Skip to main content
NCBI home page
HLRP: Health Literacy Research and Practice logoLink to HLRP: Health Literacy Research and Practice
. 2023 May 8;7(2):e89–e98. doi: 10.3928/24748307-20230419-01

Patterns of Change in Parental Health Literacy in Relation to Children's Oral Health

Sarah J Schmiege ✉, Luohua Jiang, Judith Albino, Rachel L Johnson, Anne R Wilson, Angela G Brega
PMCID: PMC10169466  PMID: 37162254

Abstract

Background:

Although health literacy (HL) skills may change over time, most research treats HL as a constant, using baseline HL to predict other health-related constructs. Few studies have explored change in HL over time.

Objective:

We examined person-level differences in HL trajectories. We identified subgroups (latent classes) based on longitudinal assessments of HL and examined the association of class membership with demographic and oral health variables.

Methods:

We used four measurement waves of parental HL data, reflecting the risk of limited HL, collected as part of an intervention to reduce dental decay in American Indian children (N = 579 parent-child dyads at baseline). Repeated measures latent class analysis (RMLCA) models were estimated to identify subgroups of HL trajectories over time. We examined class membership in association with baseline demographics and with 36-month assessments of parental oral health knowledge, beliefs, and behaviors as well as pediatric oral health.

Key Results:

A four-class model best fit the data. The largest class (high HL; 49.7% of the sample) was characterized by high levels of HL at all waves. A second class (improving HL; 17.7%) improved over all waves. The remaining two classes were characterized as moderate HL (20%) and low HL (12.6%) and maintained relatively stable HL levels over time. Higher educational attainment was associated with membership in the high HL and improving HL classes. Older age among this young-adult sample and higher income also were associated with high HL class membership. Parents in the high HL and improving HL classes exhibited more favorable performance on measures of oral health knowledge, beliefs, and behavioral adherence than did those in the other classes. Class membership was not associated with pediatric oral health.

Conclusions:

RMLCA demonstrated person-level variability in HL trajectories. Longitudinal patterns were associated with baseline demographics and prospectively with parental oral health knowledge, beliefs, and behaviors, but not with pediatric oral health. [HLRP: Health Literacy Research and Practice. 2023;7(2):e89–e98.]

Plain Language Summary:

Longitudinal changes in HL are understudied. We categorized parents of American Indian children into latent classes based on level of HL (high, medium, and low) and changes in HL over a 3-year period. Both high and increasing HL were associated with higher education and income and with more favorable values, knowledge, beliefs, and behavior related to children's oral health.

Health literacy (HL) is commonly defined as “the capacity to obtain, process, and understand basic health information and services needed to make appropriate health decisions” (Ratzan & Parker, 2000). Skills underlying HL include ability to read, write, and engage in effective verbal communications on health-related topics, and understand and use numbers involved in health management (Institute of Medicine Committee on Health Literacy, 2004). Although HL and its component skills have potential to change over time (Institute of Medicine Committee on Health Literacy, 2004), HL is treated as a constant in most research studies. Customarily, investigators examine the association of baseline HL with other health-related constructs at baseline or longitudinally.

Rarely do researchers examine change in HL itself (Curtis et al., 2021; Edwards et al., 2012; Morris at al., 2013). Two longitudinal studies involving adults demonstrated a decline in HL over time—Curtis et al., 2021: age range: 55 to 74 years, mean 63; and Morris et al., 2013: age range: 22–92 years, mean 64.. This finding is consistent with evidence that HL limitations are common among older adults (Kutner et al., 2006) and that HL is associated with cognitive functioning (Baker et al., 2008; Wolf et al., 2012). Only one known study has explored the potential for HL to improve over time, showing improved HL among participants in a depression intervention who took part in adult basic education classes (Weiss et al., 2006).

We investigated longitudinal change in HL of parents with young children enrolled in a randomized controlled trial (RCT) designed to improve oral health outcomes among American Indian children (Batliner et al., 2014; Batliner et al., 2018), who experience a substantially higher rate of dental disease compared to their Black, Hispanic, and White peers (Indian Health Service, 2013; Phipps & Ricks, 2015; Phipps et al., 2012; Phipps et al., 2019). In conducting this secondary analysis, aims included: (1) identifying patterns of change over time in parental HL, (2) examining the association of sociodemographic characteristics with change patterns, and (3) determining how patterns of change in HL are associated with parental oral health knowledge, beliefs, and behaviors and pediatric oral health.

We used repeated measures latent class analysis (RMLCA) to classify parents into subgroups based on patterns of change in HL over four waves of data collection. The measure of HL used in this study assessed parents' self-reported confidence in their ability to read and complete health-related forms and reflects the risk of having limited HL. Because latent class methods are data driven, we did not have specific hypotheses regarding the number and nature of anticipated classes. However, we expected individual variability in change as reflected by emergence of more than one latent class. Based on prior research showing that income and educational attainment are associated with HL (Kutner et al., 2006), we anticipated these sociodemographic characteristics would be linked to change in HL. Given prior data linking stronger HL with more optimal oral health beliefs and behaviors (Brega et al., 2020; Brega, Johnson, Jiang, et al., 2021; Brega, Johnson, Schmiege, Jiang, et al., 2021; Brega, Johnson, Schmiege, Wilson, et al., 2021; Brega et al., 2016; Vann et al., 2010; Vilella et al., 2016), we anticipated improvement in HL would be associated with better performance on the parental oral health constructs and pediatric oral health.

Methods

Participants and Procedures

Data were collected as part of the Promoting Behavioral Change for Oral Health in American Indian Mothers and Children (PBC) clinical trial, described in detail in earlier reports (Batliner et al., 2014; Batliner et al., 2018). The PBC study tested a motivational interviewing intervention (Borrelli et al., 2015; Miller & Rose, 2009) aimed at reducing dental decay in American Indian children residing on or near a specific Indian reservation in the Northern Plains. Parent-child dyads (N = 579) were enrolled, randomized, and reassessed annually for 3 years (Batliner et al., 2018).

The participating Tribe's research review board and the Colorado Multiple Institutional Review Board approved this secondary analysis and the original study. Written informed consent and HIPAA authorization were obtained from parents prior to participation.

Measurement

Overview

At all measurement waves, parents completed the Basic Research Factors Questionnaire (BRFQ) (Albino et al., 2017), which included questions assessing HL; sociodemographic characteristics; and parental oral health knowledge, beliefs, and behavior. All assessed BRFQ items have been validated in American Indian populations (Brega et al., 2012; Wilson et al., 2016; Wilson et al., 2018). Dental evaluations were completed at ages 12, 24, and 36 months (Batliner et al., 2014; Batliner et al., 2018) by anonymized/licensed dental providers and calibrated using standardized criteria (Pitts, 2001, 2004; Warren et al., 2015).

Measurement Time Points

The RMLCA used HL data collected at four waves (baseline, 12, 24, and 36 months). Participant demographics were measured at baseline. Parental oral health knowledge, beliefs, and behavior as well as pediatric oral health were assessed at 36 months.

Measures

Health literacy. We measured HL as the sum of three items examining participants' self-reported confidence in reading and completing medical forms. As previously described, the questions were adapted from existing items known to accurately identify patients with inadequate HL (Brega et al., 2020). Because these screening items were subjective rather than performance based, they are best understood as measuring the risk of limited HL. The possible range of the HL score was 3 to 15, with larger numbers indicating stronger HL skills (baseline: M [mean] = 11.78, SD [standard deviation] = 2.43).

Parental oral health knowledge. Seventeen items assessed parents' knowledge of pediatric oral health and recommended parental oral health behaviors. Each item was coded as correct or incorrect/don't know. We computed the percentage of questions answered correctly (36 months: M = 75.7, SD = 12.9).

Parental oral health beliefs. We assessed five constructs from the expanded Health Belief Model (HBM) (Janz & Becker, 1984; Rosenstock, 1974; Rosenstock et al., 1988). Applied to oral health, the expanded HBM suggests parents are more likely to engage in positive oral health behaviors if they perceive their children to be susceptible to cavities, believe cavities to be a severe problem, have few barriers to and many benefits of recommended oral health behaviors, and have confidence (self-efficacy) in their capability to engage in those practices. Three to five items assessed each of four constructs: perceived susceptibility (36 months: M = 2.8, SD = 1), severity (36 months: M = 4.3, SD = 0.8), barriers (36 months: M = 2.1, SD = 0.9), and benefits (36 months: M = 4.4, SD = 0.7). All items were measured on a 1 to 5 scale with larger numbers reflecting greater endorsement of the construct. We computed the mean for each construct.

Self-efficacy was assessed using 14 items. Consistent with prior research (Brega et al., 2020; Brega, Johnson, Jiang, et al., 2021; Brega, Johnson, Schmiege, Jiang, et al., 2021; Brega, Johnson, Schmiege, Wilson, et al., 2021), we computed self-efficacy as the number of items for which the highest score was chosen (possible range 0–14; 36 months: M = 7.9, SD = 4.1).

Adherence to recommended parental oral health behaviors. Thirteen items assessed adherence to oral health behaviors recommended for parents of young children. Responses were coded as adherent or nonadherent, with the behavioral adherence score computed as the percentage of behaviors for which a parent was adherent (36 months: M = 52, SD = 20).

Pediatric oral health. We calculated the number of decayed, missing, and filled tooth surfaces (DMFS) for each child as an objective assessment of pediatric oral health. The scoring of DMFS was based on dental evaluation data collected by calibrated dental examiners who were masked to group assignment (Warren et al., 2015). In our analyses, we used the 36-month DMFS value, which reflected cumulative disease burden over the first 3 years of life (36 months: M = 10.2, SD = 16.1).

Sociodemographic characteristics. Characteristics of parents included age, highest grade completed, household income, gender, race and ethnicity, and employment status.

Statistical Analysis

Analyses were performed using Mplus (version 8.3) (Muthén & Muthén, 1998–2017) and R version 4.0.5 (R Core Team, 2022). Analyses combined participants across treatment groups to maximize sample size. We included treatment group as a covariate and found no impact on model conclusions.

We used RMLCA to identify subgroups of parents with similar patterns of change in HL over time. Each parent was assigned to a specific latent class of participants who shared similar patterns. We estimated the four waves of HL as indicators in a standard latent class model. Model estimation followed a traditional empirical approach, where we started with a 1-class model and increased the number of classes until the best-fitting model was observed. Relative fit indices included the Bayesian Information Criterion (BIC) and the Consistent Akaike's Information Criterion (CAIC) (lower is better for both). Fit was also assessed using a likelihood ratio test, which compares the k-class model under consideration to a k-1 class, where a significant p value indicates that the estimated model fits better than a model with one fewer class (Masyn, 2013; Nylund et al., 2007). We selected the best-fitting model based on the cumulative information provided by model fit indices and the requirement that each class include a minimum of 5% of the sample.

Class membership was examined in association with baseline sociodemographic characteristics using analysis of variance (continuously measured demographics) and Fisher's exact tests (categorically measured demographics that included some small cell sizes) and with oral health variables (knowledge, beliefs, behavior, DMFS) using regression. In regression models, 36-month oral health variables were regressed on dummy-coded indicators of class membership, adjusting for age, education, income, and treatment group. Models appropriately addressed the response distribution for each oral health variable, with linear regression used for most variables and a negative binomial distribution for DMFS, which is highly skewed with overdispersion. All modeling was performed most likely using class membership with participants assigned to the class most reflective of their HL trajectory. Although this “classify/analyze” approach has better interpretability because it is easily applied using standard regression methods, it does not address classification uncertainty (Clark & Muthén, 2009). Therefore, a sensitivity analysis was conducted to confirm model conclusions using a weighting procedure (Asparouhov & Muthén, 2014) to account for classification uncertainty. As model conclusions were unchanged, the simpler approach was retained for interpretability.

Results

Participant Demographics

Participant demographics have been fully presented in prior work (Batliner et al., 2018; Brega et al., 2020), but are summarized. Mean parental baseline age was 25.1 years (SD = 5.5). Almost all parents were women (97%) and American Indian (96%). A small number of parents (5%) reported being Latino/a/e. More than one-third (35.9%) reported completing some college coursework and more than one-half reported having an income less than $10,000 (51.2%) and being unemployed (52.1%).

Health Literacy Trajectory Patterns

Fit indices for 1- to 4-class models are shown in Table 1. The smallest class in the 5-class solution (not shown) only contained 2.8% of the sample and, therefore, this model was not considered further as a candidate model. The BIC and CAIC values were lowest for the 4-class solution relative to the other candidate models. Additionally, there was evidence of statistical improvement of the 4-class model relative to the 3-class model based on the marginally significant likelihood ratio test (p = .068). We selected the 4-class solution as the final model based on fit statistics and meeting the minimum class size requirement.

Table 1.

Fit Statistics for 1-Class to 4-Class Model Solutions

graphic file with name 10.3928_24748307-20230419-01-table1.jpg

Model (K-Class) Smallest Class Size (% of Sample) BIC CAIC LMR-LRT LMR-LRT p Value
1-Class 100 9,080 9,088 - -
2-Class 34.2 8,549 8,562 545.6 <.0001
3-Class 18.7 8,501 8,519 77.02 .016
4-Class 12.6 8,494 8,517 37.54 .068
Open in a new tab

Note. Lower values are better for BIC and CAIC. Significant p values for likelihood ratio tests reflect improved fit of given model compared to a model with one fewer class. BIC = Bayesian information criteria; CAIC = consistent Akaike's information criteria; LMR-LRT = Lo-Mendell-Rubin likelihood ratio test.

Figure 1 depicts estimated HL means and standard errors for each class across the four waves, using the four-class solution. The smallest class (low HL; 12.6% of the sample) demonstrated the lowest HL at all waves and the largest class (high HL; 49.7%) demonstrated the highest HL at all waves. A moderate HL class (20%) fell between the low and high classes at all time points. Of note, 17.7% of the sample was classified into the improving HL class, which demonstrated relatively lower HL at baseline with a steady increase over time. All classes significantly changed from baseline to 12 months, with increased HL in the low HL and improving HL classes and small decreases in the moderate HL and high HL classes. The Improving HL class was the only class to show statistically significant changes across later waves.

Figure 1.

Figure 1.

Open in a new tab

Class means and standard errors of health literacy over time for the 4-Class solution.

Class Membership in Relation to Baseline Demographics

Table 2 shows parental demographics stratified by class membership. Classes did not differ in terms of study group but differed by parent age, education, and income (all omnibus p values < .001). Post-hoc pairwise comparisons clarify precisely how the classes differed (denoted in Table 2). Compared to all other classes, participants in the high HL class were older and had higher educational status. Those in the high HL class were more likely to be in the highest income category relative to parents in the low HL and moderate HL classes but did not differ in terms of income from the improving HL class. Those in the improving HL class reported higher educational attainment relative to those in the low HL and moderate HL classes.

Table 2.

Baseline Parental Demographic Characteristics of Latent Classes

graphic file with name 10.3928_24748307-20230419-01-table2.jpg

Parental Demographics Low (N =54) Moderate (N = 108) Improving (N = 88) High (N = 238) p Value
M (SD)
Age 22.1 (4.4)a 23.5 (5)a 23.8 (5.3)a 26.7 (5.4)b <.001
n (%)
Education
  High school degree or less 49 (90.7) 88 (81.5) 58 (65.9) 118 (49.6) <.001
  Some college 5 (9.3)a 20 (18.5)a 30 (34.1)b 120 (50.4)c
Income
  <$10K 30 (55.6) 54 (50) 48 (54.5) 118 (49.6) <.001
  $10–$20K 2 (3.7) 8 (7.4) 7 (8) 27 (11.3)
  ≥$20K 3 (5.6)a 7 (6.5)a 13 (14.8)a,b 54 (22.7)b
  Data missing 19 (35.2) 39 (36.1) 20 (22.7) 39 (16.4)
Study group
  Control 28 (51.9) 56 (51.9) 49 (55.7) 118 (49.6) .810
  Intervention 26 (48.1) 52 (48.1) 39 (44.3) 120 (50.4)
Open in a new tab

Note. Differences were tested with one-way analysis of variance for continuous variables and Fisher's exact tests for categorical variables due to small sample sizes in some cells. Pairwise contrasts were tested following a significant omnibus p value.

a,b,c

Nonsignificant pairwise differences between classes.

Class Membership in Relation to 36-Month Oral Health Variables

Table 3 shows class-stratified mean (SD) values for oral health variables and the results of regression-based significance testing. All p values reflect omnibus tests of class differences. If a significant omnibus p value was observed, pairwise comparisons were made between the four classes. After adjusting for covariates, class membership was significantly associated with oral health knowledge, perceived barriers, perceived severity, perceived susceptibility, self-efficacy, and behavioral adherence. Specifically, those in the high HL and improving HL classes reported greater knowledge, lower perceived barriers, higher perceived severity, lower perceived susceptibility, greater self-efficacy, and greater behavioral adherence relative to those in the other two classes. The low HL and moderate HL classes only significantly differed from one another on perceived barriers, with lower perceived barriers reported by participants in the moderate HL class relative to the low HL class.

Table 3.

Class Membership in Relation to the Oral Health Variables at the 36-Month Study Visit

graphic file with name 10.3928_24748307-20230419-01-table3.jpg

Oral Health Variables Low (N = 54) Moderate (N = 108) Improving (N = 88) High (N = 238) p Value
M (SD)
Perceived barriers 71.9 (14.8)a 75.8 (15.7)a 82.0 (11)b 85.2 (10.7)b <.001
Perceived benefits 2.7 (0.7)a 2.4 (0.8)b 2.0 (0.7)c 2.0 (0.7)c <.001
Perceived severity 4.3 (0.9) 4.3 (0.7) 4.5 (0.6) 4.5 (0.7) .099
Perceived susceptibility 3.8 (0.9)a 4.0 (0.9)a 4.5 (0.7)b 4.5 (0.7)b <.001
Self-efficacy 3.1 (1)a 3.2 (1)a 2.9 (0.9)b 2.6 (1)b <.001
Behavioral adherence 6.0 (4.6)a 6.9 (4.2)a 8.0 (4)b 8.7 (3.7)b <.001
Decayed, missing, and filled tooth surfaces 43.8 (17.5)a 48.9 (20)a 54.4 (19.9)b 54.4 (20)b .005
Open in a new tab

Note. The 36-month mean (SD) is presented for each oral health variable by class. P values are from linear or negative binomial regressions, as appropriate, and are adjusted for age, education, income, and treatment group. Pairwise contrasts were tested following a significant omnibus p value.

a,b,c

Nonsignificant pairwise differences between classes. To illustrate, the high and improving classes did not differ from one another on any outcomes and the low and moderate classes only differed on perceived barriers.

Discussion

In this secondary analysis, we focused on (1) whether and how parental HL changes over time; (2) personal characteristics associated with different patterns of change; and (3) how change in HL is associated with important oral health constructs. Using RMLCA to classify parents into subgroups reflecting change patterns, we identified four distinct classes. These were distinguishable in their relative HL values (low, moderate, high) and change over time. Although all classes experienced statistically significant changes in HL from baseline to 12 months, the clinical significance of some changes may be negligible, and most classes did not significantly change after 12 months. In contrast, parents in the improving HL class showed improved HL at each time point from baseline to 36 months. This level of increase is likely to be both clinically and practically important. Overall, there was heterogeneity in patterns of change in HL over time, with sustained, clinically meaningful change only observed in a subset of parents.

Class membership was associated with baseline demographics and parental oral health knowledge, beliefs, and behaviors at 36 months. Previous work has shown associations of greater educational attainment and higher income with stronger HL and has indicated that adults age 25 to 39 years have the highest average HL scores (Kutner et al., 2006). Research has also shown higher HL is associated with more optimal oral health knowledge, beliefs, and behaviors (Brega et al., 2020; Brega, Johnson, Jiang, et al., 2021; Brega, Johnson, Schmiege, Jiang, et al., 2021; Brega, Johnson, Schmiege, Wilson, et al., 2021; Brega et al., 2016; Vann et al., 2010; Vilella et al., 2016). The associations of class membership with baseline demographics and oral health variables observed in the current analyses are consistent with those prior conclusions. Notably, the High HL class was the oldest in this sample mainly comprised of young adults (mean age = 26.7), was the most educated, and reported the highest income. This class demonstrated more optimal performance on oral health knowledge, beliefs (except perceived susceptibility), and behavioral adherence variables.

As the first examination of change over time in parental HL, this study expands understanding of trajectories of change in HL relative to demographics and oral health variables. Notably, while the improving HL class was younger and had lower baseline educational achievement than the high HL class, it did not differ from the high HL class on any oral health constructs examined. However, the improving HL class demonstrated significantly lower baseline HL than the high and moderate HL classes. Had we considered only baseline HL, as is done in most HL research, increases in HL over time observed in the Improving HL class would have been obscured. Given large differences in the oral health constructs observed between the Improving HL and Low HL groups, assuming all parents with limited HL at the outset represent a unified class would have masked the true association of HL with oral health constructs for these participants. This underscores the value of examining change in HL over time and the trajectories of change on oral health outcomes.

It is important to consider why changes in HL were observed for some participants but not others. Although the PBC study was not designed to explain changes in HL over time, we conducted post-hoc analyses to explore changes in educational attainment over time in relation to class membership. These analyses (presented in Table A) provided initial evidence that improved education over time may be associated with improved HL. Future research should further examine this, including the potential for overlap between education and other demographic variables. In addition to demonstrating that HL can improve over time, our findings draw attention to the stability of HL limitations in some adults. Roughly one-third of the sample fell into the low or moderate HL classes, which maintained stable and relatively limited levels of HL over the 3-year project period. Although participants in these classes were similar in age and income at baseline to those in the improving HL class, they were less likely to have had any college education at baseline and less likely to report improved educational attainment over time. Together these results suggest that formal education may play a crucial role in the continued growth of HL in adulthood.

Table A.

Exploration of Change Over Time in Education Status by Class Membership

graphic file with name 10.3928_24748307-20230419-01-table4.jpg

Change in Educational Attainment* Low (N = 54) N (%) Moderate (N = 108) N (%) Improving (N = 88) N (%) High (N = 238) N (%) P Value**
Baseline to 12 months 0.089
  Improved 14 (25.9%) 23 (22.8%) 19 (22.4%) 33 (14.5%)
  Not Improved 40 (74.1%) 78 (77.2%) 66 (77.6%) 194 (85.5%)
Baseline to 24 months 0.015
  Improved 15 (28.3%) 29 (27.9%) 33 (38.8%) 46 (20.7%)
  Not Improved 38 (71.7%) 75 (72.1%) 52 (61.2%) 176 (79.3%)
Baseline to 36 months 0.003
  Improved 16 (29.6%) 38 (35.5%) 42 (47.7%) 61 (26.0%)
  Not Improved 38 (70.4%) 69 (64.5%) 46 (52.3%) 174 (74.0%)
Open in a new tab
*

Educational attainment was coded as improved versus not improved from baseline to each later wave based on a version of the variable that included thirteen possible options ranging from 6th grade to advanced/graduate degree. Improvement was operationalized as any improvement from baseline to each later wave (e.g., 6th to 7th grade; 11th grade to GED attainment).

**

Differences were tested using chi-square tests

These findings also highlight the important role that the health care system can play in supporting parents with limited HL. Although we often define and measure HL as a personal characteristic, researchers have long acknowledged that the adequacy of a person's HL is influenced by the demands placed on those skills (Rudd et al., 2012). Health care providers and systems can support patients of all HL levels by reducing those demands. Numerous studies show that health-related documents, including some oral health materials, are commonly written above the reading level of many adults in the U.S. (Amini et al., 2007; Blinkhorn & Verity, 1979). By ensuring that patient-facing materials are easy to read and follow other principles of clear communication (Baur & Prue, 2014; Doak et al., 1996; Shoemaker et al., 2014), health care systems can help patients to better understand and act upon the written information they receive. Providing information in formats other than the written word may also be beneficial, particularly for those with limited HL. For example, in an oral health intervention with pregnant individuals, participants with low oral health literacy gained more knowledge than participants with stronger HL skills when educational information was presented verbally, rather than in writing (Vilella et al., 2017). Other studies suggest that patients benefit when providers seek to confirm comprehension of health information delivered verbally (Bertakis, 1977; Fink et al., 2010; Schillinger et al., 2003). Through these and other methods (Brega et al., 2015), health care providers and systems can improve their communication with and support for parents with limited HL.

One study strength was using RMLCA over standard longitudinal analysis methods to capture individual differences in HL change patterns. This approach allowed us to assess whether and how HL changed and for whom it changed. The large changes observed in the improving HL class, relative to modest fluctuations in the other classes, likely would have been obscured using standard longitudinal methods. This approach may have allowed for stronger prediction of outcomes. Person-centered approaches, like RMLCA, are particularly useful for uncovering complex individual difference patterns in outcomes.

Study Limitations

The longitudinal assessments were not fully observational as they were collected as part of an RCT. However, the intervention was not designed to change HL and our analysis approach ruled out the potential effect of treatment group on conclusions. Class labels reflect mean levels relative to other classes and should not be interpreted in absolute terms. The study population was characterized by economic hardship and limited access to dental care, influencing the generalizability of the observed HL trajectories to the general population and limiting our ability to fully understand why some participants changed whereas others did not. Future research should focus on understanding naturally occurring trajectories of HL in a sample diverse in characteristics such as age, education, and income.

Other limitations relate to the measurement of HL. There was a ceiling effect regarding the HL measure; the stable high scores within the high HL class should therefore be interpreted within the confines of our measurement tool. In addition, the HL measure focused on parents' ability to read and write in the context of health, only one of the important skills sets believed to provide the foundation of HL. It is possible that our results would have been different if we had also assessed other components of HL, such as numeracy or verbal communication skills. Likewise, our self-report measure of HL was, by its nature, subjective. As a result, the measure captured the risk of HL limitations as opposed to a precise, performance-based indicator of HL skills. Although the items from which our measure was adapted have repeatedly been shown to be well correlated with performance-based measures (Bishop et al., 2016; Chew et al., 2004; Chew et al., 2008; Sarkar et al., 2011; Wallace et al., 2007; Wallace et al., 2006) and our measure has been validated in American Indian and Alaska Native populations (Brega et al., 2012), it is possible that our results would have been different if a performance-based measure was used. Studies on numeracy, for instance, suggest that people overestimate their numerical skills, self- reporting greater skills than test-based measures would indicate (Woloshin et al., 2005). It is possible that our subjective measure might similarly overestimate actual reading and writing skills. However, as we used the same measure across different time points, the magnitude of overestimates should be similar across time and the patterns of the longitudinal trajectories should be similar even without the overestimation. In addition, a subjective measure of HL may be associated with outcomes in a different way than performance-based measures. For instance, it is possible that prior care received as well as positive or negative experiences with the health care system may impact a parent's perceived ability to understand and use health information, whereas objective measures may be less affected by contact with the health care system.

To summarize, person-level variability in HL was measured over time using longitudinal latent class methods. Changes over time in HL were not consistent across individuals. Although some individuals improved in HL, perhaps due to increases in education, most participants remained relatively stable, as measured by self-reported risk of limited HL. Longitudinal trajectory patterns were associated with baseline demographics, where the class with consistently high HL was older, had more education, and had higher income relative to other classes. The improving HL class most closely mimicked the high HL class on all oral health variables, demonstrating the value of longitudinally measured HL in predicting health outcomes.

Acknowledgments

The authors thank the families who participated in the randomized controlled trial that was the source of the data used in this analysis.

Funding Statement

Grant: This research was supported by the National Institute of Dental and Craniofacial Research (NIDCR) of the National Institutes of Health (NIH) grant (R01DE027077). The data used in this secondary analysis were collected as part of a randomized controlled trial funded by the NIDCR (U54DE019259). Development of the Basic Research Factors Questionnaire was also supported by the NIDCR (U54DE019285, U54DE019275, and U54DE019259).

References

  1. Albino , J. , Tiwari , T. , Gansky , S. A. , Henshaw , M. M. , Barker , J. C. , Brega , A. G. , Gregorich , S. E. , Heaton , B. , Batliner , T. S. , Borrelli , B. , Geltman , P. , Kressin , N. R. , Weintraub , J. A. , Finlayson , T. L. , Garcia , R. I. , & the Early Childhood Caries Collaborating Centers . ( 2017. ). The Basic Research Factors Questionnaire for studying early childhood caries . BMC Oral Health , 17 ( 83 ), 1 – 12 . 10.1186/s12903-017-0374-5 [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Amini , H. , Casamassimo , P. S. , Lin , H. L. , & Hayes , J. R. ( 2007. ). Readability of the American Academy of Pediatric Dentistry patient education materials . Pediatric Dentistry , 29 ( 5 ), 431 – 435 . PMID: [PubMed] [Google Scholar]
  3. Asparouhov , T. , & Muthén , B . ( 2014. ). Auxiliary variables in mixture modeling: Using the BCH method in Mplus to estimate a distal outcome model and an arbitrary secondary model . http://statmodel.com/examples/webnotes/webnote21.pdf [Google Scholar]
  4. Baker , D. W. , Wolf , M. S. , Feinglass , J. , & Thompson , J. A. ( 2008. ). Health literacy, cognitive abilities, and mortality among elderly persons . Journal of General Internal Medicine , 23 ( 6 ), 723 – 726 . 10.1007/s11606-008-0566-4 [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Batliner , T. , Fehringer , K. A. , Tiwari , T. , Henderson , W. G. , Wilson , A. , Brega , A. G. , & Albino , J. ( 2014. ). Motivational interviewing with American Indian mothers to prevent early childhood caries: Study design and methodology of a randomized control trial. [Electronic Resource] . Trials , 15 ( 1 ), 125 . 10.1186/1745-6215-15-125 [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Batliner , T. S. , Tiwari , T. , Henderson , W. G. , Wilson , A. R. , Gregorich , S. E. , Fehringer , K. A. , Brega , A. G. , Swyers , E. , Zacher , T. , Harper , M. M. , Plunkett , K. , Santo , W. , Cheng , N. F. , Shain , S. , Rasmussen , M. , Manson , S. M. , & Albino , J. ( 2018. ). Randomized trial of motivational interviewing to prevent early childhood caries in American Indian children . JDR Clinical and Translational Research , 3 ( 4 ), 366 – 375 . 10.1177/2380084418787785 [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Baur , C. , & Prue , C. ( 2014. ). The CDC Clear Communication Index is a new evidence-based tool to prepare and review health information . Health Promotion Practice , 15 ( 5 ), 629 – 637 . 10.1177/1524839914538969 [DOI] [PubMed] [Google Scholar]
  8. Bertakis , K. D. ( 1977. ). The communication of information from physician to patient: A method for increasing patient retention and satisfaction . The Journal of Family Practice , 5 ( 2 ), 217 – 222 . [PubMed] [Google Scholar]
  9. Bishop , W. P. , Craddock Lee , S. J. , Skinner , C. S. , Jones , T. M. , McCallister , K. , & Tiro , J. A. ( 2016. ). Validity of single-item screening for limited health literacy in English and Spanish speakers . American Journal of Public Health , 106 ( 5 ), 889 – 892 . 10.2105/AJPH.2016.303092 [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Blinkhorn , A. S. , & Verity , J. M. ( 1979. ). Assessment of the readability of dental health education literature . Community Dentistry and Oral Epidemiology , 7 ( 4 ), 195 – 198 . 10.1111/j.1600-0528.1979.tb01215.x [DOI] [PubMed] [Google Scholar]
  11. Borrelli , B. , Tooley , E. M. , & Scott-Sheldon , L. A. ( 2015. ). Motivational interviewing for parent-child health interventions: A systematic review and meta-analysis . Pediatric Dentistry , 37 ( 3 ), 254 – 265 . PMID: [PubMed] [Google Scholar]
  12. Brega, A. G., Jiang, L., Beals, J., Manson, S. M., Acton, K. J., Roubideaux, Y., & Special Diabetes Program for Indians Healthy Heart Demonstration Project . (2012). Special diabetes program for Indians: Reliability and validity of brief measures of print literacy and numeracy. Ethnicity and Disease, 22(2), 207–214. https://pubmed.ncbi.nlm.nih.gov/22764644/ PMID: [PubMed] [Google Scholar]
  13. Brega , A. G. , Barnard , J. , Mabachi , N. M. , Weiss , B. D. , DeWalt , D. A. , Brach , C. , Cifuentes , M. , Albright , K. , West , D. R. ( 2015. ). AHRQ health literacy universal precautions toolkit . ( 2nd ed. ). https://www.ahrq.gov/sites/default/files/publications/files/healthlittoolkit2_3.pdf [DOI] [PMC free article] [PubMed]
  14. Brega , A. G. , Thomas , J. F. , Henderson , W. G. , Batliner , T. S. , Quissell , D. O. , Braun , P. A. , Wilson , A. , Bryant , L. L. , Nadeau , K. J. , & Albino , J. ( 2016. ). Association of parental health literacy with oral health of Navajo Nation preschoolers . Health Education Research , 31 ( 1 ), 70 – 81 . 10.1093/her/cyv055 [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Brega , A. G. , Jiang , L. , Johnson , R. L. , Wilson , A. R. , Schmiege , S. J. , & Albino , J. ( 2020. ). Health literacy and parental oral health knowledge, beliefs, behavior, and status among parents of American Indian newborns . Journal of Racial and Ethnic Health Disparities , 7 ( 4 ), 598 – 608 . 10.1007/s40615-019-00688-4 [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Brega , A. G. , Johnson , R. L. , Jiang , L. , Wilson , A. R. , Schmiege , S. J. , & Albino , J. ( 2021. ). Influence of parental health literacy on change over time in the oral health of American Indian children . International Journal of Environmental Research and Public Health , 18 ( 11 ), 5633 . 10.3390/ijerph18115633 [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Brega , A. G. , Johnson , R. L. , Schmiege , S. J. , Jiang , L. , Wilson , A. R. , & Albino , J. ( 2021. ). Longitudinal association of health literacy with parental oral health behavior . HLRP: Health Literacy Research and Practice , 5 ( 4 ), e333 – e341 . 10.3928/24748307-20211105-01 [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Brega , A. G. , Johnson , R. L. , Schmiege , S. J. , Wilson , A. R. , Jiang , L. , & Albino , J. ( 2021. ). Pathways through which health literacy is linked to parental oral health behavior in an American Indian tribe . Annals of Behavioral Medicine , 55 ( 11 ), 1144 – 1155 . 10.1093/abm/kaab006 [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Chew , L. D. , Bradley , K. A. , & Boyko , E. J. ( 2004. ). Brief questions to identify patients with inadequate health literacy . Family Medicine , 36 ( 8 ), 588 – 594 . https://knilt.arcc.albany.edu/images/3/30/Liechty_2011.pdf [PubMed] [Google Scholar]
  20. Chew , L. D. , Griffin , J. M. , Partin , M. R. , Noorbaloochi , S. , Grill , J. P. , Snyder , A. , Bradley , K. A. , Nugent , S. M. , Baines , A. D. , & Vanryn , M. ( 2008. ). Validation of screening questions for limited health literacy in a large VA outpatient population . Journal of General Internal Medicine , 23 ( 5 ), 561 – 566 . 10.1007/s11606-008-0520-5 [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Clark , S. , & Muthén , B . ( 2009. ). Relating latent class analysis results to variables not included in the analysis . http://www.statmodel.com/download/relatinglca.pdf [Google Scholar]
  22. Curtis , L. M. , Kwasny , M. J. , Opsasnick , L. , O'Conor , R. M. , Yoshino-Benavente , J. , Eifler , M. , Federman , A. D. , Altschul , D. , & Wolf , M. S. ( 2021. ). Change in health literacy over a decade in a prospective cohort of community-dwelling older adults . Journal of General Internal Medicine , 36 ( 4 ), 916 – 922 . 10.1007/s11606-020-06423-8 [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Doak , C. C. , Doak , L. G. , & Root , J. H. ( 1996. ). Teaching patients with low literacy skills . ( 2nd ed. ). J. B. Lippincott Company; . 10.1097/00000446-199612000-00022 [DOI] [Google Scholar]
  24. Edwards , M. , Wood , F. , Davies , M. , & Edwards , A. ( 2012. ). The development of health literacy in patients with a long-term health condition: The health literacy pathway model . BMC Public Health , 12 , 130 . 10.1186/1471-2458-12-130 [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Fink , A. S. , Prochazka , A. V. , Henderson , W. G. , Bartenfeld , D. , Nyirenda , C. , Webb , A. , Berger , D. H. , Itani , K. , Whitehill , T. , Edwards , J. , Wilson , M. , Karsonovich , C. , & Parmelee , P. ( 2010. ). Enhancement of surgical informed consent by addition of repeat back: A multicenter, randomized controlled clinical trial . Annals of Surgery , 252 ( 1 ), 27 – 36 . 10.1097/SLA.0b013e3181e3ec61 [DOI] [PubMed] [Google Scholar]
  26. Indian Health Service . ( 2013. ). The 2010 Indian Health Service Oral Health Survey of American Indian and Alaska Native Preschool Children . United States Department of Health and Human Services; . http://www.ihs.gov/DOH/documents/IHS%20Oral%20Health%20Report%2004-17-2014.pdf [Google Scholar]
  27. Institute of Medicine Committee on Health Literacy . ( 2004. ). Health literacy: a prescription to end confusion . The National Academies Press; . https://www.nap.edu/catalog/10883/health-literacy-a-prescription-to-end-confusion [PubMed] [Google Scholar]
  28. Janz , N. K. , & Becker , M. H. ( 1984. ). The Health Belief Model: A decade later . Health Education Quarterly , 11 ( 1 ), 1 – 47 . 10.1177/109019818401100101 [DOI] [PubMed] [Google Scholar]
  29. Kutner , M. , Greenberg , E. , Jin , Y. , & Paulsen , C . ( 2006. ). The Health Literacy of America's Adults: Results from the 2003 National Assessment of Adult Literacy (NCES 2006–483) . U.S. Department of Education; ., http://nces.ed.gov/pubs2006/2006483.pdf [Google Scholar]
  30. Masyn , K. E. ( 2013. ). Latent class analysis and finite mixture modeling . In Little T. D. (Ed.), Statistical Analysis (pp. 551 – 611 ). The Oxford Handbook of Quantitative Methods . Oxford University Press; . [Google Scholar]
  31. Miller , W. R. , & Rose , G. S. ( 2009. ). Toward a theory of motivational interviewing . The American Psychologist , 64 ( 6 ), 527 – 537 . 10.1037/a0016830 [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Morris , N. S. , Maclean , C. D. , & Littenberg , B. ( 2013. ). Change in health literacy over 2 years in older adults with diabetes . The Diabetes Educator , 39 ( 5 ), 638 – 646 . 10.1177/0145721713496871 [DOI] [PubMed] [Google Scholar]
  33. Muthén , L. K. , & Muthén , B. O. ( 1998–2017. ). Mplus user's guide . ( 8th ed. ). https://www.statmodel.com/download/usersguide/MplusUser-GuideVer_8.pdf
  34. Nylund , K. L. , Asparouhov , T. , & Muthén , B. ( 2007. ). Deciding on the number of classes in latent class analysis and growth mixture modeling. A Monte Carlo simulation study . Structural Equation Modeling , 14 , 535 – 569 . 10.1080/10705510701575396 [DOI] [Google Scholar]
  35. Phipps , K. R. , & Ricks , T. L. ( 2015. ). Indian Health Service Data Brief: The Oral Health of American Indian and Alaska Native Children Aged 1–5. Years: Results of the 2014 IHS Oral Health Survey . United States Department of Health and Human Services, Indian Health Service; . https://www.ihs.gov/doh/documents/IHS_Data_Brief_1-5_Year-Old.pdf [Google Scholar]
  36. Phipps , K. R. , Ricks , T. L. , Manz , M. C. , & Blahut , P. ( 2012. ). Prevalence and severity of dental caries among American Indian and Alaska Native preschool children . Journal of Public Health Dentistry , 72 ( 3 ), 208 – 215 . 10.1111/j.1752-7325.2012.00331.x [DOI] [PubMed] [Google Scholar]
  37. Phipps , K. R. , Ricks , T. L. , Mork , N. P. , & Lozon , T. L. ( 2019. ). Indian Health Service Data Brief: The Oral Health of American Indian and Alaska Native Children Aged 1–5. Years: Results of the 2018–19. IHS Oral Health Survey . United States Department of Health and Human Services, Indian Health Service; . https://www.ihs.gov/doh/documents/surveillance/2018-19%20Data%20Brief%20of%201-5%20Year-Old%20AI-AN%20Preschool%20Children.pdf [Google Scholar]
  38. Pitts , N. B. ( 2001. ). Clinical diagnosis of dental caries: A European perspective . Journal of Dental Education , 65 ( 10 ), 972 – 978 . 10.1002/j.0022-0337.2001.65.10.tb03472.x [DOI] [PubMed] [Google Scholar]
  39. Pitts , N. B. ( 2004. ). Modern concepts of caries measurement . Journal of Dental Research , 83 ( Suppl. 1 ), 43 – 47 . 10.1177/154405910408301s09 [DOI] [PubMed] [Google Scholar]
  40. R Core Team . ( 2022. ). R: A language and environment for statistical computing . https://www.R-project.org
  41. Ratzan , S. C. , & Parker , R. M. ( 2000. ). Introduction . In Selden C. R. , Zorn M. , Ratzan S. C. , & Parker R. M. (Eds.), National Library of Medicine Current Bibliographies in Medicine: Health Literacy. NLM Pub. No. CBM 2000-1 . National Institutes of Health, United States Department of Health and Human Services; . [Google Scholar]
  42. Rosenstock , I. M. ( 1974. ). The Health Belief Model and preventive health behavior . Health Education Monographs , 2 ( 4 ), 354 – 386 . https://journals.sagepub.com/doi/pdf/10.1177/109019817400200405 10.1177/109019817400200405 [DOI] [PubMed] [Google Scholar]
  43. Rosenstock , I. M. , Strecher , V. J. , & Becker , M. H. ( 1988. ). Social learning theory and the Health Belief Model . Health Education & Behavior , 15 , 175 – 183 . [DOI] [PubMed] [Google Scholar]
  44. Rudd , R. E. , McCray , A. T. , & Nutbeam , D . ( 2012. ). Health Literacy and Definition of Terms . In Begoray D. L. , Gillis D. , & Rowlands G. (Eds.), Health Literacy in Context: International Perspectives (pp. 13 – 32 ). Nova Science Publishers, Inc; . [Google Scholar]
  45. Sarkar , U. , Schillinger , D. , Lopez , A. , & Sudore , R. ( 2011. ). Validation of self-reported health literacy questions among diverse English and Spanish-speaking populations . Journal of General Internal Medicine , 26 ( 3 ), 265 – 271 . 10.1007/s11606-010-1552-1 [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. Schillinger , D. , Piette , J. , Grumbach , K. , Wang , F. , Wilson , C. , Daher , C. , Leong-Grotz , K. , Castro , C. , & Bindman , A. B. ( 2003. ). Closing the loop. Physician communication with diabetic patients who have low health literacy . Archives of Internal Medicine , 163 ( 1 ), 83 – 90 . 10.1001/archinte.163.1.83 [DOI] [PubMed] [Google Scholar]
  47. Shoemaker , S. J. , Wolf , M. S. , & Brach , C . ( 2014. ). The patient education materials assessment tool (PMET) and user's guide . Agency for Healthcare Research and Quality; . https://www.ahrq.gov/health-literacy/patient-education/pemat.html [Google Scholar]
  48. Vann , W. F. , Jr , Lee , J. Y. , Baker , D. , & Divaris , K. ( 2010. ). Oral health literacy among female caregivers: Impact on oral health outcomes in early childhood . Journal of Dental Research , 89 ( 12 ), 1395 – 1400 . 10.1177/0022034510379601 [DOI] [PMC free article] [PubMed] [Google Scholar]
  49. Vilella , K. D. , Alves , S. G. , de Souza , J. F. , Fraiz , F. C. , & Assuncao , L. R. ( 2016. ). The association of oral health literacy and oral health knowledge with social determinants in pregnant Brazilian women . Journal of Community Health , 41 ( 5 ), 1027 – 1032 . 10.1007/s10900-016-0186-6 [DOI] [PubMed] [Google Scholar]
  50. Vilella , K. D. , Fraiz , F. C. , Benelli , E. M. , & Assuncao , L. R. ( 2017. ). Oral health literacy and retention of health information among pregnant women: A randomised controlled trial . Oral Health & Preventive Dentistry , 15 ( 1 ), 41 – 48 . 10.3290/j.ohpd.a37712 [DOI] [PubMed] [Google Scholar]
  51. Wallace , L. S. , Cassada , D. C. , Rogers , E. S. , Freeman , M. B. , Grandas , O. H. , Stevens , S. L. , & Goldman , M. H. ( 2007. ). Can screening items identify surgery patients at risk of limited health literacy? The Journal of Surgical Research , 140 ( 2 ), 208 – 213 . 10.1016/j.jss.2007.01.029 [DOI] [PubMed] [Google Scholar]
  52. Wallace , L. S. , Rogers , E. S. , Roskos , S. E. , Holiday , D. B. , & Weiss , B. D. ( 2006. ). Brief report: Screening items to identify patients with limited health literacy skills . Journal of General Internal Medicine , 21 ( 8 ), 874 – 877 . 10.1111/j.1525-1497.2006.00532.x [DOI] [PMC free article] [PubMed] [Google Scholar]
  53. Warren , J. J. , Weber-Gasparoni , K. , Tinanoff , N. , Batliner , T. S. , Jue , B. , Santo , W. , Garcia , R. I. , Gansky , S. A. , & the Early Childhood Caries Collaborating Centers . ( 2015. ). Examination criteria and calibration procedures for prevention trials of the Early Childhood Caries Collaborating Centers . Journal of Public Health Dentistry , 75 ( 4 ), 317 – 326 . 10.1111/jphd.12102 [DOI] [PMC free article] [PubMed] [Google Scholar]
  54. Weiss , B. D. , Francis , L. , Senf , J. H. , Heist , K. , & Hargraves , R. ( 2006. ). Literacy education as treatment for depression in patients with limited literacy and depression: A randomized controlled trial . Journal of General Internal Medicine , 21 ( 8 ), 823 – 828 . 10.1111/j.1525-1497.2006.00531.x [DOI] [PMC free article] [PubMed] [Google Scholar]
  55. Wilson , A. R. , Brega , A. G. , Campagna , E. , Braun , P. A. , Henderson , W. G. , Bryant , L. L. , Batliner , T. S. , Quissell , D. O. , Albino , J. ( 2016. ). Validation and impact of caregivers' oral health knowledge and behavior on children's oral health status . Pediatric Dentistry , 38 , 47 – 54 . PMID: [PMC free article] [PubMed] [Google Scholar]
  56. Wilson , A. R. , Brega , A. G. , Thomas , J. F. , Henderson , W. G. , Lind , K. E. , Braun , P. A. , Batliner , T. S. , & Albino , J. ( 2018. ). Validity of measures assessing oral health beliefs of American Indian parents . Journal of Racial and Ethnic Health Disparities , 5 ( 6 ), 1254 – 1263 . 10.1007/s40615-018-0472-3 [DOI] [PMC free article] [PubMed] [Google Scholar]
  57. Wolf , M. S. , Curtis , L. M. , Wilson , E. A. , Revelle , W. , Waite , K. R. , Smith , S. G. , Weintraub , S. , Borosh , B. , Rapp , D. N. , Park , D. C. , Deary , I. C. , & Baker , D. W. ( 2012. ). Literacy, cognitive function, and health: Results of the LitCog study . Journal of General Internal Medicine , 27 ( 10 ), 1300 – 1307 . 10.1007/s11606-012-2079-4 [DOI] [PMC free article] [PubMed] [Google Scholar]
  58. Woloshin , S. , Schwartz , L. M. , & Welch , H. G. ( 2005. ). Patients and medical statistics. Interest, confidence, and ability . Journal of General Internal Medicine , 20 ( 11 ), 996 – 1000 . 10.1007/s11606-005-0245-7 [DOI] [PMC free article] [PubMed] [Google Scholar]

Articles from HLRP: Health Literacy Research and Practice are provided here courtesy of SLACK, Incorporated

RESOURCES