To describe when and by whom concern is first expressed for children referred to rehabilitation because of neuromotor problems.
Study Design and Setting
We conducted a survey of parents of 92 children (aged 0-6 years) who were on the waiting list for physical or occupational therapy services at rehabilitation centers in Montréal, Québec. We compared age of child at initial concern with who first expressed concern for children who were considered at risk due to their perinatal history of prematurity and those who were not born prematurely but were later diagnosed as having neuromotor problems.
Parents were interviewed regarding their child’s medical history and utilization of health care services.
Parents were concerned later than physicians were regarding their child’s development (mean difference, 8.2 months; 95% confidence interval [CI], 3.7-12.6 months). There was no significant difference in time of recognition of problems between the premature (10.2 months) and full-term (11.9 months) groups. Even after controlling for risk group, parental concern occurred later than physician concern (β coefficient, 7.3; 95% CI, 2.5-12.2). The child’s age at the time of initial concern was associated with the child’s age at referral to rehabilitation (β coefficient, 0.04; 95% CI, 0.01-0.06).
Early recognition is important if a child is to benefit from early rehabilitation. It may be important to improve primary care screening of children for neuromotor problems and to increase parental awareness regarding normal motor development of their children. Prompt, simultaneous referral to medical evaluation and rehabilitation resources may decrease delays in rehabilitation.
Early rehabilitation intervention is believed beneficial in assisting children with neuromotor problems to develop to their full potential and prevent secondary complications.1 Therefore, early identification is important to initiate rehabilitation services as soon as possible. In many cases, not achieving age-appropriate motor milestones is a cause for concern and may be a marker of morbidity. For example, Johnson et al2 found that 56% of high-risk infants who were not walking by age 18 months had an associated abnormality diagnosed before age 3 years. Recognition of a neuromotor problem by either the parent or the physician may depend on the nature of the problem itself (eg, its severity), organizational characteristics (eg, access to primary and specialty care for investigation), and family-related factors (eg, socioeconomic status, language, and place of residence).
Concern is often related to the history of the child because parents of high-risk children (eg, very low-birth-weight children) tend to underestimate their infant’s development by considering it even slower than the norm.3 Moreover, physicians tend to identify those with more severe delays than those with milder problems.4 Furthermore, the severity of the observed delay is the main reason for referral by a community pediatrician to a specialist for investigation and confirmation of the diagnosis.5
Parents may be accurate in their ability to detect developmental problems in their children. Parental sensitivity regarding their child’s development has been found to be 80%, and specificity, 94%.6 A study7 of 31 high-risk infants (aged 3-9 months) who were assessed by their parents (using a 4-item questionnaire) and a professional physical therapist (using the Bayley Scales of Infant Development) indicated a sensitivity of 80% and a specificity of 91% for parental assessment of motor development. Although parental knowledge of child development in general has been previously linked with greater age, higher educational level, and higher income in the parents,8 Glascoe et al6 did not find an association between parental concerns and demographic factors.
Even after initial concern is expressed, there may be substantial waiting times before services are received (both diagnostic/evaluative and rehabilitation services). Shevell et al5 found that the mean wait was 15.5 months between initial parental concern and specialty assessment by a neurologist or developmental pediatrician in their sample of patients in Montreal, Quebec. The mean age at initial parental concern for children with neurodevelopmental problems was 22.9 months (SD, 11.4 months) and ranged from 11.7 months for a motor delay to 27.3 months for a speech delay. Parental concern did not increase the probability of referral of children with neurodevelopmental problems to rehabilitation or for further medical consultation.5,9
One study10 reported that the diagnosis of the referring physician was confirmed in two thirds of cases referred to a pediatric neurologist. Establishing a diagnosis is important because determination of an etiology for children with developmental difficulties is highly associated with the receipt of occupational and physical therapy rehabilitation services.11 Factors associated with shorter waiting times between referral of patients from the Montreal Children’s Hospital, Montreal, and receipt of rehabilitation services included younger age of the child, living in the city as opposed to the suburbs, and, possibly, higher parental empowerment.12
The objective of our study was to determine the child’s age at the time of initial recognition of the problem (“age at initial concern”) by either the parent or the physician in a cohort of children with neuromotor disabilities (aged 0-6 years) who were referred to rehabilitation. Factors related to age at initial concern and who expressed that concern were explored, including whether the child was born prematurely (ie, considered to be at higher risk for developmental problems), who expressed initial concern (parent or physician), parental empowerment, burden of care, socioeconomic status, child sex, presence of older children in the family, language, and place of residence. In addition, we evaluated whether age at initial concern influenced age at referral to rehabilitation services.
This study was carried out in the context of a larger study assessing access to rehabilitation services among children with physical disabilities. The study population for the larger study included 224 parents of children (aged 0-6 years) who were referred for neuromotor problems (gross and/or fine-motor developmental problems) from 2 Montreal pediatric tertiary care hospitals to outpatient physical or occupational therapy services at local rehabilitation centers from September 1, 2002, to February 28, 2004. Children with predominantly developmental language or cognitive problems were excluded. Of the 224 patients, 144 who were still waiting for services from the local rehabilitation center provided us with a medical history; of these, 92 were referred for neuromotor problems and did not have an identifiable pathologic condition (eg, genetic syndrome, seizure disorder, or other neurological problem). This last group of 92 children constituted the final sample. Families were identified from the hospital central referral database, which included each child’s medical diagnosis, date of birth, address, and telephone number. Parents were contacted and asked to participate in the study; those who consented were interviewed in person by one of 2 research therapists (M.C. and L.G.) working on the project.
The interview included questions regarding the medical history of the child, the child’s diagnosis (according to the parent), who first noticed that there was a problem (“Who was the first person or professional who expressed concern regarding your child’s development?”) and when this was noticed (“When did this person express concern, ie, at what age of the child?”), current use of services, burden of care (as a proxy for severity), family empowerment, date of referral to rehabilitation services, and demographic data. Burden of care (severity) was measured by the WeeFIM (Functional Independence Measure for Children),13,14 a reliable and valid tool that is scored on 3 subscales (self-care, mobility, and cognition) in addition to a total score, with lower normalized scores or quotients representing higher levels of disability. Scores may be categorized as mild (76-100), moderate (50-75), or severe (<50).15 Family empowerment was measured by the service system subscale of the Family Empowerment Scale,16 a reliable and valid scale that measures empowerment on a scale of 0 (least empowered) to 60 (most empowered).
Analysis consisted of descriptive statistics to report means and standard deviations for continuous variables, and proportions for categorical variables. Children were classified as being at higher risk for neuromotor problems if they were born prematurely. Unpaired, 2-tailed t tests were used to evaluate differences in age at initial concern and age at referral to rehabilitation services according to risk group, socioeconomic status (household income and maternal education), sex, language (English or French), place of residence (city or suburb), severity, presence of older children in the family, and who expressed initial concern about the child (physician or parent). Prematurity was defined by the referring medical diagnosis. WeeFIM scores were computed using the child’s corrected age. However, we used the child’s uncorrected age for time of initial concern and time of referral to rehabilitation because children born prematurely are usually monitored more closely from birth and because it is this specific time that is of interest and not the time from the child’s corrected age.
We used χ2 tests to evaluate differences in who expressed initial concern about the child according to risk group, socioeconomic status, child’s sex, language, and place of residence. We used linear regression analysis to model age at initial concern as a function of risk group, who expressed initial concern, severity, socioeconomic status, presence of older children in the family, child’s sex, language, and place of residence. Logistic regression was used to model the dichotomous dependent variable: the person who expressed initial concern (parent or physician) as a function of the various factors. Linear regression was used to determine whether age at initial concern and age at referral to rehabilitation were associated, adjusted for risk group, severity, and family-specific and demographic variables.
The study was approved by the research ethics boards of the 2 pediatric hospitals. All parents who agreed to participate provided informed consent.
The characteristics of our sample of 92 children and their parents are given in Table 1. The mean age of our sample was 2.8 years (SD, 1.4 years; age range, 0.3-4.9 years) and the average age at initial concern was 1.0 year (SD, 0.9 year; age range, 0-3.0 years). The mean age at referral to an outpatient rehabilitation center was 2.6 years (SD, 1.4 years; age range, 0.05-4.8 years). WeeFIM quotients for the total scores averaged 73.6 (SD, 15.8; range, 19.8-116.6). The mean parental empowerment score on the services subscale of the family empowerment scale was high, averaging 49.6 (SD, 7.0; range, 27-60).
On the whole, children whose parents were the ones to express initial concern were older than those whom physicians detected initially, and concern was expressed earlier for female children. Neuromotor problems in children who were born prematurely were recognized only slightly earlier than those who were born at term, and the result was not statistically significant. Parental vs physician concern differed by whether the child was premature (P = .002): only 4 (20%) of the 20 children in the premature group were detected by the parent, whereas 42 (58%) of the 72 children in the full-term group were detected by the parent. Neither severity nor socioeconomic status (household income and maternal educational level) was associated with who was initially concerned; however, problems of children in French-speaking families, more so than those in English-speaking families, tended to be recognized by the physician rather than the parent (P = .005). Moreover, the children from the French pediatric hospital were recognized as having problems earlier than those from the English pediatric hospital (8.7 months vs 14.1 months; 95% confidence interval [CI] for mean difference, 0.6-10.0 months). Correspondingly, the age at initial concern was earlier for French-speaking families than for English-speaking ones (8.9 months vs 15.9 months; 95% CI for mean difference, 2.4-11.7 months). Table 2 summarizes aspects of the univariate analyses.
Multiple linear regression indicated that older age at initial concern was associated with parental (as opposed to physician) recognition of the problem (β coefficient, 7.3; 95% CI, 2.5-12.2) and with male sex of the child (β coefficient, 5.0; 95% CI, 0.2-9.9). A logistic model was calculated with person concerned as the dependent variable (parent vs physician) (Table 3). Parents were less likely than physicians to be concerned about children who were in the premature group (odds ratio, 0.12; 95% CI, 0.03-0.47). Neuromotor problems in children whose families spoke English rather than French tended to be detected by parents (odds ratio, 4.4; 95% CI, 1.4-13.8).
Finally, we found that age at initial concern was associated with age at referral to rehabilitation services even after controlling for prematurity, severity, maternal education, child sex, presence of older children in the family, city vs suburban residence, language, and empowerment. In addition, initial parental concern was associated with later referral to rehabilitation services (Table 4).
In our study, parents recognized neuromotor problems in their children at older ages than did physicians. This was true even when controlling for prematurity. Earlier recognition of problems was associated with earlier referral to rehabilitation services.
As in most epidemiological studies, there were several limitations. First, our entire sample of children was referred to rehabilitation services at local rehabilitation centers (they were on the waiting list for such services). Nevertheless, the mean age at initial concern in the present study (1 year) was comparable to that reported in a large American representative sample of children referred to early intervention services for developmental delay (11 months).17 However, the children in the American study were limited to those who began early intervention by 31 months of age, while our study included children who were referred to rehabilitation services until 6 years of age, which may explain the slightly higher mean age at initial concern in our study. The age at initial concern in the present study was younger than the 19.2 months described by Shevell et al5 in another Canadian study. Our subjects were all referred to rehabilitation services (either by pediatricians, neurologists, orthopedists, or other specialists who had established a need for rehabilitation services for the child), whereas their subjects were sent for diagnostic consultation to a neurologist, possibly indicating that the sample in the present study had more severe delays. A second explanation for the difference is that our study included children with neuromotor disabilities who may have had concomitant cognitive or speech and language problems but must also have had a physical disability to be included in the study, whereas those in the study by Shevell et al5 could have had any type of developmental delay. In most cases, neuromotor disabilities manifest earlier than speech, cognitive, and behavioral problems do.18
Parent self-reports may have introduced an information bias. However, because of the young age of the children, recall was relatively recent, possibly minimizing the effect of such bias. Severity (WeeFIM score) was assessed at the interview and not at the time of initial concern. Manifestation of severity may be more evident at older ages. If this is true, then it is doubtful that severity at the time of initial concern would be associated with any of the outcomes assessed, especially since severity manifested later (ie, at the time of referral to rehabilitation) was not associated with either who expressed concern, the age at initial concern, or the age at referral to rehabilitation services. In addition, the WeeFIM may lack sensitivity to assess severity for children at very young ages (6-12 months) (D.E.F., M.C., L.G., M.-N.S., L.A., and J.G., unpublished data, May 2004). Similarly, parental empowerment was evaluated at the time of referral to rehabilitation and may have changed from the time of initial concern. In agreement with Glascoe et al,6 we found no association between parental concern and demographic variables. We also did not find an association between parental empowerment and referral to rehabilitation. This is concordant with the literature reporting that the expression of parental concern did not increase the likelihood of referral to diagnostic5,9 or treatment9 services. Although socioeconomic status (maternal education and household income) was not related to age at initial concern, Bailey and associates17 noted that families with higher socioeconomic status were more likely to discuss their concerns with a medical professional than were families in lower socioeconomic levels.
We found that problems were detected earlier for girls, whereas Shevell et al5 observed a male preponderance with respect to referral to specialists for diagnostic evaluation. However, in their sample, boys tended to have more developmental language disorders and autistic spectrum disorders. Our study included children with motor disabilities, and we excluded children for whom cognitive, language, or autistic spectrum disorders were the main reason for referral. Sices et al9 surveyed physicians by using vignettes and reported that physicians were more likely to refer 18-month-old girls with language delays to audiology than they were to refer boys. However, they found no differences for hypothetical referral of 18-month-old girls and boys with gross motor delays.
The problems of children from French-speaking families tended to be recognized by the physician rather than by the parent and sooner than those of children from English-speaking families. Children from the French pediatric hospital were diagnosed as having developmental problems at a younger age than those from the English pediatric hospital, but this difference was no longer statistically significant in the linear regression analysis. Although we did not ask parents whether they have a regular family physician or pediatrician, those who have a regular physician may be more likely to have problems recognized and diagnosed sooner. Another possibility is that French-speaking families or physicians from the French hospital tend to be concerned about neuromotor problems earlier as opposed to taking a “wait-and-see” attitude.
We found no difference in age at initial concern between children who were considered to be at higher risk for developmental disabilities (born prematurely) and those born at term. Our initial hypothesis was that the neuromotor problems of children who were born prematurely would be recognized sooner. The small sample size may account for the lack of a significant difference between the time of initial concern in these 2 groups. Nevertheless, even after controlling for prematurity, severity, sex, socioeconomic status, empowerment, and demographic factors, we found that age at initial recognition of a problem was related to age at referral to rehabilitation services. Moreover, if the parent was the one to express initial concern, referral to rehabilitation services tended to be later. The converse of this is that children whose problems were detected initially by physicians tended to be sent for rehabilitation earlier. This is consistent with the finding of Majnemer et al11 that children who had an established diagnosis tended to be referred to rehabilitation services. We found an average delay of approximately 20 months between initial concern and referral to rehabilitation. This delay may result from health services organization–related factors such as lengthy delays for appointments to specialists5; physician-related factors such as perception of severity, diagnostic uncertainty, and beliefs regarding the efficacy of physical therapy19; and parental-related factors such as denial or nonadherence to medical recommendations.20
Early recognition and diagnosis is important if a child is to benefit from early rehabilitation. Since parental concern tended to occur later than physician concern and was associated with longer delays for rehabilitation, a possible solution may be to implement systematic developmental screening in primary care.21 The use of short developmental screening questionnaires completed by parents could help in early identification of children who are at a higher risk of developmental disabilities. Parental awareness regarding normal child development should also be increased, and parental concerns should be dealt with appropriately. Moreover, delays in evaluation by specialists and consultants, who may begin the process of therapy, should be minimized to enhance the functional abilities of these children. One solution may be to move away from a “linear” model, which involves initial concern prompting medical evaluation and, only then, referral to rehabilitation services, toward a “dual-track” model, in which a referral is made to medical evaluation and rehabilitation resources simultaneously.
Correspondence: Debbie Ehrmann Feldman, PT, PhD, Université de Montréal, Canada, École de Réadaptation, CP 6128 Succ Centre-Ville, Montréal, Québec, Canada H3C 3J7 (firstname.lastname@example.org).
Accepted for Publication: April 11, 2005.
Funding/Support: This study was supported in part by Fonds de la recherche en santé du Québec, Montréal, and the Canadian Institutes of Health Research, Ottawa, Ontario. Dr Feldman was the recipient of a new investigator award from the Institute of Gender and Health of the Canadian Institutes of Health Research, Ottawa, and is currently funded by The Arthritis Society, Toronto, Ontario.
Acknowledgment: We thank Barbara Mazer, PhD, for reviewing the manuscript.
Ehrmann Feldman D, Couture M, Grilli L, Simard M, Azoulay L, Gosselin J. When and by Whom Is Concern First Expressed for Children With Neuromotor Problems?. Arch Pediatr Adolesc Med. 2005;159(9):882-886. doi:10.1001/archpedi.159.9.882