Rosemary Frasso's presentation from the
Penn Urban Doctoral Symposium
May 13, 2011
Co-sponsored with Penn’s Urban Studies program, this symposium celebrates the work of graduating urban-focused doctoral candidates. Graduates present and discuss their dissertation findings. Luncheon attended by the students, their families and their committees follows.
USPS® Forced Meter Migration - How to Know if Your Postage Meter Will Soon be...
Exploring the Association between Maternal Health Literacy and Pediatric Healthcare Utilization
1. Exploring the Association between Maternal Health Literacy and Pediatric Healthcare Utilization:Is Low Health Literacy a Barrier of Concern? Rosemary Frasso Dissertation Committee Chair ~ Phyllis Solomon, PhD Steve Marcus PhD Ian Bennett, MD, PhD Agency for Healthcare Research and Quality Dissertation Grant 1 R36 HS017471-01
2. Agenda Background and Significance Methods Results Discussion Limitations and Lessons Learned Next Steps 2
3. Background Health Literacy (HL) “the degree to which individuals have the capacity to obtain, process, and understand basic health information and services needed to make appropriate health decisions” DHHS, 2000 3
4. Background We know Women with low health literacy have poor health outcomes and underutilize preventive care 4
5. Background We know Appropriate use of pediatric preventive care is associated with significant reductions in morbidity and mortality and has been shown to reduce healthcare costs and decrease hospital admissions 5
6. Significance We don’t know Are children of mothers with low health literacy at a disadvantage similar to that of their mothers? Conflicting evidence about the impact of maternal HL on pediatric social and health outcomes Pati et al (2011) -TANF/Vaccination compliance 6
7. Methods Mix methods Quantitative (secondary data analysis) Qualitative 14 semi-structured interviews 11 different mothers with varied HL 1 critical case exploration MOTHERS FROM THE PARENT STUDY (REALM) CHILDREN FROM THE MEDICAID CLAIM DATA 185 DYADS 7
8. Parent Study Community based prospective cohort study of mothers and infants in Philadelphia Investigating the contextual, social, behavioral, and family context of maternal child health (extensive surveys) Followed from prenatal period to 24 months post partum >5000 participants 1034 had health literacy assessments REALM / STOFHLA Funded by the CDC and National Institute of Child Health and Development 8 CDC (TS 312 15/15; Culhane) and NICHD (1R01 D36462 01A; Elo and Culhane)
9. Operationalized Independent Variables Predisposing Enabling Need 9 Demographic characteristics, such as race, age, and maternal education have been shown to impact parent driven pediatric health service use Here Andersen grouped personal and family factors including social supports, income, insurance & physical access to providers Need, the strongest predictor of health service use based on how people view their own functional capacity, symptoms, & general state of health (and that of the children they care for)
10. Dependent Variables Operationalized 10 The primary outcome measure of preventive care utilization is the overall number of documented well-child visits in the first two years of life The AAP recommends 7 WCV in year 1 of life and 3 in year 2 of life ED. SCV, % Compliance CPT and ICD-9 Codes were used to identify these visits in the Medicaid claims files Well- Child Visits Year 1 /Year 2 (WCV) Sick - Child Visits Year 1 /Year 2 (SCV) ED Visits Year 1 /Year 2 (EDV) % Compliant Year 1 /Year 2 (WCV) AAP, 2011
13. Quantitative Analysis Revealed HL was not associated with the number of well-child visits, sick-child visits, ED visits or % compliance with a minimum number of visits in year 1 and year 2 of life HL did not prove to mediate or moderate the relationships between any of predisposing and enabling factors under study and our outcomes of interest Higher health literacy was not protective in this population 13
16. Qualitative Analysis Revealed Women with low HL and women higher HL encountered an overlapping set of challenges when navigating the healthcare system Several themes emerged and were used to elaborate on Andersen’s Model and shed light on the quantitative findings and a critical case emerged 16
20. Time in community /strength of ties Social support Literacy Strategies for working around low literacy Access to sources of health information Internet Access to information Communication Having an advocate Continuity of care Prior satisfaction with a healthcare provider Disabling Factors Power imbalance Lack of an advocate Limited or no access to health information Administrative/logistic hassles † Work gets in the way (unemployment) 17
22. Special Thanks To Phyllis Solomon, PhD Steven Marcus, PhD Ian Bennett, MD, PhD Leny Mathew, MS Jennifer Culhane, PhD, MPH All the members of Dr. Culhane’s Paper Group Sara Cullen, MSW And of course my terrific kids for all their love, patience and support 19
23. Some References (others available upon request) Agency for Healthcare Research and Quality. (2011, March 28). Low Health Literacy Linked to Higher Risk of Death and More Emergency Room Visits and Hospitalizations. Retrieved from http://www.ahrq.gov/news/press/pr2011/lowhlitpr.htm Andersen, R. M. (1995). Revisiting the behavioral model and access to medical care: Does it matter? J Health Soc Behav, 36(1), 1-10. Baron, R. M., & Kenny, D. A. (1986). The moderator-mediator variable distinction in social psychological research: Conceptual Berkman, N. D., Dewalt, D. A., Pignone, M. P., Sheridan, S. L., Lohr, K. N., Lux, L., Bonito, A. J. (2004). Literacy and health outcomes. Evid Rep Technol Assess (Summ), (87), 1-8. American Academy of Pediatrics, Committee on Practice and Ambulatory Medicine. (2000). Recommedations for preventive pediatric health care. Hakim, R. B., & Bye, B. V. (2001). Effectiveness of compliance with pediatric preventive care guidelines among Medicaid beneficiaries. Pediatrics, 108(1), 90-97. Shulman, S. (2006). Poor preventive care achievement and program retention among low birth weight infant Medicaid enrollees. Pediatrics, 118(5), e1509-1515. doi: peds.2004-0489 [pii] 10.1542/peds.2004-0489 20
26. Measuring Health Literacy Rapid Assessment of Adult Literacy in Medicine REALM 66 items Word familiarity Approximately three minutes Short Test of Functional Health Literacy in Adults STOFHLA 36 items Functional health literacy Approximately 7 minutes 23 (Baker, Williams, Parker, Gazmararian, & Nurss, 1999; Davis, Bocchini, et al., 1996; Davis, et al., 1993; Davis, et al., 1994; Moon, et al., 1998)
28. Missed Opportunities “Yes, but I, but I’m like, OK, it prevents cancer, is there any side effects from it, he’s like no, no, it’s in the pamphlet and I’m like, I see that, I understand that, I understand it was on paper, but it’s different when you hear it from someone. And I just wish he would’ve had more of a conversation about it, ‘cause it was like, no, I, everybody’s getting it, and I’m like OK?” Note: This participant declined the HPV vaccine for her daughter, who she generally relies on to translate written materials. 25
29. Benefit of Mix Methods New Concepts Lay informants Pseudo experts Experts Confirmation Parity Employment Unexpected findings (dyslexia example) “I don’t take advice from family or friends as much as I would a doctor” “I would call the hospital…... I got reprimanded for calling” “My cousin is in nursing school” “Friends, but their kids are younger so they don’t know” 26
34. Dependent Variables Operationalized We also documented the percent of dyads that were compliant with a minimum number of WCV per year There is a precedent in the literature for setting this bench-mark at 4 for year 1 and 2 for year 2 of life, slightly lower than the AAP recommendations CPT and ICD-9 Codes were used to identify these visits in the Medicaid claims files Compliant Year 1 (WCV) Compliant Year 2 (WCV) Shulmen, 2006 31
35. Hypotheses Mothers with low health literacy (< 6th grade) will be less likely than mothers with marginal to higher health literacy (> 7th grade) to meet pediatric preventive care recommendations. 32
36. Hypotheses Children of mothers with low health literacy (< 6th grade) will be more likely than children of mothers with marginal to high health literacy (> 7th grade) to visit an emergency room. be seen by a provider for a sick-child visit. 33
37. Hypotheses Low maternal health literacy will mediate the relationship between negative predisposing and enabling factors and timely receipt of pediatric preventive care. pediatric emergency room visits. the number of sick-child visits. 34
38. Hypotheses Low maternal health literacy will moderate the relationship between negative predisposing and enabling factors and timely receipt of pediatric preventive care. pediatric emergency room visits. the number of sick-child visits. 35
39. Quantitative Analysis Categorical Variable Chi-squared test of independence Continuous Variables Wilcoxon rank-sum test (Mann-Whitney-Wilcoxon) or Kruskal-Wallis non-parametric test Linear Regression Moderation Analysis Mediation Analysis Likelihood ratio test STATA Data Analysis Statistical Software 36
40. Qualitative Methods In depth semi structured interviews Issues of interest Need factors / perceived need Health beliefs Social support/relationships Ability to navigate the healthcare system Transcribed verbatim Coded using NVIVO8 (QSR) software guided by Andersen’s model 37
41. Guiding Qualitative Hypotheses Qualitative interviews will show that mothers with low health literacy will report different issues related to access to preventive pediatric care than mothers with marginal to high health literacy. Additionally, they will perceive different barriers to care and will suggest different interventions to reduce these barriers. 38
56. Pathway 1 Percent of Participants in Education Group by Health Literacy Level REALM Percent of Participants in Each Age Group by Health Literacy Level Maternal Education REALM Maternal Age 42
107. Pathway 4Moderation 55 Parity Well-Child Care Sick-Child CareED Visits % Compliance (Minimum # of visits/year) REALM (categorical) Parity X REALM Planned Analysis Barron & Kenny, 1986
108. Quantitative Analysis Revealed HL was not associated with the number of well-child visits, sick-child visits, ED visits or % compliance with a minimum number of visits in year 1 and year 2 of life When we controlled for health literacy we saw no impact on establish associations between a set of independent variables and our outcome variables 56
109. Associations of Interest 66% of the women in our sample had completed high school or GED however only 50% had a REALM score > 9th grade 100% of the women in the highest HL group were born in the US while that was the case for only 80% of the women in the low HL group 57
110. Qualitative Analysis Revealed Women with low HL and women higher HL encountered an overlapping set of challenges when navigating the healthcare system Confirmed the quantitative findings (for the most part kids are getting the minimum number of visits) Several themes emerged and were used to elaborate on Andersen’s Model and shed light on the quantitative findings Critical case emerged 58
111. Need Factors Views & evaluation of the child’s functional capacity, symptoms, & general state of health Informed by health beliefs, values about health and illness & attitudes towards about health services and knowledge about health Personal /Family Resources Income People @ home Financial Support Employment Insurance Community Resources Demographics Race/Ethnicity Nativity Age Education Literacy Language Housing Social Structure Marital Status Parity 59 Discussion
115. Time in community /strength of ties .Social support Literacy Strategies for working around low literacy Access to sources of health information Internet Access to information Communication Having an advocate Continuity of care Prior satisfaction with a healthcare provider Disabling Factors Power imbalance Lack of an advocate Limited or no access to health information Administrative/logistic hassles † Work gets in the way (unemployment) 60
116. Benefit of Mix Methods New Concepts Lay informants Pseudo experts Experts Confirmation Parity Employment Unexpected findings (dyslexia example) “I don’t take advice from family or friends as much as I would a doctor” “I would call the hospital…... I got reprimanded for calling” “My cousin is in nursing school” “Friends, but their kids are younger so they don’t know” 61
120. Time in community /strength of ties .Social support Literacy Strategies for working around low literacy Access to sources of health information Internet Access to information Communication Having an advocate Continuity of care Prior satisfaction with a healthcare provider Disabling Factors Power imbalance Lack of an advocate Limited or no access to health information Administrative/logistic hassles † Work gets in the way (unemployment) 62
125. Time in community /strength of ties Social support Literacy Strategies for working around low literacy Access to sources of health information Internet Access to information Communication Having an advocate Continuity of care Prior satisfaction with a healthcare provider Disabling Factors Power imbalance Lack of an advocate Limited or no access to health information Administrative/logistic hassles † Work gets in the way (unemployment) 64
126. Missed Opportunities “Yes, but I, but I’m like, OK, it prevents cancer, is there any side effects from it, he’s like no, no, it’s in the pamphlet and I’m like, I see that, I understand that, I understand it was on paper, but it’s different when you hear it from someone. And I just wish he would’ve had more of a conversation about it, ‘cause it was like, no, I, everybody’s getting it, and I’m like OK?” Note: This participant declined the HPV vaccine for her daughter, who she generally relies on to translate written materials. 65
127. Health Literacy 50% of the group was compliant in year 1 and this did not vary by health literacy 44% of the group was compliant in year 2 and again no variation by health literacy Does health literacy matter in this population? 66
128. Limitations & Lessons Learned Sample size / power Inclusion criteria 10 month of eligibility Locating the poorest readers REALM Was it reliable in the population? Does it need to be validated in the context of LD? Did not take full advantage of the available data* Limited generalizability and transferability 67
129. Next Steps Augment the current analysis and submit a paper for publication Abstract has been sent to APHA Critical Case was presented at Health Literacy Annual Research Conference Need to further explore how a learning disability impacts health literacy 68
130. Next Steps Continued Now that the dyads have been established we plan to revisit the survey data in order to explore additional research questions (many of which were brought to light in the qualitative arm of the study) For example: Feeling respected by a provider (qualitative) Mastery Scale and Coping Questions (parent study) “Sometimes I feel that I am being pushed around in life” “There is little I can do to change many of the important things in life” For example: Depression/mental illness (qualitative) Depression (parent study) 69
131. Some References (others available upon request) Agency for Healthcare Research and Quality. (2011, March 28). Low Health Literacy Linked to Higher Risk of Death and More Emergency Room Visits and Hospitalizations. Retrieved from http://www.ahrq.gov/news/press/pr2011/lowhlitpr.htm Andersen, R. M. (1995). Revisiting the behavioral model and access to medical care: Does it matter? J Health Soc Behav, 36(1), 1-10. Baron, R. M., & Kenny, D. A. (1986). The moderator-mediator variable distinction in social psychological research: Conceptual Berkman, N. D., Dewalt, D. A., Pignone, M. P., Sheridan, S. L., Lohr, K. N., Lux, L., Bonito, A. J. (2004). Literacy and health outcomes. Evid Rep Technol Assess (Summ), (87), 1-8. American Academy of Pediatrics, Committee on Practice and Ambulatory Medicine. (2000). Recommedations for preventive pediatric health care. Hakim, R. B., & Bye, B. V. (2001). Effectiveness of compliance with pediatric preventive care guidelines among Medicaid beneficiaries. Pediatrics, 108(1), 90-97. Shulman, S. (2006). Poor preventive care achievement and program retention among low birth weight infant Medicaid enrollees. Pediatrics, 118(5), e1509-1515. doi: peds.2004-0489 [pii] 10.1542/peds.2004-0489 70
Linking the records: Mothers ~ birth certificate ~ children’s SSN ~ Medicaid claims Is maternal HL associated pediatric healthcare use? If and to what extent does maternal HL mediate or moderate relationships between the predisposing and enabling factors and the dependent variables understudy?
We employed a HEDIS measure, appropriate to our age group of interest, which explicitly described which Current Procedural Terminology (CPT) and International Classification of Diseases 9thRevision, Clinical Modification (ICD-9-CM) procedures and diagnostic codes indicate well-child preventive visits (National Committee for Quality Assurance, 2011; Zuckerman, et al., 2004). Sick –child visits are non-routine visits to a provider (for illness or injury) ED visits are any visit to an emergency department for care (for illness or injury) There is a precedent in the literature for setting this bench-mark at 4 for year 1 and 2 for year 2 of life, slightly lower than the AAP recommendations
This participant’s learning disability, in this context was a negative predisposing (denoted by circle A in diagram 4.5) factor and her decision not to disclose her reading disability to the clinician or ask her daughter to help her interpret the written education material can be categorized as a negative enabling factor (denoted by circle B in diagram 4.5) as well as factors that compromised her ability to assess the health care needs of her daughter (denoted by circle C in diagram 4.5). Ultimately these factors diminished the quality of the healthcare experience.
Not just about reading but there are a set of well validated measure that are used to screen for low health literacy REALM and STOFHLA consistent – so we focuses on the REALM
In statistics, the Kruskal–Wallis one-way analysis of variance by ranks (named after William Kruskal and W. Allen Wallis) is a non-parametric method for testing equality of population medians among groups. It is identical to a one-way analysis of variance with the data replaced by their ranks. It is an extension of the Mann–Whitney U test to 3 or more groups
Some 84% of our sample self identified as non-Hispanic Black, 5% White, 8% Latina and 4% (or one participant) did not identify with any of these racial/ethnic groups. Of the 185 women in our sample only 20 (11%) were born outside of the United States, and English was the first language for all but 18 (10%) participants. Age was captured at enrollment in the parent study, which was during the first prenatal visit. The majority of our participants (45%) were between 20 and 24 years of age, 28% were under age 20 and 27% were 25 and older. All participants reported having housing at the time of enrollment but the housing arrangements varied as did the level and type of financial assistance received for housing (both formal and informal). In our sample 135 (73%) women rented an apartment or house with no financial assistance, 26 (14%) women reported owning their own home, the remaining 13% lived with family and friends and had a variety of informal financial arrangements with the home-mates. Most of the women in our sample were unmarried at the time of enrollment in the study (83%) and 106 women (57%) were pregnant for the first time or with the first child they ultimately delivered (Table 4.1).
A variable functions as a mediator when it meets the followingconditions: (a) variations in levels of the independent variablesignificantly account for variations in the presumed mediator(i.e., Path a), (b) variations in the mediator significantly accountfor variations in the dependent variable (i.e., Path b), and(c) when Paths a and b are controlled, a previously significantrelation between the independent and dependent variables is nolonger significant, with the strongest demonstration of mediationoccurring when Path c is zero.
Controlled for REALM even though there was not sig association between REALM and the DV Or, instead of adding each of these chi2 values and the p-values, you can out a foot note that the likelihood ratio test was used to evaluate the additional predictive power of the reading score in the regression models. None of the p-values were significant and hence adding the reading score did not increase the strength of fit of any of the models.
impact of the noise intensity as a predictor (Path a), the impactof controllability as a moderator (Path b), and the interactionor product of these two (Path c). The moderator hypothesis issupported if the interaction (Path c) is significant. There mayalso be significant main effects for the predictor and the moderator(Paths a and b), but these are not directly relevant conceptuallyto testing the moderator hypothesis.In addition to these basic considerations, it is desirable thatthe moderator variable be uncorrelated with both the predictorand the criterion (the dependent variable) to provide a clearlyinterpretable interaction term. Another property of the moderatorvariable apparent from Figure 1 is that, unlike the mediator-predictor relation (where the predictor is causally antecedentto the mediator), moderators and predictors are at the samelevel in regard to their role as causal variables antecedent orexogenous to certain criterion effects. That is, moderator variablesalways function as independent variables, whereas mediatingevents shift roles from effects to causes, depending on thefocus oftbe analysis.
General Educational Development Test (GED).
This participant’s learning disability, in this context was a negative predisposing (denoted by circle A in diagram 4.5) factor and her decision not to disclose her reading disability to the clinician or ask her daughter to help her interpret the written education material can be categorized as a negative enabling factor (denoted by circle B in diagram 4.5) as well as factors that compromised her ability to assess the health care needs of her daughter (denoted by circle C in diagram 4.5). Ultimately these factors diminished the quality of the healthcare experience.
This participant’s learning disability, in this context was a negative predisposing (denoted by circle A in diagram 4.5) factor and her decision not to disclose her reading disability to the clinician or ask her daughter to help her interpret the written education material can be categorized as a negative enabling factor (denoted by circle B in diagram 4.5) as well as factors that compromised her ability to assess the health care needs of her daughter (denoted by circle C in diagram 4.5). Ultimately these factors diminished the quality of the healthcare experience.
This participant’s learning disability, in this context was a negative predisposing (denoted by circle A in diagram 4.5) factor and her decision not to disclose her reading disability to the clinician or ask her daughter to help her interpret the written education material can be categorized as a negative enabling factor (denoted by circle B in diagram 4.5) as well as factors that compromised her ability to assess the health care needs of her daughter (denoted by circle C in diagram 4.5). Ultimately these factors diminished the quality of the healthcare experience.
This participant’s learning disability, in this context was a negative predisposing (denoted by circle A in diagram 4.5) factor and her decision not to disclose her reading disability to the clinician or ask her daughter to help her interpret the written education material can be categorized as a negative enabling factor (denoted by circle B in diagram 4.5) as well as factors that compromised her ability to assess the health care needs of her daughter (denoted by circle C in diagram 4.5). Ultimately these factors diminished the quality of the healthcare experience.