Health Literacy and Diabetic Patients Health Education Study
DRAFT Summary
Background
Health literacy is defined as he ability to obtain, understand, and use the information needed to make wise health choices is known as health literacy (DHHS, 2005). Low literacy among members of such populations as older adults, people with poor reading skills, those with limited mastery of the English language, members of ethnic and cultural minorities, and immigrants is likely a major contributor to health disparities in this country, according to Healthy People 2010.
People with low health literacy often lack not only the ability to read well but also the knowledge about the body, its functioning, and the nature and causes of different types of disease (DHHS 2005). Beyond reading and other communication skills, as well as knowledge of relevant health topics, making sense of health information and the healthcare system also requires numerical skills, such as disease risk or the normal range of values such as blood pressure or cholesterol (DHHS 2005).
“Adequate functional health literacy means being able to apply literacy skills to health related materials such as prescriptions, appointment cards, medicine labels, and directions for home health care” (Parker et al., 1995). Research based on this definition has shown, e.g. that poor functional health literacy poses a major barrier to educating patients with chronic diseases (Williams et al., 1998), and may represent a major cost to the health care industry through inadequate or inappropriate use of medicines (National Academy on an Aging Society-Center for Health Care Strategies, 1998).
Many barriers—including poverty, limited education, low reading levels, low self-efficacy, and inadequate English-language skills—stand in the way of developing health literacy (DHHS 2005). Tests such as the Rapid Estimate of Adult Literacy in Medicine (REALM) and the Test of Functional Health Literacy in Adults (TOFHLA) can identify patients with low literacy skills (Davis et al., 1998). Patient age, education level, cultural background, prestige and power of the provider, and practical clues are all factors that can limit literacy.
There are many steps that can improve the health of millions of Americans by addressing health literacy barriers (DHHS 2005). These range from collaboration between professional societies, government agencies, and the educational system to increasing the public’s access to accurate and understandable health information through the mass media, the development of culturally and linguistically appropriate materials, and the collaboration of professional groups.
Health literacy is clearly dependent upon levels of fundamental literacy and associated cognitive development. Individuals with undeveloped skills in reading or writing will not only have less exposure to traditional health education, but also less developed skills to act upon the information received; for these reasons, strategies to promote health literacy will remain inextricably tied to more general strategies to promote literacy (Nutbeam 2000).
According to the Joint Commission on Accreditation of Healthcare Organizations' (JCAHO) standards on patient and family health education (2002), it is incumbent upon health professionals to identify learning and educational needs, use appropriate educational resources, and to assess the patient's and family's ability to comprehend, use and apply information taught. However, amid the demands of disease management, understanding complex clinical presentations, knowing multiple differential diagnoses, interpreting laboratory and diagnostic tests, literacy assessment may pale in comparison (Artinian 2003).
Methods
This study was conducted in two primary healthcare facilities, the Internal Medicine and Family Practice clinics of the Moses Cone Health System in Greensboro, NC. The Internal Medicine Clinic serves approximately 900 diabetic patients and the Family Practice Clinic serves approximately 625 diabetics annually. Potential subjects in the study were identified by accessing the hospital system’s computerized database. Patients were eligible if they were currently diagnosed with either type I or type II diabetes. Patients who did not speak English and did not have an interpreter on site were excluded.
Between September 2007 and May 2008, two research assistants enrolled all eligible, willing, and able patients who were present at the clinic for an appointment. Patients who were fluent in English, or patients who spoke a different language and had an interpreter, were asked to participate in the study. Written and oral consent were obtained from patients before participation and a $10 CVS gift card was offered for patients who completed the study. The protocol was approved by the Institutional Review Boards of The University of North Carolina at Greensboro and Moses Cone Health System.
Measures
To measure health literacy, we used the English version of the Short Test of Functional Health Literacy in Adults (STOFHLA), a 36-item reading comprehension test that uses a modified Cloze procedure. Every fifth to seventh word of the passage is omitted and 4 multiple-choice options are provided. Participants were given 7 minutes to complete this test. The STOFHLA was scored on a scale of 0 to 36. The test scores were used to categorize patient health literacy levels as inadequate (between 0 and 12), marginally adequate (between 13 and 24), or adequate (25 or above). Participants were then administered a test of diabetes knowledge which included 9 multiple choice questions. These items were taken from the Michigan Test of Diabetes Knowledge. Upon completion of the test, participants were given one of three types of diabetes education materials (1) a DVD tutorial (2) a low literacy notebook or (3) the written pamphlet that was being distributed within the clinic at the time. Immediately after the intervention, the participants took the knowledge quiz again. The scores of both the previous and post tests were analyzed to observe whether patient scores were influenced by the education materials provided to determine the best method of intervention. As a side note, the research plan originally called for a fourth intervention of a computer-based tutorial. After 100 subjects were enrolled, only 3 had completed the computer-based intervention, due to subject refusal to participate in this method. Even when the researchers offered assistance with computer use and sat with the participants, few subjects were willing to attempt to navigate the diabetes education website. This intervention was subsequently dropped for the remainder of subject recruitment.
Demographic Data
One hundred and seventy five subjects participated in the study. Of those, 158 subjects were included for analysis. The three subjects who completed the computer tutorial were excluded, as were subjects who did not complete all of three tests (STOFHLA, pre-test, post-test). The age of participants ranged from 19 to 82, with a mean age of 53. One hundred twenty-two subjects were female (77%); 36 were male (23%). More minorities participated in this study than would be expected, given the overall population. The subjects self-identified as black (63%), white (36%), and Hispanic (1%). Participants were randomly given one of the three interventions DVD (26%), low literacy materials (37%) or regular written materials (36%).
STOFHLA
The STOFHLA scores were calculated into percentile scores for ease of comparison to the diabetes knowledge test scores. STOFHLA scores ranged from 11% to 100%, with a mean score of 78%. Categorically, 13% of subjects scored in the “inadequate” range, 13% scored in the “marginal” range, and 74% scored in the “adequate” range. Overall, study participants had higher than expected levels of health literacy. A previous study of health literacy levels in the same clinics, using the same instrument showed that health literacy levels mirrored the general, national population, where about half of the population has adequate health literacy, 30% has marginal health literacy and 20% had inadequate health literacy. It is expected that the lower rate of low health literacy in the study population is directly related to the intensity of the reading demand made on subjects to participate. The researchers noticed many potential subjects who chose not to participate, most likely due to difficulties with reading.
Diabetes Knowledge Test
The mean pre-test score was 63.92, with a range of 14 to 100 and a standard deviation of 17.6. At post-test, the mean score was 71.27, with a standard deviation of 17.447, and a range of 21 to 100. The overall change from pre to post was significant at p=.000 using a paired samples t-test measure. The average gain from pre to post test was 7.25 percentile points, with a range of -22 to 60.
Change scores were calculated for each subject using the Reliable Change Index (RCI) to ensure the strength of the changes observed.
RCI = (posttest-pretest) / SEmeas
A series of one way analysis of variance (ANOVA) tests were conducted for change by intervention type, gender, and literacy level. None of these tests produced significant results, though change by intervention type approached significance (p=.074).
Analysis by… / Sum of Squares / df / Mean square / F / SignificanceIntervention type / 651.760 / 2 / 325.88 / 2.653 / .074
Gender / 19.353 / 1 / 19.353 / .153 / .696
Literacy Level / 265.605 / 2 / 132.802 / 1.060 / .349
To further examine the effects of the three intervention types, independent samples t-tests were performed to compare written materials vs. the DVD, where RCI was the measure of change. Participants who read the low literacy and regular written materials were grouped and compared to those who watched the DVD; the results were not significant (p=.696). A second t-test was performed comparing low literacy written materials to the regular written materials, with the low literacy materials performing significantly better than the regular materials (p=.039). A third t-test was performed comparing low literacy written materials to the educational DVD; the difference was not significant (p=.160).
Summary
While overall, statistically significant improvements in diabetes knowledge were observed for the entire study population, statistically significant differences were also observed between subjects exposed to the low literacy written educational materials and the regular written materials. Between the low literacy materials and the audiovisual (DVD) materials, there was no statistical significance, though subjects in the low literacy group did improve more than those in the DVD group. These results suggest that low literacy written materials may be the most effective means to improving health knowledge. It is important to note that all of the materials utilized in this study produced improvement in knowledge scores. This indicates that targeted education in many forms may produce short-term gains in knowledge. What remains unknown is whether the knowledge is kept in long-term memory and whether the knowledge is translated into health behavior practice. We intend to test these questions in the next phase of this study.
The scope of the diabetes knowledge test may also be a key factor in subjects’ learning. By pre-testing the subjects on core knowledge items, their memories may have been primed to seek answers to the pre-test questions. This may have increased their motivation to learn and perform on the post-test. This factor may actually strengthen the results of the between-intervention study, as the knowledge was available in all three formats, but was recalled with greatest accuracy by the group that received the low-literacy materials.
In terms of practical significance of this study, a gain in core knowledge about caring for diabetes should translate into better disease outcomes. All of the study subjects had a diabetes diagnosis for one year or longer, meaning that they should have received ongoing care and education about the management of their disease. Few subjects (n=4) knew all of the core diabetes knowledge competencies at pretest, while significantly more (n=13) had acquired this knowledge at posttest. On average, subjects learned at least one additional fact about caring for their diabetes, which could potentially improve health. While further study about the implications of these findings is needed, attention to patient health literacy in general, and emphasis on easily understandable education materials in particular, may play a significant role in improving the health of patients with chronic illness.
Report prepared by Jen Kimbrough, PhD, Associate Director at the Center for Youth, Family & Community Partnerships at UNCG. Phone 336-217-9737 Email Thanks to Mandy Benson, MPH and Mohammed Khalaf for data collection and assistance in preparation of this report.