Heidi D. Nelson, M.D., M.P.H.^a,b,d; Peggy Nygren, M.A.^a,d; Miranda Walker, B.A.^a,d; Rita Panoscha, M.D.^a,c,d

The authors of this article are responsible for its contents, including any clinical or treatment recommendations. No statement in this article should be construed as an official position of the Agency for Healthcare Research and Quality or the U.S. Department of Health and Human Services.

Address correspondence to: Heidi D. Nelson, M.D., M.P.H., Oregon Health & Science University, Mail Code BICC 504, 3181 SW Sam Jackson Park Road, Portland, Oregon 97239

This article first appeared in Pediatrics. Select for copyright, source, and reprint information.

Background. Speech and language development is a useful indicator of a child's overall development and cognitive ability and is related to school success. Identification of children at risk for developmental delay or related problems may lead to intervention services and family assistance at a young age, when the chances for improvement are best. However, optimal methods for screening for speech and language delay have not been identified, and screening is inconsistently practiced in primary care.

Purpose. We sought to evaluate the strengths and limits of evidence about the effectiveness of screening and interventions for speech and language delay in preschool-aged children to determine the balance of benefits and adverse effects of routine screening in primary care for the development of guidelines by the U.S. Preventive Services Task Force. The target population includes all children up to 5 years old without previously known conditions associated with speech and language delay, such as hearing and neurologic impairments.

Methods. Studies were identified from MEDLINE®, PsycINFO, and CINAHL databases (1966 to November 19, 2004), systematic reviews, reference lists, and experts. The evidence review included only English-language, published articles that are available through libraries. Only randomized, controlled trials were considered for examining the effectiveness of interventions. Outcome measures were considered if they were obtained at any time or age after screening and/or intervention as long as the initial assessment occurred while the child was ≤ 5 years old. Outcomes included speech and language measures and other functional and health outcomes such as social behavior. A total of 745 full-text articles met our eligibility criteria and were reviewed. Data were extracted from each included study, summarized descriptively, and rated for quality by using criteria specific to different study designs developed by the U.S. Preventive Services Task Force.

Results. The use of risk factors for selective screening has not been evaluated, and a list of specific risk factors to guide primary care physicians has not been developed or tested. Sixteen studies about potential risk factors for speech and language delay in children enrolled heterogeneous populations, had dissimilar inclusion and exclusion criteria, and measured different risk factors and outcomes. The most consistently reported risk factors included a family history of speech and language delay, male gender, and perinatal factors. Other risk factors reported less consistently included educational levels of the mother and father, childhood illnesses, birth order, and family size.

The performance characteristics of evaluation techniques that take ≤ 10 minutes to administer were described in 24 studies relevant to screening. Studies that were rated good-to-fair quality reported wide ranges of sensitivity and specificity when compared with reference (sensitivity 17%-100%; specificity 45%-100%). Most of the evaluations, however, were not designed for screening purposes, the instruments measured different domains, and the study populations and setting were often outside of primary care. No "gold standard" has been developed and tested for screening, reference standards varied across studies, few studies compared the performance of ≥ 2 screening techniques in 1 population, and comparisons of a single screening technique across different populations are lacking.

Fourteen good- and fair-quality randomized, controlled trials of interventions reported significantly improved speech and language outcomes compared with control groups. Improvement was demonstrated in several domains including articulation, phonology, expressive language, receptive language, lexical acquisition, and syntax among children in all age groups studies and across multiple therapeutic settings. Improvement in other functional outcomes such as socialization skills, self-esteem, and improved play themes were demonstrated in some, but not all, of the 4 studies that measured them. In general, studies of interventions were small and heterogeneous, may be subject to plateau effects, and reported short-term outcomes based on various instruments and measures.As a result, long-term outcomes are not known, interventions could not be compared directly, and generalization is questionable.

Conclusions. Use of risk factors to guide selective screening is not supported by studies. Several aspects of screening have been inadequately studied to determine optimal methods, including which instrument to use, the age at which to screen, and which interval is most useful.  Trials of interventions demonstrate improvement in some outcome measures, but conclusions and generalizability are limited. Data are not available addressing the effectiveness of screening in primary care, role of enhanced surveillance by primary care physicians before referral for diagnostic evaluation,  non-speech and language and long-term benefits of interventions, and adverse effects of screening and interventions.

Keywords: speech and language delay and disorders, preschool children, screening, interventions.

Speech and language development is considered by experts to be a useful indicator of a child's overall development and cognitive ability¹ and is related to school success.^2-7 Identification of children at risk for developmental delay or related problems may lead to intervention services and family assistance at a young age when chances for improvement are best.¹ This rationale supports preschool screening for speech and language delay, or primary language impairment/disorder, as a part of routine well-child care.

Several types of speech and language delay and disorders have been described,⁸ although terminology varies (Table 1). Expressive language delay may exist without receptive language delay but often they occur together in children as a mixed expressive/receptive language delay. Some children also have disordered language. Language problems can involve difficulty with grammar (syntax), words or vocabulary (semantics), the rules and system for speech sound production (phonology), units of word meaning (morphology) and the use of language particularly in social contexts (pragmatics). Speech problems may include stuttering or dysfluency, articulation disorders, or unusual voice quality. Language and speech problems can exist together or by themselves.

Prevalence rates for speech and language delay have been reported across wide ranges. A recent Cochrane review summarized prevalence data on speech delay, language delay, and combined delay in preschool and school-aged children.⁹ For preschool-aged children, 2 to 4.5 years old, studies that evaluated combined speech and language delay reported prevalence rates ranging from 5% to 8%,^10,11 and studies of language delay have reported prevalence rates from 2.3% to 19%.^9,12-15 Untreated speech and language delay in preschool children has shown variable persistence rates (from 0% to 100%), with most studies reporting 40% to 60%.⁹ In 1 study, two-thirds of preschool children who were referred for speech and language therapy and given no direct intervention proved eligible for therapy 12 months later.¹⁶

Preschool-aged children with speech and language delay may be at increased risk for learning disabilities once they reach school age.¹⁷ They may have difficulty reading in grade school,² exhibit poor reading skills at age 7 or 8,^3-5 and have difficulty with written language,⁶ in particular. This may lead to overall academic underachievement⁷ and, in some cases, lower IQ scores¹⁸ As adults, children with phonological difficulties may hold lower skilled jobs than their non-language-impaired siblings.¹⁹ In addition to persisting speech- and language-related underachievement (verbal, reading, spelling), language-delayed children have also shown more behavior problems and impaired psychosocial adjustment.^20,21

Assessing children for speech and language delay and disorders can involve a number of approaches, although there is no uniformly accepted screening technique for use in the primary care setting. Milestones for speech and language development in young children are generally acknowledged.²² Concerns for delay arise if there are no verbalizations by the age of 1 year, if speech is not clear, or if speech or language is different from that of other children of the same age. Parent questionnaires and parent concern are often used to detect delay.²³ Most formal instruments were designed for diagnostic purposes and have not been widely evaluated for screening. Instruments constructed to assess multiple developmental components, such as the Ages and Stages Questionnaire,²⁴ Clinical Adaptive Test/Clinical Linguistic and Auditory Milestone Scale,²⁵ and Denver Developmental Screening Test,²⁶ include speech and language components. Instruments designed for specific communication domains include the McArthur Communicative Development Inventory,²⁷ Ward Infant Language Screening Test, Assessment, Acceleration, and Remediation (WILSTAAR),²⁸ Fluharty Preschool Speech and Language Screening Test,²⁹ Early Language Milestone Scale,³⁰ and several others.

A specific diagnosis is most often made by a speech and language specialist using a battery of instruments. Once a child has been diagnosed with a speech and/or language delay, interventions may be prescribed. Therapy takes place in various settings including speech and language specialty clinics, home, and schools or classrooms. Direct therapy or group therapy provided by a clinician, caretaker, or teacher can be child centered and/or include peer and family components. The duration of the intervention varies. Intervention strategies focus on 1 or more domains depending on individual needs, such as expressive language, receptive language, phonology, syntax, and lexical acquisition. Therapies can include naming objects, modeling and prompting, individual or group play, discrimination tasks, reading, and conversation.

It is not clear how consistently clinicians screen for speech and language delay in primary care practice. In 1 study, 43% of parents reported that their young child (aged 10 to 35 months) did not receive any type of developmental assessment at their well-child visit, and 30% of parents reported that their child's physician had not discussed how the child communicates.³¹ Potential barriers to screening include lack of time, no clear protocols, and the competing demands of the primary care visit.

This evidence review focuses on the strengths and limits of evidence about the effectiveness of screening and interventions for speech and language delay in preschool age children. Its objective is to determine the balance of benefits and adverse effects of routine screening in primary care for the development of guidelines by the U.S. Preventive Services Task Force (USPSTF). The target population includes all children up to age 5 years without previously known conditions associated with speech and language delay, such as hearing and neurological impairments. The evidence synthesis emphasizes the patient's perspective in the choice of tests, interventions, outcome measures, and potential adverse effects, and focuses on those that are available and easily interpreted in the context of primary care.It also considers the generalizability of efficacy studies performed in controlled or academic settings and interprets the use of the tests and interventions in community-based populations seeking primary health care.

Analytic Framework and Key Questions

Evidence reviews for the USPSTF follow a specific methodology³² beginning with the development of an analytic framework and key questions in collaboration with members of the USPSTF. The analytic framework represents an outline of the evidence review and includes the patient population, interventions, outcomes, and adverse effects of the screening process (Figure 1, 17 KB). Corresponding key questions examine a chain of evidence about the effectiveness, accuracy, and feasibility of screening children age 5 years and younger for speech and language delay in primary care settings (key questions 1 and 2), adverse effects of screening (key question 3), the role of enhanced surveillance in primary care (key question 4), effectiveness of interventions for children identified with delay (key questions 5, 6, and 7), and adverse effects of interventions (key question 8).

Studies addressing key question 1, corresponding to the overarching arrow in the analytic framework, would include all components in the continuum of the screening process, including the screening evaluation, diagnostic evaluation for children identified with delay by the screening evaluation, interventions for children diagnosed with delay, and outcome measures allowing determination of the effectiveness of the overall screening process. Enhanced surveillance in primary care relates to the practice of closely observing children who may have clinical concern for delay but not of the degree warranting a referral ("watchful waiting"). Outcome measures in this review include speech and language specific outcomes as well as non-speech and language health and functional outcomes such as social behavior, self-esteem, family function, peer interaction, and school performance. Key questions 5 examines whether speech and language interventions lead to improved speech and language outcomes. Key question 6 examines whether speech and language interventions lead to improved non-speech and language outcomes. Key question 7 evaluates the subsequent effects of improved speech and language, such as improved school performance at a later age.

Literature Search and Selection

Relevant studies were identified from multiple searches of MEDLINE®, PsycINFO, and CINAHL databases (1966 to November 19, 2004). Search terms were determined by investigators and a research librarian and are described elsewhere.³³ Articles were also obtained from recent systematic reviews,^34,35 reference lists of pertinent studies, reviews, editorials, and Web sites, and by consulting experts.In addition, investigators attempted to collect instruments and accompanying manuals, however, these materials are not generally available and must be purchased, which limited the evidence review to published articles.

Investigators reviewed all abstracts identified by the searches and determined eligibility of full-text articles based on several criteria. Eligible articles had English-language abstracts, were applicable to U.S. clinical practice, and provided primary data relevant to key questions. Studies of children with previously diagnosed conditions known to cause speech and language delay (e.g., autism, mental retardation, Fragile X, hearing loss, degenerative and other neurological disorders) were not included because the scope of this review is screening children without known diagnoses.

Studies of risk factors were included if they focused on children age 5 years or younger, reported associations between predictor variables and speech and language outcomes, and were relevant to selecting candidates for screening. Otitis media as a risk factor for speech and language delay is a complex and controversial area and was not included in this review.

Studies of techniques to assess speech and language were included if they focused on children aged 5 years and younger, could be applied to a primary care setting, used clearly defined measures, compared the screening technique to an acceptable reference standard, and reported data allowing calculation of sensitivity and specificity. Techniques that take 10 minutes or less to complete that could be administered in a primary care setting by nonspecialists are most relevant to screening and are described in this report. Instruments taking more than 10 minutes and up to 30 minutes or for which administration time was not reported are described elsewhere.³³ In general, if the instrument was administered by primary care physicians, nurses, research associates, or other nonspecialists for the study, it was assumed that it could be administered by nonspecialists in a clinic. For questionable cases, experts in the field were consulted to help determine appropriateness for primary care. Studies of broader developmental screening instruments, such as the Ages and Stages Questionnaire and Denver Developmental Screening Test, were included if they provided outcomes related to speech and language delay specifically.

Only RCTs were considered for examining the effectiveness of interventions. Outcome measures were considered if they were obtained at any time or age after screening and/or intervention as long as the initial assessment occurred while the child was aged 5 years or younger. Outcomes included speech and language measures as well as other functional and health outcomes as previously described.

Data Extraction and Synthesis

Investigators reviewed 5,377 abstracts identified by the searches. A total of 690 full-text articles from searches and an additional 55 non-duplicate articles from reference lists and experts met eligibility criteria and were reviewed. Data were extracted from each study, entered into evidence tables, and summarized by descriptive methods. For some studies of screening instruments, sensitivity and specificity were calculated by the investigators if adequate data were presented in the paper. No statistical analyses were performed because of heterogeneity of studies. Investigators independently rated the quality of studies using criteria specific to different study designs developed by the USPSTF (Appendix A).³²

The quality of the study does not necessarily indicate the quality of an instrument or intervention but may influence interpretation of the results of the study. Select Appendix B for a list of excluded studies of instruments and reasons for exclusion.

Key Question 1. Does Screening for Speech and Language Delay Result in Improved Speech and Language as well as Improved Other Non-speech and Language Outcomes?

No studies directly addressed this question.

Key Question 2. Do Screening Evaluations in the Primary Care Setting Accurately Identify Children for Diagnostic Evaluation and Interventions?

Key Question 2a. Does Identification of Risk Factors Improve Screening?

Nine studies conducted in English speaking populations,^36-44 and 7 studies from non-English speaking populations^45-51 met inclusion criteria (Table 2). The most consistently reported risk factors include a family history of speech and language delay, male gender, and perinatal risk factors; however, their role in screening is unclear. A list of specific risk factors to guide primary care physicians in selective screening has not been developed or tested.

English-language studies include case control,^37,39-41,43 cross sectional,^36,38,42 and prospective cohort⁴⁴ designs. Most studies evaluated risk for language delay with or without speech delay, and 1 restricted the evaluation to expressive language only.⁴⁴ Family history was the most consistent significantly associated risk factor in 5 of 7 studies that examined it.^37,39,41-43 Family history was defined as family members who were late talking or had language disorders, speech problems, or learning problems. Male gender was a significant factor in all 3 of the studies examining it.^37,39,42 Three^37,41,43 of 5 studies reported an association between lower maternal education level and language delay, while 3^41-43 of 4 studies evaluating paternal education level reported a similar relationship. Other associated risk factors that were reported less consistently included childhood illnesses,^36,40 born late in the family birth order,⁴² family size,³⁹ older parents³⁹ or younger mother⁴³ at birth, and low socioeconomic status or minority race.⁴⁰ One study that evaluated history of asthma found no association with speech and language delay.³⁹

The 7 studies assessing risk in non-English speaking populations included case-control,⁴⁷ cross-sectional,⁴⁵ prospective-cohort,^48-51 and concurrent-comparison⁴⁶ designs. Studies evaluated several types of delay including vocabulary,⁴⁶ speech,⁴⁵ stuttering,⁴⁷ language,^48-51 and learning.^49-51 Significant associations were reported in the 2 studies evaluating family history,^45,48 and 1 of 2 studies evaluating male gender.⁵¹ Three of 4 non-English language studies, including a cohort of more than 8,000 children in Finland,⁵¹ reported significant associations with perinatal risk factors such as prematurity,^50,51 birth difficulties,⁴⁵ low birth weight,^50,51 and sucking habits.⁴⁵ An association with perinatal risk factors was not found in the 1 English language study that examined low birth weight.⁴³ Other associated risk factors reported less consistently include parental education level,^49,50 and family factors such as size and overcrowding.^50,51 These studies did not find associations with mother's stuttering or speaking style or rate,⁴⁷ mother's age,⁵¹ or child temperament.⁴⁶

Key Questions 2b & 2c. What Are Screening Techniques and How Do They Differ by Age? What Is the Accuracy of Screening Techniques and How Does It Vary by Age?

A total of 22 articles reporting performance characteristics of 24 evaluations met inclusion criteria.³³ Studies utilized several different standardized and nonstandardized instruments (Table 3), although many were not designed specifically for screening purposes. Results of instruments were compared with those of a variety of reference standards and no gold standard was acknowledged or used across studies, which limited comparisons between them.

The studies provided limited demographic details of subjects, and most included predominantly white children with similar proportions of boys and girls. One study enrolled predominantly black children⁵² and another, children from rural areas.⁵³ Study sizes ranged from 25⁵⁴ to 2,590¹¹ subjects. Testing was conducted in general health clinics, specialty clinics, day care centers, schools, and homes by pediatricians, nurses, speech and language specialists, psychologists, health visitors, medical or graduate students, teachers, parents, and research assistants. Studies are summarized below by age categories according to the youngest ages included, although many studies included children in overlapping categories.

Ages 0 to 2 years. Eleven studies from 10 publications utilized instruments taking 10 minutes or less to administer for children up to 2 years old including the Early Language Milestone Scale,^30,55 Parent Evaluation of Developmental Status,⁵⁶ Denver Developmental Screening Test II (language component),⁵⁷ Pediatric Language Acquisition Screening Tool for Early Referral,⁵² Clinical Linguistic and Auditory Milestone Scale,⁵⁸ Language Development Survey,^59-61 Development Profile II,⁵⁷ and the Bayley Infant Neurodevelopmental Screener⁶² (Table 4). Of these studies, 6 tested expressive and/or receptive language,^{30,52,55,57,62} 3 expressive vocabulary,^59-61 1, expressive language and articulation,⁵⁶ and 1, syntax and pragmatics.⁵⁸

For the 10 fair- and good-quality studies that provided data to determine sensitivity and specificity, sensitivity ranged from 22% to 97% and specificity from 66% to 97%.^30,52,56-62 Four studies reported sensitivity and specificity of 80% or more using the Early Language Milestone Scale,³⁰ the Language Development Survey,^59-60 and the Clinical Linguistic and Auditory Milestone Scale.⁵⁸ The study of the Clinical Linguistic and Auditory Milestone Scale also determined sensitivity and specificity by age, and reported higher sensitivity/specificity at age 14 to 24 months (83%/93%) than 25 to 36 months (68%/89%) for receptive function, but lower sensitivity/specificity at age 14 to 24 months (50%/91%) than 25 to 36 months (88%/98%) for expressive function.⁵⁸ A study testing expressive vocabulary using the Language Development Survey indicated higher sensitivity/specificity at age 2 years (83%/97%) than at age 3 years (67%/93%).⁶⁰

Ages 2 to 3 years. Ten studies in 9 publications used instruments taking 10 minutes or less to administer for children aged 2 to 3 years including the Parent Language Checklist,¹¹ Structured Screening Test,⁶³ Levett-Muir Language Screening Test,⁶⁴ Fluharty Preschool Speech and Language Screening Test,^53,65 Screening Kit of Language Development,⁶⁶ Hackney Early Language Screening Test,^54,67 and Early Language Milestone Scale⁶⁸ (Table 5). All studies tested expressive and/or receptive language.^{11,53,54,63-68} In addition, 3 studies tested articulation^53,65 and 1 tested syntax and phonology.⁶⁴

For the 8 fair and good-quality studies providing data to determine sensitivity and specificity, sensitivity ranged from 17% to 100% and specificity from 45% to 100%. Two studies reported sensitivity and specificity of 80% or better using the Levett-Muir Language Screening Test⁶⁴ and the Screening Kit of Language Development.⁶⁶ The study of the Screening Kit of Language Development reported comparable sensitivity/specificity at ages 30 to 36 months (100%/98%), 37 to 42 months (100%/91%), and 43 to 48 months (100%/93%).⁶⁶

Ages 3 to 5 years. Three studies used instruments taking 10 minutes or less to administer including the Fluharty Preschool Speech and Language Screening Test,⁶⁹ Test for Examining Expressive Morphology,⁷⁰ and the Sentence Repetition Screening Test⁷¹ (Table 6). Of these, 2 studies tested expressive and receptive language and articulation,^69,71 and 1 tested expressive vocabulary and syntax.⁷⁰ The 2 fair-quality studies reported sensitivity ranging from 57% to 62% and specificity from 80% to 95%.^66,69,71

Systematic review. A Cochrane systematic review of 45 studies, including most of the studies cited above, summarized the sensitivity and specificity of instruments taking 30 minutes or less to administer.³⁴ Sensitivity of instruments for normally developing children ranged from 17% to 100%, and for children from clinical settings it ranged from 30% to 100%. Specificity ranged from 43% to 100%, and 14% to 100% respectively. Studies considered to be of higher quality tended to have higher specificity than sensitivity (t=4.41, P<0.001), however, high false-positive and false-negative rates were reported often.³⁴

Key Question 2d. What Are the Optimal Ages and Frequency for Screening?

No studies addressed this question.

Key Question 3. What Are the Adverse Effects of Screening?

No studies addressed this question. Potential adverse effects include false-positive and false-negative results. False-positive results can erroneously label children with normal speech and language as impaired, potentially leading to anxiety for children and families and further testing and interventions. False-negative results would miss identifying children with impairment, potentially leading to progressive speech and language delay and other long-term effects including communication, social, and academic problems.In addition, once delay is identified, children may be unable to access services because of unavailability or lack of insurance coverage.

Key Question 4. What Is the Role of Enhanced Surveillance by Primary Care Clinicians?

No studies addressed this question.

Key Question 5. Do Interventions for Speech and Language Delay Improve Speech and Language Outcomes?

Twenty-five RCTs in 24 publications met inclusion criteria including 1 rated good,⁷² 13 rated fair,^73-85 and 11 rated poor quality (Table 7).^77,86-95 Studies were considered poor quality if they reported important differences between intervention and comparison groups at baseline, did not use intention-to-treat analysis, no method of randomization was reported, and there were fewer than 10 subjects in intervention or comparison groups. Limitations of studies, in general, include small numbers of participants (only 4 studies enrolled more than 50 subjects), lack of consideration of potential confounders, and disparate methods of assessment, intervention, and outcome measurement. As a result, conclusions about effectiveness are limited. Although children in the studies ranged from 18 to 75 months old, most studies included children age 2 to 4 years old and results do not allow for determination of optimal ages of intervention.

Studies evaluated the effects of individual or group therapy directed by clinicians and/or parents that focused on specific speech and language domains. These include expressive and receptive language, articulation, phonology, lexical acquisition, and syntax. Several studies used established approaches to therapy, such as the WILSTAAR program⁹⁶ and the HANEN principles.^78,79,85,93 Others used more theoretical approaches, such as focused stimulation,^{78,79,86,87,93} auditory discrimination,^83,90 imitation or modeling procedures,^76,92 auditory processing or work mapping,⁸⁵ and play narrative language.^80,81  Some interventions focused on specific words and sounds, used unconventional methods, or targeted a specific deficit.

Outcomes were measured by subjective reports from parents,^77,78,80,85 and by scores on standardized instruments, such as the Reynell Expressive and Receptive Scales,^74,77 the Preschool Language Scale,^72,75,85 and the MacArthur Communicative Development Inventories.^80,93 The most widely used outcome measure was mean length utterances, used by 6 studies.^{73,75,77,80,85}

Studies rated good or fair quality are described below by age categories according to the youngest ages included, although many studies included children in overlapping categories

Ages 0 to 2 years. No studies examined this age group exclusively, although 1 good-quality study enrolled children 18 to 42 months old.⁷² The clinician-directed, 12-month intervention consisted of 10-minute weekly sessions focusing on multiple language domains, expressive and receptive language, and phonology. Treatment for receptive auditory comprehension led to significant improvement for the intervention group compared with control group, however, results did not differ between groups for several expressive and phonology outcomes.⁷²

Ages 2 to 3 years. One good-⁷² and 6 fair-quality studies^77-80,84,85 evaluated speech and language interventions for children who were 2 to 3 years old. Studies reported improvement on a variety of communication domains including clinician-directed treatment for expressive and receptive language,⁸⁰ parent-directed therapy for expressive delay,^77,78 and clinician-directed receptive auditory comprehension.⁷² Lexical acquisition was improved with both clinician-directed therapy^84,91 and group therapy approaches.⁸⁴ In 3 studies, there were no between group differences for clinician-directed expressive^72,85 or receptive language therapy,^72,85 parent-directed expressive or receptive therapy,⁸⁵ or parent-directed phonology treatment.⁷⁹

Ages 3 to 5 years. Five fair-quality studies reported significant improvements for children 3 to 5 years old undergoing interventions compared with controls,^{73,74,76,81,82} while 2 studies reported no differences.^75,83 Both group-based interventions⁸¹ and clinician-directed interventions⁷⁴ were successful in improving expressive and receptive competencies.

Systematic review. A Cochrane systematic review included a meta-analysis utilizing data from 25 RCTs of interventions for speech and language delay for children up to adolescence.³⁵ Twenty-three of these studies also met criteria for this review and were included in Table 7,^72-92,95 and 2 trials were unpublished. The review reported results in terms of standard mean differences (SMD) in scores for a number of domains (phonology, syntax, and vocabulary). Effectiveness was considered significant for both the phonological (SMD=0.44; 95% CI, 0.01-0.86) and vocabulary (SMD=0.89; 95% CI, 0.21-1.56) interventions. Less effective was the receptive intervention (SMD=-0.04; 95% CI, 0.64-0.56), and results were mixed for the expressive syntax intervention (SMD=1.02; 95% CI, 0.04-2.01). In the analysis, when interventions were comparable in duration and intensity, there were no differences between interventions when administered by trained parents or clinicians for expressive delays. Use of normal-language peers as part of the intervention strategy also proved beneficial.⁸¹

Key Question 6. Do Interventions for Speech and Language Delay Improve Other Non-Speech and Language Outcomes?

Four good-⁷² or fair-quality^80,81,85 intervention studies included functional outcomes other than speech and language. Increased toddler socialization skills,⁸⁰ improved child self-esteem,⁸⁵ and improved play themes⁸¹ were reported for children in intervention groups in 3 studies. Improved parent-related functional outcomes included decreased stress⁸⁰ and increased positive feelings toward their children.⁸⁵ Functional outcomes that were studied but did not show significant treatment effects included well being, levels of play and attention, and socialization skills in 1 study.⁷²

Key Question 7. Does Improvement in Speech and Language Outcomes Lead to Improved Additional Outcomes?

No studies addressed this question.

Key Question 8. What Are the Adverse Effects of Interventions?

No studies addressed this question. Potential adverse effects of treatment programs include the impact of time and cost of interventions on clinicians, parents, children, and siblings. Loss of time for play and family activities, stigmatization, and labeling may also be potential adverse effects.

Studies are not available addressing the overarching key question about the effectiveness of screening (key question 1), adverse effects of screening (key question 3), the role of enhanced surveillance in primary care (key question 4), long-term effectiveness of interventions on non-speech and language outcomes for children identified with delay (key questions 7), and adverse effects of interventions (key question 8). No studies determine the optimal ages and frequency for screening (key question 2d). Relevant studies are available regarding the use of risk factors for screening (key question 2a), techniques for screening (key question 2b and 2c), and effectiveness of interventions on short-term speech and language and non-speech and language outcomes for children identified with delay (key questions 5 and 6).

The use of risk factors for selective screening has not been evaluated and a list of specific risk factors to guide primary care physicians has not been developed or tested. Sixteen studies about potential risk factors for speech and language delay in children enrolled heterogeneous populations, had dissimilar inclusion and exclusion criteria, and measured different risk factors and outcomes. The most consistently reported risk factors included a family history of speech and language delay, male gender, and perinatal factors. Other risk factors that were reported less consistently included educational levels of the mother and father, childhood illnesses, birth order, and family size.

Although brief evaluations are available and have been used in a number of settings with administration by professional and nonprofessional individuals, including parents, the optimal method of screening for speech and language delay has not been established. The performance characteristics of evaluation techniques taking 10 minutes or less to administer were described in 24 studies relevant to screening. Studies rated good-to-fair quality reported wide ranges of sensitivity and specificity when compared with reference standards (sensitivity 31% to 100%; specificity 45% to 100%). In these studies, the instruments providing the highest sensitivity and specificity included the Early Language Milestone Scale, Clinical Linguistic and Auditory Milestone Scale, Language Development Survey, Screening Kit of Language Development, and the Levett-Muir Language Screening Test. Most of the evaluations, however, were not designed for screening purposes, the instruments measured different domains, and the study populations and settings were often outside primary care. No gold standard has been developed and tested for screening, reference standards varied across studies, few studies compared the performance of 2 or more screening techniques in 1 population, and comparisons of a single screening technique across different populations are lacking.

RCTs of multiple types of interventions reported significantly improved speech and language outcomes compared with control groups. Improvement was demonstrated in several domains including articulation, phonology, expressive language, receptive language, lexical acquisition, and syntax among children in all age groups studied and across multiple therapeutic settings. However, studies were small, heterogeneous, may be subject to plateau effects, and reported short-term outcomes based on various instruments and measures. As a result, long-term outcomes are not known, interventions could not be directly compared to determine optimal approaches, and generalizability is questionable.

There are many limitations of the literature relevant to screening for speech and language delay in preschool-aged children including lack of studies specific to screening as well as difficulties inherent in this area of research. This evidence review is limited by use of only published studies of instruments and interventions. Data about performance characteristics of instruments, in particular, are not generally accessible and are often only available in manuals that must purchased. Interventions vary widely and may not be generalizable. In addition, studies from countries with different health care systems, such as the U.K., may not translate well to U.S. practice.

Although speech and language development is multi-dimensional, the individual constructs that comprise it are often assessed separately. Numerous evaluation instruments and interventions that accommodate children across a wide range of developmental stages have been developed to identify and treat specific abnormalities of these functions. As a result, studies include many different instruments and interventions that are most often designed for purposes other than screening. Also, studies of interventions typically focus on 1 or a few interventions. In clinical practice, children are provided with individualized therapies consisting of multiple interventions. The effectiveness of these complex interventions may be difficult to evaluate. Adapting results of this heterogeneous literature to determine benefits and adverse effects of screening is problematic. Also, behavioral interventions are difficult to conduct in long-term randomized trials, and it is not possible to blind parents or clinicians. Randomizing children to therapy or control groups when clinical practice standards support therapy raises ethical dilemmas.

Speech and language delay is defined by measurements on diagnostic instruments in terms of a position on a normal distribution. Measures and terminology are inconsistently used and there is no recognized gold standard. This is challenging when defining cases and determining performance characteristics of screening instruments in studies.

Identification of speech and language delay may be associated with benefits and adverse effects that would not be captured by studies of clinical or health outcomes. The process of screening alerts physicians and caretakers to developmental milestones and focuses attention on the child's development, potentially leading to increased surveillance, feelings of caregiver support, and improved child self esteem. Alternatively, caretakers and children may experience increased anxiety and stress during the screening and evaluation process. Detection of other conditions during the course of speech and language evaluation, such as hearing loss, is an unmeasured benefit if appropriate interventions can improve the child's status.

Future research should focus on determining optimal approaches of identifying preschool children with speech and language delay in primary care settings who would be appropriate candidates for further evaluations and possibly speech and language interventions. These approaches should be integrated into routine developmental surveillance practices of clinicians caring for children.⁹⁷ Studies that evaluate the effectiveness of validated brief screening instruments that include child and caretaker components could lead to a more standardized approach. Studies of specific speech and language components of currently available broad developmental screening instruments, such as Ages and Stages, would be useful. Incorporation of risk factors and parent report in studies of screening approaches could provide information about their added value. Additional studies that compare screening instruments and methods in large primary care populations could lead to defining gold standards and acceptable referral criteria. Evaluating these criteria in different populations of children would minimize cultural and language biases.

Additional work about the effectiveness of interventions, including speech and language domain-specific results, may provide new insights. School-based efforts could be designed to complement strategies developed for young children improving long-term outcomes. Results of these studies may help determine optimal ages and intervals for screening. Functional long-term outcomes such as school performance, high school graduation rates, in-grade retention, special education placement/duration, and social adjustment need to be addressed more thoroughly. Cost-effectiveness evaluations of effective approaches that consider cost of treatment, the time that caregivers spend in transit to treatment locations, the time they spend participating in the program on site or in the home, and long-term outcomes, among other factors, would be useful.

Acknowledgments

The investigators thank Andrew Hamilton, M.L.S., M.S., for conducting the literature searches, expert reviewers for commenting on draft versions of this report, and members of the USPSTF who served as leads for this project including Alfred Berg, M.D., M.P.H.; Paul Frame, M.D.; Leon Gordis, M.D., Dr.P.H.; Jonathan Klein, M.D., M.P.H.; Virginia Moyer, M.D., M.P.H.; and Barbara Yawn, M.D., M.Sc.

This study was conducted by the Oregon Evidence-based Practice Center under contract to the Agency for Healthcare Research and Quality, Contract #290-02-0024, Task Order Number 2, Rockville, MD.

Author Affiliations

All Oregon Health and Science University, Portland, Oregon.

[a] Department of Medical Informatics and Clinical Epidemiology.

[b] Department of Medicine.

[c] Department of Pediatrics.

[d] The Oregon Evidence-based Practice Center.

Copyright and Source Information

This document is in the public domain within the United States.

Requests for linking or to incorporate content in electronic resources should be sent via the USPSTF contact form.

Source: Nelson HD, Nygren P, Walker M, Panoscha R. Screening for Speech and Language Delay in Preschool Children: Systematic Evidence Review for the US Preventive Services Task Force. Pediatrics 2006;117:e298-e319.

1. Does screening for speech & language delay result in improved speech & language as well as improved other non speech & language outcomes?
2. Do screening evaluations in the primary care setting accurately identify children for diagnostic evaluation and interventions?
   1. Does identification of risk factors improve screening?
   2. What are screening techniques and how do they differ by age?
   3. What is the accuracy of screening techniques and how does it vary by age?
   4. What are the optimal ages and frequency for screening?
3. What are the adverse effects of screening?
4. What is the role of enhanced surveillance by primary care clinicians?
5. Do interventions for speech & language delay improve speech & language outcomes?
6. Do interventions for speech & language delay improve other non speech & language outcomes?
7. Does improvement in speech & language outcomes lead to improved additional outcomes?
8. What are the adverse effects of interventions?

* Autism, mental retardation, Fragile X, hearing loss, degenerative and other neurologic disorders.
† School performance, family function, social behavior, and others.

Diagnostic Accuracy Studies

Criteria

• Screening test relevant, available for primary care, adequately described.
• Study uses a credible reference standard, performed regardless of test results.
• Reference standard interpreted independently of screening test.
• Handles indeterminate results in a reasonable manner.
• Spectrum of patients included in study.
• Sample size.
• Administration of reliable screening test.

Definition of Ratings Based on Above Criteria

Good: Evaluates relevant available screening test; uses a credible reference standard; interprets reference standard independently of screening test; reliability of test assessed; has few or handles indeterminate results in a reasonable manner; includes large number (more than 100) broad-spectrum patients with and without disease.

Fair: Evaluates relevant available screening test; uses reasonable although not best standard; interprets reference standard independent of screening test; moderate sample size (50 to 100 subjects) and a "medium" spectrum of patients.

Poor: Has important limitations such as: uses inappropriate reference standard; screening test improperly administered; biased ascertainment of reference standard; very small sample size of very narrow selected spectrum of patients.

Randomized Controlled Trials (RCTs) and Cohort Studies

Criteria

• Initial assembly of comparable groups: RCTs—adequate randomization, including concealment and whether potential confounders were distributed equally among groups; cohort studies—consideration of potential confounders with either restriction or measurement for adjustment in the analysis; consideration of inception cohorts.
• Maintenance of comparable groups (includes attrition, cross-overs, adherence, contamination).
• Important differential loss to followup or overall high loss to followup.
• Measurements: equal, reliable, and valid (includes masking of outcome assessment).
• Clear definition of interventions.
• Important outcomes considered.
• Analysis: adjustment for potential confounders for cohort studies, or intention-to-treat analysis for RCTs (i.e. analysis in which all participants in a trial are analyzed according to the intervention to which they were allocated, regardless of whether or not they completed the intervention).

Definition of Ratings Based on Above Criteria

Good: Meets all criteria: Comparable groups are assembled initially and maintained throughout the study (followup at least 80 percent); reliable and valid measurement instruments are used and applied equally to the groups; interventions are spelled out clearly; important outcomes are considered; and appropriate attention to confounders in analysis.

Fair: Studies will be graded "fair" if any or all of the following problems occur, without the important limitations noted in the "poor" category below: Generally comparable groups are assembled initially but some question remains whether some (although not major) differences occurred in followup; measurement instruments are acceptable (although not the best) and generally applied equally; some but not all important outcomes are considered; and some but not all potential confounders are accounted for.

Poor: Studies will be graded "poor" if any of the following major limitations exists: Groups assembled initially are not close to being comparable or maintained throughout the study; unreliable or invalid measurement instruments are used or not applied at all equally among groups (including not masking outcome assessment); and key confounders are given little or no attention.

Key

* Speech and language are part of a broader screening instrument.