Health Literacy Among Medicare Enrollees in a Managed Care Organization

Context Elderly patients may have limited ability to read and comprehend medical information pertinent to their health.

Objective To determine the prevalence of low functional health literacy among community-dwelling Medicare enrollees in a national managed care organization.

Design Cross-sectional survey.

Setting Four Prudential HealthCare plans (Cleveland, Ohio; Houston, Tex; south Florida; Tampa, Fla).

Participants A total of 3260 new Medicare enrollees aged 65 years or older were interviewed in person between June and December 1997 (853 in Cleveland, 498 in Houston, 975 in south Florida, 934 in Tampa); 2956 spoke English and 304 spoke Spanish as their native language.

Main Outcome Measure Functional health literacy as measured by the Short Test of Functional Health Literacy in Adults.

Results Overall, 33.9% of English-speaking and 53.9% of Spanish-speaking respondents had inadequate or marginal health literacy. The prevalence of inadequate or marginal functional health literacy among English speakers ranged from 26.8% to 44.0%. In multivariate analysis, study location, race/language, age, years of school completed, occupation, and cognitive impairment were significantly associated with inadequate or marginal literacy. Reading ability declined dramatically with age, even after adjusting for years of school completed and cognitive impairment. The adjusted odds ratio for having inadequate or marginal health literacy was 8.62 (95% confidence interval, 5.55-13.38) for enrollees aged 85 years or older compared with individuals aged 65 to 69 years.

Conclusions Elderly managed care enrollees may not have the literacy skills necessary to function adequately in the health care environment. Low health literacy may impair elderly patients' understanding of health messages and limit their ability to care for their medical problems.

In 1993, the National Adult Literacy Survey (NALS) reported that 44% of adults aged 65 years or older scored in the lowest reading level (level 1), meaning they could not perform the basic reading tasks necessary to fully function in society.^1,2 However, because the test used in the NALS did not include health-related items, it is unclear how many elderly persons cannot read adequately to function in health care settings. One study measuring patients' functional health literacy at a public hospital found that 81% of English-speaking patients aged 60 years or older had inadequate health literacy.³ These patients could not correctly read basic items commonly encountered in the health care setting, such as prescription bottles and appointment slips.³ Another study of low-income community-dwelling adults aged 60 to 94 years found mean reading skills at a fifth-grade level.⁴

Recent research has examined the impact of patients' literacy skills on their health and health care. Numerous studies document that health materials such as discharge instructions,^5-8 consent forms,^9-13 and medical education brochures^14-16 often are written at levels exceeding patients' reading skills. Patients with low health literacy and chronic diseases, such as diabetes, asthma, or hypertension, have less knowledge of their disease and its treatment and fewer correct self-management skills than literate patients.^17,18 These factors may explain why patients with inadequate functional health literacy are more likely to be hospitalized than those with adequate health literacy.¹⁹ While many patients with inadequate literacy are unaware of their deficiency,¹ others feel significant shame and are unwilling to disclose their reading problem to health care professionals.^20,21

Previous research on functional health literacy has focused on indigent patients,³ but the results raise concerns about older patients' ability to meet the reading demands they face in the health care setting. Limited health literacy, when coupled with the physical disabilities and chronic illnesses common in the elderly, adds to the barriers faced by elderly patients as they attempt to actively participate in their health care.²² This is of particular concern as more individuals covered by Medicare are enrolling in managed care arrangements, thus adapting to a new health care system.

We conducted this study to determine what proportion of Medicare enrollees in a national managed care organization have low literacy skills in the health care setting (ie, poor ability to read and comprehend the things most commonly encountered in the health care setting, such as prescription bottles, appointment slips, and informed consent forms). We also sought to determine enrollee characteristics associated with low functional health literacy.

The study protocol was approved by the Prudential Center for Health Care Research Institutional Review Board. We selected 4 of the 8 locations where Prudential HealthCare had Medicare managed care enrollees (Cleveland, Ohio; Houston, Tex; south Florida, including Fort Lauderdale and Miami; and Tampa, Fla). These 4 locations were selected based on their projected enrollment, low disenrollment rate, and geographic diversity. One project coordinator and 8 interviewers were hired at each site and participated in intensive training. Several interviewers at the Houston, south Florida, and Tampa locations were bilingual in English and Spanish.

A letter of introduction describing the study was sent to each member who was aged 65 years or older 3 months after he/she enrolled in Prudential HealthCare. One week after the letters were sent, an interviewer called each enrollee to determine eligibility. Individuals who indicated that they were not comfortable speaking either English or Spanish, were blind, had severely impaired vision not correctable with eyeglasses, or were living in a nursing home were excluded. We also excluded enrollees who missed 1 or more screening questions for severe cognitive impairment (not able to correctly identify year, month, state, year of their birth, or home address).

Eligible individuals who agreed to participate completed a 1-hour in-person orally administered survey. Written informed consent was obtained from all participants prior to beginning the interview. The survey consisted of questions to determine demographics, self-rated health,²³ physical functioning,²³ chronic conditions,²⁴ health care use, mental health,^23,25 cognitive impairment,²⁶ social support,²⁷ and health behaviors.²⁸

The last section of the survey assessed enrollees' health literacy using the Short Test of Functional Health Literacy in Adults (S-TOFHLA),²⁹ which takes no more than 12 minutes to administer and is available in both English and Spanish versions.³⁰ The S-TOFHLA uses actual materials that patients might encounter in the health care setting and consists of 2 parts. The reading comprehension section is a 36-item test using the modified Cloze procedure.³¹ This section measures patients' ability to read and understand 2 prose passages written at grade levels of 4.3 (instructions for preparation for an upper gastrointestinal tract radiographic procedure) and 10.4 (Medicaid "Rights and Responsibilities" passage) based on the Gunning-Fog index.³² The numeracy section is a 4-item test using actual hospital forms and labeled prescription vials. This section tests a patient's ability to comprehend directions for taking medicines, monitoring blood glucose level, and keeping clinic appointments. Each item in the reading comprehension is multiplied by 2 (×36 items) to create a score from 0 to 72 and each numeracy question is multiplied by 7 (×4 items) to create a score from 0 to 28. The sum of the 2 sections yields the S-TOFHLA score, which ranges from 0 to 100.

Scores on the S-TOFHLA are classified and interpreted as follows: inadequate health literacy (scores of 0-53) indicate individuals will often misread the simplest materials, including prescription bottles and appointment slips and the instructions for preparation for an upper gastrointestinal tract radiographic procedure. Marginal health literacy (scores of 54-66) indicate individuals perform better on the simplest tasks but have difficulty comprehending the Medicaid rights and responsibilities passage. Adequate health literacy (scores of 67-100) indicate individuals will successfully complete most of the tasks required to function in the health care setting, although many still have difficulty comprehending more difficult information (ie, materials written at higher than a 10th-grade reading level).

Selected enrollee characteristics were examined, including race/language, sex, age, education, income, occupation, self-reported general health status, number of medications taken per day, presence of at least 1 target chronic condition (chronic obstructive pulmonary disease, coronary heart disease, heart failure, hypertension, or diabetes), and cognitive impairment. We determined occupation by asking respondents what kind of work they did for the longest period of time during their adult life. Responses to this question were coded according to US census occupation codes.^33,34 We measured cognitive impairment in the survey using the Mini-Mental State Examination instrument.²⁶ Based on cut points previously established,³⁵ we identified individuals as having severe (0-17), mild-to-moderate (18-23), or no (24-30) cognitive impairment.

To determine differences between respondents and nonrespondents, we compared age and sex, available from enrollment files, between these groups. In addition, we linked ZIP codes of respondents' and nonrespondents' residences to census data to determine differences in socioeconomic status (mean per capita income, percentage of residents that were black, and percentage of residents with low educational attainment).³⁶

Analysis of survey data consisted of comparing the distribution of selected enrollee characteristics by study location, the distribution of health literacy scores by study location and language, and the proportion of incorrect responses to specific numeracy and reading items on the S-TOFHLA for the 3 categories of health literacy (inadequate, marginal, and adequate). We also examined the distribution of functional health literacy category by selected enrollee characteristics. χ² Analyses were conducted to determine significant differences between selected characteristics and study location and between selected characteristics and health literacy level.³⁷ A P value of .05 was used to determine statistical significance. Significant variables in the bivariate analyses were included in the multiple logistic regression analysis to determine the association between selected characteristics and health literacy.

To analyze data across study locations, we conducted weighted analyses using SUDAAN³⁸ software to adjust for differences at each location in number of eligible members, sample size, and response rate. Because weighted analyses yielded results similar to unweighted analyses, we present only the latter. All unweighted analyses were conducted using SAS.³⁹

From the original sample (n=8409), 938 were unable to be contacted and 7471 individuals were contacted 3 months after they joined Prudential HealthCare. Of these, 3247 refused to participate and 737 did not meet eligibility criteria, leaving 3487 enrollees who were eligible and agreed to participate in the survey. A total of 143 people did not keep their interview appointment and 84 did not complete the survey, the health literacy testing, or both. The final sample included 3260 patients. Nonresponders were slightly older than responders (7.5% were 85 years or older compared with 5.4% of responders; P=.009) but the sex distribution was similar. Nonresponders were also more likely to live in a ZIP code with a higher median income, higher educational attainment, and lower proportion of blacks.

The majority of the respondents were white, female, between 65 and 74 years old, with at least a high school education, and currently earning more than $15,000 per year (Table 1). Primary occupation was broadly distributed. All of these characteristics differed by study location (P<.001), except for sex (P=.13). More than one quarter of respondents described their health as fair or poor, 43.6% were taking more than 3 medications per day, 66.5% had at least 1 of the 5 target chronic conditions (chronic obstructive pulmonary disease, coronary heart disease, heart failure, hypertension, or diabetes), and 2.1% had severe cognitive impairment.

Overall, 23.5% of English-speaking and 34.2% of Spanish-speaking respondents had inadequate health literacy (Table 2) and another 10.4% and 19.7%, respectively, had marginal health literacy. Rates of health literacy varied according to study location and language (P<.001 for both). For English speakers, the highest rate of inadequate health literacy was in Cleveland (34.1%), followed by Houston (28.0%), south Florida (17.3%), and Tampa (16.6%). For Spanish speakers, the highest rate of inadequate health literacy was in Tampa (60.0%), followed by south Florida (34.3%) and Houston (21.2%).

The percentages of incorrect responses to specific items on the S-TOFHLA for individuals classified as having inadequate, marginal, and adequate functional health literacy are shown in Table 3. Respondents with inadequate functional health literacy often misread simple prescription instructions, information regarding the results of blood sugar tests, and the simplest reading comprehension passage with instructions for preparation for an upper gastrointestinal tract radiographic procedure (grade level of 4.3 on Gunning-Fog index). Those with marginal health literacy performed better on all these tasks but showed poor comprehension of blood glucose tests, instructions for taking medication on an empty stomach, and the Medicaid rights and responsibilities reading comprehension passage (grade level of 10.4 on Gunning-Fog index). Individuals with adequate health literacy did well on most tasks, although some had difficulty interpreting more difficult numeracy tasks. For instance, 23.5% did not understand a blood glucose range and 17.3% had poor comprehension of the Medicaid passage.

Several enrollee characteristics were also related to health literacy level (Table 4). Characteristics associated with higher rates of inadequate health literacy included black race, older age, fewer years of school completed, and having a work history in "blue collar" occupations (P<.001). For example, 18.9% of whites had inadequate health literacy compared with 29.5% of English-speaking Hispanics, 34.3% of Spanish-speaking Hispanics, and 52.1% of blacks. The relationship between age and health literacy showed a strong trend, with the prevalence of inadequate health literacy steadily increasing from 15.6% of individuals aged 65 to 69 years to 58.0% of those aged 85 years or older. Individuals who rated their health as fair/poor were twice as likely to have inadequate health literacy compared with individuals who rated their health as good/excellent (38.7% vs 19.2%, respectively; P<.001), and individuals who had at least 1 of the target chronic conditions had slightly higher rates of inadequate health literacy than individuals with none of these chronic conditions (25.8% vs 22.1%, respectively; P = .03).

In multivariate analyses, study location, race/language, age, years of school completed, occupation, and cognitive impairment were all significantly associated with inadequate or marginal health literacy (Table 5). For example, the adjusted odds ratio for having inadequate or marginal health literacy was 32.81 (95% confidence interval [CI], 9.68-111.16) for individuals with a severe cognitive impairment and 5.24 (95% CI, 4.21-6.53) for those with a mild to moderate cognitive impairment. Age was also strongly related to health literacy skills, even when adjusting for education and cognitive impairment; the adjusted odds ratio for having inadequate or marginal health literacy was 8.62 (95% CI, 5.55-13.38) for enrollees aged 85 years or older.

To our knowledge, this is the first study examining functional health literacy in a population of Medicare enrollees in a national managed care organization. We found that among Medicare managed care enrollees in 4 areas, more than one third of respondents had inadequate or marginal health literacy. This figure is somewhat lower than statistics on general reading ability from the 1993 NALS, which reported that 44% of adults aged 65 years or older were at level 1—the lowest reading level. This difference may reflect higher socioeconomic status among this study cohort and differences in reading difficulty between the S-TOFHLA and NALS instruments.

The distribution of age, sex, and education in our study population is similar to that for all Medicare enrollees and individuals older than 65 years, based on US census data. Because of the study sites selected, our study population had a higher proportion of Hispanic individuals (11.2% vs 1.1%) as well as slightly more individuals in lower income brackets than the national Medicare population. In addition, our response rate was comparable with other surveys of older adults.^40-42 Thus, we believe our results should be generalizable to elderly populations in other locations.

We found striking differences in the prevalence of inadequate health literacy across the 4 study sites. Much of this was because of differences in race/language and socioeconomic status across the sites. However, even after adjusting for other variables, participants in Cleveland were more likely to have low health literacy. These geographic differences suggest that practitioners and health care delivery systems need to assess health literacy levels in their own setting rather than relying on national data or estimates from 1 location. Blacks and Hispanics had higher rates of low health literacy, even after adjusting for years of school completed, which probably reflects the poor educational experiences of many minorities during their youth. Although health literacy and years of school completed were strongly associated, almost 17% of respondents with a high school education and 10% with more than a high school education had inadequate health literacy. This finding is consistent with previous research showing that years of school completed is an inaccurate indicator of someone's true educational attainment.^1,3,30,43

The markedly higher prevalence of inadequate and marginal health literacy among participants aged 85 years or older was not anticipated. The NALS and previous studies of functional health literacy have found that the proportion of people who are functionally illiterate increases with age. However, these studies combined all individuals aged 65 years or older into 1 group and did not address whether the prevalence of poor reading ability continues to increase after age 65 years. Our results show that the proportion of people with inadequate or marginal health literacy continues to increase beyond age 65 years, even after adjusting for the number of years of school completed. Although there have been no longitudinal studies of individuals' reading ability, this finding suggests that reading ability declines with age.

There are several possible explanations for the increased prevalence of reading difficulties with advancing age. First, older individuals are more likely to have a dementing illness that could affect their reading ability. Second, older individuals may have had more difficulty completing the S-TOFHLA within the allotted time. However, the inverse relationship between age and reading ability was observed when the numeracy items were analyzed separately and these items had no time limit for their completion. Third, because the S-TOFHLA was administered at the end of the interview, older individuals may have been fatigued and may have given less attention to the task than younger individuals.

Despite the strengths of our study, we have only an estimate of the differences between responders and nonresponders to our survey. It would have been useful to have additional data for the nonresponders, particularly on education and language. Although the differences we found between responders and nonresponders (ie, nonresponders were older and had higher education and income levels than responders) could potentially bias results, it is difficult to determine the direction of the bias. The older age of nonresponders suggests that our figures may underestimate the prevalence of low health literacy, whereas higher education and income among nonresponders (estimated by ZIP code of residence) suggest that our figures may overestimate the prevalence of low health literacy.

Results from this study have implications for all levels of the health care delivery system—patient, clinician, and organization. First, to function adequately in the health care environment, patients need to be able to read consent forms, medicine labels and package inserts, and other written health care information; understand written and oral information from physicians, nurses, pharmacists, and insurance companies; and act on necessary instructions, such as those on medication labels and appointment slips.² Moreover, patients entering managed care settings have certain responsibilities that they may not have experienced with traditional fee-for-service plans, such as selecting their primary care physician, determining how to obtain a referral, and using designated pharmacies for discounted services. An individual's health literacy may directly affect his/her ability to negotiate the system and, thus, health outcomes.

Second, clinicians need to be aware of the prevalence of health literacy problems and need to identify patients with poor health literacy skills. One study of otolaryngologists and radiation oncologists indicated that physicians perceive low health literacy as a problem but lack the data needed to enable them to quantify its effect on treatment outcomes.⁴⁴ Physicians and other health care personnel can play an important role in the identification of individuals with reading difficulties. As a first step, practitioners can ask their patients how many years of school they completed. If a patient did not complete high school, the clinician could ask additional questions that may provide clues to patients having reading problems (ie, not knowing the name of the medication they are taking), followed by administration of formal screening tests for patients they have concerns with.²⁰

Third, health care organizations should be aware that inadequate health literacy may adversely affect costs and delivery of care. To adequately serve patients, managed care organizations must know what proportion have limited health literacy skills, particularly among elderly patients who frequently use the health care system and require significant education to manage their chronic health problems. With this knowledge, programs can be designed to try to increase the effectiveness of educational materials, improve health outcomes, and decrease preventable hospitalizations.

Managed care organizations provide an ideal setting in which to address many of the health literacy–related issues that affect the various levels of the health care delivery system. For instance, the results from this study could be used to help design an intervention program for elderly enrollees, specifically targeting certain subgroups that are at high risk (eg, older age, less education). Possible intervention strategies include use of audiotape and videotape recorders with medical instructions or providing instructions with visual cues rather than written instructions. These types of efforts lend themselves to partnerships with medical professional organizations, pharmacists, community groups, literacy councils, managed care organizations, other health care delivery organizations (eg, hospitals, medical groups), and health policy agencies to appropriately address the impact of literacy on health.