This exploratory study suggests that patients' reading ability is independently associated with their risk of hospitalization. Patients with inadequate literacy had more than twice the risk of being hospitalized during 1994 and 1995. After adjusting for demographics, self-reported health, economic indicators, and health insurance coverage, patients with inadequate literacy were still more likely to be hospitalized (adjusted OR 1.69; 95% CI 1.13, 2.53). Inadequate literacy had an even stronger association among patients who said they had been hospitalized in the year before study entry (adjusted OR 3.15; 95% CI 1.45, 6.85). This latter group may have more complex treatment plans, or they may have more precarious health so that even minor misunderstandings about medication instructions or care plans may result in deteriorating health and hospital admission.
When thinking about possible explanations for the relation between literacy and hospital admission, it is important to recognize that inadequate literacy is not only a reading problem. Patients with inadequate literacy are likely to struggle with oral as well as written communication. For example, patients with poor reading ability may have more difficulty understanding oral instructions because of limited vocabulary or difficulty following complex sentence structure. They may also have limited problem-solving abilities or be less likely to change behavior on the basis of new information. Thus, inadequate health literacy may be a marker for a complicated array of problems with provider-patient communication and health behaviors that affect the risk of hospital admission but are not directly related to reading ability.
Previous research suggests that inadequate literacy could directly affect patients' health and the risk of hospital admission. Patients with inadequate functional health literacy are more likely to be unable to read or to misread directions on prescription labels.2
This can lead to patients taking either too much or too little of their prescribed medications.3
Patients with inadequate functional health literacy are also less likely to know basic elements of their care plan for diabetes and hypertension (e.g., low-salt diet, symptoms of hypoglycemia, normal range for blood pressure or blood glucose level).17
If patients lack knowledge of their medications and self-management techniques, they may be more vulnerable to persistent health problems or even have worsening health that eventually results in hospital admission.
The association between health literacy and hospital admission could also be explained by inadequate adjustment for health status differences between patients with adequate and those with inadequate literacy. Although we adjusted for differences in patients' self-reported health, it is possible that other indicators of health status (e.g., chronic conditions, health-related quality of life) would increase the effect of health status in the model and decrease the effect of inadequate literacy. However, self-reported overall health is significantly associated with other measures of self-reported health and the presence of chronic disease.33, 34
Therefore, the addition of other health measures to the model would be unlikely to have a large enough independent effect to explain entirely the increased risk of hospital admission among patients with inadequate literacy. In addition, some of the increased risk of hospital admission for patients with inadequate literacy could result from a higher prevalence of adverse health behaviors (i.e., smoking, alcohol or drug use, or diet). Previous studies have shown a strong link between the number of years of school completed and health behaviors.35
We did not collect information on health behaviors, so we could not adjust for these factors.
Health literacy was associated more strongly with hospital admission than with the number of years of school completed. This should not be surprising. The number of years of school completed represents education attempted, whereas health literacy is a more valid indicator of educational attainment(i.e., what was actually learned during the years of schooling). For example, 67 (18.6%) of 361 high school graduates in this study had inadequate literacy, and 40 (11.1%) had marginal literacy according to the TOFHLA (see ). In addition to being a more accurate measure of educational attainment, health literacy may be an indicator of someone's ability to acquire new information and complete complex, cognitive tasks. Learning ability, rather than reading ability per se, may be the true mediator of the relation between health literacy and hospital admission, and this may not be captured by the years of school completed.
There are several important limitations to this study. First, we only used data on hospital admissions at Grady Memorial Hospital. If patients with adequate reading skills were more likely to obtain health insurance or shift their care to sites outside the Grady Health System, then analyzing hospital admissions at Grady Memorial Hospital alone would make it appear that patients with adequate literacy were less likely to be hospitalized. If this were the case, then we would expect the relation between literacy and hospitalization to be stronger for 1995 than for 1994, because more patients with adequate literacy would have migrated out of the Grady Health System by 1995. However, there was no difference in the relation between literacy and hospital admission for the two study years and no difference according to patients' reported regular source of care (data not shown). This suggests that out-migration by patients with better reading ability does not explain our findings.
Second, we enrolled patients who were seeking medical care in the walk-in clinic or the emergency department. More than half the patients said they were not followed by a regular physician or that they used the emergency department as their regular source of care. Persons with inadequate reading skills may have more difficulty than those with better reading ability when trying to manage their medical problems and coordinate their medical care without the assistance of a regular provider. In addition, patients in primary care settings may have more resources available to ameliorate the problems created by inadequate literacy (i.e., a physician knowledgeable of the patient's reading problems, nursing staff, or health educators). The effect of inadequate literacy in other settings may therefore be less. The relation between literacy, health care use, and health outcomes will probably vary depending on the patient population, practice setting, and characteristics of the health care system.
Patients with inadequate literacy form a uniquely vulnerable group. They have worse health, fewer economic resources, and less ability to successfully navigate the health care system and complete personal health care tasks. This issue is of particular concern for the elderly. In 1993, 36 million Americans were enrolled in Medicare.36
According to the 1993 National Adult Literacy Survey, 44% of Americans over the age of 65 were functionally illiterate, indicating there were approximately 16 million functionally illiterate Medicare beneficiaries.1
If these individuals have higher than expected hospital costs, and if these excess hospitalizations could have been prevented by improved communication and education, then Medicare hospital costs could be reduced substantially. Further studies are necessary to determine whether the results of this study are generalizable to other patient populations, whether the relation results directly from problems with patient-provider communication, and whether innovative approaches to improve communication and patient education can improve outcomes.37–49