To evaluate the affect of families' attitudes about the appropriateness of discussing psychosocial concerns on pediatric providers' identification of psychosocial problems.
These data were collected as part of the Greater New Haven Child Health Study, New Haven, Conn. The study design was a prospective cohort.
Families were recruited from a stratified random sample of all primary care practices in the greater New Haven area. Nineteen of 23 invited practices agreed to participate including 2 prepaid practices, 2 neighborhood health centers, and 7 fee-for-service group and 8 fee-for-service solo practices.
All families of children aged 4 to 8 years who attended these practices during 2 separate 3-week periods (1 in fall 1987 and 1 in spring 1988) were invited to participate in the study. Families were invited to participate only once, on the first contact with any eligible child, using approved procedures. Of 2006 eligible families, 1886 (94%) chose to participate.
Main Outcome Measure
The outcome variable for these analyses is the identification of any behavioral, emotional, or developmental problem by the pediatrician on the 13-category checklist. Overall, pediatric clinicians identified 27.5% of children with 1 or more psychosocial problems.
Our data suggest that there is a great deal of discrepancy between what parents report is appropriate to do when their children have psychosocial problems and what they actually do when they recognize such problems in their children. Most (81.1%) believed it was appropriate to discuss 4 or more of the 6 hypothetical situations with their children's physician, while only 40.9% actually did discuss any of these problems with a physician when a problem occurred. Given the correlates of parents who intended to discuss such problems (higher education, older age, Euro-American ethnicity, higher income, married, availability of medical insurance) the possibility that parents are providing socially acceptable responses to such questions seems likely. Further, our data indicate that parents' actual reports of discussions of psychosocial problems is unrelated to whether physicians identified those problems in children.
Pediatricians' judgments about the presence of psychosocial problems in their young patients seem to be based on their own observations rather than on what parents report. Physician-parent communication about psychosocial problems will be increasingly important as primary care physicians assume their role as gatekeepers to more expensive services such as mental health interventions.
COMMUNITY-BASED epidemiological studies of children have documented high rates of mental health problems and use of specialty mental health services that is lower than what would be expected from the community prevalence rates.1- 4 These findings, coupled with the documentation by Regier et al5 that 15% of the US population with mental disorders were most often treated (60%) in the general medical sector, have focused attention on the role of the general pediatric medical sector as the de facto mental health care system for children.5,6 Consequently, considerable research has focused on the prevalence of mental health problems in pediatric primary care practice, the accuracy of pediatricians' diagnoses when compared with either a standardized instrument or a mental health specialist, and the treatment or referral of mental health problems once they are identified.7,8 The results of these studies are reasonably consistent: a large proportion of children have identifiable mental health problems that are undiagnosed and, therefore, untreated. A recent study suggests that, when using a classification system appropriate for pediatric primary care (ie, using a system that emphasizes the behavioral, familial, social, and school-related concerns that the Task Force on Pediatric Education recommended as part of the education for pediatricians rather than psychiatric morbidity) clinicians recognize problems in many school-aged children.6,9 Furthermore, prior studies examining the identification of psychosocial issues in primary care pediatric practices have ignored practice- and visit-related characteristics that are likely to affect pediatricians' abilities to detect psychosocial problems, although Horwitz et al6 documented that physicians were more likely to recognize mental health problems during longer well-child visits rather than acute care visits and in children who were well-known by the physicians.
More surprising is the absence of attention to family factors that may influence whether parents or other caretakers discuss behavioral, emotional, social, or developmental issues with their pediatricians. Prior studies have assumed that families want to discuss social and behavioral issues with their pediatricians and that underrecognition of problems is largely due to characteristics of the physician or practice.10,11 This assumption has been fueled by studies documenting families' interests in discussing such problems and the influence pediatricians' interview styles have on mothers' discussions of psychosocial problems.12- 14 Consequently, previous studies have largely confined themselves to the examination of sociodemographic characteristics that influence whether a psychosocial problem is identified. In general, family sociodemographic characteristics, child's gender, and type of problem have been associated with identification of psychosocial problems.1,4
Previous studies have not examined whether parents' attitudes about the appropriateness of requesting assistance for psychosocial problems influences pediatricians' abilities to detect and subsequently treat such problems. This void exists despite 2 domains of research. First, research on attitudes of adults visiting family practitioners clearly demonstrates that propensity to discuss psychosocial issues is related to physicians' recognition of these problems and that propensity to discuss such issues varies systematically by social and cultural characteristics.10 Second, data suggest that parents are more likely to discuss certain types of problems with professionals and that some parents are more likely than others to discuss psychosocial concerns.1,12
The goal of these analyses was to evaluate the affect of families' attitudes about the appropriateness of discussing psychosocial concerns on pediatric providers' identification of psychosocial problems. Specifically, this research was designed to answer 4 questions: (1) What proportion of parents believe it is appropriate to discuss psychosocial issues with their physicians? (2) What proportion of caretakers who report their children as having had a psychosocial problem actually had discussed it with their physician? (3) What sociodemographic and health-related factors are associated with the propensity to discuss psychosocial problems? and (4) Do the reports of actual discussions of psychosocial problems relate to physicians' recognition of these problems?
The families in this study were interviewed as part of the Greater New Haven Child Health Study (New Haven, Conn). They came from a stratified random sample of all primary care pediatric practices in the greater New Haven area. These 19 practices (83% of all those invited) included prepaid practices (n=2), neighborhood health centers (n=2), fee-for-service group practices (n=7), and solo practices (n=8). All families of children aged 4 to 8 years who attended these practices for either acute or well-child care during 2 separate 3-week periods (1 in fall 1987 and 1 in spring 1988) were invited to participate in the study. Families were invited to participate only once, on the first contact with any eligible child, using procedures approved by the Yale University School of Medicine's Institutional Review Board. The total number of eligible families was 2006; 1886 (94%) chose to participate.6
Prior to the pediatric encounter, children were screened for emotional and behavioral problems using the parent-completed version of the Child Behavior Checklist.15 After the index medical visit, the child's pediatrician completed a 13-category checklist that focused on behavioral and developmental problems and was based on a World Health Organization–sponsored primary care, child-oriented classification system by Burns et al.16 All 929 families whose children were identified on either the Child Behavior Checklist or the physician form as having a psychosocial problem, and a randomly chosen 464 families of children who did not screen positive were invited into the interview portion of the study. A total of 1148 families (83%) agreed to participate in this portion of the study. Adult informants (usually the child's biological mother) in these 1148 families were administered a 90-minute semistructured interview schedule and children were given a 30-minute series of age-appropriate tasks. There were no statistically significant differences in income or ethnicity between participants and nonparticipants. To adjust for the oversampling of children with psychosocial problems, all analyses are reported using sampling weights (weighted sample size is 1841). The weighting of the interviewed sample makes it representative of the families with visits to community-based, primary care pediatric practices in the greater New Haven area.
The family's demographic variables included in these analyses were caretaker's age, level of education, employment, and marital status, as well as availability of medical insurance for the index child. Poverty status was examined by creating a variable that was dichotomized, above or below the federal poverty index . This measure is based on the Consumer Price Index and is a combination of family income and size.17 Ethnicity was trichotomized as follows: Euro-American, African American, and Hispanic. Health-related variables included self-perceived physical and mental health status, each rated on a 4-point Likert scale ranging from poor to excellent. Additionally, mothers' reports of their children's vulnerability were assessed using the child vulnerability scale.18 The child's demographic characteristics included age, sex, and birth order.
Propensity to seek care for psychosocial problems was measured through a modification of an instrument developed by DelVecchio Good et al.10 The original citation contained no reliability or validity information and, although no formal test-retest reliability or accuracy work was undertaken for this study, debriefing of respondents following the pretesting of the interview schedule indicated that they understood the questions about the propensity to seek care. The measure was based on questions that asked whether caretakers thought it appropriate (scored yes or no) to discuss with their pediatricians psychosocial stress in a variety of areas (child's school functioning, peer relations, troublesome habits, family problems, behavioral difficulties, and disciplinary problems). For example, adult informants were asked, "During the past 12 months has [CHILD] had difficulties with habits you wish he/she could change such as thumbsucking, eating too much, or eating too little?" Positive responses to discussing psychosocial problems were summed creating a score ranging from 0 to 6. A second series of questions were then asked about whether the children had experienced any of these problems in the preceding 6 months. If children had experienced any of these problems, a third series of questions asked parents whether they had actually discussed the recent psychosocial problem with their children's pediatrician.
The main outcome variable for these analyses is the identification of any behavioral, emotional, or developmental problem by the pediatricians on the 13-category checklist. Overall, pediatric clinicians identified 27.5% of the children visiting their practices as having 1 or more psychosocial problems.
Bivariate analysis of categorical variables was undertaken using the Pearson χ2. When variables were ordered, such as education or self-evaluated health status, the Mantel-Haenzel χ2 was used. Multivariate analysis evaluating the association of propensity to seek care for psychosocial problems with clinicians' identification of such problems was undertaken using logistic regression because the outcome measure, pediatricians' identification of psychosocial problems, was binary.
The sociodemographic characteristics of the families and children are given in Table 1. Children were evenly split by sex (49.4% male) and age (49.9% between 4 and 5 years), and usually lived with their married biological parents (73.5%).
As in previous studies examining parents' attitudes towards discussing psychosocial concerns with their children's physicians, our results show that, when presented with the 6 hypothetical problems, parents consistently reported that they would discuss the problems with their children's pediatrician (Table 2). Examining the 6 situations, 91.4% of parents thought it appropriate to discuss family problems with medical providers, while 78.8% thought it appropriate to discuss disciplinary issues (reported propensity). Overall, 81.1% of the parents would discuss 4 or more of these issues with their pediatricians. When parents were asked if the child in the study had 1 of these 6 problems (column 2, Table 2), 34.2% responded that the child had had a problem with an undesirable habit (eg, thumbsucking), while only 9% reported a social or peer problem. When parents who reported that a child had 1 of these problems were asked if they had discussed it with their pediatrician the rates of actual interchanges with physicians were considerably lower, ranging from a high of 59.0% for habits to a low of 21.7% for social/peer relationship problems (actual reports). On average, only 40.9% of parents who had experienced a problem discussed it with their children's pediatricians.
We next examined the sociodemographic and health-related factors associated with reported propensity to discuss psychosocial issues and actual reports of the discussions of such problems (Table 3). Looking first at reported propensity, these results are similar to other studies examining features related to discussing such concerns with physicians. Older (>30 years), married caregivers of Euro-American decent with higher education, medical insurance, and family incomes above the poverty line are more likely to report a willingness to discuss such problems as are caregivers whose children are younger and first born. However, when examining features related to those who actually did speak to a provider about 1 of these problems, only marital status and caregiver's educational level remained even marginally statistically significant.
Table 4 gives the results of the logistic regression model predicting the identification of a psychosocial problem by a primary care physician. Older caregivers, families with incomes below the poverty level, and parents with less favorable ratings of their own mental health were more likely to have a child diagnosed with a psychosocial problem than younger caregivers with higher incomes and better self-rated mental health. Child characteristics related to identification included male sex and health status. Finally, as shown in earlier bivariate results, type of visit and provider knowledge of the child also were related to having a medical care physician identify a psychosocial problem. There was no association between caregivers' reported propensity to discuss or actual report of a discussion of psychosocial issues with the providers' identification of psychosocial problems.
Although previous studies have shown that parents report being enthusiastic about discussing developmental, behavioral, and emotional issues with their children's physician, to our knowledge, no previous study has investigated parents' actions when they recognized that their children had problems or the parental and child characteristics that were associated with which parents did discuss these issues with their children's medical providers.
Our data suggest that there is a great deal of discrepancy between what parents report is appropriate to do when their children have psychosocial problems and what they actually do when they recognize such problems in their children. Most (81.1%) believed it was appropriate to discuss 4 or more of the 6 hypothetical situations with their children's pediatrician, while only 40.9% actually did discuss any of these problems with a physician when one occurred. Further, given the correlates of parents who intended to discuss such problems (eg, higher educational level, older age, Euro-American ethnicity, higher income, married, availability of medical insurance), the possibility that parents are providing what they believe are socially acceptable responses to such questions seems likely. However, when we turn to the correlates of parents who report actually discussing such a problem with their child's medical provider, only marital status remains statistically significant. Examining the range of reports of discussions with providers across types of problems shows that the type of problem a child has may influence parent reporting. Parents are far more likely to discuss annoying habits (eg, nail biting), learning difficulties, behavior problems, disciplinary issues, and family difficulties rather than social/peer relationship problems with their pediatric providers (Table 2). Costello and Janiszewski1 found that children receiving services were more likely to evidence behaviors that were upsetting to adults. Similar findings were documented by Cohen et al4 who noted that adolescents with diagnoses of conduct and oppositional defiant disorder were more likely to receive services than those with other disorders.
The finding that parents' actual reports of discussions of psychosocial problems was unrelated to whether providers identified those problems in children is surprising. It may be that parents do not effectively communicate their concerns to medical providers, or providers may rely heavily on their own observations rather than on what parents are reporting when making judgments about psychosocial problems. Further, it may be that medical encounters are too short to explore these issues, that pediatric medical care providers are uncomfortable discussing behavioral, emotional, and social problems, or that parents not having experienced effective responses from physicians about psychosocial concerns in past health visits no longer communicate their concerns.12,13,19 Whatever the cause of this lack of communication, it is clearly an area worth investigating. Improvements in communication around psychosocial issues are particularly important in this era of managed care because primary care physicians are becoming gatekeepers to more expensive services such as mental health interventions.
Accepted for publication November 19, 1997.
This research was funded by a National Institute of Mental Health, award R01 MH41638-04 (Dr Leaf).
We sincerely thank the pediatricians, nurse practitioners, office personnel, and families who so generously gave their time to this research effort as well as 2 anonymous reviewers whose comments on an earlier draft significantly improved the manuscript.
Editor's Note: Here's another study documenting the importance of listening to the patient (parent). With managed cost, it is difficult to find the time to listen, much less act on what is heard. Perhaps that's the idea: if I don't know about it, it doesn't exist.—Catherine D. DeAngelis, MD
Corresponding author: Sarah McCue Horwitz, PhD, Department of Epidemiology and Public Health, Yale University School of Medicine, 60 College St, Box 208034, New Haven, CT 06520-8034.
Horwitz SM, Leaf PJ, Leventhal JM. Identification of Psychosocial Problems in Pediatric Primary CareDo Family Attitudes Make a Difference?. Arch Pediatr Adolesc Med. 1998;152(4):367-371. doi:10.1001/archpedi.152.4.367