Factors that influence health literacy in patients with coronary artery disease

Objective: to investigate the factors that exert an influence on health literacy in patients with coronary artery disease. Methods: a crosssectional study, including 122 patients with coronary diseases (60.7% male; 62.07 ± 8.8 years old). Health literacy and specific knowledge about the disease were evaluated through interviews with the participants by means of the Short Test of Functional Health Literacy in Adults and the Short version of the coronary artery disease education questionnaire. The data were described by means of central tendency measures and frequencies. The factors that exert an influence on health literacy were determined by means of a linear regression model. The significance level adopted was 5%. The study was approved by the Research Ethics Committee. Results: age and arterial hypertension presented an inverse and significant relationship with health literacy. On the other hand, higher schooling levels and having a job were associated with better scores in the health literacy instrument. Specific knowledge about the disease did not exert any influence on health literacy. The variables included in the regression model explained 55.3% of inadequate literacy. Conclusion: this study, knowledge about the disease exerts no influence on health literacy: however, the professionals should consider the sociodemographic and clinical factors to plan the interventions.


Introduction
Complexity of the varied information and treatment options is acknowledged as a barrier for a successful implementation of health interventions.
Using unnecessarily complicated language or inaccurate translations in educational materials, or other education sources for patients generally leads to poorly informed decision-making and decreased participation of the patients in interventions with potential benefits for health.
Such complexity exerts a negative impact on health results, especially among patients with low health literacy (HL) (1)(2) levels.
Health literacy is defined as a person's ability to acquire, process and understand all the health-related information required to make suitable decisions and achieve positive health results. It involves acquiring specific skills to perform activities of daily living related to health and tasks, as well as making decisions to improve health outcomes (3)(4) .
The prevalence of low HL is relatively high at the global level, particularly among people with low socioeconomic status (5) . A number of research studies have evidenced that low HL is an independent health determinant and that it is associated with worse health results, such as increased hospitalizations, use of emergency services, low adherence to medication regimes and higher mortality rates. In addition, patients with low HL have more difficulty understanding health information, less knowledge about their disease and less support to discuss health problems, in addition to feeling less comfortable communicating with health professionals and discouraged or even embarrassed, to ask questions in order to clarify the information they received (5)(6) . On the other hand, the literature shows higher HL levels are associated with better health selfmanagement indicators such as adherence to healthy lifestyles and lower rates of obesity, smoking and hospital readmissions (7)(8) .
Inadequate HL is acknowledged as a barrier for health maintenance and prevention of coronary artery disease (CAD) and is associated with non-adoption of self-care behaviors for management of the disease (9)(10)(11) . In Brazil, a developing country with profound social inequalities and weaknesses in its health and education systems, studies on HL are still scarce, specifically in patients with CAD.
To the present date, no national studies analyzing HL in this group of patients have been found, however, data from other groups of patients show that the HL level in the country is low or limited (12)(13)(14)(15) .
Specifically in patients with CAD, the literature evidences that the prevalence of inadequate HL is considered high, with a frequency of up to 74.5% and that it has been associated with unfavorable outcomes for health, such as less knowledge about the disease and non-adherence to the treatment (9)(10) . A systematic review that synthesized the literature in relation to HL in people with CAD showed knowledge about the disease, age, schooling, socioeconomic disadvantaged status, non-white ethnicity and comorbidities are associated with low HL (9) .
According to the conceptual model on HL, widely discussed in the literature, text comprehension, numbering, vocabulary and prior knowledge are necessary resources to effectively deal with health information (4) .
However, the influence of specific knowledge about CAD on HL is not fully understood. Although some studies have shown the relationship between HL and specific knowledge about the disease in different groups of patients (11,(16)(17) , other surveys that included patients with chronic and infectious diseases failed to demonstrate that specific knowledge about the disease influences HL (18)(19)(20)(21) . A metaanalysis that included more than 18,000 participants with diabetes showed that higher HL levels were associated with higher levels of knowledge about the disease (r=0.308, p<0.001) and with better clinical outcomes, such as lower levels of glycated hemoglobin (HbA1C) (r =-0.048, p = 0.027) (18) .
Genuine patient-centered care, especially in patients with CAD, will only be attained if high HL levels are taken into account (22) . Knowing the factors associated with HL will increase success: in the implementation of health interventions in CAD management, directing and adapting the communication between health professionals and patients and planning of educational interventions. Thus, the current study investigated the factors that exert an influence on health literacy in patients with CAD.

Type of study
This is an observational and cross-sectional study with a quantitative approach that is part of a main project aimed at analyzing the association between self-care and HL in patients with CAD.

Period
The data collection period was from August to October 2019.

Population and sample
The population consisted of patients with CAD registered in the tertiary-level Coronary diseases outpatient service, a reference in Cardiology from the city of São Paulo, Brazil.
For the main study that aimed at "Analyzing the association between self-care ability, health literacy and knowledge about the disease in patients with CAD", the minimum sample size was 84 participants, calculated by means of the G Power statistical program, version 3.1 (24) , based on an infinite population, moderate correlation (r = 0.30), 80% test power and 5% type I error. The correlation value adopted for sample size calculation allows analyzing the existence of a correlation between the variables of interest, without excessively increasing the sample size to the point of rendering the study unfeasible.
The individuals included in the study were those aged at least 18 years old, with a medical diagnosis of CAD documented in their medical charts and who stated being able to read. The Mini-Mental State Examination (MMSE) was used as pre-screening tool to determine presence of cognitive impairment, when the score obtained by the participants was below 20. This pre-screening is necessary because impairment in cognitive functions might result in poor performance in the other assessments due to difficulties understanding the tasks. This procedure is welladopted in the literature (15) . The participants excluded from the study were those with deficits in visual, auditory and/ or verbal communication skills that precluded application of the data collection instruments.

Procedures for data collection
The potential study participants, that is, those aged at least 18 years old and with a medical diagnosis of CAD, were identified based on the coronary diseases outpatient service schedule of appointments. Subsequently, the main researcher herself invited the potential participants to the study. Those who agreed signed the informed consent form and answered MMSE. For the participants included in the study, individual meetings were held to apply the other data collection instruments, as described below.

Instruments for data collection
The sociodemographic and clinical data were collected from the medical charts or through the participants' selfreports. The data of interest for this study were sex, age, race, schooling, marital status, professional status, per capita income of the patients, number of medications, smoking habit, comorbidities (hypertension, diabetes and dyslipidemia) and previously-received health education.
The HL level was assessed by means of the Short Test of Functional Health Literacy in Adults (S-TOFHLA), validated for use in Brazil (12) . Choice of this instrument was due to the fact that it assesses both reading comprehension and numbering. S-TOFHLA contains 36 reading comprehension items organized into two passages: A and B. Passage A includes information related to a gastrointestinal examination and Passage B refers to a term of rights and responsibilities of a patient admitted to a hospital. Each passage has a space (each space corresponds to an item) and, below each space, there are five or six word options, from which the participant has to select one to complete the sentence and make sense of the it. Two points are assigned to each word that is properly selected, so that the total score is 72 points.
In turn, the numbering assessment consists in presenting four cards for the participant to interpret the information. The first card is a medication label. The second refers to how to interpret a glycaemia value. Card three deals with the date of next appointment, considering the one printed in the card. Finally, in card four, the participant has to calculate the time to take a medication.
Two points are assigned to each correct answer, so that the numbering component has a total score of 28 points.
It is recommended that the reading comprehension and numbering components should be finished in seven and five minutes, respectively, therefore, application of the instrument should be timed.
The total S-TOFHLA score varies from 0 to 100 and comprises the sum of the reading comprehension (from 0 to 72) and numbering (from 0 to 28) scores. Based on the score obtained, the participants' HL level is classified as inadequate (from 0 to 53), marginal or borderline (from 54 to 66) or adequate (from 67 to 100). The original instrument's internal consistency was 0.68 for the four numbering items and 0.97 for the 36 items from both passages of the reading comprehension test (12,25) . psychosocial risk. Each of the four areas has four items that are randomly arranged. The answer options are as follows: "True", "False" and "I don't know". Each correct answer equals one point; therefore, the total score varies from 0 to 20. The higher the score, the better the patient's knowledge about CAD. The Brazilian version of the instrument presented adequate evidence in terms of reproducibility and its intraclass correlation coefficient was greater than 0.70 for all the items (26) .

Data treatment and analysis
The data were analyzed with the aid of the R software, Normality of the regression analysis residuals was evaluated by means of Q-Q plots. Multicollinearity of the independent variables was assessed by means of the variance inflation factor. The significance level adopted in all the tests was 5%, with 95% confidence intervals.
Internal consistency of the instruments used to measure HL and knowledge about CAD in the sample of this study was calculated by means of Cronbach's alpha (27) .

Ethical aspects
The study was approved by the Research Ethics Committee at the hospital and the Nursing School of the

Results
The sociodemographic data and clinical characteristics of the sample are detailed in Table 1.      However, several studies show that inadequate HL is common in patients with cardiovascular diseases in general (12)(13)(14)(15) , particularly in those with CAD (9)(10) .
In this study, the factors that influenced HL were age, schooling, employment and hypertension. There was a reduction in HL with increasing age and diagnosis In addition, normal distribution was observed for the residuals of the variables included in the regression model by means of Q-Q graph, as shown in Figure 2. of hypertension, while individuals with some professional occupation and with more years of study presented higher HL levels. A systematic review also showed that the HL level was worse in unemployed people, of advanced age, with low schooling levels and multiple comorbidities (9) .
Other authors agree that HL is a multidimensional concept influenced by personal, family, environmental and social factors. They also consider that the social factors tend to exert a strong influence not only on development of the skills required for HL, but also on the way in which patients use health information (4,(28)(29) .
The literature shows that low HL is associated with poor blood pressure control (30) . Although there is evidence that hypertensive people with adequate HL have better adherence to the treatment (29,31) , a study showed that the contribution of HL to explain such adherence was minimal (30) .
In relation to age, the older adults' low performance in tests that evaluate HL can be attributed to decline in cognition, hearing and vision, as well as to less access to educational information in the past (31) . A study showed that participants with lower S-TOFHLA scores, classified as inadequate or marginal, were significantly older, prone to be dependent on daily activities and to perform worse in cognitive domains tests, such as MMSE (32) .
In another study that included a sample of 575 patients with heart failure, the authors showed that older patients were more likely to have low HL and that the health outcomes (readmissions and mortality) were worse in those aged over 65 years and with low HL. In addition to that, the authors were able to show that HL mediates the relationship between age and health results (33) .
With regard to schooling, the findings of this study are consistent with the literature, where it is observed that individuals with lower schooling levels have lower HL scores, worse reading skills, less autonomy to seek health information in different sources and difficulty understanding and judging what is most appropriate for their own well-being (25,34) . A cross-sectional study that included 252 patients with cardiovascular diseases showed that the lower the schooling level, the lower the HL level.
The authors showed that patients without schooling had lower scores in all nine dimensions of the HL instrument used in the study (Health Literacy Questionnaire -HLQ) (34) . In a systematic review, education was considered a predictive factor for HL in patients with heart failure. The studies included in that review showed that the patients who had less than High School were more likely to have low HL levels (35) .
Individuals with Higher Education tend to have better reading skills, more autonomy to seek health information in different sources and better ability to understand and judge what is most appropriate for their own well-being, in addition to having other positive attitudes and behaviors (25,34) . It is possible that individuals with high schooling levels feel safer and have more clarity when communicating their needs to health professionals (34)(35) .
In relation to employment, work status is considered as the economic inclusion degree that grants or restricts people's access to the necessary resources (36) . Thus, being employed can be seen as an indicator of socioeconomic status. Actually, socioeconomic status does not directly affect health status. However, it was recognized as an important determinant of health-related results (37) . Those who are unemployed, retired or on sick leave and who earn low incomes and are devoid of additional health resources such as health insurance, tend to have low adherence to preventive practices and less contact with health-related information (37) . In addition, as they are not part of the workforce, they may feel disabled and vulnerable. As a result, they delegate some activities to family members, such as scheduling appointments and exams or controlling drug therapy and end up distancing themselves from the health system, thus limiting communication with professionals and, in some cases, not making decisions related to their own health (38) . Recently, a study also showed that HL is a mediator of the relationship between socioeconomic level and health status (37) .
Regarding knowledge about the disease, CADE-Q SV has been used in few studies, which hinders comparisons.
In the sample under study, the mean CADE-Q SV score was 12.3 ± 2.5 (possible range from 0 to 20). In the validation study of this instrument in Brazil, the authors found a mean CADE-Q SV score of 13.1 (26) . with specific issues related to body functioning and the disease would contribute to HL (4) .
Corroborating the findings of this study, the results of A meta-analysis of more than 18,000 patients with diabetes showed that higher HL levels were significantly associated with better knowledge about diabetes and, in particular, the performance-based HL measures were the best predictors for knowledge about diabetes (18) .
Interestingly, tools for assessing HL with a numbering section led to a significantly smaller effect size than those that did not include it (18) .
In this study, most of the participants answered the numbering questions correctly, which may suggest that the content or the random answer were easy for the sample under study. As a result, the HL score may have been overestimated, affecting the Cronbach's alpha value.
This is not to say that the instrument is not suitable for HL assessments but, rather, that the size of our sample was not sufficient to calculate psychometric parameters.
In this sense, it is not possible to draw conclusions regarding validity of the instrument based on our results. In addition, the health system as a whole should be reorganized to assist patients according to their HL levels to reduce the complexity inherent in navigating across the health services, whether for simple clarification of doubts about their health condition, for scheduling, and for referral to services and other specialized professionals.
The interventions need to be focused on helping patients organize the information in a meaningful and simple way so they can put interventions into practice to improve decision-making about their health management.
There is a need for studies on HL levels in different scenarios, as well as for HL assessments to be routinely performed by the health professionals during the first appointment with the patients so that interventions can be designed according to the individual and socioeconomic characteristics and to each patient's degree of understanding.