Background. Family-centeredness, compassion, and trust are 3 attributes of the clinician-parent relationship in the medical home. Among adults, these attributes are associated with patients' adherence to clinicians' advice.
Objectives. The objectives were (1) to measure medical home attributes related to the clinician-parent relationship, (2) to measure provision of anticipatory guidance regarding injury and illness prevention, (3) to relate anticipatory guidance to parental behavior changes, and (4) to relate medical home attributes to anticipatory guidance and parental behavior changes.
Methods. A cross-sectional study of data collected among at-risk families when children were 1 year of age, in a randomized, controlled trial of a home-visiting program to prevent child abuse and neglect, was performed. Modified subscales of the Primary Care Assessment Survey were used to measure parental ratings of clinicians' family-centeredness, compassion, and trust. Parental reports of provision of anticipatory guidance regarding injury and illness prevention topics (smoke alarms, infant walkers, car seats, hot water temperature, stair guards, sunscreen, firearm safety, and bottle propping) and behavior changes were recorded.
Results. Of the 564 mothers interviewed when their children were 1 year of age, 402 (71%) had a primary care provider and had complete data for anticipatory guidance items. By definition, poverty, partner violence, poor maternal mental health, and maternal substance abuse were common in the study sample. Maternal ratings of clinicians' family-centeredness, compassion, and trust were fairly high but ranged widely and varied among population subgroups. Families reported anticipatory guidance for a mean of 4.6 ± 2.2 topics relevant for discussion. Each medical home attribute was positively associated with parental reports of completeness of anticipatory guidance, ie, family-centeredness (β = .026, SE = .004), compassion (β = .019, SE = .005), and trust (β = .016, SE = .005). Parents' perceptions of behavior changes were positively associated with trust (β = .018, SE = .006). Analyses were adjusted for potential confounding by randomized, controlled trial group assignment, receipt of ≥5 well-child visits, and baseline attributes.
Conclusions. Among at-risk families, we found an association between parental ratings of the medical home and parental reports of the completeness of anticipatory guidance regarding selected injury and illness prevention topics. Parents' trust of the clinician was associated with parent-reported behavior changes for discussed topics.
- medical home
- anticipatory guidance
- behavior changes
- clinician-parent relationship
- at-risk families
The clinician-patient relationship is central to primary care.1 In its 1994 report, the Institute of Medicine recognized the importance of primary care clinicians “developing a sustained partnership with patients” and “practicing in the context of family and community.”1 This description expanded primary care beyond accessible, comprehensive, coordinated, continuous, and accountable health care.2
The American Academy of Pediatrics (AAP) definition of the medical home also affirms the importance of the clinician-family relationship in children's health care.3,4 The medical home consists of many elements, including accessibility, family-centeredness, comprehensiveness, continuity, coordination, cultural effectiveness, and compassion. The clinician-family relationship in the medical home is developed through periodic health supervision visits from infancy through adolescence. Essential to the medical home concept is the trust that develops between the clinician and the family during these visits.
Anticipatory guidance is an element of the medical home under comprehensive care,3 and its content and frequency are recommended by the AAP5 and Maternal and Child Health Bureau.6 Because the greatest childhood morbidity and mortality causes after the first 1 year of life are unintentional injuries,7 anticipatory guidance regarding injury prevention is an especially important aspect of the medical home. Because the amount of recommended guidance has grown, however, clinicians in practice face time barriers in discussing many topics.8–10
Research among adult patients has demonstrated the association of the clinician-patient relationship with patient adherence to clinician counseling and behavior changes. Safran et al11 found that patient ratings of clinicians' knowledge of the patient and high patient trust in the clinician were positively associated with adherence to the clinician's advice and desired behavior changes. Thom et al12 reported that patients who trust their clinician are more likely to intend to follow the clinician's advice. In pediatrics, lack of provision of anticipatory guidance was associated recently with parental reports of poor family-centered care in the medical home.13
Additional pediatric research on the links among the clinician-family relationship in the medical home, provision of anticipatory guidance, and parental behavior changes regarding the topics discussed is needed. The objectives of this study were (1) to measure medical home attributes related to the clinician-parent relationship, (2) to measure provision of anticipatory guidance regarding injury prevention, (3) to relate anticipatory guidance to parental behavior changes, and (4) to relate medical home attributes to anticipatory guidance and parental behavior changes.
This was a cross-sectional study of data collected in a randomized, controlled trial (RCT) to evaluate a home visiting program for families at risk for abuse of their newborns.14–18 Families of newborns were enrolled in the RCT from November 1994 through December 1995. By design, more families were allocated to the experimental group (receiving home visitation). Data on baseline characteristics were collected at birth by evaluation staff members, as described in detail previously.14 The data used for the current cross-sectional study were collected through structured maternal interviews and review of pediatric primary care records when the child was 1 year of age (year 1).
The study population included families that were assessed, through population-based screening and assessment, as being at risk for abuse or neglect of their newborns and were enrolled in a RCT to evaluate the Hawaii Healthy Start Program (HSP).14–18 The goals of HSP are (1) to improve family functioning, (2) to promote child health and development, in part by linking families with a medical home, and (3) to prevent child abuse and neglect. Prior analyses showed that the Hawaii HSP improved access to a medical home through health insurance coverage and a regular provider but there was no program impact on the other attributes of the medical home.16 Therefore, in the analyses reported here, the control and intervention groups were combined.
Because we were interested in the clinician-parent relationship, we limited analysis to children who had a primary care provider at the year 1 interview. Of the 643 families enrolled in the RCT of the Hawaii HSP, 564 (88%) were interviewed at year 1. Of those interviewed, 457 (81%) received the survey version that elicited parental recall of anticipatory guidance and, of those families, 418 identified 1 individual as the child's physician and still had custody of the child. Of those 418 families, 402 families (96%) had complete data on anticipatory guidance items and 16 families (4%) had missing data. Those 402 families were used in the cross-sectional study reported here.
The study was reviewed and approved by the institutional review boards of the Johns Hopkins University School of Medicine (Baltimore, MD), the Hawaii State Department of Health (Honolulu, HI), and the hospitals at which study families were recruited. Signed informed consent was obtained from all parents before study enrollment.
Primary Care Provider, Health Insurance, and Number of Well-Child Visits
At the year 1 interview, mothers were asked whether a particular person was their child's doctor and how long that person had been the child's doctor. Mothers were also asked whether and for how long the child had health insurance.
To quantify the opportunities for anticipatory guidance in the first 1 year of life, pediatric primary care records were reviewed to count the number of well-child visits (WCVs) the children had experienced from birth through 12 months of age. The AAP recommends that children undergo a minimum of 5 WCVs by their first birthday, including the newborn and 2-, 4-, 6-, and 9-month visits.19 For the analyses, results were dichotomized as ≥5 WCVs or <5 WCVs.
Medical Home: Clinician-Parent Relationship Measures
Modified subscales of the Primary Care Assessment Survey (PCAS) described by Safran et al20 were used to measure attributes of the clinician-parent relationship in the medical home. As previously described,16 the PCAS was modified for application to pediatric care, and factor analysis was conducted to organize the items for our study sample. Three subscales measuring the clinician-parent relationship were developed, ie, family-centeredness, compassion, and trust. Table 1 presents the individual subscale items. Family-centeredness was measured with a 4-item contextual knowledge subscale of the PCAS (Cronbach α = .85). Compassion was measured with a 5-item interpersonal treatment subscale (Cronbach α = .93). Trust was measured with a 4-item trust subscale (Cronbach α = .65). Family-centeredness, compassion, and trust scores were calculated through addition of the individual response choices. The scores were recoded so that the possible range for each was 0 to 20, with higher scores indicating more favorable ratings. The responses to the PCAS trust subscale statement that asked directly about trust and parental behavior changes (“I trust my infant's doctor so much that I always try to follow his/her advice”) were dichotomized as yes (strongly agree or agree) or no (not sure, disagree, or strongly disagree).
Medical Home: Anticipatory Guidance Measures
Mothers were asked about 8 anticipatory guidance topics related to injury and illness prevention, ie, presence of working smoke alarms in the home, use of infant walkers, use of car seats, hot water heater temperature turned down to ≤120°F, use of stair guards, use of sunscreen, presence of guns in the home, and bottle propping at bedtime. At the time of the study, the bottle propping item was recommended in Guidelines for Health Supervision II,21 sunscreen use was a national health objective,22 and the remaining anticipatory guidance items were recommended by the AAP Injury Prevention Program for the first 1 year of life.23 The mother was asked, “Has the infant's doctor ever talked to you about [each topic (eg, infant car seat use)]?” If the mother answered yes to this question, then we considered that the topic had been discussed by the clinician.
To determine whether behavior changes were needed for anticipatory guidance topics (thus relevant for discussion with the clinician) and whether parents changed their behavior, the interviewer asked, “Since the infant was born, have you [tried to change behavior regarding each anticipatory guidance topic (eg, used an infant car seat)]?” There were 4 possible responses to this question, ie, (1) not applicable because not needed (eg, do not own or use a car), (2) not applicable because already doing/did it (eg, have been using a car seat since the birth), (3) no (eg, not using an infant car seat), and (4) yes (eg, started using a car seat). The topic was considered relevant for discussion with the clinician if the mother answered that she had not changed behavior (response 3) or had changed behavior (response 4). The topic was considered not relevant for discussion if the mother answered that the topic was not applicable because it was not needed (response 1) or because she was already doing it (response 2). Not applicable because not needed was not an answer choice for smoke detectors, bottle propping, and sunscreen use, because it was determined during the survey design that these items were needed by all families. To quantify parental behavior changes, we considered that the parent changed behavior if the mother answered that she changed behavior (response 4).
Summary measures were calculated for the proportion of anticipatory guidance topics that were relevant, the proportion of relevant topics for which anticipatory guidance was provided, and the proportion of relevant anticipatory guidance topics for which guidance was provided and the parent changed behavior. These summary measures were used in the statistical analyses.
The baseline characteristics of the 402 families included in the cross-sectional study were compared, with Pearson's χ2 test (poverty, parity, mental health, ethnicity, substance abuse, partner violence, and control versus intervention arm) or Student's t test (maternal age), with those of the 162 families interviewed at year 1 but excluded from our analysis. Descriptive statistics were used to measure medical home attributes and provision of anticipatory guidance regarding injury prevention. The associations of maternal attributes with medical home attributes, provision of anticipatory guidance, and parental behavior changes were tested with Student's t test (poverty, parity, mental health, substance abuse, and partner violence) or one-way analysis of variance (ethnicity). We tested for differences in baseline maternal attributes according to RCT group assignment with Student's t test (maternal age and length of insurance coverage), Pearson's χ2 test (poverty, parity, mental health, substance abuse, partner violence, and maternal employment), or analysis of variance (ethnicity).
Student's t test was used to measure the association of receipt of ≥5 WCVs with medical home attributes, anticipatory guidance, and parental behavior changes. Spearman's ρ was used to measure the bivariate correlations of medical home attributes with anticipatory guidance and parental behavior changes. Pearson's χ2 test was used to measure the association of receipt of ≥5 WCVs with relevance, discussion, and behavior changes for individual anticipatory guidance items. Paired-sample Student's t test was used to measure the association of parental behavior changes for discussed topics and nondiscussed topics. In addition, the trust scores were divided into quartiles and one-way analysis of variance was used to test for differences in the average numbers of relevant items changed. Student's t test was used to measure the association of the dichotomized trust item with parental behavior changes.
Simple linear regression analyses were used to test the association of each medical home attribute (family-centeredness, compassion, and trust) with provision of anticipatory guidance and parental behavior changes. Multivariate linear regression analyses were used to test for confounding by RCT group assignment, receipt of ≥5 WCVs, and baseline characteristics associated with anticipatory guidance, behavior changes, and medical home attributes.
Statistical analyses were completed with SPSS version 11.0 software (SPSS, Chicago, IL). Statistical significance was defined as P < .05, and trend was defined as P < .10.
As noted earlier, the study sample was limited to families with a pediatric primary care clinician at the time of the year 1 interview. Eighty-two percent of children had the same clinician since birth. The mean duration of health insurance in the child's first 1 year of life was 11.7 ± 1.3 months. Two hundred thirty-eight children (59%) in the sample had ≥5 WCVs. The distribution of the number of WCVs was as follows: 15% with 0 to 2 WCVs, 26% with 3 or 4 WCVs, 21% with 5 WCVs, and 38% with ≥6 WCVs.
Poverty, partner violence, poor maternal general mental health, and maternal substance abuse were common in the study sample at the time of the child's birth (Table 2). There were no significant differences in maternal baseline attributes between families interviewed at year 1 and included in the cross-sectional study sample (N = 402) and those not included (N = 162).
RCT Group Assignment
In this sample (N = 402), 61% of families were in the intervention group of the RCT (received home visiting) and 39% of families were in the control group (did not receive home visiting). There were no significant differences or trends in maternal age, length of insurance coverage, poverty, mental health, substance abuse, partner violence, or ethnicity according to RCT group. Mothers in the intervention group were more likely to have worked in the previous year (55% vs 45%, P < .10) and were less likely to be first-time mothers (40% vs 49%, P < .10). Therefore, in subsequent analyses with adjustment for potential confounding by RCT group assignment, adjustment was also performed for maternal employment in the previous year and parity.
Medical Home Attributes
Maternal ratings of the clinician-parent relationship in the medical home were fairly high but ranged widely. The family-centeredness score had a mean of 13.5 ± 3.9 and an interquartile range of 11.0 to 16.0. The compassion score had a mean of 16.1 ± 3.5 and an interquartile range of 13.6 to 19.2. The trust score had a mean of 15.1 ± 3.4 and an interquartile range of 13.3 to 17.3.
In bivariate analyses, maternal ratings of the medical home did not differ significantly according to the number of WCVs. Families with ≥5 WCVs did not rate the clinician differently than did families with <5 WCVs with respect to compassion (P = .99), family-centeredness (P = .13), or trust (P = .14).
There were some differences in ratings of medical home attributes among population subgroups. Mothers were significantly more likely to rate clinicians higher for family-centered care if the mother had a previous child (P < .05) (Table 3). There were trends for mothers to rate clinicians higher for family-centeredness if the family was below the federal baseline poverty level (P = .05) or the mother had good mental health (P = .06). Mothers rated clinicians significantly higher for compassionate care if the family was below the federal baseline poverty level (P < .05), had a previous child (P < .05), or was Native Hawaiian or Pacific Islander, rather than Asian (P < .05). There was a trend for parents to rate clinicians higher for trust if there was no history of partner violence (P = .06).
Parental Reports of Anticipatory Guidance
The number of anticipatory guidance items relevant for discussion with the clinician varied, with a mean of 4.6 ± 2.2 and a range of 0 to 8 relevant items. Parental reports of provision of relevant guidance were not associated with specific maternal baseline attributes (Table 3). The relevance and completeness of anticipatory guidance varied widely according to topic (Table 4). Overall, parents reported anticipatory guidance for a mean of 62% of topics relevant for discussion (data not shown). For the total sample, relevance ranged from 86% for bottle propping to 12% for guns in the home. Smoke detectors and guns in the home were discussed with the fewest families for which the topics were relevant (35% and 42% of families, respectively). Car seat use and bottle propping were discussed with the most families for which the topics were relevant (77% and 83% of families, respectively).
In bivariate analyses, families with ≥5 WCVs were more likely to report that hot water temperature (P < .05) and car seat use (P = .07) were relevant for discussion than were families with <5 WCVs. The provision of anticipatory guidance regarding topics relevant for discussion did not vary according to the number of WCVs (63% of topics for families with ≥5 WCVs vs 59% of topics for families with <5 WCVs, P = .21). However, families with ≥5 WCVs were more likely to receive anticipatory guidance regarding infant walkers (62% vs 49%, P < .05).
Relationship of Medical Home Attributes to Reported Anticipatory Guidance
In bivariate analyses, each medical home attribute was positively correlated with parental reports of completeness of anticipatory guidance (family-centeredness: r = 0.31, P < .001; compassion: r = 0.20, P < .001; trust: r = 0.14, P < .01). In simple linear regression analyses, family-centeredness (β = .026, SE = .004, P < .001), compassion (β = .019, SE = .005, P < .001), and trust (β = .016, SE = .005, P = .001) were each significantly associated with parental reports of completeness of anticipatory guidance. In multivariate models with adjustment for variables individually and simultaneously, there was no evidence of confounding by RCT group assignment, receipt of ≥5 WCVs, or baseline characteristics associated with medical home attributes at the bivariate level (family centeredness: poverty and mental health; compassion: ethnicity and poverty; trust: partner violence) (Table 3). Therefore, we did not need to adjust for these variables, and the results from the simple linear regression models were accurate. Each 1-point increase in the family-centeredness rating was associated with a 2.6% increase in parent-reported completeness of anticipatory guidance, each 1-point increase in the compassion rating was associated with a 1.9% increase in parent-reported completeness of anticipatory guidance, and each 1-point increase in the trust rating was associated with a 1.6% increase in parent-reported completeness of anticipatory guidance.
Relationship of Reported Anticipatory Guidance to Reported Behavior Changes
Mothers reported changing their behavior for 62% of relevant anticipatory guidance topics that were discussed by the clinician and for 33% of relevant topics that were not discussed by the clinician (P < .001; 95% confidence interval of the difference of 22% to 35%). Parental behavior changes varied according to anticipatory guidance topic (Table 4). When relevant topics were discussed, mothers reported changing their behavior in 41% (infant walkers) to 94% (car seats) of cases.
The association between provision of guidance and reported behavior changes did not vary according to the number of WCVs. Families that received ≥5 WCVs reported behavior changes for 62% of topics discussed and 33% of topics not discussed by the clinician (P < .001), and families that received <5 WCVs reported behavior changes for 60% of topics discussed and 33% of topics not discussed (P < .001). For specific discussed topics, families with ≥5 WCVs were more likely to change behavior for hot water temperature (95% vs 86%, P < .10).
Relationship of Medical Home Attributes to Reported Behavior Changes
In bivariate analyses, parent-reported behavior changes for relevant and discussed topics were positively associated with trust (r = 0.18, P < .01) but not with family-centeredness (r = −0.04, P = .43) or compassion (r = 0.03, P = .57). Mothers who agreed with the statement, “I trust my infant's doctor so much that I always try to follow his/her advice,” reported behavior changes for 65% of topics that the clinician discussed; parents who disagreed reported behavior changes for 48% of discussed topics (P < .05; 95% confidence interval of the difference of 4% to 29%).
In simple linear regression analyses, parental reports of behavior changes for relevant and discussed topics were positively associated with trust (β = .018, SE = .006, P < .01) but not with family-centeredness (β = −.003, SE = .005, P = .52) or compassion (β = .003, SE = .006, P = .64). In multivariate models with adjustment for variables individually and simultaneously, there was no evidence of confounding by RCT group assignment, receipt of ≥5 WCVs, or baseline characteristics associated with behavior changes and medical home attributes at the bivariate level (behavior changes: partner violence; family centeredness: poverty and mental health; compassion: race and poverty; trust: partner violence) (Table 3). Therefore, we did not need to adjust for these variables, and the results from the simple linear regression models were accurate. Each 1-point increase in the trust rating was associated with a 1.8% increase in parental reports of behavior changes.
There was a dose-response relationship between trust and parental behavior changes. Mothers in the lowest quartile for trust reported changing behavior for an average of 54% of relevant topics discussed, those in the second lowest quartile changed 55%, and those in the second highest quartile changed 64%. Those in the highest trust quartile reported changing behavior for an average of 72% of relevant topics discussed (P < .01).
Parental reports of more favorable clinician-parent relationships in the medical home were associated with perceptions of more complete anticipatory guidance regarding injury and illness prevention topics. This finding underscores the importance of Accreditation Council for Graduate Medical Education competencies for residents to demonstrate the ability to develop family-centered, compassionate, trusting relationships with families24 and the importance of the clinician-parent relationship and the medical home in providing quality preventive care to children. Halfon et al13 recently demonstrated the association of family-centered care with provision of anticipatory guidance in early childhood. Among adults, trust of the physician has been associated with completeness of preventive counseling.12 To our knowledge, our study is the first time to demonstrate that compassion and trust, in pediatric care, are linked to completeness of anticipatory guidance.
Mothers who reported greater trust in their children's clinicians were more likely to report behavior changes attributable to the clinicians' advice. Among adults, patient behavior changes in response to preventive counseling by physicians were most strongly associated with patient ratings of trust in the physician, the physician's “whole-person” knowledge of the patient (equivalent to family-centeredness in pediatrics), and interpersonal treatment (equivalent to compassion).11 In fact, adherence to advice was 2.5 times greater when adult patients rated physicians highly with respect to knowledge of the patient and trust. Additional research demonstrated that adult patients who trust physicians are more likely to adhere to the physician's advice and follow preventive recommendations12 or take prescribed medications.
The sample for this study included children at risk for child abuse and neglect, as measured with population-based screening and assessment. Family attributes that place children at risk for maltreatment also put them at risk for chronic family stress, poor parenting, and dysfunctional parent-child relationships, which can lead to substantial deficits in cognitive and socioemotional development and increased risks of unintentional injury and illness.25,26 Therefore, at-risk children may benefit more from the clinician-family relationship and delivery of preventive services, including anticipatory guidance, in the medical home to prevent illnesses and injuries and to prevent or detect early behavioral and developmental problems. Because many mothers with at-risk children may have difficulties forging relationships because of risk factors (eg, poor mental health, substance abuse, or partner violence), it is promising that in this study they gave overall high ratings to their relationships with their children's clinicians.
Parents who reported receipt of anticipatory guidance were more likely to report behavior changes for discussed injury and illness prevention topics. This supports previously reported findings. Two critical reviews27,28 of injury prevention office counseling evaluated the effectiveness of anticipatory guidance to affect behavior and injuries and concluded that there were associations between office counseling and decreased hot water temperature,27,28 increased installation of smoke detectors,27,28 increased use of car seats,27,28 and decreased incidences of falls and motor vehicle occupant injuries.27 However, several studies found that anticipatory guidance had little effect on television viewing,29 firearm storage,30 bicycle helmet ownership,28 childproofing in the home,28,31 and medical care for injuries.28
Our study found that parents reported behavior changes for one third of topics that were not discussed by the pediatrician. Furthermore, even when parents rated the clinician low with respect to trust, they changed behavior for almost one half of discussed topics. This may be attributable to the role of “influential others” in parents' lives affecting their parenting behaviors.32 Previous research showed that the likelihood of childhood injury decreases as the number of sources of parenting education increases, including both professionals and nonprofessionals.33 These influential others may be child educators or caregivers, home visitors, parenting program staff members, family members, or friends. Pediatric clinicians represent only one voice among many and need to acknowledge other influences on parental behavior. This is also an important area for future pediatric research, to determine whether parents are receiving the same messages from clinicians and from influential others regarding prevention, to determine which messages have more weight if they conflict, to determine how parents reconcile conflicting messages, and to determine how influential others can reinforce messages supplied by primary care providers.
Although all children in our sample had a primary care provider, 82% had the same provider for the entire first 1 year of life, and we limited our analysis to anticipatory guidance items relevant for discussion, we found that anticipatory guidance regarding the items selected for this study was often incomplete. On average, mothers did not recall receiving guidance on more than one third of relevant items. This could be attributable to poor parental recall or actual incomplete guidance because of time constraints during WCVs and the need to address other pressing health issues during the visits. Because of time restrictions for WCVs and a growing body of recommended anticipatory guidance topics, it is rare (and nearly impossible) for all recommended topics to be discussed.8–10,34–39 Previous research found that topic discussion is influenced by maternal beliefs and expectations regarding guidance40,41 and physician attitudes regarding the importance of a topic and efficacy in discussing the topic.8 Our finding that firearm safety was one of the items least often discussed is consistent with findings of national surveys of parents and pediatricians.8,35 It may also be attributable to low rates of firearm ownership in Hawaii, compared with other states.42
Parental ratings of the clinician-parent relationship and parental reports of completeness of anticipatory guidance and behavior changes did not differ among families with and without the recommended number of WCVs in the first 1 year of life. This finding may demonstrate that the clinician-parent relationship can develop from the first office visit encounters. However, several other factors must be considered. First, we used medical record data to determine the number of WCVs in the first 1 year of life. There might have been misclassification of WCVs as sick/acute care visits, with underestimation of the number of WCVs in the first 1 year of life. Sick/acute care visit encounters might have influenced the clinician-parent relationship and would not be measured with WCV determinations. Also, patients might have received anticipatory guidance during sick/acute care visits, which would decrease the influence of the number of WCVs on anticipatory guidance. In addition, as we determined by examining the distribution of the number of WCVs, families received a large proportion of recommended WCVs; only 15% of families had <3 WCVs. Therefore, our sample was positively skewed with respect to WCV distribution, and the cutoff of ≥5 WCVs we used, which was based on AAP recommendations, might not represent the threshold for receipt of adequate anticipatory guidance.
Overall, mothers gave high ratings to their relationships with the clinicians. The quality of these relationships may affect whether families voluntarily switch physicians; poor physician-patient relationship quality was the leading predictor of adult patients voluntarily switching physicians within a 3-year period.43
Clinician-parent relationship ratings differed among family subgroups. Mothers with a previous child were more likely to rate clinicians higher for family-centeredness and compassion. This may be attributable to the mother having a longer relationship with the clinician for care of her older children, but this factor was not measured in this study. Longitudinality (presence and use of a regular source of care over time) for an individual is a domain of primary care.44 To our knowledge, however, increased longitudinality because of older siblings and its influence on the quality of primary care for younger children have not been studied and represent an area for future research.
Impoverished families were more likely to rate clinicians higher for family-centeredness and compassion. This finding is surprising because of concerns regarding health care disparities among patients with low socioeconomic status.45 Previous studies of disparities attributable to poverty concentrated on poorer access to health care.46,47 In our sample, all patients had access to a primary care provider and almost all were insured for the entire study period; therefore, we were studying a distinct subset of impoverished patients. It is promising that, when impoverished families had financial access to primary care, they gave high ratings to relationships with clinicians.
Our findings that Asian parents were more likely to rate clinicians lower with respect to compassion, compared with Native Hawaiian or Pacific Islander families, were consistent with results from previous research on the quality of care for Asian Americans. In previous research, Asian American adults rated primary care quality poorer than did white, black, or Latino subjects,48,49 and Asian American parents reported poorer-quality relationships with their children's primary care clinicians than did white parents.50 Additional research is needed for elucidation of these ethnic differences in clinician-patient relationships, including exploration of differences attributable to cultural beliefs, English-speaking ability, and physician ethnicity.
We found that mothers with poor mental health rated clinicians lower for family-centered care. This finding is similar to results from the 2000 National Survey of Early Childhood Health.51 In general, we know that identification of maternal mental health problems by pediatricians is poor.52 Mothers with poor mental health might have rated their children's clinicians poorly with respect to family-centeredness because of the clinicians' failure to recognize or address the mothers' problems.
Mothers with a history of partner violence gave lower trust scores to their children's clinicians and were less likely to report behavior changes. Previous research demonstrated that abused women were more likely to report poor communication with their physicians and this poor communication was associated with dissatisfaction with care.53 The trust measure included items relevant to communication, disclosure of information, and unanswered questions. Communication is a bidirectional process, and mothers with a history of partner violence might have had poor communication with and distrust of clinicians because of previous experience with violent relationships. Also, it might be more difficult for mothers with a history of partner violence to makes changes in the home without causing stress. Eckenrode et al54 found that the effects of home visitation on child abuse and neglect decreased as incidents of domestic violence in the home increased.
Our study sample was drawn from a RCT evaluating a home visitation program for at-risk families. The study showed no program impact on attributes of the medical home.16 Nevertheless, we controlled for RCT group assignment and baseline factors associated with group assignment as potential confounders in our analyses. We did not find RCT group assignment to be a confounder in this study.
We limited our sample to families with complete survey data for anticipatory guidance items (N = 402). With selection of the families with complete data, a selection bias might have been introduced. However, the baseline characteristics did not differ between the 402 families included in the sample and the remaining 162 families interviewed at year 1.
We relied on parental reports to measure provision of anticipatory guidance and parental behavior changes. Self-reporting is prone to recall bias and social desirability bias and may be inaccurate. Because we did not validate parental reports of anticipatory guidance and behavior changes, the findings reported here should be considered within the context of other studies with parent-reported measures.55–64 Perhaps parental reporting of anticipatory guidance is an accurate way of measuring what information was conveyed to the parent during the office visit. If a parent cannot recall the guidance, then we probably should not consider that it occurred enough to have an impact. In addition, in a recent study of the validity of parents' self-reported home safety behaviors by Chen et al,64 parents were found to over-report safety practices. These overestimations were uniform across the sample; therefore, the authors concluded that self–reporting might be valid for determining relative differences among groups.
In addition, anticipatory guidance topics were limited to injury and illness prevention. Unintentional injuries are the leading cause of death for children >1 year of age,65 and children 15 to 17 months of age have the highest overall injury rate among children <15 years of age.66 Injury prevention guidance in the first 1 year of life is important so that parents can adopt practices before children reach 1 year of age and are at highest risk for injuries. However, multiple other anticipatory guidance topics are recommended for the first 1 year of life,5,6 and different topics may be related differently to the clinician-parent relationship and behavior changes. Finally, the study sample was limited to families at risk of child abuse and neglect, and the findings may not be generalizable to the general population of parents of infants.
Among at-risk families, we found an association between parental ratings of the medical home and parental reports of the completeness of anticipatory guidance regarding selected injury and illness prevention topics. Parents' trust of the clinician was associated with parent-reported behavior changes for discussed topics. Additional research is necessary to determine whether the clinician-parent relationship affects the rates of childhood injuries and illnesses and whether the relationship plays the same role with anticipatory guidance topics concerning child development, as well as the role of advice from influential others in parental behavior changes and the factors involved in ethnic differences in clinician-parent relationship ratings.
This project was supported by the Robert Wood Johnson Foundation, Annie E. Casey Foundation, David and Lucile Packard Foundation, Hawaii State Department of Health, and US Maternal and Child Health Bureau grants MCJ-240637 and MCJ-240838. C.S.N. is supported by a Faculty Development in Primary Care Award training grant from the Health Resources and Services Administration, Bureau of Health Professions (grant 5 D14 HP 00118).
Special thanks go to Janet Serwint, MD, and Kimberly Stone, MD, for thoughtful review of this manuscript.
- Accepted June 7, 2004.
- Address correspondence to Anne K. Duggan, ScD, Johns Hopkins University School of Medicine, General Pediatrics Research Center, 1620 McElderry St, Reed Hall 203, Baltimore, MD 21205. E-mail:
This work was presented in part at the Pediatric Academic Societies Annual Meeting; May 3, 2003; Seattle, WA.
No conflict of interest declared.
- ↵Institute of Medicine. Defining Primary Care: An Interim Report. Washington, DC: National Academy Press; 1994
- ↵Institute of Medicine. A Manpower Policy for Primary Health Care: Report of a Study. Washington, DC: National Academy Press; 1978
- ↵American Academy of Pediatrics, Medical Home Initiatives for Children With Special Needs Project Advisory Committee. The medical home. Pediatrics.2002;110 :184– 186
- ↵American Academy of Pediatrics, Committee on Psychosocial Aspects of Child and Family Health. Guidelines for Health Supervision III. Elk Grove Village, IL: American Academy of Pediatrics; 2002
- ↵Green M, Palfrey JS, Clark EM, Anastasi JM, eds. Bright Futures: Guidelines for Health Supervision of Infants, Children, and Adolescents. 2nd rev ed. Arlington, VA: National Center for Education in Maternal and Child Health; 2002
- Reisinger KS, Bires JA. Anticipatory guidance in pediatric practice. Pediatrics.1980;66 :889– 892
- ↵Norkin Goldstein E, Dworkin PH, Bernstein B. Time devoted to anticipatory guidance during child health supervision visits: how are we doing? Ambul Child Health.1999;5 :113– 120
- ↵Thom DH, Kravitz RL, Bell RA, et al. Patient trust in the physician: relationship to patient requests. Fam Pract.2002;19 :476– 483
- ↵Halfon N, Mistry R, Inkelas M, et al. Missed opportunities for anticipatory guidance in pediatric health services for young children [abstract]. Pediatr Res.2003;53 :225A
- ↵Duggan AK, Higman S, Fuddy L, et al. RCT of home visiting: impact on promoting a medical home for environmentally at-risk children. Presented at 2002 Pediatric Academic Societies Annual Meeting; May 5, 2002; Baltimore, MD
- ↵American Academy of Pediatrics. Guidelines for Health Supervision II. Elk Grove Village, IL: American Academy of Pediatrics; 1988
- ↵Korsch BM, Nelson KG, Reinhart JB, et al. Feeding skills in infancy and early childhood. In: Guidelines for Health Supervision II. Elk Grove Village, IL: American Academy of Pediatrics; 1988:127
- ↵US Department of Health and Human Services. Healthy People 2000: National Promotion and Disease Prevention Objectives. Washington, DC: US Department of Health and Human Services; 1990
- ↵American Academy of Pediatrics, Committee on Injury and Poison Prevention. Office-based counseling for injury prevention. Pediatrics.1994;94 :566– 567
- ↵Accreditation Council for Graduate Medical Education. Program requirements for residency education in pediatrics. Available at: www.acgme.org/rq/320pr701.pdf. Accessed August 26, 2003
- ↵Bass JL, Christoffel KK, Widome M, et al. Childhood injury prevention counseling in primary care settings: a critical review of the literature. Pediatrics.1993;92 :544– 550
- ↵Grossman DC, Cummings P, Koepsell TD, et al. Firearm safety counseling in primary care pediatrics: a randomized, controlled trial. Pediatrics.2000;106 :22– 26
- ↵Glanz K, Rimer BK, Lewis FM, eds. Health Behavior and Health Education: Theory, Research, and Practice. 3rd ed. San Francisco, CA: Jossey-Bass; 2002
- ↵Jordan EA, Duggan AK, Hardy JB. Injuries in children of adolescent mothers: home safety education associated with decreased injury risk. Pediatrics.1993;91 :481– 487
- ↵Galuska DA, Fulton JE, Powell KE, et al. Pediatrician counseling about preventive health topics: results from the Physicians' Practices Survey, 1998–1999. Pediatrics.2002;109(5) . Available at: www.pediatrics.org/cgi/content/full/109/5/e83
- Stevens GD, Inkelas M, Kuo AA, et al. Child injury prevention: disparities in physician guidance and parent practice [abstract]. Pediatr Res.2003;53 :218A
- Morrongiello BA, Hillier L, Bass M. “What I said” versus “what you heard”: a comparison of physicians' and parents' reporting of anticipatory guidance on child safety issues. Inj Prev.1995;1 :223– 227
- ↵Cheng TL, Savageau JA, DeWitt TG, et al. Expectations, goals, and perceived effectiveness of child health supervision: a study of mothers in a pediatric practice. Clin Pediatr (Phila).1996;35 :129– 137
- ↵Starfield B. Primary Care: Concept, Evaluation, and Policy. New York, NY: Oxford University Press; 1992
- ↵Kuo AA, Inkelas M, Wright J, et al. Developing medical homes: are pediatricians practicing family-centered care [abstract]? Pediatr Res.2003;53 :224A
- ↵Heneghan AM, Silver EJ, Bauman LJ, et al. Do pediatricians recognize mothers with depressive symptoms? Pediatrics.2000;106 :1367– 1373
- ↵Daly KA, Lindgren B, Giebink GS. Validity of parental report of a child's medical history in otitis media research. Am J Epidemiol.1994;139 :1116– 1121
- Alho OP. The validity of questionnaire reports of a history of acute otitis media. Am J Epidemiol.1990;132 :1164– 1170
- Suarez L, Simpson DM, Smith DR. Errors and correlates in parental recall of child immunizations: effects on vaccination coverage estimates. Pediatrics.1997;99(5) . Available at: www.pediatrics.org/cgi/content/full/99/5/e3
- AbdelSalam HH, Sokal MM. Accuracy of parental reporting of immunization. Clin Pediatr (Phila).2004;43 :83– 85
- O'Sullivan JJ, Pearce MS, Parker L. Parental recall of birth weight: how accurate is it? Arch Dis Child.2000;82 :202– 203
- ↵Chen LH, Gielen AC, McDonald EM. Validity of self-reported home safety practices. Inj Prev.2003;9 :73– 75
- ↵National Center for Injury Prevention and Control. Injury Fact Book 2001–2002. Atlanta, GA: Centers for Disease Control and Prevention; 2001
- ↵Agran PF, Anderson C, Winn D, Trent R, Walton-Haynes L, Thayer S. Rates of pediatric injuries by 3-month intervals for children 0 to 3 years of age. Pediatrics.2003;111(6) . Available at: www.pediatrics.org/cgi/content/full/111/6/e683
- Copyright © 2005 by the American Academy of Pediatrics