above the means of the cohort as a whole, and those for the profound group, consistently below. Nevertheless, the two groups overlapped somewhat, especially in the early tests. At 15 months the score of one subject in the profound group exceeded those of three subjects in the high group; there was no overlap at 2 and 3 years, but at 4 years, the score of this same girl exceeded (by 1 point) that of one subject in the high group. Thereafter no overlap was seen. About half the children with very high or very low early (10-month) scores continued in those positions. The majority of the remaining subjects had scores that varied somewhat, but one girl’s scores changed dramatically, moving from an early position in the high group to the profound group at age 2 1. This girl became phobic and withdrawn at about 9 years of age, but to date no other factor such as health or family circumstances can be adduced to account for this massive change. The reasons for the differences between the groups are hard to tease out. Of those in the high or low groups at age 21, females predominated in the high

LONGITUDINAL RESEARCH IN DOWN SYNDROME

207

group and males in the low group. Social class was similar in each, four being middle class, one working class, and one placed out of home in early life. Heart, vision, and hearing problems did not differ between the groups (the young woman with the highest score at age 21 was deaf and wore a hearing aid); however, the profoundly disabled group had more health problems; at age 21 they had had an average of three hospital appointments and two spells as an inpatient in the last year and 40 days as an inpatient since age 1 1, compared with average figures of 1.3,0.2, and 5 , respectively, in the high group. Three subjects in the low group had two or more severe medical problems (such as very poor vision or hearing, fits, and heart or kidney problems); one subject in this group had two, one had three, and one had four such problems compared with only one subject in the high group who had two of these severe problems. Although these problems may have contributed it seems unlikely that they would be sufficient to account for the low group’s more severe impairment. Graphs of IQ scores over time for groups of people with Down syndrome, such as those in Fig. 1, commonly show a relatively smooth, declining curve. This is composed of means derived from the scores of individuals, which may deviate considerably from this pattern. Although IQs are relatively stable in Down syndrome populations, with high correlations from one age to another, individuals do not necessarily make steady progress but may vary in their performance from one age to another. It is not possible to predict with certainty how any one will develop, although approximately half may be expected to remain at or near the positions they originally occupied relative to the group. 2. DISCUSSION

Five main points emerge from this part of the investigation. First, mean IQs declined steadily from infancy to childhood, increasing slightly at age 21. Second, this group trend did not necessarily mirror the curve of scores for individuals, some of whom demonstrated large changes. Third, the deleterious effect of out-of-home rearing was much less evident at ages 11 and 21 than it had been earlier. Fourth, females had consistently higher mean scores than did males. Fifth, neither social class nor parental education was significantly associated in this group with IQ. Turning to a more general discussion of these points, the curve of scores for the whole group seen in Fig. 1 is, in its early stages, similar to that shown in other nonintervention studies (Bayley, Rhodes, & Gooch, 1966; Dicks Mireaux, 1972; Fishler et al., 1964; Shipe & Shotwell, 1965; Shotwell & Shipe, 1964). This decline has been variously suggested as resulting from deterioration of cerebral function, artifacts of test construction or content, or specific neuromotor and sensory disabilities emerging with age, but no definitive conclusion has been reached (see Gibson, 1978, pp. 30-34). Intervention studies have shown a less marked decline in scores (Cunningham,

208

Janet Carr

1987; Schnell, 1984), this effect being commonly reported while the intervention was ongoing. Cunningham (1987) has documented the failure of such effects to survive the termination of the intervention, although children in the intervention group were found to be healthier and more likely to have attended a mixed/mainstream preschool; also, their mothers were more likely to be employed and willing to leave the child unsupervised for long periods. These findings, both those on child health and schooling and those directly concerned with the mother, reflect findings from other long-term intervention studies, in the light of which it has been postulated that a more lasting effect of intervention may be seen in parental attitudes and motivation than can be observed in the children themselves (Clarke & Clarke, 1981, p. 354, 1989, pp. 292-293). At the young adult level mean Leiter IQ is similar to that reported from a group with an average CA of 28 (Holmes & Cam, 1991). Berry, Groenweg, Gibson, and Brown (1984) present results from 28 adults, between 15 and 42 years of age, (mean CA = 21), who gained a mean mental age of 5 t on Raven’s Coloured Progressive Matrices. This corresponds exactly to the 21-year mean mental age in Surrey, when the six profoundly disabled young people who would not have been able to attempt Raven’s test are excluded. In the Berry et al. study mean mental age rose from 54 to 74 six years later, at a mean CA of 27 years, more than twice the increase seen from 11 to 21 years in Surrey. In the former study, however, this gain was demonstrated on identical tests. Although the possible remedial effects of the program in which the group was involved must be taken into account, it appears that a real gain has been demonstrated. In another study nearly half of a group of 39 adults with Down syndrome, reassessed after 6 i years, had increases in mental age, most of the increase coming from those in their twenties and thirties (Fenner, Hewitt, & Torpy, 1987). These findings demonstrate the importance of charting intelligence levels of adults with Down syndrome throughout their third and fourth decades of life and beyond, especially if in later years comparisons are to be drawn with previous levels to establish the process of aging in this population. If IQ can be shown reliably to increase beyond the second decade, this needs to be taken into account when documenting any later decline. Identifying differences between the scores of children brought up within their own homes and those reared elsewhere was a prime concern of some of the early studies (Bayley et al., 1966; Dameron, 1963; Shipe & Shotwell, 1965; Shotwell & Shipe, 1964; Stedman & Eichorn, 1964). In all these studies marked differences were seen between the groups, to the advantage of those brought up within their own homes, as indeed was found in the early stages in Surrey. By 11 years, however, and again at 21 years, there was no difference in nonverbal IQs, although there was a suggestion of some superiority in language for those brought up at home (Carr, 1988a). It should be remembered that the non-homereared Surrey children were brought up in small homes and foster homes, not in

LONGITUDINAL RESEARCH IN DOWN SYNDROME

209

large institutions. Nevertheless, the specific effect on IQ of in-home rearing may be less pronounced in the long term than it is in the early ages. The Surrey finding of higher scores by females is mirrored in other longitudinal (Cunningham, 1987; Schnell, 1984) and nonlongitudinal (Clements, Bates, & Hafer, 1976; Connolly, 1978; Gath & Gumley, 1984) studies. Gibson (1978) suggests that this difference is due to selective mortality, the more severely disabled females dying earlier (p. 109), but this conjecture could not be substantiated in the Surrey study (Carr, 1988a). Others have pointed to the greater linguistic facility of females as an explanation for these differences (Cunningham, 1987; Schnell, 1984). Cunningham suggests that this “may not be some inherent factor but related to interactional style within families” (p. 175), paralleling ideas put forward about the same phenomenon in nondisabled children (Newson & Newson, 1977, p. 186). This hypothesis deserves exploration. Nevertheless, in Surrey the differences between the scores of the sexes increased steadily over the years and was largest at age 21, when the test used required no language, either verbal or expressive; the females’ superiority on tests of language was not significant when allowance was made for IQ. No study, so far as is known, has shown males to be superior. Bearing in mind the distribution of scores seen in Fig. 2 and despite the need, in view of the small numbers, for this finding to be replicated, the possibility must be entertained that female sex confers some advantage to populations with Down syndrome. In Surrey, neither social class nor parental education was associated with IQ. In contrast, Cunningham (1987) found both factors to be significantly associated with child mental age, parents of higher social class and with higher educational levels having children with higher mental ages. In Surrey, the direction of the social class effect, although not significant, was the reverse of that seen by Cunningham, with a slight advantage to children and adults from working-class families. Despite measures of parental education being similar in the two studies, the significant effects on IQ seen in Manchester could not be seen at any age in Surrey. In seeking an explanation of these discordancies, we note that an obvious difference between the two studies lies in the proportion of those with profound disabilities: none in the Manchester study (Cunningham, personal communication) and 5% at 4 years rising to 15% at 2 1 in the Surrey study. The Surrey figures were recalculated omitting these children, resulting in a slight shift toward the positive end of the scale; for example, correlations of IQ with parental educational level went from between -0.27 and 0.025 to between -0.18 and 0.012, all remaining very small and nonsignificant. It therefore seems that the presence in the Surrey study of people with profound disabilities is not sufficient to explain these differences. Another difference is that of size, the Manchester group being roughly four times larger than its Surrey counterpart; however, the numbers in the latter are

210

Janet Carr

similar to the numbers in other studies in which positive relationships with parental abilities have been reported (Fraser and Sadovnick, 1976). Although larger numbers may have resulted in associations more like those found in Manchester, nothing in the figures as they stand suggests that this is likely to be the case. It may be that such findings will not be seen invariably and that populations of people with Down syndrome, like the individuals within such a population, may vary considerably from each other.

B.

Self-Help and Independence

The degree to which a child or adult can take responsibility for fulfilling his or her own personal needs can have a major effect on the burden of care for the family and on the person’s own sense of independence. In the Surrey study the skills were not observed directly, but questions were asked about them in the course of structured interviews carried out with the parent (usually the mother) or principal caregiver at 15 months and 4, 11, and 21 years. The data presented here focus on four areas-feeding, washing, dressing, and toiletingand on a combined score labeled total independence. 1. RESULTS Figure 3 indicates the proportion of people independent and those totally dependent in each area at 4, 1 1, and 21 years of age. Fairly steady progress was made over the years; most progress was made by age 11 although further improvement was still possible beyond that age. Nevertheless, by age 21 only 18 young people, 44% of the whole group, were fully independent in all four areas. Where full independence had not been achieved washing was the most frequent difficulty for nearly half (49%), problems with hair washing accounting for most of that. After 4 years those totally dependent were very severely and profoundly disabled individuals. Correlations of each self-help score with the others at ages 1 1 and 21 show that at 1 1 years dressing was significantly related to feeding and washing and these to each other ( r = 0.43 to 0.521, but toileting was not related to any other score. Total independence was highly significantly related to feeding, washing, and dressing ( r = 0.72 to 0.81); however, by age 21 all areas, including toileting, were strongly related to each other ( r = 0.70 to 0.86). IQs at 2, 3, 4, 11, and 21 years were each correlated with each self-help score at ages 11 and 2 1. At age 11 all previous IQs were significantly related to total independence, but only IQs at 3 and 4 years were related to washing and, at 3 years, to dressing. None of the previous IQs was related to feeding or toileting. All correlation coefficients, although significant, were small, between 0.43 and 0.56. At 21 years, IQs at all the previous ages were significantly related to all areas. Between 2 and 4 years, the coefficients were between 0.43 and 0.55; from 11 years onward, between 0.6 and 0.78.

21 1

LONGITUDINAL RESEARCH IN DOWN SYNDROME

Feeding (knife & fork) independent

Bathing

loo[ 80 60 40 20

0

-3

'O0

a,

80

C a,

6o

ea,

a

Dressing

4o 20

0 l o80 o[

60 40 20 0

loo

80

- All four areas -

60 40

-

20

-

0

1//,//1

FIG. 3. Independence and dependence in four self-help areas at 4, I I , and 21 years (percentages).

212

Junet Curr

If regression analysis is used to explain the scores at 4 years, IQ at age 3 has some effect on scores for feeding and nighttime continence, but even with sex and social class these account for less than 23% of the variance. IQ has no perceptible effect on total independence. At 1 1 years, total independence is quite well accounted for by 1Q at age 4 (42%), and this is hardly added to by the inclusion of sex and social class. Using 1 I-year IQ in place of that at age 4 results in a dramatic improvement, with 77% of the variance now accounted for; however, substituting the 4-year independence score gives a much reduced percentage, to less than 10%. So total independence at age 1 I can be largely accounted for by IQ, and of the measures available at 4 years, IQ has much greater predictive power than has independence at that time. At 21 years, total independence is well explained by 1 I-year IQ ( 6 5 % )and sex and social class add only another 4%. Using IQ at age 21 in place of that at age 1 1 increases the proportion explained by 9%, but substituting independence at I I increases this to 85%. Of the individual skills at age 21, IQ at age 1 1 made the greatest contribution apart from self-management at the toilet, which is more influenced by the same factor at age I I . With IQ allowed for, no effect could be seen of sex or social class but those young people brought up from infancy out of their own homes had a higher mean score for dressing and for total independence. In general, neither sex nor social class had any denionstrable effect on selfhelp skills, either singly or in combination, at any point. Total independence was difficult to explain at age 4, and was predominantly influenced by IQ at age I 1 and by the previous level of independence at age 21. It seems then that the timing of the acquisition of these skills may be quite variable in young children so that the level they achieve by the middle school years is little related to their achievement in the preschool years but more closely governed by their innate ability. By the middle school age, however, the pattern of children’s skills is more clearly defined and, although still strongly associated with ability level, predicts with considerable accuracy the level of independence they will demonstrate as young adults. 2. DISCUSSION The data from the Surrey study may be compared with those from a number of other studies that have included examination of self-help skills in children with Down syndrome, although few of these studies have been longitudinal. Buckley and Sacks (1987) examined the self-help skills of 90 children with Down syndrome: 40 girls and SO boys, aged 1 1 to 17, divided roughly equally into those over and under the age of 14. There are references to “improvement with age” ( pp. 24, 34) but the study is not a longitudinal one. In general, the authors are careful not to imply progression from one age to another, but discuss the levels of achievement of the “younger” and “older” children. The only significant dif-

LONGITUDINAL RESEARCH IN DOWN SYNDROME

213

ferences between the age groups concerned the preparation of snacks, and the summed self-help scores, both in favor of the older group. Neither sex, health, nor family position had a significant effect on skill development. Elsewhere, a significant increase in self-help with increasing CA has been shown in adolescents with Down syndrome, although not in those with autism (Loveland & Kelley, 1988). Two other British studies give some information on self-help skills. Gath (1985) had not assessed these at earlier ages but, in the 23 children remaining at 9 years, scores on the AAMD Adaptive Behavior scales (covering self-help as well as social function and communication) were related to IQ (r = 0.44) and to birth weight ( r = 0.56) but not to any other factor considered. Byme, Cunningham, and Sloper (1988) do not discuss feeding, washing, or dressing, and toileting and sleep are considered as behavior problems (raising the question of what should be regarded as a behavior problem). Sleep and toileting headed the list of such problems, shown by between 20 and 40% of the children, but longitudinal data are not given. In a later paper (’hmer, Sloper, Knussen, & Cunningham, 1991) an Index of Self-Sufficiency, includlng self-help, domestic, and community skills, was related to a variety of child and family factors. Overall scores for 127 children aged 6 to 14 were significantly correlated with MA (r = 0.71) and IQ (r = 0.52). Also positively related were a lower level of behavioral problems, more social contacts and activities for the child, and the mother’s use of practical coping methods. Cullen, Cronk, Pueschel, Schnell, and Reed (1984) used the Vineland Social Maturity Scale to evaluate self-help skills in a group of 70 to 80 Down syndrome children aged from 6 to 36 months, seen biannually. Mean social quotient declined, though not significantly, from 84 at 6 months to about 70 by 24 to 36 months. The presence of moderate and severe congenital heart disease was a powerful factor, associated with a 3-month deficit on the Vineland at 36 months. Excluding those with moderate and severe heart disease, those with good muscle tone, and those whose parents were able to “follow through” on the guidance given them did better than those with poor muscle tone and those whose parents did not “follow through” ( p. 62). Feeding milestones were particularly carefully assessed and were shown to be passed in the same developmental sequence, although delayed, as in normal children. Sex had no effect, cardiac status affected finger feeding and independent spoon use and, more particularly, chewing and weaning, but parents’ good “follow-through” affected positively all nine feeding variables. The authors considered the question of the effect of “follow-through,” pointing out that the group rated adequate in this respect contained more children with good muscle tone and without significant heart disease. Despite this, “followthrough” was shown to have an independent effect; however, IQ may also have had an effect. IQ is related to “follow-through” (pp. 218-222), and although no

214

Junet Carr

correlations between IQ on the one hand and Vineland and feeding scores on the other are given, the Vineland and feeding scores were all significantly correlated, being “frequently higher than those among the feeding milestones” themselves (p. 248). The implication is that good “follow-through’’ was instrumental in bringing about better self-help skills, but it is also possible that the reverse obtained and that parents were more willing to exert themselves and to “follow through” where they saw the child as responsive and as having a good measure of learning ability. This study demonstrates the extra difficulties imposed by cardiac and muscle weakness on the acquisition of self-help skills by a young child with Down syndrome, and the advantages that accrue when parents enthusiastically embrace the advice they are offered. These effects were demonstrated in the course of a continuing program, and there is no evidence as to how far they would endure beyond its discontinuation. In both Surrey and Manchester (Turner et al., 1991) strong associations were found between self-help and IQ. None of the studies discussed has shown an effect of sex on self-help. This seems somewhat unexpected. If “interactional style within families” is thought to play a part in the higher intellectual abilities of females it might be expected to do the same for self-care skills; however, no such effect is found. Perhaps family interactional style affects intellectual and self-help skills differently, or this finding may diminish somewhat the force of the argument for its effect on intellectual ability.

C. Behavior and Management Important as is the developmental level of a child with Down syndrome, personality, temperament, and behavior are of almost equal salience in determining how well he or she will fit into a family and be accepted by society. Mothers were asked about behavior difficulties-aggressiveness, tantrums, naughtiness, and destructiveness-and how they dealt with these. When their children were ages 1 1 and 21, mothers were also questioned about personality, cooperativeness, and manageability; about behavior problems such as rebelliousness, pestering, self-injury, and public disturbances; and about their disciplinary methods, use of rewards, and attitudes to discipline. 1. RESULTS

Looking at the consistency of behaviors over time, nearly half (45%) of the 1 1-year-olds who had tantrums and half of those aggressive at that time were similarly rated at 21 (though only two of those said to be aggressive at age 4 were said to be so at ages 1 1 and 21). Two-thirds (65%) of those rebellious at age I 1 were said to be so also at age 21, though now only occasionally. Nearly half (47%) of those who pestered for attention but only 19% of those who showed difficult behavior in public at age 11 were still doing so at 21. Few children who at age 1 1 neither pestered nor rebelled nor were difficult in public showed these

LONGITUDINAL RESEARCH IN DOWN SYNDROME

215

behaviors at age 2 1. Self-injury was infrequent in this population, affecting less than one-sixth at any one time; only five ever were head-bangers and in no case was this a serious problem. Four of the five, however, continued to show some degree of self-injury; in the case of two profoundly disabled boys, this led to severe self-scratching. Two who had shown other forms of self-injury at age 11 continued it in a minor form at age 21. Thus, 11 subjects (a quarter of the total) in this population showed self-injurious behavior at some time and this persisted to some degree in 6 subjects, being severe and frequent in 2 profoundly disabled young men. Social class appeared not to be a relevant factor in persistence into adulthood of behavior difficulties, around half being implicated in each group (middle class = 55%, working class = 53%). Rather more females showed persistent behavior problems than males (males = 55%, females = 76%), more females continued to pester (males = 1, females = 6 ) , and more males continued to be difficult in public (males = 3, females = 0). Some differences in IQ appear when sex is allowed for; of the 9 males in whom behavior difficulties persisted, 6 (67%) were below the mean male IQ, compared with 3 of the 12 females (25%) who were below their mean. None of these differences reaches significance. Five of the six profoundly disabled young people had persistent problems, two showing severe self-injury and one of these other problems as well. To sum up, in this population about half of those who were aggressive, who pestered for attention, or who had tantrums and somewhat more than half of those who were rebellious at age 1 1 continued to show these behaviors at age 21. Difficult behavior in public tended to die out. Neither sex nor social class had a significant effect and these behaviors were somewhat more likely to persist in the more severely disabled men and in the more able women. Data were collected at 15 months and 4, 11, and 21 years on how parents handled their offspring’s behavior and their attitude to and use of disciplinary methods (e.g., supervision, smacking, sending to bed early, depriving of sweets or TV, and use of threats and rewards; see Carr, 1992). These factors were individually related to cooperativeness and manageability at ages 11 and 21 and to frequency of behavior problems at age 21. Manageability, rated at 4, 11, and 21 years of age, and cooperativeness, rated at 11 and 21 years of age, were also correlated across ages. The expectation was that although there might or might not be some stability in behaviors across ages, a major effect would be seen in the relationships between the disciplinary approaches used by mothers and the outcome in terms of cooperativeness, manageability, and presence or absence of behavior problems in their children. Considering first the relationships between behaviors across ages, young people with Down syndrome were found easier to manage at age 21 if they had had few tantrums at 15 months, and were more cooperative if they had been easy to manage at age 4 and, at 11, had kept out of mischief and been cooperative.

216

Janet Carr

There was then a tendency for those who were not subject to tantrums, who kept out of mischief, and who were cooperative and easily managed at early stages to be more cooperative and easy to manage later on, indicating some consistency of behavior across ages. Turning to the relationships between the measures of parental discipline and later cooperativeness, manageability, and presence or absence of behavior problems, only one (marginally) significant result was obtained: fewer behavior problems were seen at age 21 where at 4 years there had been little physical punishment for any child in the family. Given the number of associations examined this may have been a chance finding. No other factor nor the combination of individual factors into disciplinary scales at each age could be shown to have any association with later behavior. Finally, associations were explored with a factor labeled mother’s attirude. This was derived from responses to a question asked at ages 11 and 21: “How would you describe N as a person now?” Replies were grouped into those consisting of only positive, both positive and negative, and only negative comments. Because there were so few negative comments, they were combined with the positive and negative group, and this bipolar factor at ages 11 and 21 was related to cooperativeness at each age. Mother’s attitude at age 11 was related to cooperativeness at both 11 and 21 years; mother’s attitude at age 11 was not related to her attitude at 2 1, nor was attitude at 21 related to cooperativeness at 2 1, nor was cooperativeness at 11 related to mother’s attitude at 2 1, suggesting that it was not the child’s cooperativeness at the earlier age that led to a more positive attitude on the part of the mother later. Further, although, as we saw earlier, cooperativeness at age 1 1 was associated with cooperativeness at age 2 1, the association of mother’s attitude at 1 I with cooperativeness at 21 is stronger. These data may be regarded as only indicative, but they suggest that a positive attitude by the mother toward her Down syndrome child may play some part in the child’s later cooperativeness. In summary, in this cohort some consistencies were seen in behavioral ratings from one age to another: almost no effect could be demonstrated for parental disciplinary practices, and cooperativeness at a later age may have been influenced by the mother’s attitude earlier. 2. DISCUSSION Few data exist from other studies of children with learning disabilities with which the currently discussed data may be compared. Richardson, Koller, and Katz (1985), following up a group of young people with mild learning disabilities at ages 16 and 22, did not consider the effect of parental discipline, but concluded that behavior disorder was most likely attributable to the stressful conditions in which the young people grew up. (Similar conclusions, relating child behavior disorders to family discord, are reported from studies of nondisabled

LONGITUDINAL RESEARCH IN DOWN SYNDROME

217

children; see Rutter, 1976.) Byrne et al. (1988), in their longitudinal study of children with Down syndrome in Manchester, found some consistency of behavior, in that 75% of mothers who were concerned about their children’s behavior between ages 2 and 10 were still concerned about their behavior 2 to 3 years later. Efforts to relate the mother’s disciplinary methods to outcome resulted in only two significant associations: less problem behavior where the mother felt happy about her handling of the child ( r = 0.23) and more problems where she had threatened to send the child away ( r = 0.2; Sloper, personal communication). Data from large populations of children without disabilities suggest that poor adjustment in early adult life is related to difficult temperament, behavior disorder, and adjustment of the child and with parental conflict at age 3 (Chess & Thomas, 1984, p. 99). Also, difficult 1 I-year-olds were likely to be difficult at age 16 and later, a child frequently physically punished (once a week or more) at age 1 1 was more likely to be difficult at age 16 and to go on to acquire a criminal record than those not so punished, and this association was maintained even when sex, family size, and social class were controlled (Newson & Newson, 1989). Comparing these data with those from the Surrey study, both the Manchester and, especially, the Newsons’ studies concur with the present one in finding some agreement between the occurrence of problem behaviors at different ages. In the Newsons’ studies, these behaviors were generally less frequent in girls; the same tendency was seen in Surrey apart from pestering, which was more frequent in females. Neither of the relationships seen in Manchester, fewer problems where the mother felt confident and more problems if she has threatened to send the child away, could be detected, but some similarity was seen to the Newson’s findings on the relationship with physical punishment. Twenty-one percent of 1 1year-olds with Down syndrome were physically punished once a week or more often; at 2 I , all relationships-with being difficult to manage, uncooperative and having two or more behavior problems-were in the same direction as in the Newsons’ study, of more problems in those physically punished more often, although the figures do not reach significance. It seems, first, that there is evidence of some consistency in the behavior of children, with or without disabilities and that easy children tend to be easygoing as they get older and those with behavioral difficulties at one age are likely to be difficult later (although there are also many exceptions to this). Second, child behavior problems are associated with instability of upbringing and family discord in both groups, although there was insufficient of either unstable upbringing or family discord in the Surrey study to bear this out. Third, no evidence can be found in the studies discussed to show that frequent physical punishment leads to improvement in the child’s behavior. The evidence, so far as it goes, points to the contrary. Many mothers of children with learning difficulties who use physical punishment do so because they say, “It’s the only thing he understands.” If,

218

Janet Carr

however, it is ineffective in the long run we may need to help the mothers to look for other methods. Besides methods based on learning principles (which have not been evaluated in large-scale longitudinal studies), it may be useful to help mothers see, enjoy, and respond to positive aspects of their children and to generate more positive aspects of the “interpersonal context” of childrearing (Caldwell, 1964), which may have a beneficial influence on later behavior. Of great interest would be the discovery of factors that militated against the ernergence of behavior problems, in the way that Cunningham (1987) suggests that higher parental education “can overcome the boys’ ‘disadvantage’ compared to girls” (p. 174). In their work with nondisabled children Newson and Newson ( 1989) indicate that greater paternal involvement, apart from that connected with discipline, reduced the risk of future behavior disorder; with limited data it was not possible to replicate this finding in the Surrey study.

V.

CONCLUSIONS

Like other groups of children with Down syndrome, those in Surrey had relatively high scores in infancy, the scores decreasing rapidly in early childhood although, in this instance, more slowly after the age of 10 months. As the decline to 1 1 years had been a steady one, the rise at age 21 was the more unexpected. The scores of individuals do not, by any means, follow a uniform downward trend, so an increase might occur by chance in small groups of subjects (Connolly, 1978; Kostrzewski, 1974). A substantial rise has been shown by Berry et al. (1984) in a group of adults, although it is not clear to what extent this was due to the remedial program in which the group was involved. Nevertheless, the downward curve of scores, seen in numerous studies in early childhood, may not inevitably continue. It has long been accepted that people with Down syndrome continue to be able to learn and to develop skills well after leaving school; it may be that, in the course of the learning that occurs in childhood and adolescence, they also acquire techniques, strategies, and insights that can be applied in psychological tests. If this were the case one might expect such learning to be more readily applicable in verbal rather than nonverbal tests; but in the Surrey study the gain was shown on a nonverbal test (the Leiter scale) following one with some verbal content (Merrill-Palmer Scale), and in the study by Berry et al. (1984) the gain on a nonverbal test (RCPM) was substantially greater than that on a verbal test (PPVT-R). Turning from a discussion of the data from the aggregated group to those of individuals within it, one of the most striking features of the study is the wide variation in ability shown by people with Down syndrome. Although recognizing that “IQ” scores do not have the same meaning at the extremities as they do over

LONGITUDINAL RESEARCH IN DOWN SYNDROME

219

the middle part of the range, the distribution of scores, of about 60 points, is of the same order in the Surrey group as for the majority of the more to the less able in a normal population (IQs 70-130). Certainly the difference experienced in a social encounter is at least as great in the former as in the latter case. So a point of particular interest has been to try to discover what determines an individual’s ability level. These efforts have been largely unsuccessful. The association with gender is well established but factors such as parental education and social class, seen as important in other studies, were not found to be so here. Neither has a diligent search among other measured factors produced any answers: although anecdotally the most profoundly disabled could be seen to have suffered from more adverse health factors than the most able, it was not possible to establish that these, or any other, factors had had a significant effect on the scores of the whole group. It is tempting to wonder whether there are differences within the condition itself, as yet unidentified, that may be at least partly responsible for the wide differences, some of which remain quite stable over time, between individuals with an apparently identical condition. Where the development of self-help skills is concerned, overall children with Down syndrome acquire these in the same order as nondisabled children, but more slowly, so that even as young adults nearly two-thirds were not able to look after their own self-care entirely. Most skills were acquired by age 11, after which further acquisition, although still possible, was less often seen. A major finding has been the importance of IQ, the relationship to skill learning being very similar to that relating to academic subjects (Carr, 1988a). Both kinds of skills were more easily acquired by able than by less able children with Down syndrome, although other factors, such as the opportunities and encouragement provided, were seen in individual cases to have been important. Cullen et al. ( 1984) have shown children with physical disabilities to have more cognitive difficulties in the early years; although this is eminently plausible it remains to be seen if these effects are maintained later in life. Although in Surrey the self-help scores of the females were slightly higher, no study has found sex to be a significant factor where these skills are concerned. It may be that the self-help areas under consideration are so basic that there can be no question of their being seen as more appropriate to one or the other sex; all need as much expertise as possible, both to enhance their own feelings of competence and independence and to maximize their capability of functioning in and being accepted by society. It is interesting to speculate how the skills may best be taught to people with Down syndrome. Cullen et al. (1984) point to the importance of the mother’s commitment; the indication in Surrey of the small advantage to the young people who grew up out of their own homes suggests that, if parents are not on hand to help, young people may be better able to develop their competence. Berger and Cunningham (1983) showed that infants’ vocal and

220

Janet Carr

social skills blossomed when their mothers allowed them more time to respond rather than rushing in with help and stimulation. It might be that a similar approach to self-help teaching, with parents and caregivers keeping a lower profile, would pay dividends in this area too. It should, however, be noted that these speculations were not investigated in the Surrey study. One of the most useful functions a longitudinal study like the present one might be expected to serve is a demonstration of the disciplinary methods that are most successful with this population. Many mothers felt uncertain as to how best they should handle their children and uncertain as to whether the methods they used with their other children would be appropriate and, if not, what they should do. Apart from some indication that with these young people, as with others, frequent physical punishment is ineffective and that a harmonious family and positive attitude on the part of the mother may be helpful, almost no confident advice may be proffered. It may, however, be of some comfort to mothers to know that problem behaviors were fewer in this population at age 21 than they had been at age 1 I . Many commented on how much easier their young people were as adults; this was due partly to the gains made in self-help (“he can d o so much more for himself now”) and partly to the decline in some behaviors, for example, being difficult in public, which at age 21 was shown by only three of the profoundly disabled young people. Finally, we may consider whether and how the three areas discussed-IQ, social competence, and behavior-are interrelated. At any one point in time subjects with higher 1Qs were also the more competent socially. Despite the decline in IQ, in this population, at least to age 1 1, social competence increased over time, as did mental age, both reflecting greater social maturity. Regarding behavior, the trend here also was toward improvement with age; this may similarly have been related to better social competence seen in increasing mental age and self-help skills, notwithstanding the tendency, seen also in other studies of normal and disabled populations, for certain personality traits to persist from early to adult life. Since this study began great changes have taken place for people with Down syndrome. Babies whose fanlilies feel unable to look after them are no longer placed in institutions, much progress has been made in the treatment of heart and respiratory conditions, the long-term future for any infant born with Down syndrome today is envisaged as in the community and not, as the Surrey mothers anticipated with horror, in the back ward of a subnormality hospital. All this places limitations on the usefulness of a single cohort study like the present one and the lessons that can be learned from it; circumstances have changed so much that conclusions drawn from the earlier group may not apply to a later one. Nevertheless, it is hoped that the project has something in it of interest to later generations; certainly, the intention is to continue following the present cohort into their thirties and, perhaps, beyond.

LONGITUDINAL RESEARCH IN DOWN SYNDROME

22 1

ACKNOWLEDGMENTS I am grateful to Professors Ann and Alan Clarke, Dr. Beate Hermelin, and Dr. Sheila Hewett for reading and commenting on an earlier draft of this article, to Sean McGuigan for statistical assistance, and to Hermione Holmes for typing repeated redrafts. REFERENCES Aronson. M.. & Fallstrom, K. (1977). Immediate and long-term effects of developmental training in children with Down’s syndrome. Developmental Medicine & Child Neurology. 19. 489-494. Bayley, N. (1955). On the growth of intelligence. American Psychologisr, 10, 12-15. Bayley, N. (1964). Manual of directions for infant scales of development (Restricted Research Forms). Bethesda, MD: National Institute of Mental Health. Bayley, N.. Rhodes, L., & Gooch, B. (1966). A comparison of the development of institutionalized and home-reared mongoloids. California Mental Health Research Digesr. 4 , 104- 105. Bennett, F. C., Sells, C. J., & Brand, C. (1979). Influence on measured intelligence in Down’s syndrome. American Journal of Mental Deficiency. 133. 700-703. Berger, J., & Cunningham, C. (1983). Early social interactions between infants with Down’s syndrome and their parents. Health Visitor, 56, 58-60. Berry, P., Groenweg, G., Gibson, D., & Brown, R. I. (1984). Mental development of adults with Down syndrome. American Journal of Mental Deficiency, 89. 252-256. Berry, P., Gunn, V. P., & Andrews, R. J. (1984). Development of Down’s syndrome children from birth to five years. In J. M. Berg (Ed.), Perspectives & progress in menral rerardation: Vol. 1. Social. psychological & educational aspects (pp. 167- 177). Austin, TX: PRO-Ed. Buckley. S . , & Sacks, B. (1987). The adolescent with Down’s syndrome: Life for the teenager and f;w the fami1.y. Portsmouth Polytechnic, Portsmouth Down’s Syndrome Trust. Byrne, E. A,, Cunningham, C. C., & Sloper, P. (1988). Families and their children wirh Down’s svndrome: One feature in common. London: Routledge. Caldwell, B. M. (1964). The effects of infant care. In M. L. Hoffman & L. W. Hoffman (Eds.), Review ofchilddevelopment research (Vol. I , pp. 261-298). New York: Russell Sage Foundation

Carr, 1. (1975). Young children wirh Down’s syndrome. London: Butterworths. Carr, J. (1988a). Six weeks to twenty-one years old: A longitudinal study of children with Down’s syndrome and their families. Journal of Child Psychology & Psychiatry, 29. 407-43 I , Carr, J. (1988b). Down syndrome. In J. Hogg, J. Sebba, & L. Lambe (Eds.), Profound rerardation and multiple impairment (pp. 40-53). London: Chapman & Hall. Carr, J. (1992). Assets and deficits in the behaviour of people with Down’s syndrome-A longitudinal study. In B. Tizard (Ed.), Human resilience and vulnerability (pp. 55-69). London: Jessica Kingsley Publishers. Carr, J., & Hewett, S . (1982). Children with Down’s syndrome growing up. Association of Child Psychology & Psychiatry News, Spring, 10- 13. Chess, S., & Thomas, A. (1984). Origins and evolutions of behavior disorders: From infancy to adult life. New York: Brunner-Mazel. Clarke, A. D. B., & Clarke, A. M. (1981). “Sleeper effects” in development: Fact or artifact? Developmental Review, 1 . 344-360. Clarke, A. M., & Clarke, A. D. B. (1989). The later cognitive effects of early intervention. lnrelligence. 13, 289-297. Clements, P. R.. Bates, M. V., & Hafer. M. (1976). Variability within Down’s syndrome (trisomy 21): Empirically observed sex differences. Mental Rerardarion, 14. 30-31.

222

Janet Carr

Connolly, J. A. (1978). Intelligence levels of Down’s syndrome children. American Journal of Mental Deficiency. 83. 193- 196. de Coriat, L. F., Theslenco, L.. & Waksman, J. (1967). The effects of psychomotor stimulation on the IQ of young children with trisomy 21. In B. Richards (Ed.), Proceedings of the first congress of the International Association for the Scientific Study of Mental Deficiency (pp. 377-385). Reigate, U.K.: Michael Jackson. Cowie, V. (1970). A study of the early development of rnongols. Oxford: Pergamon Press. Cullen, S. M., Cronk, C. E., Pueschel, S. M., Schnell, R. R., & Reed, R. B. (1984). Social development and feeding milestones. In S. M. Pueschel (Ed.), The young child with Down syndrome (pp.227-252). New York: Human Sciences Press. Cunningham, C. C. (1982). Psychological and educational aspects of handicap. In F. Cockburn & R. Gitzelman (Eds.), Inborn errors of metabolism (pp. 237-253). Lancaster, England: MTP Press Ltd. Cunningham, C. C. (1983). Early development and its facilitation in infants with Down’s syndrome (Final Report to DHSS, Part 1). Cunningham, C. C. (1987). Early intervention in Down’s syndrome. In G. Hosking & G. Murphy (Eds.), Prevention of mental handicap: A world view (RSM Services International Congress & Symposium Series No. 112, pp. 169-182). London: Royal Society of Medicine Services Ltd. Dameron, L. E. (1963). Development of intelligence of infants with mongolism. Child Development, 34, 733-738. Dicks-Mireaux, M. J. (1972). Mental development of infants with Down’s syndrome. American Journal of Mental Deficiency, 77, 26-32. Farran, D. C. (1990). Effects of intervention with disabled and disadvantaged children. In S. J. Meisels & J. P. Shonkoff (Eds.), Handbook of early childhood intervention (pp. 501-539). Cambridge, U.K.: Cambridge University Press. Fenner, M. E., Hewitt, K. E., & Torpy, D. M. (1987). Down’s syndrome: Intellectual and behavioural functioning during adulthood. Journal of Mental Deficiency Research, 31, 241 -249. Fishler, K., Share, J., & Koch, R. (1964). Adaptation of Gesell developmental scales for evaluation of development in children with Down’s syndrome (mongolism). American Journal of Mental Deficiency, 68, 642-646. Fraser, F. C., & Sadovnick, A. D. (1976). Correlation of IQ in subjects with Down syndrome and their parents and sibs. Journal of Mental Deficiency Research, 20, 179-182. Gath, A. (1985). Down’s syndrome in the first nine years. In A. R. Nichol (Ed.), Longitudinal studies in child psychology and psychiatry ( pp. 203-2 19). LondonKhichester: Wiley. Gath, A,, & Gumley, D. (1984). Down’s syndrome and the family: Follow-up of children first seen in infancy. Developmental Medicine and Child Neurology, 27, 392-400. Gibson, D. (1978). Down’s syndrome: The psychology of mongolism. London: Cambridge University Press. Gibson, D., & Harris. A. (1988). Aggregated early intervention effects: Patterning and longevity of benefits. Journal of Mental Deficiency Research, 32, 1-17. Grifiths, R. (1954). The abilities ofbabies. A study in mental measurement. London: University of London Press. Hewett, S. (1970). The family and the handicapped child. London: Allen & Unwin. Holmes, N., & Cam. J. (1991). The pattern of care in families of adults with a mental handicap: A comparison of families of autistic adults and Down’s syndrome adults. Journal of Autism & Developmental Disorders, 2 I 159- 174. Koch, R., Share, J., Webb, A,, & Graliker, B. V. ( I 963). The predictability of Gesell developmental scales in mongolism. Journal of Pediatrics, 62, 93-97. Kostrzewski, J. (1974). The dynamics of intellectual development in individuals with complete and incomplete trisomy of chromosome group G in the karyotype. Polish Psychological Bullerin, 5 . 153- 158.

LONGITUDINAL RESEARCH I N DOWN SYNDROME

223

Leiter, R. G . (1980). Leiter International Performance Scale instruction manual. Chicago: Stoelting co. Loveland, K. A,, & Kelley, M. L. (1988). Development of adaptive behaviour in adolescents and young adults with autism and Down’s syndrome. American Journal on Mental Retardation, 93, 84-92. Ludlow, J. R., & Allen, L. M. (1979). The effect of early intervention and pre-school stimulus on the development of the Down’s syndrome child. Journal of Child Psychology & Psychiatry. 23, 29-44. Melyn, M. A., & White, D. T. (1973). Mental and developmental milestones of noninstitutionalised Down’s syndrome children. Pediatrics 52, 542-545. Merrill-Palmer Scale of Mental Tests (1948). New York: Harcourt Brace & World. Newson, J., & Newson, E. (1963). Infant cure in an urban community. London: Allen & Unwin. Newson, 1.. & Newson, E. (1977). Perspectives on school at seven years old. Oxford: George Allen & Unwin. Newson, J., & Newson, E. (1989). The extent of parental physical punishment in the U K . London: APPROACH (Association for the Protection of All Children). Piper, M. C . , & Pless, I. B. (1980). Early intervention for infants with Down syndrome: A controlled trial. Pediatrics, 65, 463-468. Pueschel, S. (1984). The young child with Down’s syndrome. New York: Human Sciences Press. Reynell, J. (1969). Reynell Developmental Language Scale. Slough, Bucks.: National Foundation for Educational Research. Richardson, S . A,, Koller, H.,& Katz, M. (1985). Relationship of upbringing to later behavior disturbance of mildly mentally retarded young people. American Journal of Mental Deficiency, 90, 1-8. Rutter, M. ( 1976). Helping troubled children. Harmondsworth: Penguin. Rutter, M., Tizard, I., & Whitmore, K. (1970). Education, health and behaviour. London: Longmans. Schnell, R. R. (1984). Rychomotor development. In S. M. Pueschel (Ed.), The young child with Down syndrome (pp. 207-226). New York: Human Sciences Press. Sharav, T., & Shlomo, L. (1986). Stimulation of infants with Down’s syndrome: Long-term effects. Mental Retardation, 24, 8 1-86, Share, J., Webb, A,. & Koch, R. (1961). A preliminary investigation of the early developmental status of mongoloid infants. American Journal of Mental Deficiency, 66, 238-241. Shipe, D., & Shotwell, A. M. (1965). Effect of out-of-home care on mongoloid children: A continuation study. American Journal of Mental Dejiciency, 69, 649-652. Shotwell, A. M., & Shipe, D. (1964). Effect of out-of-home care on the intellectual and social development of mongoloid children. American Journal of Mental Deficiency, 68, 693-699. Stedman, D. J., & Eichom, D. H. (1964). A comparison of the growth and development of institutionalized and home-reared mongoloids during infancy and early childhood. American Journal of Mental Deficiency. 69. 391-401. Turner, S . , Sloper, P., Knussen. C., & Cunningham, C. (1991). Factors relating to self-sufficiency in children with Down’s syndrome. Journal of Mental Deficiency Research, 35. 13-24. Wing, L. (1980). MRC Handicaps, Behaviour & Skills (HBS) Schedule in epidemiological research. Acta Psychiatrica Scandinavia, 62(Suppl. 285), 241-247.