Screening for Developmental Delay in the Setting of a Community Pediatric Clinic: A Prospective Assessment of Parent-Report Questionnaires
OBJECTIVES. Our goal for this study was to prospectively test whether parent-completed questionnaires can be effectively used in the setting of a busy ambulatory pediatric clinic to accurately screen for developmental impairments. Specific objectives included (1) assessing the feasibility of using parent-report instruments in the setting of a community pediatric clinic, (2) evaluating the accuracy of 2 available screening tests (the Ages and Stages Questionnaire and Child Development Inventory), and (3) ascertaining if the pediatrician's clinical judgment could be used as a potential modifier.
METHODS. Subjects were recruited from the patient population of a community clinic providing primary ambulatory pediatric care. Subjects without previous developmental delay or concerns noted were contacted at the time of their routine 18-month-old visit. Those subjects who agreed to participate were randomly assigned to 1 of 2 groups and completed either the Ages and Stages Questionnaire or Child Development Inventory. The child's pediatrician also completed a brief questionnaire regarding his or her opinion of the child's development. Those children for whom concerns were identified by either questionnaire underwent additional detailed testing by the Battelle Development Inventory, the “gold standard” for the purposes of this study. An equal number of children scoring within the norms of the screening measures also underwent testing with the Battelle Development Inventory.
RESULTS. Of the 356 parents contacted, 317 parents (90%) agreed to participate. Most parents correctly completed the Ages and Stages Questionnaire (81%) and the Child Development Inventory (75%). Predictive values were calculated for the Ages and Stages Questionnaire and the Child Development Inventory (sensitivity: 0.67 and 0.50; specificity: 0.39 and 0.86; positive predictive value: 34% and 50%; negative predictive value: 71% and 86%, respectively). Incorporating the physician's opinion regarding the developmental status of the child did not improve the accuracy of the screening questionnaires.
CONCLUSIONS. Three important conclusions were reached: (1) parent-completed questionnaires can be feasibly used in the setting of a pediatric clinic; (2) the pediatrician's opinion had little effect in ameliorating the accuracy of either questionnaire; and (3) single-point accuracy of these screening instruments in a community setting did not meet the requisite standard for development screening tests as set by current recommendations. This study raises important questions about how developmental screening can be performed, and we recommend additional research to elucidate a successful screening procedure.
An estimated 12% to 16% of the US pediatric population has a developmental disability.1 There has been increasing pressure to identify these children at an earlier age, with the current focus being on infants.2 Part of the reason is that research has demonstrated not only that intervention programs are cost-effective and have lifelong benefits but also that developmental attainment is maximized when intervention is commenced earlier. Consequently, professional societies such as the American Academy of Pediatrics (AAP) advocate the identification of developmentally delayed children before 2 years of age.2
The most effective manner to identify children with developmental delay still remains elusive, although the need to find a practical screening method has been indicated in the literature as early as 1979. Shonkoff et al concluded that “more precise techniques for pediatric developmental assessment and conclusive evaluations of specific interventions will have to be produced… . The current difficulty in defining criteria for optimal pediatric management emphasizes the need for creative, methodologically sophisticated research in the area.”3 More recently, Sices et al4 stated, “The AAP … does not provide specific guidance on how a primary care physician is to perform developmental surveillance and screening. Research on how these guidelines can be best implemented in the context of primary care practice would help standardize and enhance the value of the experience for patients and families.”4 These authors call for the formulation of guidelines, which can be used by primary care practitioners in a time- and cost-efficient manner.
Currently, most primary care practitioners incorporate developmental surveillance that relies heavily on a pediatrician's intuition and clinical judgment. Although such a method might be accurate given enough time, training, supplementary resources, and expertise, studies have demonstrated that clinical judgment alone does not accurately detect developmental delay. For example, a study from 1987 demonstrated that only 20% to 30% of developmentally impaired children were being identified before school age by primary care practitioners using developmental surveillance.5 For this reason, professional organizations advocate the use of standardized developmental screening tests.2 Developmental screening has been shown to improve the accuracy with which children's developmental delays are identified when compared with decisions based only on clinical judgment.6–8 Indeed, the sensitivity and specificity of screening instruments are usually reported between 70% and 90%.9–12 Although the use of screening tests would improve the rate and accuracy of identification, a recent survey demonstrated that only 23% of primary care clinicians used a standardized screening tool, leading the authors to conclude: “Our findings do raise…the concerns that systems of care that foster the proper use of adequate detection methods in the primary care setting continue to be elusive.”13 A previous survey documented some of the obstacles faced by primary care practitioners, with time constraints being voiced by 82% of them.14 The lack of medical staff (48%) and the burden of cost (44%) in monitoring developmental delay, for which they are not well compensated, were also listed as significant preventive factors.14
As a substitute to these time-intensive, pediatrician-administered screening tests, a role for standardized parent-completed questionnaires can be considered. Parent-completed questionnaires are as accurate as developmental screening instruments, because current research strongly supports the observation that parents, regardless of differences in socioeconomic status, geographical location, or parental well-being, can give accurate information about their child's development.15–21 Moreover, parent-completed screening tools are cost-effective in the short-term, can be completed over the telephone, in the waiting room, or by mail, and are time-efficient.22 Realizing the advantages of parent-completed questionnaires, Regalado and Halfon wrote: “The available evidence suggests that assessment of developmental issues might benefit from the wider use of structured, validated approaches… . [S]ystemic assessment of parent's concerns can play a role in identifying children with developmental problems, replacing or supplementing longer and more costly developmental screening assessments.”23 Consequently, the use of parent-completed questionnaires might be more practical in the busy pediatric clinic.
The literature concerning the topic of developmental screening with parent-completed questionnaires has several shortfalls that limit the widespread applicability of the results. Previous studies have focused on specific cohorts that had already been flagged as being delayed or consisted of high-risk infants (very low birth weight and prematurity).24–26 Sample sizes in the majority of studies, which rely on specific cohorts, are small, and the applicability of these results to the general pediatric population as a whole is uncertain.24–26 Studies, using other screening instruments as the gold standard, must also be analyzed skeptically, because validating a screening instrument with another parent-completed screening instrument can skew results.25 These studies do not offer guidelines that primary care practitioners can use to screen for developmental delays.7,24–26 Part of the reason might be that these tests are not designed with an emphasis on keeping the methods in line with the actual reality of the clinical setting. The majority of research has focused on validation or predictive value and not on actual methods or practicality. Indeed, Sonnander states that “few … studies focusing on evaluation of developmental screening programs conducted within a clinical setting were found,” whereas “empirical research into child development and the predictive value of developmental tests is extensive.”27 Thus, there is a lack of direct pragmatic research addressing the area of how a primary care practitioner should screen for developmental delay in a community.
To create an efficient system for recognition of developmental problems, we defined 3 objectives for our study: (1) to test whether questionnaires can be feasibly completed by the caregiver in the waiting room of a busy pediatric clinic; (2) to examine the accuracy of these parent-completed questionnaires when used in such a setting; and (3) to assess what role the practitioner plays in ameliorating the questionnaire's accuracy. These aspects of parent-completed screening tests need to be examined systematically to provide a rational foundation to the formulation of a feasible, clear, and accurate screening protocol.
Subjects were recruited from the patient population of the Children's Care Clinic (CCC). Located in Pierrefonds, Quebec, Canada, the CCC is an exclusively pediatrics group practice, incorporating 7 full-time pediatricians. It is community based, drawing from a suburban, largely middle class population and providing comprehensive primary pediatric care to its clientele. Its patient demographic and appointment systems are fully computerized.
Subjects were recruited at the time of their 18-month-old visit. This is a standard routine visit that coincides with the administration of a number of vaccines. The 18-month timing was also chosen because it allows for assessment of motor, language, social, and cognitive skills by standardized developmental screening instruments. Furthermore, successful identification of delays at this point in time would represent a substantial improvement over what is currently achieved (ie, referral to pediatric subspecialists and rehabilitation services typically occurs at ≥2–3 years).18 All subjects turning 18 months were contacted and eligible to participate unless there was an established significant developmental disability. Recruitment was terminated after 101 assessments were completed with the Battelle Development Inventory (BDI) because of funding constraints. Also, parents completing the questionnaire needed to have a fourth-grade reading level and understanding of English, because these questionnaires are presently available in English only.
The database at the CCC was scanned continually for subjects who were turning 18 months of age in the forthcoming 2 months. The identified caregivers were sent a cover letter, signed by the pediatricians at the CCC indicating their support, that described the study and included study contact information. A telephone call from a research assistant was made soon after receipt of the letter, during which any questions or concerns regarding the study were addressed and verbal consent or refusal was noted. If a parent refused participation, the reason for refusal was obtained and noted. All parents who refused were also asked to complete a simple demographic questionnaire over the telephone that gathered basic demographic information on employment, income, and education. If verbal consent was given, the subject's appointment date was noted, and a dossier was placed in the child's folder at the clinic. Each dossier contained:
One of 2 selected parental-report measures (the Ages and Stages Questionnaire [ASQ] or Child Development Inventory [CDI]) depending on group assignment, which was performed using random-number tables and blocking to ensure equal distribution.
A sheet with the question “Do you have any present developmental concerns regarding your child (yes/no)?” If answering “yes,” the caregiver chose from a checklist to specify if this concern related to motor (gross and/or fine), language, social, or cognitive domains or to >1 domain; the caregiver was provided space to provide any additional comments.
A simple demographic questionnaire requesting information on employment, education and income.
A consent form for participation.
A Likert-type questionnaire regarding the ease of use of the questionnaire they had completed.
Completed questionnaires were scored by a research assistant (D.R.) according to established procedures for each measure. The number of incorrectly completed questionnaires for each measure was noted. Those children who failed the parent-completed questionnaire (<2 SDs on ≥1 domain) underwent BDI testing (the gold standard for the purpose of this study), which was administered approximately 3 months after the initial screen by an experienced pediatric occupational therapist who was blinded to both the screening instruments used and the results obtained. The 3-month interval used in this study represents the 3-month referral wait typically experienced by families awaiting a more thorough developmental assessment, and at 21 months of age, the child falls in the middle of the 18- to 23-month marking scheme of the BDI, giving the most accurate estimation of his or her real abilities according to the parameters of this measure.
To investigate false-negative estimates, an equal number of children scoring within the norms of the screening questionnaires were selected as controls and also underwent BDI testing. Controls were the next child of the same gender participating in the study who had scored normally on the parent-completed screening instrument. A false-negative estimate for this approach was also obtained in a likewise manner. In this fashion, the control group was selected in an unbiased manner. Because of constraints posed by funding and time available from the occupational therapist, only 101 children underwent BDI testing, after which ongoing study recruitment was terminated. Although this is not as favorable as testing all the participants with the BDI, 101 completed BDIs represents a large enough sample to be a good indicator of the accuracy of the ASQ and CDI.
The pediatrician who provided care to the child was blinded to the actual questionnaire used and the results obtained but was asked as part of the study to answer a question identical to the one given to the parent regarding any possible concerns about the child's development. This question was answered at the time of the subject's visit by simply checking the appropriate response box(es). The results of this simple questionnaire were used in the analysis to establish what role the pediatrician could play in influencing the accuracy of the ASQ and CDI.
Parents who had agreed over the telephone to participate but did not return the questionnaire were contacted by telephone and reminded to return the questionnaire. If the questionnaire had not been returned 3 months after it was due, the parents were sent a demographic questionnaire to assess whether any parameters differentiated these parents from those who actually completed the questionnaire. Families with whom we had no personal contact by telephone were not included in the study.
Throughout this study, data obtained were stored and managed on secure computers within the Montreal Children's Hospital Division of Pediatric Neurology with separation of subject identifier variables. Ethical approval for conduct of the study was obtained from the Montreal Children's Hospital–McGill University Health Centre Research Ethics Board. Data storage and analysis were performed by using SPSS 12.0 (SPSS Inc, Chicago, IL), and all data entered were verified for errors.
The BDI,28 a psychometrically sound developmental assessment for children from birth to 8 years, is widely used in studies evaluating developmental delay and in US-mandated early intervention programs. The BDI gathers information, via a structured test format, to include interviews with the caregiver and direct observation of the child. Items are assigned an age level at which 75% of the norming population was able to perform. It is composed of 341 items that evaluate 5 different domains: personal/social, adaptive, motor, communication, and cognitive, with each domain further subdivided into subdomains. Items are scored as typically signifying fully developed skills, sometimes signifying emerging abilities, or rarely signifying absent skills and are scored at values of 2, 1, or 0, respectively. The authors of the test recommend the test for many functions including general screening, detailed assessment of children who have previously been identified as having developmental delay, or identifying strengths and weaknesses of normal or developmentally impaired children. The manual states that the assessment may take up to 2 hours to administer with children <3 years of age.
The ASQs29 are 19 parent-report questionnaires that span the age range of 4 to 60 months. Questionnaire points include 4, 6, 8, 10, 12, 14, 16, 18, 20, 22, 24, 27, 30, 33, 36, 42, 48, 54, and 60 months of age. Each questionnaire is composed of 3 sections: a brief set of demographic items; 30 questions about the infant's or child's development assessing 5 different domains equally (communication, gross motor, fine motor, problem-solving, and personal/social); and 7 open-ended questions eliciting parental concerns. The choice of responses for each item is “yes,” “sometimes,” or “not yet,” which are scored as 10, 5, or 0, respectively. The test is graded according to the domain tested and compared with an empirically derived screening cutoff score and takes ∼10 to 15 minutes to complete and ∼5 minutes to score.
Although the ASQ states that the questionnaire can be used in-clinic, the ASQ is meant for mail-out purposes to be completed at home at the specific age intervals listed above. Because of this, the instructions ask the parents to attempt every activity with their child and, if the child is noncompliant, to try at a later time. Unfortunately, this is not feasible in the context of a busy waiting room. Thus, these instructions were abridged, and parents were asked to base their responses on their observations and previous experiences with their child. The ASQ manual does state that completion of the questionnaire in the waiting room is the least accurate method, because the parent has a limited time to complete the activities, and the appointment might not fall at an age at which the best results are achieved.
The CDI30 is useful for testing children whose ages range from birth to 6 years. It consists of 300 items: 270 items composed of yes/no statements about the child's development and 30 items that evaluate various potential sensory, physical, motor, language, and behavioral problems. These items are grouped into 8 subscales: social, self-help, gross motor, fine motor, expressive language, comprehension, letters, and numbers. Each scale is scored by tallying the “yes” answers; a child who receives a score that is 1.5 SDs below the mean is graded as borderline, whereas a child scoring >2 SDs below the mean is graded as delayed. The parent needs ∼10 to 15 minutes to complete this screening instrument and ∼5 minutes to score it.
The CDI was reformatted for the purposes of this study. In its original form, the CDI is answered on a separate marking sheet such that the questions are found in one booklet and the answers are written in another. To make the process easier for parents, the CDI was retyped and presented in a format such that the parent can immediately read the question and then answer the item in the same booklet.
Over the recruitment period we received the names of 532 possible subjects from the database at the CCC. From this list, 84 children were excluded because (1) the children were no longer being followed at the clinic (60 children), (2) the children had moved (12 children), (3) the parents did not speak or read English (8 children), (4) the children had an established developmental delay (2 children), (5) the study terminated before the 18-month appointment had been booked (1 child), or (6) the child had passed away (1 child). Of the 448 remaining children, we were unable to contact by telephone 92 of them; therefore, they also were not included in the study. We successfully contacted the remaining 356 children, and of these, 317 families (90%) agreed to participate over the telephone and 39 families (10%) refused. Finally, of the 317 who had agreed over the telephone to participate at the forthcoming clinic visit, 5 returned the questionnaire incomplete and 64 did not return the questionnaire at all despite regular subsequent prompting as described. Those who had originally agreed to participate over the telephone but never returned the questionnaire were labeled “false participants.” Thus, the total number that completed the questionnaire was 248. Table 1 shows the demographic data of these 3 groups: true participants, false participants, and refusals.
The average age of the children who participated in the study was 18.4 ± 0.64 months. The data for gestational ages and gender of the children are summarized in Table 2.
Performance on Screening Tools
Of the 134 ASQs that were returned completed, 53 (40%) children failed at least one domain. The domain that the infant failed most often on the questionnaire was communication (46), followed by the problem solving, (7) gross motor, (4) social, (4) and fine motor (3) domains. Of the 53 children who failed the screening tool, 41 continued study participation and underwent BDI assessment (ie, 12 dropped out).
Conversely, of the 114 children who were assessed by the CDI, 11 failed (failure rate: 9%). Expressive language was again the most likely domain to be failed (5), followed by the gross motor, (5) fine motor, (1) self-help (1), and language comprehension (1 failure) domains. Of the 11 children who failed the CDI, 8 continued study participation and undertook BDI assessment (ie, 3 dropped out).
In total, 101 BDI assessments were performed. Twenty-nine children received a failing score. Similarly, the domain that was failed most often was communication (17), followed by the gross motor (11), cognition (6), personal-social (3), and adaptive (1) domains. Of the 101 BDIs performed, 49 were children who had previously failed one of the questionnaires, whereas 52 infants were control subjects. Of the 101 BDI tests performed, 41 failed ASQs and 8 failed CDIs were assessed with the BDI, and the remaining 52 BDI assessments were performed on control subjects.
Ease of Questionnaire Completion
Closer examination was given to the rates of completion and the parents' opinion of the questionnaires to assess feasibility and ease of use. Overall, the CDI was more likely to be either returned late (27% for CDI vs 22% for ASQ) or not returned at all (23% for CDI vs 17% for ASQ). This difference was to be expected, because the CDI contains more items to answer and, thus, is more labor intensive. Overall, 81% of the ASQs were returned completed, and 75% of the CDIs were returned completed.
The Likert-type questionnaire was used to assess the parent's opinion of the questionnaires. The options (very easy, easy, neutral, difficult, and very difficult) were given the numerical values 1 through 5, respectively. The ASQ received a mean value of 1.5 ± 0.6, whereas the CDI received a mean value of 1.6 ± 0.7. Therefore, the majority of the parents ranked the questionnaires as either very easy or easy, with the difference between the ASQ and the CDI being negligible. Table 3 summarizes these data.
Psychometric Values of the Screening Questionnaires
Using the BDI as the gold standard, we assessed the ASQ and the CDI for their psychometric properties: sensitivity, specificity, positive predictive value, and negative predictive value. Not one of these questionnaires proved to be an ideal screening instrument. The ASQ had moderate sensitivity (0.67) but poor specificity (0.39). Conversely, the CDI had poor sensitivity (0.50) but excellent specificity (0.86). Table 4 summarizes the predictive values for the ASQ, CDI, and pediatrician's questionnaire.
Psychometric Values of Incorporating the Pediatrician's Opinion With the Parent-Completed Questionnaires (ASQ and CDI)
The physician's opinion was incorporated with the parent-completed questionnaires to determine if this could be used as a potential modifier. As a result, those children (1) who had failed the parent-completed questionnaire and (2) for whom concern was listed by the pediatrician were grouped into a new category as those who had failed both screening instruments. Likewise, those (1) who had passed the parent-completed questionnaire and (2) for whom no concern was listed by the pediatrician were grouped into another category as those scoring normally on both tests. Children who did not fit the criteria for either of these 2 groups (ie, they had passed the questionnaire and failed the pediatrician's questionnaire or vice versa) were not included in the calculation. The 2 groups (one representing children who had passed both questionnaires and the other representing the children who had failed both questionnaires) were compared with the results of the BDI to obtain predictive values. Incorporating the pediatrician's opinion with the results of the questionnaire did not improve the predictive accuracy of either questionnaire substantially. Table 5 summarizes the results of comparing the combined result of the pediatrician's opinion and the screening tool with the BDI.
To be comprehensive, different analyses, such as moving the BDI cutoff to 1 SD below the mean, removing failed communication domains on the ASQ as a potential confounder, and redefining a failing score for the ASQ as 2 SDs below mean on ≥2 domains, were attempted. None of these posthoc manipulations had a beneficial effect on predictive values, and the results are not presented herein.
To determine the usefulness of parent-completed questionnaires in a practical setting, 3 issues were examined: (1) the feasibility of using parent-completed questionnaires in the waiting room; (2) the role of the pediatrician's impression in the screening process; and (3) the accuracy of 2 parent-completed screening measures (the ASQ and CDI). Each of these issues can be addressed separately in the context of our results.
The participation rate was encouraging. The majority of parents (ASQ: 95%; CDI: 92%) found the questionnaire either easy or very easy to complete. The completion rate was also high for both questionnaires. Observed completion was lower for the CDI (75%) than the ASQ (81%), which was to be expected because the CDI is longer and more detailed.
More than 75% of the contacted population was assessed with a standardized screening tool, representing an improved rate over what is currently being achieved with practitioner-administered tests. (Only 23% of primary care physicians are regularly using a standardized screening tool.)13 Most important to note is that these questionnaires were administered in the clinic at the time of the patient's appointment and did not need the assistance of either office staff or the physician. Because these tests are also cost-effective, the 3 most frequently mentioned complaints associated with a screening test administered by health care professionals (lack of time, lack of staff, and cost) were overcome.
It is also important to note that the overall completion rate may have been decreased by the additional burden placed on the parent from the 4 additional forms placed in the dossier to be completed. Nonetheless, simple endeavors can be conceptualized to increase the completion rate with active endorsement and encouragement by the primary care physician and clinic staff. Thus, it seems feasible for parents to complete developmental screening questionnaires in the waiting room while waiting to be seen by their child's physician.
Pediatrician's Opinion as a Potential Modifier
The pediatrician's opinion did not improve the predictive values of the questionnaires used. The pediatrician's opinion had good specificity but poor sensitivity; thus, the use of their clinical judgment tended to underidentify developmental impairments. Indeed, the pediatrician's questionnaire and the CDI had very similar properties.
It is disconcerting that combining the results of a standardized screening tool with the pediatrician's opinion could not produce enhanced and ultimately acceptable sensitivity and specificity values. Perhaps this indicates the lack of proper health care training in regards to childhood development. Indeed, the literature has emphasized that there is a gap between the knowledge and skills required in providing developmental services and the limited training that many clinicians receive in this area. This is a problem that seems to have persisted over decades. In 1979 Shonkoff et al lamented that “four fifths of physicians … viewed their formal training in this area as inadequate… . [A]lmost 2/3 did not feel that practical experience was an adequate substitute for formal training in developmental assessment skills.”3 This reality still persists >20 years later. A 2000 study performed by the AAP found that 64% of primary care physicians reported inadequate training in developmental assessment.31 As a result, an improvement in screening accuracy does not just necessitate more accurate developmental screening tests but also a joint effort from different facets of the professional, public, and research community advocating for more-thorough training in childhood development and impairments.
Accuracy of the ASQ and CDI
The psychometric properties of the ASQ and CDI were surprising, because several studies for both the ASQ and the CDI have advocated their accuracy in screening infants for developmental delay. A factor that may have reduced the accuracy of the ASQ and CDI in our study was the time difference between screening the child and performing the full assessment. Unlike other projects in which the gold standard was applied immediately after the screening instrument, our study imposed a 3-month waiting period representing the “real-life” delay typically experienced between screening and referral. Previous studies have demonstrated that 3 months represents a time interval within which an infant's developmental status may change. Darrah et al32 found that, when assessing fine motor skills, gross motor skills, and communication, the majority of children tested serially had 40 to 60 percentile fluctuations in performance. Changes of this magnitude took place at least once within 5 testing points (ie, 9, 11, 13, 16, and 21 months of age). Consequently, a wait of 3 months could result in a different developmental estimate. Because of the instability and discontinuity of a child's development, what might have been identified as a transient weakness at 18 months could have resolved itself by 21 months. The developmental domain of communication also seems to be too variable at the 18-month age to provide a reliable indicator between normality and abnormality. At this age, the “normal” differences in the actual times that communication milestones are reached by children are too variable,2 which demonstrates the importance of the time of actual assessment and the advantage of regular, ongoing multiple screening efforts.
Overall, neither test provided acceptable screening properties at 18 months of age; therefore, a screening protocol suitable to the setting of a clinic cannot consist solely of either instrument applied at one point in time. However, our results demonstrate that the CDI identified children with normal development with high accuracy (ie, high negative predictive value). Future studies can incorporate this particular strength within a screening procedure. Moreover, we found that a “one-shot” screening protocol might not be sufficient to identify developmental delay accurately. Indeed, the AAP31 and other researchers have come to the same conclusion. For example, Darrah et al32 concluded that screening should involve multiple time points and multiple domains before referrals are made. A similar statement was issued by the AAP Committee on Children With Disabilities.2
This study has several limitations. First, the ASQ was used as an in-clinic questionnaire; although the manual states that it can be used in this fashion, this screening instrument was originally constructed as a take-home questionnaire so that the parent could attempt to address the items over several days with the child. By limiting the time to ∼15 minutes, with the parent relying on past experiences to answer the questions, the accuracy of the instrument could have been somewhat compromised. Second, only 36 subjects underwent both the CDI and BDI. This limits the ability to evaluate the accuracy of the CDI, in particular with respect to sensitivity, because there were few subjects who actually failed the CDI. Thus, a larger sample size may be needed before drawing clearer conclusions regarding the CDI's performance as a screening tool. Finally, the community in which we were conducting the study was predominantly middle class with a high school education and some postsecondary attendance. Our ability to generalize the results to the general population, particularly to disadvantaged populations, is limited.
This study does lay the foundation for future studies on screening. First, and most important, there is the possibility of assessing the benefits of ≥2 serial screening tests at different points in time. Because the developmental trajectory, especially at younger ages, is subject to large variability, a second screening effort (either completed in-clinic or sent by mail) has the potential to improve the accuracy of the initial identification process. To decrease the amount of time used to score the questionnaire, application of computer software could be implemented (ie, questionnaire completed and scored on a hand-held device). This would reinforce the practicality of these types of screening instruments by decreasing the effort required from office staff and personnel. Finally, a study conducted on a more representative population incorporating a larger sample size would clarify issues of applicability, feasibility, and yield. With additional research, the knowledge gap between available screening tests and developing a practical protocol for their application can be bridged. Such efforts would have considerable clinical implications with respect to systematically enhancing the capability for earlier identification of children with developmental delays, which should expedite intervention and, theoretically, optimize eventual outcome.
Dr Shevell is grateful for the support of the Montreal Children's Hospital foundation during the writing of this manuscript. This project was supported by a clinical research project grant from the Montreal Children's Hospital Research Institute.
- Accepted May 23, 2006.
- Address correspondence to Michael I. Shevell, MD, CM, Montreal Children's Hospital, Room A-514, 2300 Tupper, Montreal, Quebec, Canada H3H 6P3. E-mail:
The authors have indicated they have no financial relationships relevant to this article to disclose.
- ↵Boyle CA, Decoufle P, Yeargin-Allsoop MY. Prevalence and health impact of developmental disabilities. Pediatrics.1994;93 :863– 865
- ↵American Academy of Pediatrics, Committee on Children With Disabilities. Developmental surveillance and screening of infants and young children. Pediatrics.2001;108 :192– 196
- ↵Shonkoff JP, Dworkin PH, Leviton A, Levine MD. Primary approaches to developmental disabilities. Pediatrics.1979;64 :506– 514
- ↵Ireton H, Glascoe FP. Assessing children's development using parent's reports: the Child Developmental Inventory. Clin Pediatr (Phila).1995;34 :248– 255
- ↵Dearlove J, Kearney D. How good is general practice developmental screening? BMJ.1990;300 :1177– 1180
- ↵Glascoe FP. Parents' concerns about children's development: prescreening technique or screening test. Pediatrics.1997;99 :522– 528
- Glascoe FP, Byrne KE, Ashford LG, Johnson KL, Chang B, Strickland B. Accuracy of the Denver-II in developmental screening. Pediatrics.1992;89 :1221– 1225
- ↵Squires J, Bricker D, Potter L. Revision of a parent-completed developmental screening tool: Ages and Stages Questionnaires. J Pediatr Psychol.1997;22 :313– 328
- ↵Sand N, Silverstein M, Galscoe FP, Gupta VB, Tonniges TP, O'Connor KG. Pediatricians' reported practices regarding developmental screening: do guidelines work? Do they help?Pediatrics.2005;116 :174– 179
- ↵Academy of Pediatrics, Division of Health Policy Research. Periodic survey #53: identification of children <36 months at risk for developmental problems and referral to early identification programs. 2003. Available at: www.aap.org/research/periodicsurvey/ps53exs.htm. Accessed July 14, 2004
- ↵Knobloch H, Stevens F, Malone A, Ellison P, Risemberg H. The validity of parental reporting of infant development. Pediatrics.1979;63 :872– 878
- Glascoe FP, MacLean WE, Stone WL. The importance of parent's concerns about their child's behavior. Clin Pediatr (Phila).1991;30 :8– 14
- ↵Glascoe FP. It's not what it seems: the relationship between parent's concerns and children with global delays. Clin Pediatr (Phila).1994;33 :292– 296
- Glascoe FP, Dworkin PH. The role of parents in the detection of developmental and behavioral problems. Pediatrics.1995;95 :829– 836
- Glascoe FP, Foster M, Wolraich W. An economic analysis of developmental detection methods. Pediatrics.1997;99 :830– 837
- ↵Glascoe FP. Parents' evaluation of developmental status: how well do parents' concerns identify children with behavioral and emotional problems? Clin Pediatr (Phila).2003;42 :133– 138
- ↵Dobrez D, Sasso AL, Holl J, Shalowitz M, Leon S, Budetti P. Estimating the cost of developmental and behavioral screening of preschool children in general pediatric practice. Pediatrics.2001;108 :913– 922
- ↵Montgomery ML, Saylor CF, Bell NL, Macias MM, Charles JM, Katikaneni LD. Use of the Child Development Inventory to screen high-risk populations. Clin Pediatr (Phila).1999;38 :535– 539
- ↵Newborg J, Stock JR, Wnek L, Guidubaldi J, Svinicki J. Battelle Developmental Inventory: Examiner's Manual. Allen, TX: DLM LINC Associates; 1984
- ↵Squires J, Potter L, Mounts L. ASQ User's Guide for the Ages and Stages Questionnaire. 2nd ed. Baltimore, MD: Brookes; 1997
- ↵Ireton H. Child Development Inventories Manual. Minneapolis, MN: Behavior Science Systems; 1992
- ↵Centers for Disease Control and Prevention. Barriers to developmental screening according to pediatricians: results from AAP surveys of pediatricians. Available at: www.cdc.gov/ncbddd/child/DSbarriersrpt.pdf. Accessed August 11, 2006
- Copyright © 2006 by the American Academy of Pediatrics