Can Clinical
Judgment
Detect
Children
With
Speech-Language
Problems?
Frances P. Glascoe, PhD
From the Child Development Center, Vanderbilt University, Nashville, Tennessee
ABSTRACT. Pediatricians often rely on clinical judgment
derived from observation or parental concern to identify
children with developmental problems. The less popular
but recommended alternative is to repeatedly administer
standardized screening tests. Such tests are time
consum-ing but, unlike clinical judgment, have known detection
rates. Preliminary research concerning clinical judgment
showed that clusters of parental concerns related to their
childrens’ performances on screening tests. In the present study, previous research was refined by assessment of the
meaning of parents’ concerns about their childrens’
speech-language development. In this study of 157 fami-lies seeking pediatric care, 72% of children whose speech-language screening yielded positive results had parents
who were concerned about their speech-language devel-opment. Of children with negative screening results, 83% had parents with no concerns about their speech-lan-guage development. Although standardized screening
tests should be used occasionally in the developmental
surveillance process, the findings show that the problems
of most children with developmental problems were de-tected through clinical judgment based on parental con-cern. Pediatrics 1991;87:317-322; clinical judgment, de-velopmentalproblem, speech-language development.
Speech-language impairments are among the
most common handicapping conditions of
child-hood. Fully 3% of the children in this country receive public school special education services for such disabilities.’ Because special education
statis-tics identify as speech or language impaired only
those children without mental retardation, learning disabilities, or other handicaps, 3% is a conservative estimate. In one careful epidemiological study, it was found that 11% of kindergarteners had speech-language impairments, although studies showing rates of as much as 20% are not uncommon.2’3
Received for publication Oct 23, 1989; accepted Mar 15, 1990. Reprint requests to (F.P.G.) Child Development Center, van-derbilt University, 2100 Pierce Aye, Nashville, TN 37232. PEDIATRICS (ISSN 0031 4005). Copyright © 1991 by the
American Academy of Pediatrics.
It is reasonable to ask whether a figure of 11% includes delayed but otherwise normal children who would overcome their deficits whether treated or not. Current definitions of speech-language impair-ments suggest otherwise. According to the Educa-tion for All Handicapped Children Act (PL 94- 142), speech-language impairment is defined as a disor-der, deviation, or delay in verbal, gestural, or vocal skills including articulation, fluency, voice quality,
or
language to the extent that academic learning,social adjustment, or communication skills are hindered. Although varying slightly from state to
state, numerical definitions of speech-language
im-pairments include performance significantly dis-crepant from ability (1 or more standard deviations
less than measured intelligence and at or less than
the 9th percentile (standard score 8O). Clearly, such a definition includes children whose
speech-language development is markedly deficient.
If left untreated, speech-language disabilities are strong predictors of school failure. Children with such disabilities are likely to undergo retention and/or long-term participation in special education classrooms.57 Yet, many types of speech-language impairments are correctable within 2 years. Hence, intervention before children reach first grade may prevent secondary handicapping conditions, partic-ularly learning disabilities in the areas of spoken language, listening, or reading comprehen-sion.5’6’’#{176} This, in turn, reduces the need for
lengthier and more expensive special education
services and prevents loss of human potential. These considerable savings are predicated on early detection and remediation of speech-language im-pairments.
devel-opmental handicaps as well. Mental retardation,
autism, and some types of learning disabilities all
share speech-language deficits as a core feature.2
As a consequence, the majority of children with
handicapping conditions may be expected to be detected through sensitive speech-language screen-ing.
Although pediatricians are exhorted to screen
regularly and to use standardized instruments, few
actually accomplish this widely supported goal.’3”4
Rather, most pediatricians rely on clinical
judg-ment: observing children and listening to parents’
concerns.’3 Only when a problem is suggested by
these informal methods are screening instruments
administered. Thus clinical judgment is used as a prescreening test, reducing the numbers of children given lengthier screening instruments. Unlike standardized tests that have known rates of
detec-tion, the validity of clinical judgment is unknown.
How sensitive is it in detecting children with disa-bilities? How specific is it in identifying nonhand-icapped children as such? How many children are overreferred or underreferred?
In this article, one aspect of clinical judgment,
listening to parents’ concerns, is evaluated as a
prescreening method. In previous research’5 a
re-lationship was shown between clusters of parents’ concerns and childrens’ pass/fail scores on global
tests of development. In the present study, we
de-fined this relationship more carefully by addressing the following questions: (1) Do parents’ concerns about speech-language development reflect meas-urable deficits in speech-language skills? (2) In what conditions do parents’ concerns predict their
childrens’ performances on speech-language
meas-ures? For example, are experienced or highly
edu-cated parents better able to appraise their child’s speech-language development? and (3) Can clinical judgment based on parents’ concerns function as a prescreening measure, reducing accurately the
numbers of children targeted for further screening?
METHODS
Subjects and Sites
Subjects were 157 parent-child dyads seeking
pe-diatric care in: (1) the outpatient clinics of two
urban teaching hospitals in Nashville, TN (27%);
or (2) one of three private pediatric practices, two
in the city limits (54%) and one in an outlying
county (19%). Subject selection was made according
to the availability of families. Parents of children who were between 6 months and 77 months of age
and who were not acutely ill were asked to partici-pate. Response was as follows: 6 parents declined,
2 were rejected because of language barriers, and
157 gave informed consent.
The parents were predominately from urban or suburban areas (61%), mothers (81%), married (71%), had a 12th grade or greater education (82%),
and a mean age of 29.2 years (range = 12 to 62
years ofage, standard deviation = 6.7). The children
were typically white (78%), male (54%), firstborn
(58%), siblings (67%), participated in day care or
school programs (67%), and averaged 34.7 months
of age (standard deviation = 20.3).
Measurement
Parents were administered the Parents’
Evalua-tions of Developmental Status (PEDS), an
experi-mental interview designed to elicit parents’ con-cerns about their child’s learning and develop-ment.’6 Parents are asked to respond to the
statement, “Please tell me any concerns you have
about your child’s learning and development.” A
second item probes whether parents have concerns
in each developmental area (eg, “Do you have con-cerns about the way he/she talks? moves? behaves? takes care of himself? . . . etc). Parents’ responses
are grouped into 1 1 categories: no concerns, medical concerns, and each of the various developmental
domains (eg, self-help, gross and fine motor skills,
articulation, expressive language skills, etc). Other
items elicit reasons for parents’ concerns or lack of
concerns and the types of information about child development parents use. The PEDS was validated with 200 parents seeking pediatric care and has appropriate levels of interrater and test-retest reli-ability (88%, 88% agreement).’5’7 Only data related
to parents’ concerns about their child’s articulation
and expressive language skills were used in the present study.
Children’s development was screened with one or
more instruments. All received the Battelle Devel-opmental Inventory Screen (BDIST),’8 an instru-ment standardized with a nationally stratified
sam-ple of 800 children. Although BDIST items can be
scored by parent report, observation, or direct elic-itation, with the exception of a few items
adminis-tered to children younger than 18 months of age,
only the direct measurement approach was used.
Items on the BDIST are drawn from the Battelle
Developmental Inventory, which, in turn, is highly correlated with a range of diagnostic tests including
the Bayley Scales,’9 the Peabody Picture
Vocabu-lary Test, the Wechsler Intelligence Scale for
Chil-dren-Revised, the Stanford-Binet Test, and Vine-land Social Maturity Scale.’8 The BDISTs seven
subtests provide age-equivalent and pass/fail scores
enable pass/fail scores to be adjusted to 1, 1#{189},or 2 standard deviations below the mean. In the present
study, we used only the expressive language subtest
with scores determined by the most stringent
cu-toffs (2 standard deviations less than the mean).
Children 30 months of age and older were also given the Articulation Screening Test,2#{176}a validated but not standardized measure of speech sound produc-tion. A single pass/fail score is produced for each of three age groups: 30 to 41 months, 42 to 71 months, and 72 to 108 months. Children failing either screening test were considered to have failed screening.
As a check on the adequacy of the BDIST and Articulation Screening Tests, a random sample of subjects (22 of the 157 children) were administered a battery of diagnostic speech-language tests by a certified speech-language pathologist. Measures ad-ministered were the Arizona Articulation Profi-ciency Test,2’ the Test of Language Development,22 and/or the Sequenced Inventory of Communication Development.23 Of the 22 children, the 7 who failed screening tests were diagnosed as having speech-language impairment4 (performance 9th percent-ile and 1 SD <IQ). Of the remaining 15-all of whom passed screening-14 were found to have no abnormalities in speech-language development. The screening tests were 88% sensitive in
identi-fying children with handicaps and 100% specific in
identifying children with no handicaps. Thus the
tests met standards of predictive accuracy for
screening instruments.2426 Scores on the diagnostic
tests were combined with screening test results so
that failing scores for speech-language assessment were given when children failed screening or were diagnosed as having speech-language impairment.
Procedures
All screening measures and the PEDS were ad-ministered by one of three doctoral candidates in
special education or psychology. One candidate
in-terviewed parents while another screened children. Both parents and researchers were unaware of the outcome of other measures.
RESULTS
In response to the open-ended question in the PEDS eliciting parental concerns, 22 (14%) of the
157 parents were worried about their child’s speech-language development. Typical comments included: “He can’t talk plain. . . . Others can’t understand her. . . . He only uses a few words. . . . She just
started talking. . . . She can’t say the “1” or “s”
sounds. . . . He doesn’t talk in sentences.” An ad-ditional 26 parents (17%) affirmed speech-language
concerns in response to the PEDS item in which concerns about each developmental area are elic-ited. In total, 48 (31%) of 157 parents had speech-language concerns.
Of the 157 children, 21 (13%), failed articulation measures; 10 children (6%) failed expressive lan-guage measures and 8 children (5%) failed both
measures. A total of 39 children (25%) had negative
speech-language assessment results. This high rate of identification, a 14% surplus beyond an 11% prevalence, suggests that children waiting in pedi-atric offices may not be representative of the pop-ulation of children as a whole.
Parents’ concerns or lack of concerns about speech-language development were compared with children’s performance in speech-language
assess-ment. The comparisons produced figures in keeping
with standards for prescreening measures: (1) 98 of the 118 children who passed speech-language
as-sessment had parents without speech-language
con-cerns, which produced a specificity index of 83%;
(2) 28 of the 39 children who failed speech-language
assessment had parents who were concerned about
their children’s speech and/or language, for a spec-ificity index of 72%; (3) 11 of the 39 children who failed speech-language assessment had parents without concerns, which produced a false positive rate of 17%; and (4) 20 of the 118 children who
passed speech-language assessed had parents who
were concerned about their children’s speech and/ or language, for a false-negative rate of 28%. In the Figure, the predictive usefulness of parents’ speech-language concerns is shown.
In an attempt to evaluate overreferrals and dis-cern why some parents were concerned even though their children had no abnormalities in speech-lan-guage development, we used t tests and
x2 tests
to
compare parents with validated (n = 28) vs unval-idated (n = 20) concerns (at P < .01). There were no differences between groups of parents in terms of gender, education, area of residence, age, marital
status, hours per weekday spent with children, or
Parents’ No
Concerns Yes
Speech-Language Assessment
Pass Fail
Specificity = 98/118 = 83%
Sensitivity = 28/39 72%
False Negative = 11/39 = 28%
False Positive = 20/118 = 17%
98 11
20 28
Figure. Predictive utility of parents’ concerns about
their children’s speech-language development.
did parents with validated concerns. Just as often they gave reasons for problematic development (eg, “He just won’t sit still”) or for perceiving delayed development as a normal variant (eg, “It’s just a phase. . . . She’ll outgrow it. . . .“). Finally, there
were no significant differences in the numbers of concerns (including those in other areas of devel-opment) raised by parents with validated vs unval-idated concerns.
Parents without concerns whose children had speech-language problems (n = 11) were compared with parents with validated concerns (n = 28) to determine why some parents failed to mention or notice their children’s delays. There were no differ-ences between groups of parents in terms of gender, education, age, marital status, hours per weekday spent with children, ratings of perceived severity of children’s medical histories, or numbers of other concerns. Similarly, there were no differences be-tween children in terms of gender, birth order, number of siblings, age, race, or participation in day programs. Parents with concerns about their childrens’ speech-language development tended to live in rural areas (84%) compared with uncon-cerned parents (36%; x2 = 7.22, P < .01). However, this finding did not relate to expected differences in education and employment such that sampling error is suggested. There were anticipated differ-ences between groups in the reasons given for de-velopment. Parents without concerns perceived their children as developing normally, rarely giving reasons for problematic development (x2 = 6.719, Pcz.01).
DISCUSSION
In previous research, clusters of parental con-cerns about children’s learning, ie, speech-language, fine motor, and global development, were shown to predict with 80% accuracy failing performance on screening tests of overall development.15 In the present research we refine the relationship between a specific type of concern, speech-language
devel-opment, and childrens’ performances in measures of speech-language skills. When children had sig-nificant delays in speech-language development, their parents were concerned 72% of the time. When children had normal speech-language devel-opment, 83% of the time their parents had no concerns. Parents, whether educated or inexperi-enced, were equally able to indicate whether their children had difficulty with speech-language skills. It is tempting to conclude that clinical judgment based on parents’ concerns could function as a screening test. In other words, on the basis of parental concern about speech-language develop-ment, children would be referred for diagnostic evaluations. However, almost twice as many par-ents had concerns as had children with speech-language deficits, producing an unacceptably high overreferral rate. Analysis of demographic and other variables, such as level of education,
parent-ing experience, etc, failed to provide a simple
mdi-cator distinguishing parents with validated from those with unvalidated concerns. This means that parents’ concerns are best viewed as a type of prescreening and require confirmation by standard-ized screening tests. Screening tests are also needed to enhance detection of the 20% to 80% of children whose parents do not have concerns about their children’s speech and/or language development.
Concerning developmental disabilities, the American Academy of Pediatrics recommends screening at each well-child visit, preferrably with standardized instruments.’1 In recent research,27
this recommendation is challenged and
develop-mental surveillance is advocated. Skillful observa-tion and identification of parental concern is ad-vanced as a sensitive, time- and cost-effective ap-proach to detecting children with delays. In this study, support is provided for developmental sur-veillance, modified slightly by Academy policy. Spe-cific recommendations for practicing pediatricians include: (1) administering a sensitive, standardized,
developmental screening test at least once during the preschool years to enhance detection of children whose delays are not mentioned by their parents;
(2) eliciting parents’ concerns about
speech-lan-guage development at all visits, and when concerns are present, administering developmental screen-ing tests; (3) conducting repeated developmental screening with children whose parents had unvali-dated concerns to help detect any subtle or emerg-ing language problems.
ACKNOWLEDGMENTS
to the researchers, who were willing to participate
in a study that took at least 30 minutes to complete,
and who would tolerate either a delay in seeing the
pediatrician or in leaving the office after their
ap-pointment. Parents who believed that their children
were having developmental problems may have
been more willing than others to accommodate the
research procedures. Second, all research sites
ac-cepted Medicaid and served low-income families,
perhaps skewing the sample in the direction of
children at risk for handicapping conditions. Third,
parents with concerns about development may be
more likely to seek well-child and follow-up visits.
Hence, children with speech-language or other
de-velopmental problems are well represented in
pe-diatric offices. Finally, children selected as subjects
at medical research sites are more likely than
chil-dren at other possible sites (eg, day-care centers,
schools) to have significant medical
histories-his-tories that may predispose them to developmental
problems. Any or all of these explanations may
have been operative. All support the hypothesis
that patients waiting in pediatric offices may be a
selected and nonrepresentative subgroup of
chil-dren and, as a consequence, there may be high rates
of speech-language problems. Further research is
needed on the developmental characteristics of
pe-diatric populations.
Other potential directions for research include
the following: (1) studies on the relationship of
parents’ concerns to children’s performance in each
developmental domain; (2) a longitudinal study on
whether parents with unvalidated concerns have
children with emerging speech-language
impair-ments or other disabilities; (3) research combining
other aspects of clinical judgment such as
pediatri-cians’ observations and appraisals with parents’
concerns. This may produce a brief technique for
selecting children needing diagnostic services while
eliminating the need for direct developmental
test-ing in pediatric offices.
SUMMARY
Our results suggest that clinical judgment,
de-fined as eliciting and responding to parents’
con-cerns, can detect the majority of children with
speech-language problems. Parents’ concerns met
hit-rate criteria for prescreening tests-sensitivity
of at least 70%, specificity of at least 80%, and an
overreferral rate of 1#{189}to 2 times the expected rate
of detection.2426 Recommendations for practice
in-dude routine elicitation of parents’ concerns,
ad-ministration of screening tests at least once to all
preschoolers, and when parents have concerns.
This work was supported by the Joe Kennedy Fluid Research Foundation and the Department of Pediatrics,
Vanderbilt University.
I thank William Altemeier, William MacLean, Karen
Byrne, and Wendy Stone for their assistance and Mark
Woiraich and Jerry Hickson for their suggestions on
future research.
REFERENCES
1. Algozzine B, Korinek L. Where is special education for students with high prevalence handicaps going? Except Child. 1985;51:388-394
2. Rescorla L. The Language Development Survey: a screening tool for delayed language in toddlers. J Speech Hear Disord.
1989;54:587-599
3. Beitchman JH, Nair R, Clegg M, Patel PG. Prevalence of speech and language disorders in 5-year-old kindergarten children in the Ottawa-Carleton region. J Speech Hear Disord. 1986;51:98-110
4. Student Evaluation Manual of Tennessee. Nashville, TN:
Tennessee Department of Education; 1985
5. Simner ML. The warning signs of school failure: an updated
profile of the at-risk kindergarten child. Top Early Child Education. 1983;3:3-11
6. Laughton J, Hasentab M. The Language Learning Process: Implications for Management of Disorders. Rockville, MD: Aspen Systems Corp; 1986
7. Rapin I. Disorders of higher cerebral function in preschool
children. Am J Dis Child. 1988;142:1119-1124
8. US Department of Education, Office of Special Education and Rehabilitative Services. Annual Report to Congress on the Implementation of the Education for All Handicapped Children. Washington, DC; 1988
9. Coplan J. Evaluation of the child with delayed speech or language. Pediatr Ann. 1985;14:203-208
10. Capute AJ, Accardo PJ. Linguistic and auditory milestones
during the first two years of life. Clin Pedicztr. 1978;17:847-853
11. American Academy of Pediatrics, Committee on Children
with Disabilities. Screening for developmental disabilities.
Pediatrics. 1986;78:526-528
12. Green M, ed. Guidelines for Health Supervision. Elk Grove Village, IL: American Academy of Pediatrics; 1988 13. Shonkhoff JP, Dworkin PH, Leviton A, et al. Primary
care approaches to developmental disabilities. Pediatrics.
1979;64:506-514
14. Smith RD. The use of developmental screening tests by primary-care pediatricians. J Pediatr. 1978;93:524-527
15. Glascoe FP, Altemeier WA, MacLean WE. The importance
of parents’ concerns about their child’s development. Am J Dis Child. 1989:143:955-958
16. Glascoe FP. Parents’ Evaluations of Developmental Status.
Nashville, TN: Vanderbilt University Child Development
Center; 1988
17. Glascoe FP, Altemeier WA, MacLean WE. Are parents’ concerns valid indicators of their child’s development? Am
JDis Child. 1988;142:383
18. Newborg J, Stock JR, Wnek L, Guidubaldi J, Svinicki J.
Battelle Developmental Inventory: Examiner’s ManuaL Al-len, TX: DLM-Teaching Resources; 1984
19. Boyd RD, Welge P, Sexton D, Miller JR. Concurrent validity
of the Battelle Developmental Inventory: relationship with
the Bayley Scales in young children with known or suspected
disabilities. J Early Intervent. 1989;13:14-23
20. Peters JE, Davis JS, Goolsby CM, Clements SD, Hicks TS.
21. Fudala JB, Reynolds WM. Arizona Articulation Proficiency Scale: Second Edition. Los Angeles, CA: Western Psycho-logical Services; 1986
22. Newcomer PL, Hammill DD. Test ofLanguage Development-Primary. Austin, TX: PRO-ED; 1982
23. Hedrick DL, Prather EM, Tobin AR. Sequenced Inventory of Communication Development. Seattle, WA: University of Washington Press; 1984
24. Meisels SJ. Can developmental screening tests identify
chil-dren who are developmentally at risk? Pediatrics. 1989; 83:578-585
25. Barnes KE. Preschool Screening: The Measurement and Prediction of Children At-Risk. Springfield, IL: Charles C Thomas; 1982
26. Wolery M. Child Find and Screening Issues: In: DB Bailey, M Wolery, eds. Assessing Infants and Preschoolers with Handicaps. Columbus, OH: Merrill Publishing Company;
1989
27. Dworkin PH. British and American recommendations for developmental monitoring: the role of surveillance.
Pediat-rics. 1989;84:1000-1010
WILLIAM BUTLER YEATS ON HIS SEXUAL AWAKENING AT ABOUT 17
YEARS OF AGE
William Butler Yeats (1865-1939), considered by many the greatest twentieth
century poet to write in English, wrote as follows about his first awareness of
his sexual awakening.1
The great event of a boy’s life is the awakening of sex. He will bathe many times a day, or get up at dawn and having stripped leap to and fro over a stick laid upon two chairs and hardly know, and never admit, that he had begun to take pleasure in his own nakedness, nor will he understand the change until some dream discovers it. He may never understand at all the greater change in his mind.
It all came upon me when I was close upon seventeen like the bursting of a shell.
Somnambulistic country girls, when it is upon them, throw plates about or pull them with long hairs in simulation of the polter-geist, or become mediums for some genuine spirit-mischief, surrendering to their desire of the marvellous. As I look backward, I seem to discover that my passions, my loves and my despairs, instead of being my enemies, a disturbance and an attack, became so beautiful that I had to be constantly alone to give them my whole attention. I notice that now, for the first time, what I saw when alone is
more vivid in my memory than what I did or saw in company.
REFERENCES
1. Yeats WB. The Autobiography of William Butler Yeats. New York: Collier Books; 1971:40-41
