Examining the Association Between Language Delay and BehavioralEmotional Problems in Children Under

1
Examining the Association Between Language Delay
and Behavioral/Emotional Problems in Children
Under Age 3

• Leslie Rescorla
• Professor of PsychologyBryn Mawr College
• ASHA, 2007
• rescorl_at_brynmawr.edu

2
Association Between Language Delay and Behavior
Problems in Children gt Age 5

• For children gt age 5
• Language problems are often comorbid with other
  problems (Cohen, 2001)
• About 50 of children in mental health clinics
  have a language impairment
• About 50 of children in speech/language clinics
  have a behavioral or emotional disorder

3
Association Between Language Delay and Behavior
Problems in Children Age 3

• For children age 3
• A link between language delay and behavior
  problems was found in the Dunedin (Silva, 1980)
  and London (Stevenson Richman, 1978) general
  population samples
• However, these associations were reduced when
  children with low IQ and/or neurodevelopmental
  delays were excluded

4
Association Between Language Delay and Behavior
Problems in Children lt Age 3

• Results with children lt 3 years are mixed
• Seven studies have used the Language Development
  Survey (LDS) and the Child Behavior Checklist
  (CBCL/1.5-5)
• 4 from general population samples
• 3 from clinic samples
• Will present these 7 studies in some depth, after
  summarizing findings from six other studies that
  used various other measures

5
Recruited Late Talker Sample (1)

• Caulfield, Fischel, DeBaryshe, Whitehurst
  (1989) recruited 34 late talkers (LTs)
• LTs were compared with typically developing (TD)
  peers matched on age, gender, SES, and receptive
  language
• On Eyberg Behavior Inventory and an interview,
  mothers of LTs reported more shyness and more
  bedtime problems than mothers of peers
• LTs were more resistant and difficult than peers
  in lab clean up situation

6
Recruited Late Talker Sample (2)

• Paul, Spangle-Looney Dahm (1991) recruited 21
  LTs
• 29 of the LTs had receptive language delay as
  well as expressive language delay
• LTs were compared with peers matched on age,
  gender, and SES
• 62 of LTs had socialization delays on the
  Vineland Adaptive Behavior Scale
• At follow-up 12 to 18 months later, about 50 of
  the LTs still had socialization delays on the
  Vineland

7
Urban General Population Sample (3)

• Horwitz et al. (2003) identified LTs with the CDI
  
• 15, 18, and 28 scored lt 10th percentile on
  the CDI in 24-29, 30-35, 36-39 month groups (N
  639)
• LTs scored lower than TDs on Infant-Toddler
  Social-Emotional Assessment (ITSEA) Competence
  but not higher on Internalizing or Dysregulation
• 30-39 group had higher Externalizing scores (OR
  4.2)
• ITSEA problem scores did not predict language
  delay in logistic regression analysis
• Low ITSEA Competence ( attention, compliance,
  imitation/play, mastery motivation, empathy,
  prosocial peer relations) significantly predicted
  language delay in logistic regression (OR 3.2)

8
Early Intervention Sample (4)

• Irwin, Carter, Briggs-Gowan (2002) recruited a
  sample of 14 LTs from an Early Intervention
  program
• LTs were matched with 14 TD children from a
  general population sample on age, sex, and
  ethnicity
• LTs had higher scores than TDs on ITSEA
  Depression/Withdrawal and Social Relatedness
• LTs also scored higher than TD children on CBCL
  Withdrawn
• Fitzgerald (2003) suggests LTs included some
  children with PDD, although procedures to exclude
  children with PDD were described by Irwin et al.

9
General Population Twin Sample (5)

• Plomin et al. (2002) study with 4,000 twin pairs
• Correlations between CDI score and Rutter Total
  Problems at 2 -.05 for both genders, at 3
  -.13 (B) -.09 (G)
• Correlations with nonverbal ability and Total
  Problems at 2 -15 (B) -.10 (G), at 3 -25
  (B) -.22 (G)
• rs also low between CDI and TP for those scoring
  lt10th percentile on CDI (.10 at 2, .20 at 3)
• At 2, those scoring lt10th percentile on CDI
  scored lt .20 SDs above the mean on Total Problems
• At 3, those scoring lt10th percentile on CDI
  scored about .30 above the mean on Total Problems

10
Early Intervention Clinic Sample (6)

• Ross Weinberg (2006) studied 109 children age
  18 to 36 months evaluated in an Early
  Intervention clinic
• Children with both receptive and expressive
  language delays on PLS-3 had lower scores on
  Bayley Rating Scale Orientation/Engagement and
  Emotional/Regulation factors than children with
  expressive delay only or with normal language
• Children with mixed language delays also had
  significantly more neurodevelopmental delays
• Both mixed and expressive language delayed groups
  scored lower than TD children on the Vineland
  Socialization scale

11
Summary of the Six Studies

• Associations between language delay and
  behavioral/emotional problems in children lt 3 are
  generally modest, if significant at all
• Both internalizing and externalizing problems
  have shown some association with language delay
• Associations between language delay and behavior
  problems increase as children approach age 3
• Associations may be stronger between language
  delay and measures of socialization and adaptive
  behavior than measures of behavior problems
• Selection factors may increase association
  between language delay and behavior problems in
  recruited LT samples, due to Berksons bias.
• Next, review seven LDS/CBCL studies

12
Language Development Survey

• Language Development Survey (LDS)
• checklist completed by parents of children 18-35
  months
• 310 vocabulary words arranged by semantic
  category
• vocabulary score sum of words reported to be
  used by child spontaneously (not imitated or
  comprehended only).
• Does child combine words into phrases?
• If yes, parent writes in childs five longest
  best phrases
• mean number of words calculated for the five
  phrases
• LDS norms
• Vocabulary score norms - separately by gender for
  18-23, 24-29, and 30-35 months age groups
• Mean length of phrases norms - for 24-29 30-35
  months age groups with genders combined

13
(No Transcript)
14
(No Transcript)
15
Psychometric Properties of LDS

• Reliability
• Test-retest reliability .97-.99
• Cronbachs alpha .99
• Validity
• Correlations with expressive language tests
  .66-.87
• Sensitivity generally gt80, specificity gt 85
• Predictive validity from age 2 (Rescorla, 2002)
  (N59)
• age 7 vocabulary .63
• age 8 grammar .41
• age 8 age 9 reading .39 .34

16
Assessment of Behavioral and Emotional Problems

• CBCL/1.5 - 5 and C-TRF
• 99 problem items (0, 1, 2 scale) (not true,
  somewhat or sometimes true, very true or often
  true)
• write in concerns and best things about the child
• normed in general population sample
• forms yield scores on empirically based
  syndromes, DSM-oriented scales, and
  Internalizing, Externalizing, and Total Problems

17
(No Transcript)
18
Syndromes for Preschool Forms

• Emotionally Reactive
• Anxious/Depressed
• Somatic Complaints
• Withdrawn
• Attention Problems
• Aggressive Behavior
• Sleep Problems (CBCL only)

19
DSM-Oriented Scales for Preschool Forms

• Affective Problems
• Anxiety Problems
• Pervasive Developmental Problems
• Attention Deficit/Hyperactivity Problems
• Oppositional Defiant Problems

20
(No Transcript)
21
(No Transcript)
22
Psychometric Properties of the CBCL/1.5-5

• Reliability
• test-retest mean r .85 (.68 to .90 across
  scales)
• internal consistency alphas .89 to .95 for INT,
  EXT, TP range from .63 to .92 for syndromes and
  DSM-oriented scales
• Validity
• Demographically-matched referred and nonreferred
  samples (N 563, per group) differed
  significantly on all CBCL scales and items
• ES 22 for Total Problems, 20 for INT, 8 for
  EXT

23
(No Transcript)
24
(No Transcript)
25
LDS/CBCL Studies in General Population Samples (1)

• Carson et al. (1998)
• LTs (N 17) and TD peers (N 47) who were
  identified using the LDS in a mail-based
  screening in Wyoming were seen for assessment
• LTs had higher CBCL Total Problems, Sleep
  Problems, and Other Problems scores than TD peers
• But only 38 of possible LTs invited to
  participate came in for the evaluation which
  suggests a selection bias favoring children with
  more problems

26
LDS/CBCL Studies in General Population Samples (2)

• Rescorla Alley (2001)
• Assessed 33 LTs identified with the LDS in a
  household screening (80 yield rate), plus 33
  matched TD toddlers with normal language
• LT and TD groups differed greatly on LDS
  Vocabulary (29 vs. 196 words) at screening and at
  assessment about 1 month later (43 vs. 230 words)
• Large group differences on Bayley, Reynell, and
  Vineland Adaptive Behavior score (93 vs. 110)
• No group differences on CBCL Total Problems

27
LDS/CBCL Studies in General Population Samples (3)

• Rescorla Achenbach (2002)
• Used LDS and CBCL data for 278 children in
  national survey (18-23, 24-29, 30-35 months)
• Vocabulary and Mean Phrase Length scores not
  correlated with any CBCL scores (22 tests, p lt
  .001)
• For ages 24 to 35 months, 15 had fewer than 50
  words or no combinations and 19 had CBCL TP
  scores gt 60 (borderline clinical range)
• No significant association between LDS delay and
  CBCL deviance (only 6/172 deviant on both - 3)
• None of 9 referred kids were deviant on both

28
LDS/CBCL Studies in General Population Samples (4)

• Zubrick Western Australia household survey
• 900 children 24 to 25 months screened with LDS
  and CBCL/1-5-5
• 11 had fewer than 50 words VOC Delay
• 10 had no combinations COMB Delay
• 20 had fewer than 50 words OR no combinations
  LDS delay (lack of overlap between vocabulary and
  combinations unusual)
• deviant on TP, INT, and Ext gt 84th ile
• deviant on Withdrawn gt 93rd ile

29
Zubrick Results

• LDS Vocabulary score not significantly correlated
  with any CBCL/1/5-5 scores
• Children with VOC Delay did not have higher
  scores on any CBCL problem scale
• Children with COMB Delay and LDS Delay had higher
  scores on Withdrawn (only CBCL scale with a
  significant difference)
• LDS Delay not significantly associated with
  deviance on TP, INT, or EXT
• Association with deviance on Withdrawn was modest
  (r .09) but significant (p lt .01)

30
LDS/CBCL Studies in Clinics New Jersey Sample
(5)

• Sample of 83 children age 18 to 35 months
• Study 1 in Rescorla, Ross, McClure (2007)
• Children being assessed for developmental
  problems or as part of NICU follow-up
• 64 of the sample were delayed on the LDS (lt15th
  percentile on Vocabulary or lt 20th percentile on
  Mean Phrase Length)
• Mean Total Problems score 54.44 17 scored
  gt 84th percentile on TP ( deviant)
• 31 scored gt 93rd percentile on DSM-Pervasive
  Developmental Problems (PDP) scale

31
New Jersey Sample Results

• Correlational results with Vocabulary SS
• Significant rs with CBCL Total Problems (-.38),
  Internalizing (-.44), Emotionally Reactive
  (-.24), and Withdrawn (-.56)
• BUT with children deviant on PDP excluded, only
  Withdrawn still significant (r - .42)
• Cross-tabulation of deviance on LDS/CBCL
• Only 21 of children with language delay on LDS
  were deviant on TP not significant
• But 79 of those deviant on TP had delayed
  language on LDS

32
LDS/CBCL Studies in Clinics New York Sample (6)

• Sample of 103 children age 18 to 35 months
• Study 2 in Rescorla, Ross, McClure (2007)
• Essentially the same sample as Ross Weinberg
  (2006) Early Intervention clinic study
• 7 were diagnosed with PDD, 47 had Bayley MDI
  scores lt75, and 30 were born premature
• 56 were delayed in vocabulary on the LDS
• 66 were language-delayed based on delay in
  either vocabulary or mean phrase length
• Mean CBCL Total problems scores 48.53, 16 to
  17 scored in deviant range (gt 84th percentile)

33
New York Sample Results

• Correlational results with Vocabulary SS
• The only significant LDS/CBCL association was
  with Withdrawn (r - .20, p lt .05)
• The other 9 correlations ranged from -.09 to .08
• Cross-tabulation of deviance on LDS/CBCL
• No significant associations between delay on LDS
  and deviance on Total Problems, INT, or EXT (phi
  coefficients .07 to .11)
• Only 17 of children with language delay on LDS
  were deviant on INT

34
LDS/CBCL Studies in Clinics Referred Manual
Sample (7)

• Sample of 102 children age 18 to 35 months
• Children assessed in various developmental and
  mental health clinics
• Referred children lt 3 years of age from referred
  sample in CBCL/1.5-5 manual
• Significant rs between LDS Vocabulary and
  Withdrawn, Attention Problems, Other,
  Internalizing, Total Problems, DSM-Affective, and
  DSM-PDP scales (p lt .01, rs .25 to .35)

35
Referred Manual Sample (II)

• Cross-tabulation of delay on LDS with deviance on
  CBCL problem scores
• Significant ?2 associations (p lt .01) for
  Withdrawn (OR 4.8), Internalizing (OR 3.9),
  and Total Problems (OR 3.7)
• Comorbid percentages
• delayed on LDS also deviant on CBCL
• TP 52 and INT 58 (T 60)
• WITH 27 (T 65)
• deviant on CBCL also delayed on LDS
• TP 47 and INT 44 (T 60)
• WITH 36 (T 65)

36
Conclusions Association Between Language Delay
and BehaviorProblems in Children lt Age 3

• General population samples
• Language delay on the LDS and deviance on the
  CBCL are not generally associated
• Clinical samples
• More association between language delay on the
  LDS and deviance on the CBCL
• Most consistent association is for Withdrawn
• Degree of association varies with type of clinic
  and composition of sample

37
Conclusions and Implications

• Main conclusions
• Many children lt3 with language delays have normal
  behavioral/emotional adjustment
• Many children lt3 with behavioral/emotional
  maladjustment have normal language development
• Children with both language delay and
  maladjustment are likely to have low IQs, high
  family risk, low SES, or PDD spectrum problems
• Most common problem associated with language
  delay is social withdrawal
• Children under age 3 with language delay tend to
  have delays in adaptive behavior/socialization

38
Explanatory Speculations

• Insofar as behavioral/emotional problems are a
  consequence of language delay
• it may take time for the negative effects of
  language delay to produce behavior problems
• the negative effects of language delay may
  increase when children move into group settings
  where they have to interact more with peers
• Insofar as language delay and behavior problems
  arise from shared genetic or environmental
  factors
• the emergence of language delay may precede the
  emergence of behavioral/emotional problems

39
Questions for Further Research

• Longitudinal studies are needed to examine the
  association between language delay and
  behavioral/emotional problems
• Do LT toddlers develop more behavioral or
  emotional maladjustment as they get older?
• Do LTs in day care settings develop behavioral
  problems earlier than LTs cared for at home?
• Do LTs with normal receptive language have less
  maladjustment than LTs with mixed delays?
• Do IQ and/or SES moderate the link between
  language delay and behavior problems?
• What interventions can forestall development of
  behavior problems in children with language
  delay?