Accelerat ing t he world's research.

Early identification of children's

special needs: a study in five
metropolitan communities
Judith Singer

The Journal of pediatrics

Cite this paper Downloaded from Academia.edu 

Get the citation in MLA, APA, or Chicago styles

Related papers Download a PDF Pack of t he best relat ed papers 

Child healt h care and ?t he new morbidit y?: Toward a model for t he linkage of privat e medical …
Cat hy Crossland

Missed opport unit ies in surveillance and screening syst ems t o det ect development al delay: A develo…
Zarmeneh Aly

Summary of t he Pract ice Paramet ers for t he Assessment and Treat ment of Children, Adolescent s, a…
Michelle Serue
 Early identification of children's special
needs: A study in five metropolitan
communities
Judith S. Palfrey, MD, Judith D. Singer, PhD, D e b o r a h K. Walker, EdD, a n d
John A. Butler, EdD
From the Division of Ambulatory Pediatrics, Department of Medicine, The Children's Hospital,
Boston

In a study of special e d u c a t i o n programs in five urban school systems, parent

interview data for 1726 children r e v e a l e d how e a r l y the children's problems
were identified and how the m e d i c a l system was involved in the diagnosis.
Problems included speech impairment, learning disabilities, emotional distur-
bance, mental retardation, sensory disorders, and physical and health disabil-
ities. Overall, 4.5% of the children's problems were identified at birth, and only
28.7% before the a g e of 5 years. Variation In a g e at identification d e p e n d e d on
the condition: 1 year for Down syndrome and cerebral palsy versus a 6-year
range for mental retardation. Although physicians were most likely to identify
the less common, more severe handicaps, they also identified from 15% to 25%
of learning disabilities, speech impairments, e m o t i o n a l disorders, hyperactivi-
ty, and " o t h e r " d e v e l o p m e n t problems. The type, severity, and c o m p l e x i t y of
the condition were significant predictors of physician identification. No racial,
s o c i o e c o n o m i c , or site biases were associated with whether a physician was
first to identify. A g e at identification was p r e d i c t e d by the c o m p l e x i t y of the
problem, the association with other health and d e v e l o p m e n t a l concerns,
s o c i o e c o n o m i c indicators, and whether a physician was involved in the
diagnosis. In the a b s e n c e of clear assumption of responsibility for early
identification, much terrain remains uncharted by m e d i c a l practitioners and
the schools. A better systematic sharing of responsibility for the early identifi-
cation of d e v e l o p m e n t a l l y disabling conditions is n e e d e d . (J PEDIATR
1987;111:651-9)

Few child development policies have enjoyed as much
general acceptance as early identification of disabilities.1-4
In tandem with early intervention, early identification has
been championed as a cost-effective and humane strategy
to improve children's life chances and reduce secondary
disability,s-9 Early identification has been considered such
an important indicator of the effectiveness of pediatric
health care that it has been said, "The proportion of
children with major handicaps detected under the age of

See related article, p. 722

Presented in part at the Annual Meeting of the Ambulatory
Pediatrics Association, Anaheim, California, April 1987.
Supported by The Robert Wood Johnson Foundation and the
Commonwealth Fund.
Submitted for publication March 24, 1987; accepted June 10,
1987.
Reprints requests: Judith S. Palfrey, MD, The Children's Hospi-
tal, 300 Longwood Ave., Boston, MA 02115.
12 months and of other defects detected before the child is
of school age provide good criteria of the quality of
pediatric service. ''z~
An emphasis on early identification was incorporated in
the Education for All Handicapped Children Act (P.L.
94-142), which mandates that communities should identify

651
652 Palfrey et al.
T a b l e I. Characteristics of study sites, 1982-1983 school year
Charlotte, Houston,
N.C. Texas
Elementary school population 38,003 116,040
White (%) 56.9 18.9
Black (%) 40.6 42.9
Hispanic (%) 0.0 35.0
Other (%) 2.0 3.4
Children in special education (%) 7.6 7.9
Age at which special education 5 0-3*
services begin (yr)
*For hearing-or vision-impairedand deaf-blindchildren.
all children with handicaps 11'x2 as early as possible via
"Child Find" programs. 13,~4 To this end, cooperative
arrangements have been developed between public health
agencies and public schools. Some states also have
mounted public awareness campaigns and programs
through local medical societies. The recent passage of P.L.
99-457 underscores the initiative for early identification
and early intervention.15
As part of a study of the health care and educational
services for children in special education, we collected
information on the relative contributions of medical and
educational professionals to early identification. The fol-
lowing major questions were asked: (1) Which children
were identified by the medical system, which by the
educational system? (2) Were there differences in the age
at identification or in the system that identified a child on
the basis of the diagnosis, the severity of the disorder, or
the family's socioeconomic status? (3) Did the site in
which a child lived affect the chance that the child's
disorder would be identified early?
METHODS
Sample selection. The research was carried out in five
sites, one from each of the five United States census
regions. School systems were considered as potential study
sites if they had enrollments of at least 25,000 students and
were located in states with special education enrollment of
at least 50,000. To ensure variation in child backgrounds
within and among sites, information was collected from
candidate school systems on the racial, ethnic, and socio-
economic composition of the special education popula-
tion.
The five sites selected for study were Charlotte, North
Carolina; Houston, Texas; Milwaukee, Wisconsin; Ro-
chester, New York; and Santa Clara County, California.
Typical of many American urban localities, the five sites
were racially and ethnically heterogeneous, with nonwhite
students ranging from 40% of the population in Santa
Clara County to 81% in Houston (Table I). Family income
The Journal of Pediatrics
November 1987
Milwaukee, Rochester, Santa Clara County,
Wis. N.Y. Calif.
38,407 18,846 65,557
38.1 36.9 58.1
50.5 49.0 4.3
7.5 11.3 26.6
3.8 2.5 11.6
10.6 13.4 9.4
3-5 5 3-5
also varied widely, both across and within sites, as did
levels of maternal education.
Within each site, a stratified random sample was sclcct-
ed from among all special education students in kindergar-
ten through sixth grade. Students in each site were
designated as having a special need only after a full
multidisciplinary evaluation carried out under the stipula-
tions of P.L. 94-142 using the criteria of the specific state.
The sample was constructed to overrepresent children with
less prevalent conditions while permitting generalizations
applicable to the elemcntary school special education
populations as a whole. 16 The consent rate was 72.4%.
Variable measurement and definition. A 40-minute tele-
phone interview was carried out with the child's parent or
guardian by trained interviewers from the University of
Illinois Survey Research Laboratory. The interview was
administered in English or Spanish, over 3 months in the
spring of 1983.
The child's primary handicap was defined as "the name
of the child's major problem or condition," as reported by
the parent. The child's age at diagnosis was derived from
the question "How old was the child when the diagnosis of
this condition or problem was first made?" Physician
identification was determined from the parent's response
to the question "Who made this diagnosis?" Interviewer
probes helped determine whether the person was a physi-
cian, a teacher, or someone else.
In addition to questions about the child's major problem,
parents were also asked about "trouble speaking," "trouble
hearing," "trouble seeing," "hyperactivity," "general med-
ical problems," "emotional problems," "learning prob-
lems," and "mental retardation." The total number of
limiting impairments was also tallied for each child.
Mother's education was defined as "the highest grade or
year" of schooling that she had completed. The poverty
income ratio indicated a family's financial standing rela-
tive to the poverty line after adjusting for family size.
Child race and ethnicity were obtained from school district
records.
Volume 111 Early identification of special needs 653
Number 5

I- 1 -' I ~ .... I--- ' I ~T----

Down Syndrome H
Cerebral Palsy t-E3--I
Other Neurological I I ~1- . . . . . . . . . . . . . t
General Medical

Vision t1: I I ............ t

Hearing I--11 I .....-I
Mental Retardqtion I~I I I......I
I

Speech ~-..... | I i- ....... t

Hyperactivity I. . . . . . ' ! ! F ........ -I
Emotional Problems t---I ! |-, ..... -t
Other Developmental I---] / ~ .......-I
Learning Disability t--t 'l I- ....... t
t I I ] 1 i t i t i I
0 2 4 6 8 lO
Age at FirstD/agnosis
(years}

Leq~,,e-F ....... t I t ....... t

I0 th 25 th :50 th 75 th 9 0 th percentile

Fig. 4. Two clusters of disabilities identified using Bonferroni multiple comparisons of 12 primary handicaps.

Statistical analysis.Estimates presented in this paper
incorporate a weighting procedure that compensates staffs-
tically for the oversampling of low prevalence disability
grouPs. Within each site, weights were computed to
generalize results to the entire elementary school special
education population of that site. Weights were then
cal!brated to the total number of respondents in that site so
that estimates across sch0ol systems reflect anapproxi-
mate average of individual site results.
Multiple logistic-regression analysis was used to study
differences between childre n identified by physicians and
those identified by nonphysicians. Tests of differences were
conducte d uncontrolled and then adjusted for the child's
primary handicap and number of associated limiting prob-
lems. Multiple regression analysis was used to study
factors associated with age at first identification.
The total sample size of 1726 is large enough to provide
ample statistical power (>0.80) to detect even small
effects. Within subgroups, power does diminish but still
remains high enough to safely conclude that if no relation-
ship is found it is likely not to exist in the population.
Limitations on inference. This study may be viewed as
five case studies, with identical methods applied at each
site and pooling of data across sites when appropriate.
Findings may be generalized to the entire elementary
school special education population of the five school
systems, bu t not necessarily to the entire same age disable d
population Of the United States.
The children in the study were students enrolled in
special education programs in elementary school. There-
fore, ch!ldren identified early whose problems were amelio-
rated by treatment, children who acqu!red problems after
sixth grade, and children with certain medical problems
whQ did not require special educational services were not
represented in the sample.
That the source of the data was parent report may raise
some concern about the precision of the measures (who
made the diagnosis and at what age). However, this
informaton is so salient to families that parent recall is
likely to be accurate. By phrasing the question "Who
diagnosed your child's problem?," it is recogn!zed that a
health or educational provider may have suspected the
condition and discussed it with the family earlier than the
parent report. Ttie data reflect the entire process of
654 Palfrey et al.
Table II. Relative contribution of physicians and nonphysicians to identification of handicapping conditions
Children identified
by MDs (%)
Down syndrome 98
Cerebral palsy 99
Other neurologic disorder 91
General medical 90
Vision 87
Hearing 64
Mental retardation 76
Speech 19
Hyperactivity 44
Emotional problems 23
Other developmental disorder 18
Learning disability 12
recognition and understanding of the condition by the
family. In practical terms, however, such a definition of
"identification" seems reasonable, because it is not until a
family has been fully apprised of a child's condition that
appropriate services can be provided to the child.
RESULTS
Differenees in age at identification by diagnosis. For the
conditions under study, the range for age at identification
was very wide. Four percent of the children were identified
at birth, 16.4% before age 3 years, and 28.7% before age 5
years. The period 5 to 7 years was the most active for
identification of the Problems, with 47.9% diagnosed
during the early elementarY school years. The special
needs of the remaining 23.3% of the children were detected
at age 8 years or older, with declining numbers at the later
ages. The major explanation for this variation was the
widely divergent nature of the primary handicaps included
as special needs disabilities (31.3% speech, 44.3% learning
disabilities, 8.0% emotional disturbance, 11.6% mental
retardation, 2.2% sensory impairments, and 2.6% physical
disabilities). In fact, the child's primary handicap
explained 32% of the variance in age at first identifica-
tion.
Two distinct clusters of disabilities were identified using
post hoe Bonferroni multiple comparisons of the 12 prima:
ry handicap groups (Fig. 1). The first cluster of disabilities
was composed of the "low prevalence" handicaps: Down
syndrome, cerebral palsy, other neurologic problems, gen-
eral medical conditions, vision and hearing impairments,
and mental retardation. The second cluster of disabilities
was made UP of the "hig h prevalence" disabilities: speech
impairment, hyperactivity, emotional problems, learning
disabilities, and "other" developmental problems (such as
The Journal of Pediatrics
November 1987
Mean a g e at Mean age at
identification by MD identification by non-MD
(mo) (too)
0.6 0.0
10.3 6.0
30.2 39.7
27.3 60.8
54.9 102.0
38.8 55.0
33.7 67.8
38.0 65.8
59.1 68.4
61.3 79.9
69.3 84.2
69.1 84.2
developmental delay, slow learner). Eighty-five percent of
the children in the low-prevalence group were identified
before age 5 years, whereas only 21% in the high-
prevalence group were identified before age 5 years.
Moreover, it was rare for children in the high-prevalence
group to be identified before age 3 years.
Although the two clusters did form statistically cohesive
units, within the clusters themselves there was also consid-
erable variation between and within specific disability
groups. For example, Down syndrome and cerebral palsy
usually were diagnosed during the first year of life. By
contrast, the age at first identification of mental retarda-
tion showed wide variation spanning 7 years, indicating the
heterogeneity of this designatio n . The high-prevalence
cluster also displayed substantial variability by condition.
Relative contribution of physicians and nonphysiclans to
the diagnostic process. The low-prevalence disabilities
were far more likely to be diagnosed by a physician than by
someone else, although nonphysicians did identify 25% of
the mentally retarded and 36% of the hearing impaired
children (Tabl e I I ) . By contrast, the high-prevalence
disabilities were more likely to be identified by someone
other than a physician, most commonly a teacher or other
school professional. Physicians diagnosed 0nly 44% of the
hyperactive children and less than 25% of speech impair-
ments, emotional problems, learning disabilities, and "oth-
er" developmental problems.
Further analysis indicated that the children with high -
prevalence handicaps most likely to be identified by a
physician had multiple problems (P <0.001), associated
difficulties such as hyperactivity (P <0.001), emotional
problems (P <0.001), mental retardation (P <0.001),
trouble speaking (P <0.001), trouble hearing (P <0.001),
and other medical problems (P <0.05). Racial background
Volume 111 Early identi[ication o f special needs 655
Number 5

+F - - - - non MDID
years MDID
7-
801-

6-
70t-

60t-

50F

40F
3-

2-
30# ~ Ilow
20F

1 I i I i I J l J
~0
8 12 16
Primory HS Col lege
Mother's Educat/bn (years)
Fig. 2. Relationship between disability, mother's education, physician identification (MDID), and age at identification.

was also related to physic!an identification (P <0.05), but
gender, birth order, and mother's education were all
unrelated to the likelihood of physician identification.
After statistically controlling for the child's primary
handica p (e.g., speech or learning d!sability, emotional
disturbance, hyperactivity), the relationships between
child characteristics and physician identification 9 to
persist, with the except!on of th e racial differential. How-
ever, when these relationships were examined using the
number of associated (limiting) problems as a further
control variable, all of the relationships diminished to
nonsignificance, eXcept the presence of "assOciated mental
retardation." In other wordS, the number of limiting
conditions in a given child made a substantial contribution
to the prediction of physician identification of a ct~ild's
disability.
Predictors of early age at identification. Correlations
were estimated between the age of ~dentification and three
sets of potential predictors: measures Of the severity or
complexity of the child's hand!cap; measures of race and
Socioeconomic status; and whether a physician diagnosed
the problem. In general, the more complex the problem the
earlier the diagnosis was likely to be made. For example,
age at identification was related to the number of function-
ally limiting handicaps (P <0.00 t); the presence of associ-
ated mental retardation (P <0.001), medical problems (P
<0,001), or speech problems (P <0.001); and to the degree
of activity limitation resulting from the child's primary
handicap (P <0.001).
By contrast to the situation with physician identifica-
tion, mother's education was found to be a significant
predictor of age at identification. For all disabilities, the
better educated the mother the earlier the child was
identified (0.7 months earlier for each additional year of
educatign ). Among children with low-prevalence handi-
caps, being white and flaying a higher income also were
associated with earlY identificatign (P <0.001).
Finally, physician identification and early identification
were strongly associated (P <0.001). As shown in Table I1,
for every handicap, diagnosis by a physician was made
earlier than identification by others (average difference, 2
years; P <0.Q01). The difference was most pronounced
among children with mental retardation (33.6 months) and
speech impairments (27.6 months), and least pronounced
656 Palfrey et al.
Table III. Children identified by physician, and mean age at first identification, by study site
Children identified by
a physician (%)
Children in
special education High
(%) prevalence
Charlotte, N.C. 7.6 18.0
Houston, Texas 7.9 13.0
Milwaukee, Wis. 10.6 23.2
Rochester, N.Y. 13.4 19.9
Santa Clara County, Calif 9.4 15.5
Significance test -- X2(4)= 6.95
(NS)
among children with "other" developmenta I problems
(13.2 months) and hyperactivity (9.6 months).
Fig. 2 simultaneously displays the relationship between
the three predictors---disability, mother's education, physi-
cian identification--and age at identification. Regardless
of all other factors, low-prevalence handicaps were diag-
nosed earlier than high-prevalence handicaps. Within dis-
ability groups, those identified by a physician were diag-
nosed earlier on average than those id.entified by a nonphy-
sician. Children with better educated mothers were usually
diagnosed cartier than those with poorly educated mothers.
Moreover, the effect of mother's education was particular-
ly pronounced among one subgrou p of Children: those with
low-prevalence handicaps identified bY a nonphysician.
Within this group, a child whose mother had completed
college was diagnosed nearly 2 years earlier on average
than a comparable child whose mother had only completed
eighth grade.
Effect of geographic site on early identification. Did
physicians in Certain communities play a more active role
in the identification of children with special needs than in
others? Did certain communities do a better job of
identifying these children early? Across the five communi-
ties studied, no statistically significant differences were
found in the percentage of children identified by a physi-
cian (Table III). Mean age at first identification did differ
significantly across the five sites, but after controlling
statistically for descriptors of the child's primary handicap,
these differences became nonsignificant. Thus, where a
child lived did not appear to be a major factor influencing
the timing Of identification, although there may be a
complex interaction between site, classification practices,
and access to physicians that these analyses have not
comp!etely elucidated.
DISCUSSION
Early identification involves three components: the nat-
ural history of childhood disorders , the techniques of
The Journal of Pediatrics
November 1987
Mean age at first
identification (mo)
Low High Low
prevalence prevalence prevalence
76.1 77.5 30.0
83.8 75.9 27.7
90.2 70.0 24.4
82.4 74.'I 30.8
79.8 67.8 19.8
X2(4)= 5.70 F(4,909)= 3.21 F(4,775)= 3.91
(NS) (0.125) (0.0038)
screening and diagnosis, and the delivery systems available
for applying those techniques. This study provides infor-
mation about each of these components in isolation and in
relationship to one another.
Childhood handicaps. First, the study reassures us that
many handicapping conditions (e.g., claromosomal anoma-
lies, cerebral palsy) are being diagnosed very early. The
more readily discernible the phen0typic features, the
earlier the diagnosis tends to be made. 17 Second, the data
bring into question who are the 20% of children with
high-prevalence disorders severe or distinct enough that
they are brought to the attention of the medical profession-
al. Although the concept of a continuum from hard
neurologic disorders to "softer" forms of developmental
delay has fallen out of favor in recent years, TM our data call
for a reexamination of such a possibility. Developmentally
delayed children identified at an early point by physicians
bear careful consideration. Certainly, the proxy severity
measures indicate pervasive disability. The lack of socio-
economic bias on physician identification also argues for a
relatively pure form of disability. If this is so, we should
continue to study these children (at least the earliest
identified with the severest handicaps) to determine wheth-
er there remain uncategorized genetic, metabolic, or neu-
rologic correlates of developmental disability with etiologic
or treatment implications.
In this study, an important predictor of early identifica-
tion was "multiple problems." In other investigations, too,
children with multiPle developmental or functional prob-
lems have been shown to be identified earlier and to have
more persisten t difficulties than those with one prob-
!em. 19.2~Can understanding the synergism of deveiopmen-
tal disabilities help in our management of these troubled
children and families? Certainly, when multiple problem s
are detected in a given child, the call for targeted services
should be heeded quickly.
For children with only one problem, it may be that the
dysfunction is time limited. There may be developmental
Volume 111
Number 5
"moments" when the particular functional limitation is
most salient. For example, it could be that visual percep-
tual weakness stands out only at the time letters are first
being introduced. Moreover, a child with fine motor
difficulties may be most at risk later in the school career,
when the extent and speed of written output are impor-
tant.
Seven years of age has traditionally been the point for
clinical identification of learning disabilities. We also
found the median age for diagnosis of learning disabilities
to be age 7 years. It is possible that this disability does
"emerge" at that time, but it is at least equally possible
that there are compensating strategies that protect the
child who has an isolated and specific learning problem
until that time when academic demands increase. Are the
signals there earlier, but masked by the child's other
strengths or not diagnosed because the weaknesses have
not yet resulted in actual failure?
Technology of screening and diagnosis. In this study, the
technology to diagnose Down syndrome, severe hearing
loss, and cerebral palsy appeared to be available and
successfully applied from birth on. By contrast, the low
rate of early diagnosis for speech impairments, hyperactiv-
ity, learning disorders, other developmental problems, and
emotional disorders was striking. The findings suggest a
relative inability to make these diagnoses before school
age, and virtual impossibility of making the diagnosis
before 2 years of age. The natural history of the disorders
may preclude very early diagnosis, but other explanations
include the way many of the disorders are defined, the
social realities surrounding the diagnostic process, and the
current imprecision of measurement techniques.
The definition of many of the late identified disabilities
depends on the chronicity of the problems and their social
acceptability. In the case of speech and language problems,
for example, there may be early signals in the toddler
periods, such as the lack of social turn taking, the absence
of "communication awareness," and the inability to
respond to simple commands, but most clinicians delay a
diagnosis of speech/language impairment until there has
been an obvious failure in meeting expressive milestones of
sentence utterance or until tasks at school make the
problem salient. Moreover, with a condition such as
"hyperactivity," clinicians are rarely willing to label high
activity levels as abnormal until there has been a history of
1 or 2 years of such behavior beyond the normally active
toddler period. Learning disability presents a particularly
difficult definitional dilemma, because in many states,
learning disability is defined as "2 years behind grade in
reading, math, or other subject areas., Using this defini-
tion, it is simply not possible to diagnose learning disability
before 7 years of age.
Early identification o f special needs 657
Beyond definition, how good are the techniques for
detection? With limited validation of assessment tech-
niques, physicians and psychologists often worry about
overidentification and the ethical implications of labeling a
child incorrectly.21 Well-standardized measures are avail-
able to clinicians for some disorders such as language
problems and cognitive impairment, but for other disor-
ders, including hyperactivity, early learning disabilities,
and emotional problems, there is a need for more precise
measurement of attentional problems, precursors of learn-
ing disorders, and behavioral flare-ups.
Perhaps the most distressing finding of our study was
the wide range of ages at which "mental retardation" was
first diagnosed. For the most part, mental retardation
should be readily definable and measurabl e , because tests
of IQ for the preschool period are available, reliable, and
valid. A recent report from a referral center showed that
virtually all the mentally retarded patients seen in that
context were identified by age 4 years. 22 The problem,
therefore, may not be the measures themselves, but the
system for their application and interpretation. The diag-
nosis of mental retardation is viewed by most physicians as
so stigmatizing that many are reluctant to confirm it in any
but the most extreme cases. Schools, also reluctant to label,
may nonetheless be somewhat more inclined to because the
classification may be linked to specific educational services
and school placements.
This study points to the need to look very closely at the
group of children with late identified mental retardation.
Are these children properly identified? Have they received
services previously? Have they "acquired" the retardation
status? Are there environmental, social, or medical factors
that may have contributed to the current condition? Do
racial biases account for either the designation "mentally
retarded" or for the lateness of the recognition that the
child indeed has a special need?
Delivery system. If it is true that most major handicaps
(e.g., severe mental retardation, physical defects, and
sensory impairments) can be detected by age 1 year, then
the data from the study provide a somewhat negative
comment on early detection systems, with 24% of the
low-prevalence, high-severity disorders remaining undiag-
nosed after age 3 years. Moreover, few developmental
disabilities were picked up before school entry, raising
serious concern about community capability to accomplish
early detection among toddlers and preschoolers.
A second system problem highlighted by the study is the
socioeconomic bias, Which was evident in the early identi-
fication of both low- and high-prevalence disorders. In
particular, children whose mothers had high educational
attainment were more likely to be diagnosed early than
were children whose mothers had less schooling. Although
658 Palfrey et al.
this may constitute a problem in communication and
assimilation of information about the child rather than
actual delay in recognition by the providers, there is a clear
gap in service that requires attention.
The pattern of identification we found reflects, in part,
that opportunities for identification of childhood handicaps
are based primarily in the health system from 0 to 21/2
years, in preschool from 21/~ to 5 years, and in the school
system from 5 years on. For highly educated families with
access to the full range of health and educational services,
there appear to be three opportunities for identification:
health service identification, preschool identification, and
school identification. For many families of low socioeco-
nomic status, there may well be only two opportunities for
identification: health service identification in the early
years and school identification.
This study suggests the need for an assertive approach
by pediatrics and early childhood education to the identi-
fication of developmental disability in the 3- to 5-year
period. Both fields pay lip service to "early identification,"
but there are few protocols or practice standards for this
activity. In the wake of the passage of P.L. 99-457 (The
Amendments to the Education for All Handicapped Chil-
dren Act), a first step to successful early childhood
programming will be the development of effective early
identification policies and procedures.
As a beginning toward this end, pediatricians should, at
the very least, devote at least a half-hour visit to each child
in the 3- to 5-year age range. A t that visit, vision and
hearing testing should be provided as well as a historical
review to elicit any parental concerns regarding the child's
development in language, visual perception, visual motor
integration, fine and gross motor ability, attention, activi-
ty, and memory. In addition, questions should be asked
regarding the child's social interaction with adults and
children. Sleep, appetite, toileting, and discipline problems
should also be asked about. Such an approach will lengthen
the standard health visit for this age range, but the yield in
terms of early identification will be beneficial.
The second step necessary to ensure service provision is
coordination between primary care physicians, other child
development specialists, and the schools. When problems
are detected, physicians should have a readily available
referral source for further elucidation of the problem and a
clear working relationship with the school system so that
services can be provided promptly.
This study presents major challenges to the health care
system to increase the proportion of children identified
before 1 year of age; to decrease the variability in the age
at identification for most of the conditions; to examine
carefully the designation mental retardation; to try to
develop measures for the earlier detecton of speech,
The Journal of Pediatrics
November 1987
hyperactivity, learning disabilities, and emotional prob-
lems; to pursue any possible leads regarding the cause of
physician-identified disabilities; and to erase the socioeco-
nomic inequities in the early identification of developmen-
tal disabilities. There is a long way to go before ideal
pediatric standards for early detection are attained.
We thank Drs. William Frankenburg, Harvey Levy, David
Nathan, and Jack Shonkoff for their careful review of an early
draft of the paper; and Dr. Julius Richmond for guidance.
REFERENCES
1. Meier JH. Screening, assessment, and intervention for young
children at developmental risk. In: Tjossem TD, ed. Interven-
tion strategies for high risk infants and young children.
Baltimore: University Park Press, 1976.
2. Strangler SR, Huber C J, Routh DK. Screening growth and
development of preschool children. New York: McGraw-Hill,
1980.
3. Frankenburg WK. Increasing the lead time for the preschool
aged handicapped child. In: Not all little wagons are red: the
exceptional child's early years. Reston, Va.: Council for
Exceptional Children, 1973.
4. Frankenburg WK, Erode RN, Sullivan JW. Early identifica-
tion of children at risk. New York: Plenum Press, 1985.
5. Hayden AH, McGuiness GD. Bases for early intervention. In:
Sontag E, ed. Educational programming for the severely and
profoundly handicapped. Reston, Va.: Council for Exception-
al Children, 1977.
6. Lazar I, Darlington R, eds. Lasting effects of early education:
a report from the Consortium for Longitudinal Studies.
Monogr Soc Res Child Dev 1982; 47(2-3 serial no. 195).
7. Simeonsson R, Cooper D, Scheiner A. A review and analysis
of the effectiveness of early intervention programs. Pediatrics
1982;69:635-41.
8. Clunies-Ross GC. Accelerating the development of Down's
syndrome infants and young children. J Spec Educ
1979;13:171-7.
9. Denhoff E. Current status of infant stimulation programs for
children with developmental disabilities. Pediatrics 1981;
67:32-7.
10. Egan DF, Illingworth RS, MaeKeith DC. Developmental
screening 0-5 years. Clin Dev Med 1969;40.
11. The Education for All Handicapped Children Act of 1975,
P.L. 94-142; 20 S.C. 1401 et seq: Fed Reg 42(163):42474-
42518; August 23, 1977.
12. Palfrey JS, Mervis RC, Butler JA. New directions in the
education and evaluation of handicapped children. N Engl J
Med 1978;298:819-24.
13. Martin R. Educating handicapped children: the legal man-
date. Champaign, Ill.: Research Press, 1979.
14. Jacobs FH, Walker DK. Pediatricians and The Education for
All Handicapped Children Act of 1975. Pediatrics 1978;
61(1):135-7.
15. The Education of the Handicapped Act Amendments of 1986,
P.L. 99-457.
16. Palfrey JS, Singer JD, Walker DK, Butler JA. Health and
special education: a study of new developments for handi-
capped children in five metropolitan communities. Publ
Health Rep 1986;101:379-88.
Volume 111 Early identification o f special needs 659
Number 5

17. Smith B, Phillips CJ. Identification of severe mental handi-
cap. Child Care Health Dev 1978;4:195-203.
18. LevineMD, Brooks R, Shonkoff JP. A pediatric approach to
learning disorders. New York: John Wiley, 1980.
19. Palfrey JS, Levine MD, Pierson DA. Antecedents of middle
childhood. In: Levine MD, Satz P, eds. Middle childhood.
Baltimore: University Park Press, 1984.
20. Fowler MG, Cross AW. Preschool risk factors as predictors of
early school performance. J Dev Behav Pediatr 1986;7:237-
41.
21. Keough BK, Becker LD. Early detection of learning prob-
lems: questions, cautions and guidelines. Except Child
1977;Sept:5-10.
22. Lock TM, Shapiro BK, Ross A, Capute AJ. Age of presenta-
tion in developmental disability. J Dev Behav Pediatr
1986;7:340~5.

BOUND VOLUMES AVAILABLE TO SUBSCRIBERS

Bound volumes of the 1987 issues of THE JOURNALOF PEDIATRICSare available to subscribers
(only) from the Publisher, at a cost of $44.00 ($60.00 international) for Vol. 110 (January-June)
and Vol. 111 (July-December), shipping charges included. Each bound volume contains subject
and author indexes, and all advertising is removed. Copies are shipped within 60 days after
publication of the last issue in the volume. The binding is durable buckram, with the Journal
name, volume number, and year stamped in gold on the spine. Payment must accompany all
orders. Contact The C. V. Mosby Company, Circulation Department, 11830 Westline Industrial
Dr., St. Louis, MO 63146, USA/800-325-4177, ext. 351.
Subscriptions must be in force to qualify. Bound volumes are not available in place of a regular
Journal subscription.