Anticipatory Guidance: Title and subTitle BreakWhat Information Do Parents Receive? What Information Do They Want? FREE

Our [medical] curriculum covers a certain amount of study of the anatomy and physiology of the child about which mothers never ask us, but the information which they seek has to do with that which cannot be obtained from books, but rather is that sort of knowledge which has passed from mouth to mouth down through the centuries. Instead of asking mother or grandmother what should be done, the doctor is consulted. If her confidence is to be retained, the physician must be as familiar with the proper manner of bathing a baby as he is with the treatment of pneumonia, and he may render the baby as notable a service in one instance as in the other.

The doctor is taking the place more and more of the `advice-offering neighbor,' and it behooves him to be able to advise the mother correctly.¹ (p43)

For at least a century, physicians who take care of children have been answering parents' questions about basic aspects of child rearing. In contemporary times, providing preventive advice is called anticipatory guidance, and it is an integral part of well-child care. The American Academy of Pediatrics' Guidelines for Health Supervision² and the Maternal and Child Health Bureau's Bright Futures: Guidelines for Health Supervision of Infants, Children, and Adolescents³ provide lists of age-related topics for clinicians to discuss with parents as part of well-child care. Such lists have grown substantially longer in recent years. The third edition of the American Academy of Pediatrics guidelines devotes several times more space to anticipatory guidance and covers many more topics than the second edition.⁴ It is doubtful that these organizations expect physicians to address all age-appropriate topics with every patient. They do, however, suggest that physicians should be capable of addressing all recommended guidance topics, even if they tailor what they cover to each family's particular needs. Most parents have a broad range of child-rearing concerns, and many would like to discuss them with physicians.^{5 - 9}

Although pediatricians report that they generally discuss many aspects of anticipatory guidance,^{10 - 11} few studies have examined what topics are covered and whether parents feel that they have enough information on typical child-rearing issues. The few available studies suggest that there may be large variations in the delivery of anticipatory guidance, depending on the population and clinical setting. In addition, anticipatory guidance rates found in studies based on medical records may reflect differences in documentation rather than differences in delivery of guidance. In a rural study, 39% of children receiving well-child care through the Early and Periodic Screening, Diagnosis, and Treatment Program had no documented anticipatory guidance.¹² By contrast, a study of a practice with age-specific anticipatory guidance checklists found that 99% to 100% of parents had received guidance on safety, nutrition, and family concerns.¹³ A study from a Massachusetts health maintenance organization found a wide range of discussion rates, from 6% for gun-related issues to 98% for growth and nutrition and for language development.⁵

In this article, we use a national data set to determine what subjects (from a list of 6 recommended topics) parents have discussed with a clinician, on which topics they could use more information, and how receipt of anticipatory guidance relates to their ratings of the quality of their child's clinician.

The Commonwealth Fund developed the "Survey of Parents With Young Children" to provide an overview of the health and social conditions of families with young children in the United States; details of the methods appear elsewhere.¹⁴ The survey consisted of a 25-minute telephone interview conducted from July 1995 to January 1996. A stratified random-digit dialing design was used to obtain a nationally representative sample of parents with children between 0 and 3 years. (We refer to respondents as "parents," "mothers," and "fathers" because less than 1% were nonparental guardians.) Telephone exchanges with high population densities of African American and Hispanic households were oversampled to enable subgroup analyses. When 2 parents were present, 1 was randomly selected for the interview. If more than 1 child was younger than 3 years, 1 was randomly selected as the index child for the interview. Of the people contacted by telephone, 81% completed the screening, and 68% (n = 2017 parents) who met eligibility criteria completed the interview. The sample included 1320 mothers and 697 fathers.

The survey asked, "Have you personally ever discussed this with a health professional like a doctor or nurse?" and "For each of the following topics, please tell me if you could use more information, or if you have enough information." The list of topics for these questions included the following: how to care for a newborn (asked only for children < 3 months old); how to deal with children's sleeping patterns; what to do when your child cries; how to help and encourage your child to learn; how to discipline your child (analyzed for children between 6 and 36 months old); and how and when to toilet train your child (for children between 18 and 36 months old). All 6 topics are recommended in Guidelines for Health Supervision^{2 ,4} and Bright Futures: Guidelines for Health Supervision of Infants, Children, and Adolescents.³

Some parents who had not discussed a particular topic might have had no interest in discussing it, whereas others who had discussed it might have desired more information. It is especially unfortunate when parents who could use anticipatory guidance do not receive it. We believe that these parents would be especially receptive to discussions. We classified such parents (those who had not discussed a topic but could use more information) as being in an "unaddressed need group" for that particular subject. Data on unaddressed need groups appear in Table 1.

The survey asked whether respondents would be willing to pay an additional $10 per month to receive the anticipatory guidance discussions and other services, and it asked several questions about sources of child-rearing information. The survey asked respondents to rate various aspects of the health care their child received (poor, fair, good, excellent). Few respondents rated clinicians lower than good (this skewed distribution is typical of such items, as found in the 1994 National Health Interview Survey¹⁵ ), so we divided responses into excellent vs not excellent. We also derived a dichotomous summary rating, set to 1 if the respondent gave at least 1 nonexcellent rating and set to 0 if the respondent gave only excellent ratings.

Depending on the child's age, a discussion about an anticipatory guidance topic will vary in content (eg, encouraging learning is different for parents of a 3-month-old vs a 2-year-old) and importance (eg, discipline may be a more compelling concern for parents of older children). Some topics apply only to certain age groups: 4 topics are for children younger than 3 months, 3 for 3 to 5 months, 4 for 6 to 17 months, and 5 for 18 to 36 months. The distribution of responses differs substantially across topics (Table 1). Therefore, summary measures (specifically, the proportion of topics not discussed and the proportion of topics for which parents are in unaddressed need groups) cannot be compared across age groups.

We needed a method to adjust for the child's age to provide a common metric across age groups. Therefore, we normed the summary measures by age group using percentiles similar to those used in standardized growth curves.¹⁶ We divided children into the 4 age groups listed previously and converted the summary measures into percentiles using the distribution in each group. As with percentiles in standardized growth curves, age-normed percentiles for topics not discussed have the same meaning across age groups and measure how well each child's status compares with other children in the same age group. We set the median number of missed topics for each age group at the 50th percentile. A percentile higher than 50 indicates that the child missed more topics than other children in the same age group. Such an interpretation is not available in unnormed measures. For example, when we report that the uninsured were at the 63rd percentile for the number of missed topics (Table 2), we are indicating that a typical parent whose child was uninsured had missed more topics than 63% of parents in the overall population, controlling for age. This finding does not mean that the uninsured missed 63% of the topics or that 63% of the uninsured had missed the topics.

Independent variables appear in Table 2. We included whether the child was firstborn as an indicator of the parent's potential need for information (with the expectation that parents of firstborn children would have a greater need for anticipatory guidance). We also included a variable that indicated whether respondents in 2-parent households provided more or less of the child's daily care than their spouse or partner. We believe that respondents who provide more of the care would probably be more likely to see the clinician and thus have the opportunity to receive anticipatory guidance; therefore, we repeated some analyses with the subsample of parents who provided more of the care.

We conducted univariate, bivariate, and multivariate analyses. Weighted analyses are used to account for unequal sampling probabilities to represent the overall distribution of parents in the United States with children younger than 3 years. For statistical inference, SEs and test statistics were adjusted for the design effect caused by unequal sampling probabilities and stratification, using the sandwich variance estimator.¹⁷

For bivariate analysis, we specified independent variables as dichotomous or categorical. For numerical outcome variables, we tested for association using the 1-way analysis of variance F test. For dichotomous outcomes, we used a χ² test.

A multivariate linear (logistic) regression was performed to examine the relationship between each numerical (dichotomous) outcome variable and the independent variables from Table 2, except we excluded the variables for sharing child care responsibilities (because it does not apply to single-parent households) and employment (because of collinearity with number of adults in the household). Income and parental age were entered as continuous variables.

Table 1, column 1 presents the percentage of parents who reported that they had not discussed each age-appropriate anticipatory guidance topic with a clinician. For each topic other than newborn care, more than half of the parents reported that they had never discussed it (Table 1). Thirty-seven percent of respondents had discussed no age-appropriate topics with a clinician.

Among respondents in 2-parent households who provided more of the daily care for the child, a lower percentage of parents reported no discussions for some topics, but the overall findings remained the same: for all topics except newborn care, more than half of parents still reported that they had never had a discussion. We found this same pattern when we omitted respondents who reported "emergency room/other" as their child's usual provider of care.

Columns 2 and 3 in Table 1 divide the sample into parents who had and had not discussed each topic, and present the percentage of each group who could use more information on particular topics. For sleeping patterns and discipline, parents who had discussions were significantly more likely than other parents to report that they could use more information. For the other topics, there was no significant difference based on whether parents had had discussions. When these analyses were limited to parents who provide more of the daily care for the child, the changes in percentages of who could use more information were small.

Column 4, Table 1 presents the percentage of parents who were in unaddressed need groups; ie, parents who had not discussed the topic but who could use more information. The percentage of parents in unaddressed need groups ranged from 9% for newborn care to 42% for encouraging learning. When this analysis was limited to parents who provide more of the daily care for the child, the changes in percentages were small for all topics but sleeping patterns, which decreased from 15% to 10%.

We compared rates of discussion by demographic and other characteristics using age-normed percentiles as the outcome variable. Column 1, Table 2 presents percentiles for the number of missed discussion topics. Parents of children with no insurance were much more likely to have missed discussions than parents of children with insurance. The same pattern was found for the number of topics for which parents were in an unaddressed need group (Table 2, column 2). Other notable findings include that Hispanic parents had missed the most discussions and were in the most unaddressed need groups. Higher income and usually receiving care from the same clinician were each associated with fewer missed discussions and fewer unaddressed need groups. Respondents who provide more of the daily care for their child missed fewer discussions (Table 2).

Multivariate analyses for the percentiles of missed discussions showed fewer significant predictors than the bivariate analyses. Characteristics with significant positive associations with missed discussions included the following: no insurance compared with employer-based insurance (P = .002), male respondents (P = .02), Hispanic compared with white (P = .04), and care received at a community clinic compared with a private office (P = .01). Significant negative associations were found for children who usually see the same clinician (P = .02), for college graduates (P = .02), and for people with some college education compared with high school graduates (P <.001).

Multivariate analyses for the percentiles of unaddressed need groups showed a significant positive association for Hispanic compared with white (P = .001) and a significant negative association for household income (P = .007) and for remaining with the same clinician (P = .01).

Clinicians are not parents' only source of information: 74% reported that they use "books, magazines, television, or videos to get information about how to raise" their children, and 35% had attended a class or discussion group about parenting or raising children.

When asked to list up to 2 people with whom they are most comfortable discussing how to raise their children, 52% named their spouse or partner, and 46% named their mother or mother-in-law. The next highest response was a doctor, nurse, or other health professional, at 15%.

For each aspect of care that respondents rated, more than 90% rated their clinician as excellent or good, including more than half who rated their clinician as excellent (Table 3). Fifty-seven percent gave a less-than-excellent rating for at least 1 aspect of care; there was significant variation in the ratings for only a few characteristics. Men were more likely than women to give a nonexcellent rating (60% vs 54%, P = .03). There were also significant differences by race/ethnicity, with Hispanic (68%) and other (69%) respondents more likely than African Americans (59%) or whites (54%) to give a nonexcellent rating (P<.001). In addition, differences in ratings were significant for the child's regular source of health care: private clinician's office (54%), health maintenance organization (59%), community center (71%), or emergency room/other (62%) (P = .007).

Parents who had missed more discussion topics, who could use additional information on more topics, or who were in a higher number of unaddressed need groups were more likely to give at least 1 nonexcellent rating (Table 4, column 1).

A logistic regression showed that Hispanic parents were more likely than white parents to give at least 1 nonexcellent rating (odds ratio [OR] = 1.53, P = .03), as were male respondents (OR = 1.34, P = .04) and people whose children usually received care at a community center compared with a private clinician's office (OR = 1.85, P = .02).

Sixty-four percent of respondents would be willing to pay an extra $10 per month to discuss the anticipatory guidance topics listed in Table 1 and to receive other additional services. Willingness to pay varied significantly with insurance type: employer (60%), government (75%), other (59%), or uninsured (67%) (P<.001). Parents who could use additional information on more topics and who were in a higher number of unaddressed need groups were more likely to be willing to pay extra (Table 4, column 2).

Clinicians who take care of children have a professional responsibility to provide anticipatory guidance to parents.^{2 ,4} However, for 5 of 6 standard anticipatory guidance topics covered by our study, more than half of parents of children between 0 and 3 years old reported that they had never discussed the topic with a clinician. More than one third of parents had never discussed any age-appropriate topics.

Anticipatory guidance may be more important than ever. Cultural and economic shifts during the past sev eral decades have brought many women out of the house and into the workforce. When both parents are employed, neither may have the time or energy to learn about all aspects of child rearing. Increased geographic mobility means that parents do not necessarily live near extended family members who might otherwise provide support. Because of shifting gender roles and expectations, women may be less likely than in the past to learn about child rearing while growing up. Although society is placing more responsibility on men to participate in child care, we are not familiar with any broad educational efforts to teach men how to take care of a child.

Discussions about topics covered in this study are not only important for the purpose of providing advice to parents but also serve as a means of screening for problems. A discussion of discipline could lead to identification of a parent who is unwittingly abusing a child. A discussion of sleep patterns might identify a child with night terrors whose parents were unaware that physicians can advise on such matters.

Some respondents might have missed out on discussions because they do not accompany the child to the clinician visit. Even when we limited analysis to parents who provide more of their child's daily care than their spouse or partner, we found that many had not had discussions and could use more information.

Parents might have forgotten discussions that, depending on the child's age, could have occurred up to 3 years earlier. If parents had talked about child rearing with several people, they might not have remembered with whom they had discussed a particular topic, especially if the topic was not a problem. Therefore, some parents might have underreported or overreported discussions. Regardless, many report that they could use more information, suggesting that discussions that did occur may have been inadequate. In other words, there was a significant unmet need for information. Of course, some parents might have an insatiable desire for more information and might report that they could use more information no matter how well and exhaustively the clinician, family members, or others talked with them. Some parents might also have interpreted interview items differently from the way the survey intended.

Parents who receive less anticipatory guidance fit a demographic picture similar to people in other studies found to have limited access to care: those with lower income, less education, and no insurance, and members of nonwhite racial/ethnic groups.¹⁸ The same forces that limit access may also be limiting the quality of care. Parents of children who usually receive care from the same clinician appear to be getting more anticipatory guidance. Improving both continuity of care and transmission of information between clinicians about what advice has been provided might increase the delivery of anticipatory guidance.

There are several reasons physicians and parents might not be discussing these topics. First, physicians may not have adequate training to discuss some anticipatory guidance topics.^{11 ,19 - 22} Both improved training and standardized medical record forms, which prompt physicians on age-appropriate advice, might help.²³ Incorporating anticipatory guidance into quality assessment systems would create additional incentives for physicians and health care organizations to deliver guidance.^{24 - 26}

Second, physicians may not be confident that their advice will be useful.¹¹ Few studies have examined the influence of physician counseling and anticipatory guidance on parent/patient behavior and health,^{27 - 28} and the findings have been inconsistent. Some evidence exists that anticipatory guidance for injury prevention,²⁹ violence reduction,³⁰ and infant sleep patterns³¹ can have a positive effect, but other studies have had mixed results.^{32 - 35} Anticipatory guidance may be more effective when tailored to the parent's educational level and when advice is specific rather than general.^{36 - 38} It may also have a greater effect when supplemented with handouts or videotapes.³⁶ Various creative approaches that go beyond the traditional physician-parent discussion show promise, including parenting classes and group well-child care (in which guidance is provided in a group setting).^{36 ,39 - 41} Further research is needed on how to provide the most effective anticipatory guidance.

Nonclinical sources of information may also be useful for parents. Nearly three quarters of parents in our study got child-rearing information from books, magazines, television, or videos. The Internet is a new source of information, although the accuracy is inconsistent.⁴²

It takes time to cover multiple aspects of anticipatory guidance and to answer parents' follow-up questions.⁴³ Physicians may not feel that they have enough time to provide anticipatory guidance,¹¹ particularly if they are expected to see an increasing number of patients per hour and are not specifically reimbursed for providing guidance.⁴⁴

Another reason for limited discussions is that some parents may not be interested in talking about these issues with physicians. Although more of our survey's respondents feel comfortable talking with family members than with physicians, other studies have found parents ranking physicians first (ahead of family members and other health professionals) in terms of importance and helpfulness in providing information on raising children.^{7 ,45} Another study found that physicians were the most frequently reported source of information on child care problems such as growth, health, and nutrition.⁴⁶ The difference in findings may reflect our survey's use of the word comfortable. Parents may feel more comfortable talking with family members yet trust the information more when it comes from physicians. Our survey did not ask whether parents found discussions with clinicians useful or whether parents who could use more information wanted to receive it from a clinician. However, other studies have suggested that parents do want to receive such information. In a study that asked mothers the importance of a physician or nurse bringing up a list of topics with parents during a child's first 3 years of life, a majority of mothers felt that issues such as discipline, sleep patterns, and sleep problems are important.⁴⁵ In another study, more than two thirds of mothers had nonmedical concerns, but many did not tell the pediatrician for a variety of reasons, including not knowing that the pediatrician could help with psychosocial issues and assuming that the pediatrician was too busy.⁶

Our study provides additional evidence that parents value anticipatory guidance and that parent disinterest is probably not a major reason for low reported rates of guidance. Parents who had fewer discussions gave their clinicians lower ratings. A study of mothers of children with chronic illnesses had a similar finding: receipt of anticipatory guidance was significantly associated with satisfaction.⁴⁷ Also, most parents in our study would be willing to pay for more information, which raises the possibility of improving anticipatory guidance with interventions that may increase costs. However, caution is necessary in interpreting this item. Although it is useful for gaining a general understanding of whether parents value a certain type of care, a more comprehensive survey would be needed to determine the actual amount people would pay. Nonetheless, this finding lends further credence to the notion that parents would like clinicians to provide anticipatory guidance.

Medical organizations and leaders have been calling on pediatricians to take a broad view of health care and their role as promoters of health.^{4 ,48 - 51} They encourage physicians to step beyond the biomedical model to address their patients' emotional, social, and developmental needs. This study shows that parents have unmet needs for anticipatory guidance on many aspects of caring for children. Although there are various ways of meeting these needs, physicians are among the only professionals with whom almost all parents have contact. It seems likely that the clinical setting will continue to be an important component of any strategy for meeting parents' needs for information and advice.

Accepted for publication July 24, 2000.

This work was supported in part by the Commonwealth Fund, New York, NY, and by grant U48/CCU915773 from the Centers for Disease Control and Prevention, Atlanta, Ga.

We are indebted to Linda Barthauer, MD, Neal Halfon, MD, MPH, and Kathryn Taaffe McLearn, PhD, for comments on drafts of this article; to Cung B. Pham, BA, and Myra Wong, BA, for research assistance; and to James Tebow, PhD, for assistance with the manuscript.

This work does not necessarily represent the opinions of the funding organizations or of the institutions with which the authors are affiliated.

Reprints: Mark A. Schuster, MD, PhD, RAND Corporation, 1700 Main St, Santa Monica, CA 90407 (e-mail: schuster@rand.org).