Literacy Promotion in Primary Care Pediatrics: Can We Make a Difference?

Literacy Promotion in Primary Care Pediatrics: Can We Make a Difference?

Pamela C. High

ABBREVIATION. CCLO, Child-Centered Literary Orientation.

Success as an adult depends in part on the learning and development that occur in infancy and early childhood. Infants raised in safe and stimulating environments are better learners later in life than those raised in less secure and attentive settings.[1] Because many children from socially and economically disadvantaged backgrounds lack environments that promote language and literacy growth, they are at risk for failing to develop essential literacy skills. The National Assessment of Educational Process 1998 Reading Report Card for the Nation found that children in the United States who live near or below the poverty line have lower average reading scores than their peers.[2] Of adults living in poverty, almost half read and write at the lowest literacy level.[3] This lack of basic literacy skills is linked to decreased productivity, increased rates of teenage pregnancy, and increased welfare dependence.[3,4] Low literacy levels contribute to the propagation of the cycle of poverty.[5]

Several studies have evaluated the efficacy of literacy promoting anticipatory guidance delivered by pediatric primary care providers as part of well-child care. In 1991, Needlman and his colleagues[6] indicated that family literacy orientation could be increased by a simple intervention in which pediatricians distributed children’s books to children of low-income parents at their health maintenance visits. We initially conducted a similar intervention study with multicultural, low-income families (100 intervention, 51 historical control).[7] At all scheduled well-child visits in a hospital clinic setting, primary-care providers distributed children’s books and educational materials to the parents of 6-to 38-month-old children in the intervention group. This study established the concept of Child-Centered Literacy Orientation (CCLO), a measure of a family’s ability and willingness to engage in literacy-promoting activities with young children. CCLO was considered present if parents mentioned reading in answer to open-ended questions about their child’s favorite activities, or their favorite things to do together, or if they usually shared books at bedtime. We found that 33% of parents had CCLO in the historical control group compared with 69% of families who had received the intervention. Controlling for parental education, ethnicity, reading habits, and sex and age of children, we found that CCLO was almost 5 times more likely to be present in the intervention families.[7] These data extended Needlman’s findings because the sample size was larger, and we controlled the number of books children received, the duration between the intervention and the interview, and child health status.

In a related study in pediatric continuity settings, we demonstrated that low-income families with toddlers and preschoolers often lack a CCLO.[8] Only 40% of children [is less than]5 years old had 10 or more books at home, and only 39% of families gave a positive response to one of the questions measuring CCLO. Living in 2-parent families and in those with higher adult to child ratios appeared to have protective effects. However, living in environments where English was rarely spoken and with parents who rarely read were additional risk factors for lack of both children’s books and a CCLO in the home.

In a prospective, randomized trial, pediatricians delivered a bilingual literacy promoting intervention to low-income Hispanic families with infants.[9] Ninety six percent of enrolled families were reinterviewed on average 10 months later when children were 14 to 24 months old. Parents receiving the intervention were 10 times more likely to read books with their children at least 3 days a week compared with control families, and they were also more likely to report that reading aloud was one of their 3 favorite things to do with their child. Only trends toward higher language scores in intervention children [is greater than]18 months old were found.

To expand on these studies and to assess the potential role of community-based pediatric primary care providers in enriching the literacy experiences of low-income multicultural children, we designed a prospective, randomized trial to evaluate the effectiveness of this simple and culturally appropriate literacy-promoting intervention. This study was modeled after the previous study; however, we planned to enroll a larger number of participating families from diverse cultural backgrounds and to deliver the intervention in English alone. We hypothesized that providing age-appropriate children’s books, handouts, and literacy-related anticipatory guidance to low-income parents would provide them with both the knowledge and the tools they would need to change their attitudes about the importance of reading with young children and, as a result, their behavior. Specifically, we hypothesized that this intervention would increase the frequency of parent-child reading in this population. We also hypothesized that our intervention would facilitate toddler’s acquisition of receptive and expressive vocabulary.

METHODS

Subjects

Between August 1996 and September 1997, we enrolled 205 families with infants, who presented to 1 of 4 urban community-based health centers for pediatric care. The health centers serve a low-income multiethnic population. Pediatric care is provided by 7 pediatricians and a pediatric nurse practitioner. Families were eligible if: 1) the interviewed parent could speak English well enough to participate in the initial interview; 2) the interviewed parent was the primary caregiver; and 3) the infant was between 5 and 11 months old at the time of the enrollment interview. Children were enrolled at this age, when pediatric visits occur frequently, to enable us to intervene at multiple points over a short period of time. This is also a time in infant development when parents may be particularly receptive to their pediatrician’s advice and when children start to enjoy looking at picture books. Families were excluded from the study if: 1) the infant’s birth weight was [is less than]5 pounds; 2) the infant had a significant developmental delay, congenital anomaly, or sensory deficit; or 3) the infant had been hospitalized [is greater than]14 days.

Study Design/Intervention

A single research assistant identified potentially eligible families by review of appointment schedules and patients’ charts during busy clinic sessions. Eligible parents were approached in the waiting room and asked to participate in a study of children’s play activities, interests, language development, and sleep habits. Our interest in literacy was not disclosed. Parents were informed that they would be contacted for follow-up interviews in about a year to ascertain how their child’s interests and activities may have changed. Eight parents declined participation in the study, and 15 parents could not communicate in English well enough to complete the interview. After obtaining informed consent, a research assistant conducted an 80-item structured face-to-face interview similar to one reported previously.[16-18] The interview included demographics, children’s play interests and sleeping habits, and parental language proficiencies. Parents received a bag of diapers as an incentive for participating in the interview.

At enrollment, families were randomized to intervention (n = 106) or control group (n = 99), using an alternate day schedule with those enrolled on even days of the month assigned to the intervention group. At this initial visit and at all subsequent well-child visits, pediatricians gave children in the intervention group an age-appropriate children’s board book; an age-specific handout explaining how children can benefit from, enjoy, and interact with books; and literacy promoting anticipatory guidance. The books chosen were developmentally appropriate, contained brightly colored pictures and simple language, depicted culturally diverse images, and promoted parent-child interaction. Infants presenting for the 6-, 9-, 12-, 15-, and 18-month visits received Babies,[10] Goodnight Moon,[11] Moo, Baa, La, La, La!,[12] Bedtime for Bunnies,[13] and Three Little Chicks,[14] respectively. Educational material presented briefly some benefits of reading to children, starting at a very young age and focused on interaction between the parent and the child. This 1-page handout also suggested that a good time to share books was as part of a regular bedtime routine. Handouts were written in English at a third-grade reading level and were given to the parent when their child received the book.[7] The research assistant responsible for enrollment and tracking placed books and handouts in patient charts before scheduled visits for ease of distribution of the intervention. The intervention was determined to have occurred when the child had a documented clinic visit and the book disappeared from the chart.

The pediatric providers participated in a training session describing the study design and objectives and the importance of their participation. We did not structure a standardized scenario for them to deliver, but rather encouraged them to briefly provide guidance on the benefits of reading aloud to children, reinforcing the information in the handouts. Families in the control group received routine pediatric care, without any books or handouts. The providers were asked to continue with their usual anticipatory guidance practices. Control families were not differentiated from nonenrolled families for providers, thereby decreasing the likelihood of unusual practices.

Follow-up

A month after the child had completed 3 well-child visits or, in those with fewer visits, when the child turned 22 months old, we attempted to contact families for follow-up interviews. Of the initial 205 families enrolled, 43 were lost to follow-up. Nine families were located for follow-up interviews when children were 26 to 33 months old and these were considered to be too old for inclusion in this analysis. The remaining 153 families (77 intervention and 76 control families) were reinterviewed when children were between 14 and 25 months old and are the subjects considered in this analysis.

As shown in Table 1, at follow-up mean age of children was 18.7 months for intervention and 18.3 months in the control group. Intervention families had attended 3.4 well-child visits and had received 3.2 books, compared with control families with 3.4 visits but no books. Parents were reinterviewed by 1 of 6 research assistants not involved in enrollment or tracking, and therefore unaware of the families’ group assignment. Interviews were conducted by telephone and in the home when families could not be reached by telephone. A second bag of diapers was given as incentive for participation.

TABLE 1. Demographic Characteristics and Exposure to Intervention at Follow-up (n = 153)

Control (n = 76)

Children

Age 18.3 (.4) mo

13- to 17-month-olds([dagger]) 15.4 (.2) mo

18- to 25-month-olds([dagger]) 20.4 (.3) mo

Female 45% (34)

Parents (%, n)

Female 89% (68)

Education

Not high school graduate 33% (25)(*)

High school graduate or GED 43% (33)(*)

[is greater than or equal to] 24% (18)(*)

1 year college/vocational

Ethnic group

Non-Hispanic white 16% (12)

Hispanic 41% (31)

Other 43% (33)

Born in the continental United 29% (22)(**)

States of America

Languages spoken at home

English only 25% (19)

Multilingual 64% (48)

No English 11% (8)

Speaks English well at follow-up 55% (42)

Single or separated 54% (41)

Adult to child ratio([dagger]) 1.06 (0.1)

Occupation

Employed (full or part time) 44% (33)

Student 21% (16)

Public assistance (low income)

Medicaid 86% (65)

AFDC 47% (36)

WIC 95% (72)

Visits and anticipatory guidance

Number of well-child

visits([dagger]) 3.39 (.8)

Number of books given([dagger]) 0 (0)(***)

Number books parent remembers .35 (.1)(***)

getting([dagger])

Parent memory of anticipatory

guidance

Discussed accident prevention 57% (42)

Discussed child’s nutrition 85% (64)

Reading books with your child 37% (28)

Bedtime routines 25% (19)(*)

Child received a book 12% (9)(***)

Intervention (n = 77)

Children

Age 18.7 (.4) mo

13- to 17-month-olds([dagger]) 15.6 (.2) mo

18- to 25-month-olds([dagger]) 21.0 (.3) mo

Female 45% (35)

Parents (%, n)

Female 92% (71)

Education

Not high school graduate 49% (38)(*)

High school graduate or GED 26% (20)(*)

[is greater than or equal to] 25% (19)(*)

1 year college/vocational

Ethnic group

Non-Hispanic white 21% (16)

Hispanic 53% (41)

Other 26% (20)

Born in the continental United 49% (38)(**)

States of America

Languages spoken at home

English only 26% (20)

Multilingual 62% (48)

No English 12% (9)

Speaks English well at follow-up 69% (53)

Single or separated 55% (42)

Adult to child ratio([dagger]) 1.12 (0.1)

Occupation

Employed (full or part time) 42% (32)

Student 23% (18)

Public assistance (low income)

Medicaid 74% (57)

AFDC 47% (36)

WIC 92% (71)

Visits and anticipatory guidance

Number of well-child

visits([dagger]) 3.38 (.8)

Number of books given([dagger]) 3.25 (.1)(***)

Number books parent remembers 2.65 (.2)(***)

getting([dagger])

Parent memory of anticipatory

guidance

Discussed accident prevention 66% (51)

Discussed child’s nutrition 79% (61)

Reading books with your child 76% (59)

Bedtime routines 42% (32)(*)

Child received a book 91% (70)(***)

P

Children

Age .43

13- to 17-month-olds([dagger]) .66

18- to 25-month-olds([dagger]) .25

Female .93

Parents (%, n)

Female .56

Education

Not high school graduate

High school graduate or GED

[is greater than or equal to] .05(*)

1 year college/vocational

Ethnic group

Non-Hispanic white

Hispanic

Other .08

Born in the continental United .01(**)

States of America

Languages spoken at home

English only

Multilingual

No English .96

Speaks English well at follow-up .08

Single or separated .94

Adult to child ratio([dagger]) .65

Occupation

Employed (full or part time) .76

Student .73

Public assistance (low income)

Medicaid .08

AFDC .94

WIC .53

Visits and anticipatory guidance

Number of well-child

visits([dagger]) .87

Number of books given([dagger]) <.001(***)

Number books parent remembers <.001(***)

getting([dagger])

Parent memory of anticipatory

guidance

Discussed accident prevention .23

Discussed child’s nutrition .33

Reading books with your child <.001(***)

Bedtime routines .03(*)

Child received a book <.001(***)

(*) P < .05. (**) P < .01. (***) P < .001.

[chi square] (%, n) unless specified by t = 2-sample t test with equal variance (mean, SE).

The follow-up interview consisted of a shortened version of the one conducted at baseline. Main outcome variables included the number of days/week that the parent read books with the child and the number of nights/week a parent and child shared books at bedtime. Parents were asked to list their child’s 3 favorite things to do and their 3 favorite things to do with their child. If reading books was mentioned in response to either of these questions or if families shared books at bedtime at least 6 nights/week they were identified as having a CCLO.[7,8] A minimum of 6 nights was chosen as indicative of a firmly established family routine. Parents were also asked how often they read themselves, whether they had a library card, whether they had visited the library with their child, as well as how many total books were in the home. These were considered to be literacy outcome variables less likely to be affected by, or distal to, the intervention. A modified version of the MacArthur Communicative Development Inventories (Short Form), a parent report test that provides a receptive and expressive language score, was also administered at follow-up. We selected 50 words from the MacArthur that were not present in any of the books and 50 words from the books given to all intervention children. This provided receptive (0-100) and expressive (0-100) vocabulary scores. Independent scores were generated for the 50 words present in the books and the 50 words not present in the books.[9]

Contamination of control families with anticipatory guidance was assessed by asking parents, at the conclusion of the interview, if their pediatrician had discussed reading books with their child at their most recent clinic visit. Twenty-eight parents (37%) in the control group and 59 intervention parents (77%) reported receiving such guidance. In addition, 9 control families (12%) and 70 intervention families (91%) remembered being given at least 1 book to take home by their pediatrician.

Statistical Analysis

Data were analyzed using the Stata Statistical Software: Release 5.0. (Stata Corporation, College Station, TX). Frequency counts of categorical variables provided descriptive information about the sample as a whole. [chi square] tests on dichotomous variables and 2-tailed t tests on continuous data were used to compare demographic, literacy, and language variables of the groups. Change scores were computed for literacy outcome variables measured at both enrollment and at follow-up so that each family was its own control. For example, if parents read to their child 3 days/week at enrollment and 5 days/week at follow-up, the change in this would be 2 additional days/week of reading. Hierarchical multiple linear regressions of vocabulary scores were performed correcting for parent education, country of birth, language proficiency, and child age at the first step. For 2 step regressions, intervention status was added at step 2. For 3 step regressions frequency of reading aloud was added at step 2 followed by intervention status at the third step. Power analysis showed that 150 total subjects provide sufficient power ([is greater than].80) to detect medium effects, testing differences between 2 groups (intervention and control). In addition, 88 subjects provide sufficient power (.80) to detect medium effects in a hierarchical regression with 3 sets (first set 4 predictors, second and third sets, 1 predictor each) with [R.sup.2] of the total model of 30% and [R.sup.2] increments of 10%.

Demographic Characteristics

The control and intervention groups that were comparable in their major sociodemographic characteristics with few exceptions. The mean age of the children at baseline was 7.0 ([+ or -] .2) months in both intervention and control groups and 52% were male. Half of these parents (49%) were Hispanic, 17% were non-Hispanic white, and 34% belonged to other ethnic groups. Parents were first asked if they considered themselves to be Hispanic. If their answer was “no,” they were asked if they considered themselves to be white, black, or Asian or something else including biracial. Of those who considered themselves black, more than 42% were born in West Africa, leaving only 7% of the initial parent sample US-born African-American. For this reason those who considered themselves black were included in the diverse other ethnic grouping, which also included Asian, Native American, and biracial parents. Sixty percent of parents were born outside the continental US in 1 of 27 different countries, and two-thirds of families spoke [is greater than] 1 language at home. Parents interviewed were primarily female (92%). Forty-one percent of parents were employed and 23% were students. More than half (62%) had at least a high school equivalency. At enrollment, there were no significant differences between intervention and control families in any of these demographic factors. At follow-up, as shown in Table 1, control parents were more likely to have graduated from high school than intervention parents and less likely to have been born in the continental United States.

Literacy Promoting Behaviors Proximal to the Intervention

At enrollment, when children were on average 7 months old, 2% of parents reported looking at books as 1 of their child’s 3 favorite activities. Fifteen percent said that reading together was one of their favorite activities, and 16% usually shared books at bedtime. Only 27% of families were identified as having a CCLO. Parents reported reading with their children 2.5 days/ week and sharing books as part of a bedtime routine 2.1 nights/ week. Thirty-one percent of these children had [is greater than]10 books of their own and more than a quarter had no books at all. At the time of the baseline interview, these literacy characteristics were virtually identical in intervention and control families.

At follow-up, when children were on average 18.5 months old, significant differences were found between intervention and control families in each of these literacy-related outcome variables, as shown in Table 2. Consistently, intervention families responded with more reading aloud, more enthusiasm for reading together, and more children’s books at home. Intervention families were more likely to report an increase in book-sharing as 1 of their child’s 3 favorite activities, as well as an increase in book-sharing as 1 of their 3 favorite things to do with their child. A significant effect was noted on the frequency of parent-child book-sharing. Intervention families reported a 17% increase in usually sharing books at bedtime compared with control families who had a small decline in this behavior. CCLO increased by 40% in intervention families compared with 16% in control families. Intervention families had increased reading to their toddlers by almost 2 days/week, while control families reported very little change in this activity.

TABLE 2. Literacy Characteristics Proximal and Distal to the Intervention at Follow-up and Change in These Literacy Outcomes as Compared With Baseline at Enrollment (n = 153)

Control (n = 76)

Literacy outcomes proximal to the intervention

Books 1 of child’s favorite 12% (9)(*)

Books 1 of parent’s favorite 33% (25)(**)

Shares books >6 nights/wk 13% (10)(**)

CCLO 43% (33)(**)

Parent reads to child 46% (35)(***)

[is less than or equal to] 3 d/wk

Days/wk parent reads to child([dagger]) 2.8 (.3)(***)

Nights/wk shares book at bedtime([dagger]) 2.1 (.3)(**)

Child has >10 child’s books 45% (34)(*)

Change in proximal literacy outcomes

Change in books 1 of child’s 9% (4%)(*)

3 favorites([dagger])

Change in books 1 of parent’s 18% (6%)(**)

3 favorites([dagger])

Change in shares books >6 nights/wk([dagger]) -4% (5%)(**)

Change in CCLO([dagger]) 16% (6%)(**)

Change in d/wk parent reads to child([dagger]) .16 (.3)(***)

Change in nights/wk share book -.18 (. 3)(***)

at bedtime([dagger])

Literacy outcomes distal to the intervention

Days/wk parent reads a book([dagger]) 3.05 (.3)

Parent has library card 45% (34)

Parent has taken child to library 12% (9)

Parent has used library 25% (19)

Home has <10 total books 20% (15)(*)

Change in distal literacy outcomes

Change in d/wk parent reads books([dagger]) .22 (.4)

Intervention

(n = 17)

Literacy outcomes proximal to the intervention

Books 1 of child’s favorite 27% (21)(*)

Books 1 of parent’s favorite 57% (44)(**)

Shares books >6 nights/wk 32% (32)(**)

CCLO 68% (52)(**)

Parent reads to child 78% (60)(***)

[is less than or equal to] 3 d/wk

Days/wk parent reads to child([dagger]) 4.3 (.2)(***)

Nights/wk shares book at bedtime([dagger]) 3.4 (.3)(**)

Child has >10 child’s books 61% (47)(*)

Change in proximal literacy outcomes

Change in books 1 of child’s 23% (6%)(*)

3 favorites([dagger])

Change in books 1 of parent’s 43% (7%)(**)

3 favorites([dagger])

Change in shares books >6 nights/wk([dagger]) 17% (5%)(**)

Change in CCLO([dagger]) 40% (6%)(**)

Change in d/wk parent reads to child([dagger]) 1.89 (.3)(***)

Change in nights/wk share book 1.63 (.4)(***)

at bedtime([dagger])

Literacy outcomes distal to the intervention

Days/wk parent reads a book([dagger]) 3.93 (.3)

Parent has library card 55% (42)

Parent has taken child to library 16% (12)

Parent has used library 20% (15)

Home has <10 total books 6% (5)(*)

Change in distal literacy outcomes

Change in d/wk parent reads books([dagger]) .56 (.4)

P

Literacy outcomes proximal to the intervention

Books 1 of child’s favorite .02(*)

Books 1 of parent’s favorite .003(**)

Shares books >6 nights/wk .006(**)

CCLO .003(**)

Parent reads to child <.001(***)

[is less than or equal to] 3 d/wk

Days/wk parent reads to child([dagger]) <.001(***)

Nights/wk shares book at bedtime([dagger]) .004(**)

Child has >10 child’s books .04(*)

Change in proximal literacy outcomes

Change in books 1 of child’s .04(*)

3 favorites([dagger])

Change in books 1 of parent’s <.01(**)

3 favorites([dagger])

Change in shares books >6 nights/wk([dagger]) .006(**)

Change in CCLO([dagger]) .007(**)

Change in d/wk parent reads to child([dagger]) <.001(***)

Change in nights/wk share book <.001(***)

at bedtime([dagger])

Literacy outcomes distal to the intervention

Days/wk parent reads a book([dagger]) .07

Parent has library card .23

Parent has taken child to library .50

Parent has used library .93

Home has <10 total books .02(*)

Change in distal literacy outcomes

Change in d/wk parent reads books([dagger]) .56

(*) P < .05. (**) P < .01. (***) P < .001.

[chi square] (%, n) unless specified by ([dagger]) = two-sample t test with equal variance (mean, SE).

Literacy Promoting Behaviors Distal to the Intervention

At enrollment, parents read books themselves an average of 2.9 days/week, and almost half had a library card (49%). Twenty-two percent of parents interviewed had been to the library in the past 6 months, and 12% had taken their child to the library. Almost a quarter of these families (22%) had [is less than]10 books total. At the time of the baseline interview, there were no significant differences in these literacy characteristics between intervention and control families.

At follow-up, as shown in Table 2, parents in both groups reported a slight increase in reading for themselves, but the difference between groups was not significant. Of these literacy variables less likely to be affected by the intervention, the only significant difference was found in total books in the home. Intervention families had more books than control: 22% of parents in the control group reported having [is less than]10 books total in their home compared with 6% of intervention parents.

Child Language

At follow-up, parents were asked if their child understood (receptive vocabulary) or said (expressive vocabulary) each of 100 words in the modified MacArthur Inventories. As shown in Table 3, intervention children had higher receptive vocabulary scores than controls. We divided children into 2 categories for analysis: 13 to 17 months (younger) and 18 to 25 months (older). This division was based on the trends in Golova’s bilingual study toward higher language scores in intervention children over 18 months of age[9] and seemed justified in light of the broad developmental spectrum of language found in children between 13 and 25 months old.

TABLE 3. Language Scores From MacArthur Communication and Development Inventories (Short Form) Modified Version at Follow-up (n = 150)

Control (n = 75)

Receptive Vocabulary 100 words 39.3 (2.7)(**)

Receptive vocabulary (13-17 mo) (n = 62) 40.5 (3.7)

Receptive vocabulary (18-25 mo) (n = 88) 38.3 (3.8)(**)

Receptive Vocabulary 50 words in books 17.6 (1.3)(**)

Receptive vocabulary (13-17 mo) (n = 62) 18.0 (1.7)

Receptive vocabulary (18-25 mo) (n = 88) 17.4 (1.9)(**)

Receptive Vocabulary 50 wds not in books 21.6 (1.4)(**)

Receptive vocabulary (13-17 mo) (n = 62) 22.6 (2.2)

Receptive vocabulary (18-25 mo) (n = 88) 21.0 (1.9)(**)

Expressive Vocabulary 100 words 15.9 (2.3)

Expressive vocabulary (13-17 mo) (n = 62) 13.5 (2.9)

Expressive vocabulary (18-25 mo) (n = 88) 17.5 (3.3)(*)

Expressive Vocabulary 50 words in books 7.2 (1.1)

Expressive vocabulary (13-17 mo) (n = 62) 6.0 (1.4)

Expressive vocabulary (18-25 mo) (n = 88) 8.0 (1.6)(**)

Expressive vocabulary 50 wds not in books 8.7 (1.2)

Expressive vocabulary (13-17 mo) (n = 62) 7.6 (1.5)

Expressive vocabulary (18-25 mo) (n = 88) 9.5 (1.7)(**)

Intervention (n = 75)

Receptive Vocabulary 100 words 51.0 (3.0)(**)

Receptive vocabulary (13-17 mo) (n = 62) 45.7 (4.4)

Receptive vocabulary (18-25 mo) (n = 88) 54.8 (4.1)(**)

Receptive Vocabulary 50 words in books 23.4 (1.5)(**)

Receptive vocabulary (13-17 mo) (n = 62) 19.9 (2.2)

Receptive vocabulary (18-25 mo) (n = 88) 25.3 (2.1)(**)

Receptive Vocabulary 50 wds not in books 28.0 (1.6)(**)

Receptive vocabulary (13-17 mo) (n = 62) 25.8 (2.3)

Receptive vocabulary (18-25 mo) (n = 88) 29.5 (2.1)(**)

Expressive Vocabulary 100 words 22.1 (3.1)

Expressive vocabulary (13-17 mo) (n = 62) 7.9 (1.7)

Expressive vocabulary (18-25 mo) (n = 88) 32.1 (4.6)(*)

Expressive Vocabulary 50 words in books 10.3 (1.5)

Expressive vocabulary (13-17 mo) (n = 62) 3.4 (.9)

Expressive vocabulary (18-25 mo) (n = 88) 15.3 (2.3)(**)

Expressive vocabulary 50 wds not in books 11.8 (1.6)

Expressive vocabulary (13-17 mo) (n = 62) 4.5 (.9)

Expressive vocabulary (18-25 mo) (n = 88) 16.9 (2.4)(**)

P

Receptive Vocabulary 100 words .004(**)

Receptive vocabulary (13-17 mo) (n = 62) 0.37

Receptive vocabulary (18-25 mo) (n = 88) .004(**)

Receptive Vocabulary 50 words in books .008(**)

Receptive vocabulary (13-17 mo) (n = 62) 0.48

Receptive vocabulary (18-25 mo) (n = 88) .006(**)

Receptive Vocabulary 50 wds not in books .003(**)

Receptive vocabulary (13-17 mo) (n = 62) 0.31

Receptive vocabulary (18-25 mo) (n = 88) .004(**)

Expressive Vocabulary 100 words 0.11

Expressive vocabulary (13-17 mo) (n = 62) 0.10

Expressive vocabulary (18-25 mo) (n = 88) .01(*)

Expressive Vocabulary 50 words in books 0.10

Expressive vocabulary (13-17 mo) (n = 62) 0.13

Expressive vocabulary (18-25 mo) (n = 88) .01(**)

Expressive vocabulary 50 wds not in books 0.13

Expressive vocabulary (13-17 mo) (n = 62) 0.10

Expressive vocabulary (18-25 mo) (n = 88) .01(**)

(*) P < .05. (**) P < .01. (***) P < .001. Two-sample t tests with equal variance (mean, SE).

Higher receptive and expressive vocabulary scores were found in older, but not younger, intervention toddlers. The significant effect of the intervention on both receptive and expressive vocabulary scores in older toddlers was found for both the 50 words in the books and the 50 words not in the books (Table 3). To explore this finding further and to correct for variation in demographic variables between groups, hierarchical multiple linear regressions of both receptive and expressive vocabulary scores were performed, controlling for parent education, foreign birth and language proficiency, and child age at step 1, before entering intervention status at the second step. Again significantly higher receptive and expressive vocabulary scores were found in older (Tables 4 and 6) intervention toddlers. The intervention accounted for 6% of the variability in receptive vocabulary (Table 4) and 4% of the variability in expressive vocabulary in this older toddler cohort. In similar 3-step analyses with the frequency of reading aloud to children is added to these models at step 2 and intervention status entered at step 3, effects of the intervention were no longer present (Tables 5 and 7). In these models, reading aloud to children accounted for 17% of the variability in receptive (Table 5) and 11% of the variability in expressive (Table 7) vocabulary for this older group. This finding suggests that the effect of the intervention was mediated through increased reading aloud to these older toddlers.

TABLE 4. Hierarchical Multiple Linear Regression of Receptive Vocabulary Scores in Older Toddlers (18-25 Months, n = 88) Correcting First for Parent Education, Foreign Birth, and Language Proficiency and Child Age at Step 1, Before Entering Intervention Status at Step 2

Receptive Vocabulary Older Toddlers (n = 88)

Coefficient SE

Step 1: parental education, foreign birth

and language proficiency and child age

Step 2: intervention 13.7 5.3

[Beta] P

Step 1: parental education, foreign birth .001

and language proficiency and child age

Step 2: intervention .25 .0

Adjusted [R.sup.2]

by Step

(Incremental)

Step 1: parental education, foreign birth .15

and language proficiency and child age

Step 2: intervention .06

TABLE 5. Hierarchical Multiple Linear Regression of Receptive Vocabulary Scores in Older Toddlers (18-25 Months, n = 88) Correcting First for Parent Education, Foreign Birth, and Language Proficiency and Child Age at Step 1, Before Entering Frequency of Reading Aloud at Step 2 and Then Intervention Status at Step 3 Receptive Vocabulary Older Toddlers (n = 88)

Coefficient SE

Step 1: parental education, foreign birth and

language proficiency and child age

Step 2: days per week parent reads to child 4.3 1.1

Step 3: intervention 6.4 5.2

[Beta] P

Step 1: parental education, foreign birth and .001

language proficiency and child age

Step 2: days per week parent reads to child .38 <.001

Step 3: intervention .12 .22

Adjusted [R.sup.2]

by Step

(Incremental)

Step 1: parental education, foreign birth and .15

language proficiency and child age

Step 2: days per week parent reads to child .17

Step 3: intervention .0

TABLE 6. Hierarchical Multiple Linear Regression of Expressive Vocabulary Scores in Older Toddlers (18-25 Months, n = 88) Correcting First for Parent Education, Foreign Birth, and Language Proficiency and Child Age at Step 1, Before Entering Intervention Status at Step 2

Expressive Vocabulary in Older Toddlers (n = 88)

Coefficient SE

Step 1: parental education, foreign birth and

language proficiency and child age

Step 2: intervention 11.8 5.4

[Beta] P

Step 1: parental education, foreign birth and .001

language proficiency and child age

Step 2: intervention .22 .03

Adjusted [R.sup.2]

by Step

(Incremental)

Step 1: parental education, foreign birth and .15

language proficiency and child age

Step 2: intervention .04

TABLE 7. Hierarchical Multiple Linear Regression of Expressive Vocabulary Scores in Older Toddlers (18-25 Months, n = 88) Correcting First for Parent Education, Foreign Birth, and Language Proficiency and Child Age at Step 1, Before Entering Frequency of Reading Aloud at Step 2 and Then Intervention Status at Step 3

Expressive Vocabulary in Older Toddlers (n = 88)

Coefficient SE

Step 1: parental education, foreign birth and

language proficiency and child age

Step 2: days per week parent reads to child 3.5 1.2

Step 3: intervention 5.8 5.5

[Beta] P

Step 1: parental education, foreign birth and .001

language proficiency and child age

Step 2: days per week parent reads to child .31 .003

Step 3: intervention .11 .30

Adjusted [R.sup.2]

by Step

(Incremental)

Step 1: parental education, foreign birth and .15

language proficiency and child age

Step 2: days per week parent reads to child .11

Step 3: intervention .01

DISCUSSION

In a multicultural group of low-income families, the provision of children’s books and educational materials about sharing books with children by pediatric providers as part of well-child care significantly increased CCLO. We were able to show that the intervention increased the frequency of parent-child book-sharing as well as the child and parent’s enjoyment of reading together. This is particularly important because parental involvement in children’s education is an important factor in school achievement.[15] In addition, we believe this to be the first prospective, randomized trial to demonstrate that a literacy promoting intervention delivered by pediatric primary care providers can enhance the development of children’s early oral language skills.

We found a strong association between this intervention and both receptive and expressive vocabulary in older toddlers. Regression analyses controlling demographic variables support the hypothesis that this effect was mediated through an increased emphasis on and enjoyment of shared reading experiences in intervention homes. These data further support the view that sharing books with young children encourages the development of early oral language skills. However, it should be noted that we are able to account for only 26% to 32% of the variance in language outcomes in this sample. Many other factors also play important roles in vocabulary acquisition.

How might reading aloud with children promote language and literacy? Book-sharing may impact language development and increase vocabulary in toddlers through object-word association. Researchers hypothesize that reading provides rhythmic patterning and repetition that serve as memory aides for young children, helping them to retain new vocabulary.[16] Research indicates a relationship between joint book-sharing and children’s development of language skills and acquisition of vocabulary,[17,22] which indirectly may result in acquisition of literacy. Reading aloud may also directly promote literacy acquisition through fostering emergent literacy skills such as orientation of the book, listening ability, page turning and print recognition.[23] Reading to young children, therefore, may help enhance reading achievement by strengthening the child’s language skills, by giving them emergent literacy skills and by teaching them to enjoy books and reading enough that they are willing to work as hard as is necessary to master these skills themselves.[21,24] Because this intervention increases reading aloud and thereby enhances language development, it will likely lead to easier acquisition of literacy for this at-risk group of children.

The effectiveness of this intervention may have resulted from several factors. The books were selected to appeal to both parents and children with colorful pictures of animals or children and with simple and sometimes humorous language. They were sturdy board books that often contained finger puppets or peek holes to offer multiple opportunities for parent-child interaction. The educational materials were written at a third grade reading level and bulleted, making it clear and easy for parents to read. Pediatricians’ guidance reinforced the information in the handouts. Pediatric providers in primary care settings are important sources of advice and support for families. Anticipatory guidance provided in this setting has been shown to influence parent behaviors in the areas of injury prevention[25] and sleep promotion.[26] In addition, the intervention took place at 3 or more visits and the presence of the books at home may have served as a further reminder, strengthening the message that reading aloud to young children is important. Yet another factor that may have contributed to the success of this intervention was that bedtime was recommended as a particularly good time to read with children. This provided structure and a regular time in the day for reading to occur.

The intervention did not appear to have a significant effect on the number of days/week parents themselves read books; however, the intervention may have modified parents’ beliefs regarding the importance of reading books with their children. The intervention did not have an effect on whether or not the parents had a library card, visited the library to borrow books, or had taken their children to the library. Consistent with the findings of previous studies,[7,9] these outcome variables distal to the intervention were not affected by our intervention.

One limitation of this study was our failure to achieve equivalent distributions in demographic variables in intervention and control groups at follow-up. The overrepresentation of higher parental education in control parents should bias the results in the direction contrary to the hypothesis. One might expect that more educated parents might read to their children more and, thereby, their children could have higher vocabulary scores. On the other hand, the use of change scores, as shown in Table 2, might bias the results in the opposite direction due to a ceiling effect in the more educated group with less room for improvement available. Baseline literacy characteristics were so low in both groups, however, that this may not have been an important factor. Given the extreme variability in parental ethnicity, it is not surprising that differences in language proficiency and country of origin were found between groups. As previously noted, parents had been born in 1 of 28 different countries;. Although there were no between-group differences in numbers of languages spoken in the home, there was a trend toward more comfort speaking English in intervention parents at follow-up that could skew findings in the direction of the hypothesis. Multivariate analyses were used to control for these between-group differences in parents’ country of origin, language proficiency and education as well as child age, an obvious confounder.

Another limitation of this study was the attrition of 25% of the enrolled sample. This loss did not appreciably reduce power in these analyses. The effect size remains medium (.44 vs .48) with 150 compared with 205 subjects in 2 group comparisons. This rate of attrition is not surprising for a 10- to 20-month follow-up within a low-income population in which household moves are frequent and telephone numbers are often changed or disconnected. In fact, an additional 4% of the sample was reinterviewed when children were 26 to 32 months old. No significant differences in any of the demographic or literacy variables studied were found in those families included in this analysis, compared with those lost to this first follow-up. Further, the attrition rate was equal in both intervention and control groups.

A third limitation of our study was that all outcome measures were based on data obtained by parental report. However, there is a long history of evidence supporting the use of parental report as an accurate means of assessing vocabulary and grammar in both typically and atypically developing children, especially in toddlers and preschoolers.[27-29] In addition, parents were not informed of our interests in literacy, did not know that the interviews were associated with the books that some had received, and the questions used to determine CCLO were open-ended. Finding effects of the intervention primarily in proximal literacy outcome variables, rather than in distal literacy outcome variables, supports the validity of these results. Parents were not trying to please us by providing uniformly positive literacy responses. Another limitation that relates to our language outcome measure was that we did not use a standardized and validated instrument to measure these skills. Although the instrument utilized was based on a standardized language test, it was specifically designed for this intervention. We were pleased to see effects of the intervention in both the words taken from the standardized test and in words taken from intervention books. Future studies that include home visitation with direct counts of books and the administration of standardized language testing of the children directly may provide more conclusive data to support our findings.

Another limitation of this work was that although the 6 research assistants who performed follow-up interviews were blind to a family’s group assignment, they were not blind to the study hypothesis. Parents revealed their attitudes and practices around literacy early in the interview, before providing responses to the modified MacArthur vocabulary tests. This could potentially influence interviewers to skew vocabulary data in the direction of the hypothesis. The highly structured and simple format of this instrument, our reliance on data obtained by six different individuals, our failure to find significant effects in younger toddlers, and the known reliability of similar instruments used widely in clinical and research settings, all argue against the possibility that this limitation strongly influenced our results.

This study is important because it links pediatric anticipatory guidance, including giving books to low-income infants and toddlers, with changes in parental attitudes and behaviors and also with improvement in child language outcomes. Because of the randomized prospective design of this study, it supports a causal link between early reading aloud and child language. This is important for 3 reasons: 1) The effects of the intervention on child language strongly support the conclusion that the intervention affected not only parent attitudes, but their behaviors as well, a much more important goal. 2) While many previous studies have established positive associations between reading aloud and child language, potential confounding from other sources, such as these same parents talking more with their children, has weakened claims to a causal connection. The randomized design of this intervention study largely eliminates the concern about unmeasured confounders, and thus it strongly supports the claim that reading aloud itself promotes language development. 3) The majority of previous research on reading aloud to children has addressed preschool-aged children, while this study extends this age range down into toddlerhood and infancy.

Interventions similar to the one in this study may encourage the development of literacy and language abilities in educationally at-risk low-income children and may help to equip children with the literacy skills they will need in the future, as they learn to read by and for themselves. Children must be followed for a longer period of time to assess whether interventions promoting sharing books starting at a very young age can influence children’s later kindergarten preparedness, reading abilities, language skills, and overall school achievement. Moreover, this is the first study to show that literacy promoting anticipatory guidance at well-child visits can have a significant effect on language development in toddlers. Pediatricians are in a unique position to counsel high-risk parents about the joys, pleasures, and ultimate benefits of reading with their young children.

ACKNOWLEDGMENTS

Funding for this research was provided by the Fleet Scholarship in Social Pediatrics at Brown University School of Medicine, the Eleanor Naylor Dana Charitable Trust, and the Health Resources and Services Administration, Department of Health and Human Services, Maternal and Child Health Bureau, Project No. MCJ-449505.

We would like to thank the pediatricians at the Providence Community Health Centers (Drs Stanley Block, Ellen Gurney, Margaret Lespinasse, John Moran, Boris Skurkovich, David Tapscott, and Lonnie Thomas) as well as the nurse practitioners, nurses, and administrators at the Capitol Hill, Olneyville, Allen Berry, and Fox Point Health Centers in Providence, Rhode Island, who participated in this project. We are grateful to research assistants Margarita Andrade, Alysia Turner, Arania Adolphson, and Phebe Vandersip for their dedication to the project and the extreme effort they put forth locating families for data collection. We are also indebted to Drs Natalia Golova, Anthony Alario, and Patrick Vivier who helped devise the model after which this project was designed.

REFERENCES

[1.] Starting Points: Meeting the Needs of Our Youngest Children. New York, NY: Carnegie Task Force on Meeting the Needs of Young Children; 1994. http://www.carnegie.org/starting_points/startpt1. html

[2.] US Department of Education. Office of Educational Research and Improvement. National Center for Education Statistics. Donahue PL, Voelkl KE, Campbell JR, Mazzeo J. The National Assessment of Educational Process 1998 Reading Report Card for the Nation, NCES 1999-500. Washington, DC: 1999. http://nces.ed.gov/ nationsreportcard/pubs/main1998/1999500.shtml

[3.] US Department of Education, Office of Educational Research and Improvement. National Center for Education Statistics. Adult Literacy in America. Washington, DC: 1993. http://nald.ca/fulltext/ Report2/rep15-01.htm 4. Berlin G, Sum A. Toward a More Perfect Union: Basic Skills, Poor Families and Our Economic Future. New York, NY: Ford Foundation Project on Social Welfare and the American Future; 1988

[4.] Anastasiow NJ, Levine Hanes M, Hanes ML. Language and Reading Strategies for Poverty Children. Baltimore, MD: University Park Press; 1982

[5.] Needlman R, Freid L, Morley D, Taylor S, Zuckerman B. Clinic-based intervention to promote literacy: a pilot study. Am J Dis Child. 1991; 145:881-884

[6.] High P, Hopmann M, LaGasse L, Linn H. Evaluation of a clinic-based program to promote book sharing and bedtime routines among low-income urban families with young children. Arch Pediatr Adolesc Med. 1998;152:459-465

[7.] High P, Hopmann M, LaGasse L, et al. Child-centered literacy orientation: a form of social capital? Pediatrics. 1999;103(4). URL: http://www.pediatrics.org/cgi/content/full/103/4/e55

[8.] Golova N, Alario A, Vivier P, Rodriguez M, High P. Literacy promotion for Hispanic families in a primary care setting: a randomized, controlled trial. Pediatrics. 1999;103:993-997

[9.] Fujikawa G. Babies. So Tall Board Books. New York, NY: Grosset and Dunlap, Inc; 1991. ISBN: 0-448-03084-5

[10.] Wise Brown M. Goodnight Moon. New York, NY: Harper and Row Publishers; 1991

[11.] Boynton S. Moo, Baa, La, La, La! New York, NY: Simon and Schuster Children’s Publishing Division; 1995

[12.] Sadler M, Bollen R. Bedtime for Bunnies. New York, NY: Random House; 1994

[13.] Smee N. Three Little Chicks. New York, NY: Scholastic Inc; 1994

[14.] Years of Promise: A Comprehensive Learning Strategy for America’s Children. New York, NY: Carnegie Task Force on Learning in the Primary Grades; 1996. http://www.carnegie.org/execsum.html

[15.] Moerk EL. Picture-book reading by mothers and young children and its impact on language development. J Pragmatics. 1985;9:547-566

[16.] Debaryshe, BD. Joint picture-book reading correlates of early oral language skill. J Child Lang. 1993;20:455-461

[17.] Elley W. Vocabulary acquisition from listening to stories. Reading Res Q. 1989;24:174-187

[18.] Whitehurst G, Epstein J, Angell A, Payne A, Crone D, Fischel J. Outcomes of an emergent literacy intervention in Head Start. J Educ Psychol. 1994;86:542-555

[19.] Scarborough H, Dobrich W. On the efficacy of reading to preschoolers. Dev Bev. 1994;14:245-302

[20.] Lonigan C. Reading to preschoolers exposed: is the emperor really naked? Dev Rev. 1994;14:303-323

[21.] Kagan J. The role of parents in children’s psychological development. Pediatrics. 1999;104(suppl):164-167

[22.] Teale WH. Reading to young children: its significance for literacy development. In: Goelman H, Oberg A, Smith F, eds. Awakening to Literacy. London, United Kingdom: Heinemann Educational Books; 1984

[23.] Gottfried AE, Fleming JS, Gottfried AW. Role of cognitively stimulating home environment in children’s academic intrinsic motivation: longitudinal study. Child Dev. 1998;69:1448-1460

[24.] Bass JL, Christoffel KK, Widome M, et al. Childhood injury prevention counseling in primary care settings: critical review of the literature. Pediatrics. 1993;92:544-550

[25.] Adair R, Zuckerman B, Bauchner H, Phillips B, Levenson S. Reducing night-waking in infancy: a primary care intervention. Pediatrics. 1992;89:585-588

[26.] Dales PS, Bates E, Reznick JS, Morisset C. The validity of a parent report instrument of child language at twenty months. J Child Lang. 1989;16:239-249

[27.] Dromi E. Early Lexical Development. New York, NY: Cambridge University Press; 1987

[28.] Thal DJ, O’Hanlon L, Clemmons M, Fralin L. The validity of a parent report measure of vocabulary and syntax for preschool children with language impairment. J Speech Lung Hear Res. 1999;42:482-496

Pamela C. High, MD(*[double dagger][sections]); Linda LaGasse, PhD([double dagger][sections]); Samuel Becker, BA(*); Ingrid Ahlgren, BA(*); and Adrian Gardner, BS([double dagger])

From the (*) Department of Pediatrics, Rhode Island Hospital; ([double dagger]) Brown University School of Medicine; and ([sections]) Women and Infants Hospital, Providence, Rhode Island.

Portions of these data were presented at the Pediatric Academic Societies Meeting, May 2, 1999, San Francisco, CA; and at the Society for Developmental and Behavioral Pediatrics Meeting, September 28, 1999, Seattle, WA.

Received for publication Aug 30, 1999; accepted Dec 28, 1999.

Reprint requests to (P.C.H.) Infant Development Center, Women and Infants Hospital, 101 Dudley St, Providence, RI 02905.

COPYRIGHT 2000 American Academy of Pediatrics

COPYRIGHT 2001 Gale Group