Objective. Due to the limitations of previous studies, parents' concerns have been recommended as a prescreening technique, a brief method for identifying a subset of children in need of more in-depth developmental screening. The purpose of this study was to assess whether parents' concerns could: (1) serve instead as a screening measure; (2) aid in making focused referral decisions; and (3) help pediatricians target families for developmental promotion and in-office counseling. An additional goal was to determine why most parents' concerns are accurate although some are not.
Design Survey/Setting. Public schools and day care centers in four diverse geographic sites representing the northern, central, southern and western United States.
Patients and Other Participants. A total of 408 children between 21 and 84 months of age and their parents, whose socioeconomic and demographic characteristics reflect proportions in the 1990 United States Census.
Main Outcome Measures. Licensed psychological examiners and educational diagnosticians elicited parents' concerns about children's development and measured children's development with a broad battery including measures of intelligence, language, motor, and school skills.
Results. Certain concerns, ie, motor, language, global/cognitive, and school (in children 4 years and older) had high levels of sensitivity and identified 79% of the 56 children with disabilities. Accurate referrals could be made for 70% of the 56 children. The absence of concerns or concerns in other areas, ie, socialization, self-help, or behavior, had reasonable specificity and identified 72% (N = 255) of the 352 typically developing children. Of the remaining 28% (N = 97) of parents with significant concerns but whose children did not have disabilities, most had children with substantially lower performance in almost all developmental areas than the children of the 255 parents without significant concerns. Further, more than half of the 97 children could be distinguished by a single concern (usually about expressive language) whereas the majority of accurately concerned parents had multiple concerns. A significant proportion of the 12 parents of disabled children who did not raise concerns could be identified by difficulties communicating in either English or Spanish.
Conclusion. If systematically elicited, parents' concerns approach standards for screening tests and can be used to make reasonably accurate referral decisions. Over-referrals can be significantly reduced by administered screening tests to the small group of children (16%) whose parents have a single significant concern. Those who pass screening or whose parents have nonsignificant concerns can be targeted for developmental promotion and in-office counseling. Under-referrals can be minimized by administering screening tests (with the help of an interpreter as needed) to children whose parents have communication difficulties.
Previous research illustrated a strong relationship between parents' concerns and children's developmental status; certain kinds of concerns (eg, fine motor or language) were found to sensitively detect developmental difficulties. At the same time, the absence of concerns or concerns in other areas (eg, self-help or socialization) correctly identified the majority of children without problems.1-4 Although parents' concerns approached standards for screening test accuracy (sensitivity and specificity close to 80%),5-7 they have not been recommended for screening but rather for prescreening—a very brief method that identifies a subset of children in need of screening with more thorough measures. The reasons for recommending parents' concerns as a prescreening tool came from the limitations of prior studies: sample sizes were small (100 to 157 children), all subjects resided in Tennessee, and in most studies, concerns were compared to screening tests rather than comprehensive assessments. Thus, one purpose of our study was to rectify the limitations of previous research and reconsider whether parents' concerns were sensitive and specific enough to serve as a screening tool, ie, whether referrals for further work-ups could be made accurately on the basis of parents' concerns alone. An additional goal was to assess why some parents are accurately vs inaccurately concerned or nonconcerned. Finally, because physicians are encouraged to do more with regard to development than simply screen and refer, a final purpose for this study was to view the extent to which eliciting parents' concerns can aid physicians in identifying families in need of suggestions for developmental promotion and in-office counseling.
As part of a larger validation study of several other developmental measures,8,9 four sites were selected to represent the broad geographic regions of the United States: North (Plymouth, MA); Central (Denver, CO which is within 250 miles of the geographic epicenter of the United States); South (Tampa, FL); and West (Carson City, NE). Two sites (Plymouth and Carson City) are relatively small towns and Carson City, although bordered by the city of Reno to the north, is otherwise surrounded by rural areas. The remaining two sites were exclusively urban. Within each site, one to two schools were identified that had a mixture of children from various socioeconomic backgrounds (approximately one-third of students participated in the Federal free/reduced lunch program). A single kindergarten and first grade classroom were identified at each school and students were recruited by sending informed consent letters to their parents. The 2- through 4-year siblings of any kindergarten or first grade student in each school were also recruited. In two sites, there were insufficient numbers of younger siblings, therefore recruitment was extended to children attending preschool programs in the zone of the targeted elementary schools. These preschool programs: (1) had Federal or local funding subsidies; (2) served children from varying socioeconomic backgrounds; and (3) were neither oriented for special education students nor exclusive of children with known disabilities. Overall the recruitment procedures helped ensure that the validation sample included children and families representative of the United States as a whole.
A total of 408 children participated in the study with an average of 102 at each site. They represented at least 80% of children in the targeted classrooms and their siblings. For all socioeconomic and demographic variables, distributions were similar to those found in the 1990 United States Census except where noted. Participating parents, 91% of whom were mothers, had completed an average of 13.0 grades: 16% completed fewer than 12 grades, 33% attended through grade 12, 24% attended but did not complete college, and 27% held college degrees. Of the parents 67% were employed full- or part-time and 66% were married. Twelve percent were both unmarried and unemployed. Children ranged in age from 21 to 84 months and 53% were boys. Free or reduced lunches were provided to 30% of subjects or their sibling(s). Thirty-five percent of children were not enrolled in educational programs such as preschool or day care. Racial and ethnic backgrounds varied somewhat from United States averages in that there were more Hispanic and fewer African-American participants. Distributions included: white (69%), Hispanic (20%), African-American (6%), and Asian or other races/nationalities (5%).
At each site, diagnosticians were recruited from among those employed by the public schools. In the Northern and Western sites, the diagnosticians were licensed psychological examiners or certified school psychologists and in the remaining sites were master's level educational diagnosticians. Upon receiving signed consent letters, examiners administered a battery of tests to the children and parents were asked to complete parental questionnaires at home. Diagnosticians were blinded to the results of the parental questionnaires but were asked to check for incomplete responses and to interview parents directly when needed. Children whose primary language was Spanish were tested exclusively in that language using standardized Spanish directions and their parents were given Spanish versions of the demographics questionnaire and other measures.
A broad concurrent battery was administered to each child to identify a range of developmental disabilities and to assess performance in each developmental domain. Table 1 shows the tests and provides a brief description of test content. All subjects were administered the same battery with the exception of children between 21 and 29 months who were too young to receive the selected achievement subtests from the Woodcock-Johnson. Demographic information was also collected from all parents including level of education, marital status, employment, numbers of other children, and perceptions of children's medical problems (rated as not serious, somewhat serious, or very serious).
These were elicited using the Parents' Evaluations of Developmental Status a previously validated questionnaire that requires less than 3 minutes to administer and consists of two questions.1-4 The first is open-ended, “Please tell me any concerns about how your child is learning, developing and behaving.” In previous research about 8% to 14% of parents raise concerns in response to this question.1,2 The second question probes developmental concerns in each domain, “Do you have any concerns about how she understands what you say? … talks? … .makes speech sounds? … uses hands and fingers to do things? … uses arms and legs? … .behaves? … gets along with others? … is learning to do things for himself/herself? … is learning preschool and school skills?” An additional 17% to 20% of parents have been shown to affirm concerns in response to this question.1,2Previous research illustrated that responses can be readily categorized into the various developmental domains with a high degree of inter-rater and test-retest reliability (88% and 88% agreement)1,2. The domains include: expressive and receptive language, self-help, socialization, behavior, gross and fine motor, academic, and global/cognitive skills. An additional category is used to capture medical or other issues (usually concerns about hearing).
Do Parents' Concerns Detect Children With and Without Disabilities?
To assess the accuracy of parents' concerns, children were grouped according to whether or not they met criteria for special education placement. The criteria were drawn from the Individuals with Disabilities Education Act, the Federal mandate that provides public school special education services. Table 2 shows the classification criteria. Of the 408 children, 86% (N = 352) were found to be performing within normal limits and 14% (N = 56) met eligibility criteria for special education services. These figures are similar to those of other prevalence studies.13,14 Of the 56 special education candidates, 13 had specific learning disabilities, 7 had mental retardation, 13 had physical impairments, 1 had health impairment, 1 had autism, 1 had hearing impairment, and 23 had speech-language impairments. Twenty-six of the 56 had more than one disability and 13 of the 56 were enrolled in speech-language, occupational, or physical therapy (only 1 of the 13 was enrolled in a special education classroom).
Certain concerns were highly related to developmental outcome. Concerns in the areas of receptive or expressive language, school skills (in children 4 years and older), global/cognitive skills, and fine motor skills continued to be strong predictors of developmental problems. In addition, this study showed that concerns about gross motor skills and medical/hearing status were also highly related to developmental problems. Taken together, concerns in any of these areas sensitively identified the majority of children with disabilities. Parents of normal children tended to have either none of the above concerns or concerns in other areas, ie, socialization, self-help, or behavior. This produced high levels of specificity. These results are shown in Table 3.
The accuracy of the significant concerns in detecting various types of disabilities was also scrutinized. The presence of significant concerns identified 73% of children speech-language impairments, 71% of children with mental retardation, 85% of children with specific learning disabilities, and 77% of children with physical or other impairments.
Can Parents' Concerns Help Determine Appropriate Referrals?
It is not enough to simply identify children in need of referral. Professionals involved in screening must also decide on the kinds of referrals needed. When considering eligibility for special education, children generally require one of two types of evaluations: (1) speech-language (to determine whether the criteria are met for certification as speech-language impaired); (2) and/or psychoeducational evaluations [to determine the presence of mental retardation, specific learning disabilities or physical/health impairments (a psychologist or educator must document that a medical problem interferes with the child's ability to benefit from the regular education curriculum by determining the presence of deficits in academics or adaptive behavior skills)]. To assess whether there were unique patterns of concerns associated with speech-language vs all other impairments (and hence the need for speech-language vs psychoeducational evaluations), the two broad types of diagnoses were correlated with the various kinds of significant concerns. This identified two patterns: 89% of children with speech-language impairments had parents with concerns about school performance or receptive language skills. At the same time, 62% of children with other disabilities had parents with other of the significant concerns, ie, global/cognitive development, medical/hearing, expressive language, gross or fine motor skills. Overall, 70% of children with disabilities could be accurately referred on the basis of these two patterns.
The presence or absence of significant parental concerns was then intersected with the presence or absence of children's disabilities to create four groups of parents, those who were: accurately nonconcerned, inaccurately nonconcerned, accurately concerned, and inaccurately concerned (also called highly observant for reasons explained below). These groups were then compared on all variables to assess reasons for differences in parents' concerns. The discussion of each group comparison is prefaced by the abiding research questions.
Why Do Some Parents Have Concerns Even Though Their Children Do Not Have Disabilities?
Of the 352 parents of normal children, 255 (72%) had no concerns or nonsignificant concerns (accurately nonconcerned) although 97 (28%) had significant concerns (inaccurately concerned). These two groups of parents and their children differed in several ways as shown in Tables4 and 5. Substantially more of the inaccurately concerned parents believed their children had serious or somewhat serious medical problems (χ2 = 24.1,P < .0001). Inaccurately concerned parents also raised more concerns about their children's skills in the areas of self-help, socialization, and behavior (χ2 = 35.1; 23.3; and 6.2,P < .01). Although their children did not qualify for special education services, they tended to have substantially lower scores on the socialization, fine motor, gross motor, expressive, and receptive language subtests of the Child Development Inventory (CDI). Their parents were more likely to endorse certain items on the Possible Problems list of the CDI including: “Eating problems–eats poorly or too much, etc.” and any of the four CDI items tapping perceptions of difficulties with articulation, vocal fluency or other language skills (χ2 = 16.1; 46.7; 43.8; 18.0; and 15.2,P < .001). Thus parents who at first glance appeared to be inaccurately or overly concerned were actually highly accurate observers of subtle developmental differences. For this reason, the inaccurately concerned group of parents were renamed “highly observant.”
Can Highly Observant Parents Be Discriminated From Those Who Are Accurately Concerned About True Developmental Disabilities?
When presented with a significant concern, it would be helpful to distinguish parents who are concerned about subtle developmental differences from those who are concerned about developmental disabilities (the former may not need a referral for evaluations and services although the latter clearly does). To assess differences between these groups, the 97 highly observant parents (whose children were not disabled) were compared to the 44 accurately concerned parents (whose children met criteria for special education services). Apart from the obvious differences in how their children performed on criterion measures, the two groups differed in several ways: parents with accurate concerns raised almost twice as many of the significant concerns as highly observant parents: 77% of the accurately concerned had multiple concerns and only 47% of the highly observant parents had multiple concerns. Of the highly observant parents with a single concern, 74% were worried about their child's expressive language skills. Accurately concerned parents were also more likely to have any of the nonsignificant concerns, ie, socialization, behavior, or self-help skills. Additionally, parents with accurate concerns endorsed more Possible Problems items on the CDI including: “Clumsy; walks or runs poorly, stumbles or falls” and“Disobedient; does not mind well, resists”(χ2 = 6.0; 6.9, P < .01).
Are There Differences in the Parents of Disabled Children Who Raise or Do Not Raise Significant Concerns?
Of the 56 parents whose children met criteria for special education services, 44 raised one or more of the significant concerns. Twelve had no or nonsignificant concerns. These groups differed in the following ways: parents with accurate concerns were also more likely to have nonsignificant concerns including self-help skills, socialization, and behavior. Inaccurately nonconcerned parents had children with significantly higher performance in gross and fine motor skills than did accurately concerned parents. Otherwise there were no differences between groups on the basis of sociodemographic variables, such as severity of children's problems, perceptions of children's health status, parents' education, income, family size, or prior enrollment in special education.
Can Parents With Disabled Children Who Do Not Raise Concerns Be Discriminated From Parents of Children With Typical Development Who Do Not Raise Concern?
The 255 parents who did not raise concerns and whose children were developing normally were compared to the 12 parents who also did not raise concerns but whose children had disabilities. Apart from the expected differences in children's functioning, accurately nonconcerned parents endorsed an average of 1.2 items on the CDI Problem checklist. Inaccurately nonconcerned parents raised an average of 2.3 items (F(1,266) = 16.72, P < .001). The latter were more likely to mention the following concerns: eating problems; aches and pains—earaches, stomachaches, headaches, etc.; stutters or stammers; dependent, clingy, or very upset about separating; and, does not pay attention—poor listener. Nevertheless inaccurately nonconcerned parents were not more likely to raise medical concerns or to perceive their children's medical problems as serious or somewhat serious. Finally, inaccurately nonconcerned parents were more likely not to speak English at home (χ2 = 7.31,P < .01). There were no differences in the two groups of nonconcerned parents on the basis of numbers of siblings in the home, birth order, race, eligibility for Chapter I services, parents' marital status, income, age at child's birth, level of education or in the multiplicity of psychosocial risk factors. Finally, there were no differences in the numbers of nonsignificant concerns raised, ie, social, self-help, or behavior.
Parents'concerns about language, motor, global/cognitive development, medical problems (particularly hearing status), and school skills in children 4 years and older, were highly sensitive predictors of developmental problems. The absence of concerns or concerns in other areas had high specificity in correctly identifying children without disabilities. As in previous research, parents' concerns approached standards for screening tests.1-7 Given the weight of evidence (despite variations in research settings, geographic locations, children's ages, parents' levels of education, etc), parents' concerns can be safely recommended for use as a screening tool. It is important to note, however, that across all studies, parents' concerns were systematically elicited using two standardized questions. Pediatricians wishing to make use of parents' concerns in early detection of developmental disabilities can expect a high degree of accuracy only by using the specific questions used in this study.
An important finding from the current data is that pediatricians can make reasonably focused referrals on the basis of parents' concerns. Most children with speech-language impairments (who thus require speech-language evaluations) had parents with concerns about school performance or receptive language skills. Most children with other types of disabilities had parents with other significant concerns, although this group would be over-referred somewhat for speech-language evaluations. Although 70% of children with disabilities could be accurately referred, pediatricians should supplement parents' concerns with other developmental information to decide whether children need psychoeducational evaluations with or without speech-language testing.
One of the most interesting findings in this study was that parents of children without disabilities who raised a significant concern, had children who performed substantially lower on almost all developmental measures than did the normally developing children of parents who did not raise significant concerns. Children of these highly observant parents tended to score below average in motor, receptive language, and socialization skills. Their parents also rated them as having more serious health problems and more difficulties with eating or sleeping than did the accurately nonconcerned group. This suggests that parents with concerns, but whose children appear to be developing normally, are not simply anxious or ill-informed about development. Rather, these parents appear to be noticing subtle developmental and behavioral problems. Targeting this group for developmental promotion and increased vigilance regarding developmental progress is recommended.
The data also showed that more than half of the highly observant parents had a single concern (most often about expressive language) whereas more than three-fourths of the accurately concerned parents had more than one concern. Administering screening tests that carefully measure language skills to children whose parents raise only a single significant concern should help reduce unnecessary referrals. Indeed, had this occurred in the current study, specificity would have risen to more than 80%. Further, the results of screening would also provide a baseline for monitoring the developmental progress of children whose problems are subtle and currently subclinical.
Another interesting finding, supported by the research of McCormick et al,15 is that parents' developmental concerns were often related to their perceptions that their children had health problems. In previous research on parents' concerns conducted in medical settings, this phenomena was not observed, undoubtedly because physicians had already addressed medical issues.1-3 That medical concerns led some parents to have concerns suggests that the accuracy of developmental concerns may improve when medical issues are addressed. Although the sensitivity and specificity of parents' concerns are quite similar when assessed in medical vs nonmedical settings, further research is needed on the effects of physician behavior on parents' concerns.
As with previous research, there were few significant differences in the accuracy of parents' concerns on the basis of family characteristics; parents with limited education or income were as able as more educated or wealthy parents to raise concerns that were predictive of developmental status.1-4 Similarly parents' concerns were accurate regardless of parenting experience: defined as children's birth order and numbers of children in the home. The latter continues to be surprising in that most parents, regardless of education or experience, compare their children to others as a basis for deciding whether or not to be concerned.16 This suggests that parents find opportunities outside their homes for comparing their children to others. Nevertheless, parents with four or more children were more likely to raise a significant concern that may reflect that large family size is a risk factor for developmental problems.17 The only other distinguishing characteristic of some parents who did not raise concerns about children with undetected disabilities (3 of the 12) were difficulties communicating in either English or Spanish. Although further research is needed on parental communication skills and their impact on parental concerns, it is advisable to administer screening tests (using an interpreter as needed) to children whose parents have language barriers or other communication difficulties, rather than rely on parents' concerns alone.
Implications for Practice
Parents' concerns appear effective as a screening tool, provided that concerns are systematically elicited and categorized in the manner we described.18 Once parents have mentioned a significant concern, physicians will need to scrutinize the pattern of concerns and add their clinical judgment to the process of making appropriate referrals. Administering a second screening test to children whose parents are not fluent in the physicians' language(s) or whose parents raise only a single significant concern will help minimize over- and under-referrals. This process also helps pediatricians target families for developmental promotion and in-office counseling, ie, those with nonsignificant concerns and those with a single significant concern but whose children pass screening tests.
- Received March 28, 1996.
- Accepted June 13, 1996.
Reprint requests to (F.P.G.) Division of Child Development, Department of Pediatrics, Vanderbilt University School of Medicine, 2100 Pierce Avenue, Nashville, TN 37232–3573.
- Glascoe, FP
- Glascoe FP,
- MacLean WE,
- Stone WL
- Glascoe FP
- ↵Barnes KE. Preschool Screening: The Measurement and Prediction of Children at Risk. Springfield, IL: Charles C Thomas; 1982
- ↵American Psychological Association. Standards for Educational and Psychological Tests. Washington, DC: American Psychological Association; 1985
- Frankenburg WK
- ↵Glascoe FP. Technical Report for the Brigance Screening Tests. North Billerica, MA: Curriculum Associates, Inc; 1997. In press
- Glascoe FP
- ↵Woodcock RW, Johnson MR. Woodcock-Johnson Psycho-Educational Battery-Revised. Test of Achievement. Chicago, IL: Riverside Publishing Co; 1990
- ↵Slosson R. Slosson Intelligence Test. East Aurora, NY: Slossen Educational Publications; 1983
- ↵Ireton H. Child Development Inventories. Minneapolis, MN: Behavior Science Systems; 1992
- Yeargin-Allsopp M,
- Murphy CC,
- Oakley GP,
- Sikes RK
- Algozzine B,
- Korinek L
- McCormick MC,
- Shapiro S,
- Starfield B
- Sameroff AJ,
- Seifer R,
- Barocas R,
- Zax M,
- Greenspan S
- ↵Glascoe FP. Listening to Parents: An Evidence-based Approach to Developmental and Behavioral Surveillance. Oxon, England: Radcliffe Medical Press, Ltd; In press
- Copyright © 1997 American Academy of Pediatrics