Background

The concept of health literacy is an important but complex concept, and since it was first propositioned there has been continuing debate about the definition and the approaches used to measure the levels [1–5]. The World Health Organization (WHO) defines health literacy simply as “an individual’s ability to gain access to understand and use health information” for promoting and maintaining health [6, 7], referring to the definition of health literacy given by the National Assessment of Adult Literacy, the Institute of Medicine Committee for Health Literacy and the American Medical Association (AMA) [8].

People with low health literacy tend to adopt fewer preventive services and less health information technology. As a result, they use emergency department more frequently, have poorer health outcomes and a higher risk of death [9–12]. It is particularly urgent to identify these community groups and to take measures to improve their CDHL.

There are a number of approaches which can be utilised to measure the “degree to which individuals and communities have the capacity to obtain, process, and understand basic health information and services needed to make appropriate health decisions” [13]. These include the Rapid Estimate of Adult Literacy in Medicine (REALM) [14], the Test of Functional Health Literacy in Adults (TOFHLA) [15], the Shortened Rapid Estimate of Adult Literacy in Medicine (REALM-R), the Shortened Test of Functional Health Literacy in Adults (S-TOFHLA), the Newest Vital Sign (NVS), the eHealth Literacy Scale (eHEALS), the Health Literacy Screening Question Methodologies (HLSQM) [16]. Generally, the TOFHLA and REALM tests are regarded as the gold standards for assessing the health literacy of patients [16]. The TOFHLA is used to measure patients comprehension towards health information and characterizes them as having adequate, marginal or inadequate health literacy [15]. The REALM test is the most commonly used health literacy assessment tool in the clinical setting. It tests the ability of patients to read and pronounce common medical terms [17].

While the morbidity and mortality associated with communicable diseases has been steadily declining for Beijing residents, new threats have continued to emerge in the last ten years including SARs and influenza H5N1 and H7N9 [18–23]. In order to formulate appropriate prevention/control and communication strategies it is important to understand the level of CDHL level of Beijing residents. Therefore, this new study aimed to evaluate CDHL levels of Beijing residents towards three key areas: knowledge, adoption of preventative measures/behaviours, and health skills. The distinction between the three areas is based on the health literacy definition of the WHO [6, 7].

Methods

Instrument development

The study aimed to measure and assesses health literacy. Items for the questionnaire were identified following a review of the published research. Chinese studies that explored the knowledge attitude, and practices of participants related to infectious diseases prevention and control were reviewed for possible variables. In addition, variables from the National Health Literacy Survey [24–27] that measured residents’ level of basic knowledge, adoption of preventative measures and behaviors, or health skills regarding common communicable diseases were extracted. In order to identify the most contextually appropriate variables, only studies conducted in Mainland China were utilized.

The items that were chosen for inclusion were discussed and evaluated by a small group of professionals, including three clinical doctors and three public health experts from the field of infectious diseases treatment and prevention. The three clinical doctors worked in the infectious diseases department of an infectious diseases hospital, a tertiary hospital and a secondary hospital within the Beijing region respectively, and the three public health experts were from Beijing Center for Diseases Prevention and Control (CDC), a district CDC office, and a community health center. Following the panel discussion, a total of 25 items were chosen for inclusion. The final questionnaire consisted of three sections that measured: knowledge (8 items), adoption of preventative measures and behaviors (12 items), and health skills (5 items).

The final survey was available in Chinese and took five to ten minutes to be completed. Values were assigned to each of the questions: (1) one point for a correct or positive answer; and (2) zero points if the answer to the question was incorrect or negative. For the item “How many days did you do physical activities more than 30 min per day in the past one week?” one point was assigned if participants answered “≥3 days”.

Subjects and survey design

A multistage stratified sampling approach was used to recruit participants. In 2008, The National Health Literacy Survey showed the expected percentage of people with adequate health literacy was 21.82 % among city residents. The percentage was represented by symbol π and the numerical value was 0.2182 [17]. The sample size of each subgroup was calculated by the function n = μα2 × π(1-π)/δ2 × deff. The confidence interval (CI) was ±10 %, maximum permissible error (δ) =0.1π, μα = 1.96, and the design effect of complex sampling (deff) = 1.5, which was used to adjust the effectiveness loss of complex sampling instead of random sampling. Therefore the minimum sample size of each subgroup was 2,065. Considering the differences among different age groups (18–29, 30–39, 40–49, 50–59 and 60+), a total of five subgroups were identified. The total sample size of the five subgroups was expected to be 10,325 (2,065/subgroup × 5). Taking into consideration the response rates and efficiency rates of the questionnaire, the actual sample size was increased by 5 % to a total of 10,841.

There are 16 districts in Beijing, which are divided into urban and suburban districts based on population density. The population density was >6548 people per km2 in the urban districts and ≤ 1305.4 people per km2 in the suburban districts. Six districts including three urban districts and three suburban districts were chosen to be sampled. Five towns or streets were randomly selected in each district, and five resident committees or villages were randomly selected in each town or street. In total, 150 residents’ committees or villages were confirmed as survey locations. In every location, about 73 residents were randomly selected as subjects based on the name lists of residents acquired from residents’ committees. Participants were resident’s aged ≥18 years that were able and willing to give their informed consent to participate and who had continuously lived in Beijing for more than half a year. Participants were excluded if they were foreigners, residents of Hong Kong, Macau or Taiwan, or were unable to communicate in Mandarin.

Data collection

The survey was conducted between December 2010 and January 2011. Participants were asked to complete the questionnaire by themselves or with the help of trained study staff if they had difficulty with reading or writing.

Statistical analysis

The database was established in Epidata 3.1, and analyzed using SPSS 11.5. Based on the score distribution, the average score of participants whose percentages were the closest to the top 25 % and bottom 25 % seperately were treated as the two cut points. The discrimination index (D) was calculated by taking the average score of the top 23.8 % of participants (PH, score ≥ 19) minus that of the bottom 22.9 % (PL, score ≤ 12). Reliability was calculated by internal consistency measures, using Cronbach’s alpha formula. Inter-scale correlations were calculated by the means of Pearson’s correlation coeffcients. Confirmation factor analyses were conducted to verify the scale’s construct validity. Confirmatory factor analysis was implemented by the Amos 7. Cut-point of adequate health literacy was determined via receiver operating characteristic (ROC) analyses based on their educational level. Frequencies were calculated for categorical variables. One-way ANOVA was used to compare the differences of health literacy scores between subgroups. The Pearson χ2 test was used to compare the different proportions of respondents with CDHL in different groups. Multivariate unconditional logistic regression analysis was conducted to determine factors associated with the CDHL level. The variables with P < 0.10 in univariate analysis were included in multivariate analysis. Backward logistic regression was conducted by removing variables with P > 0.10. Statistical significance was defined as P < 0.05.

Ethics approval

The study protocol was approved by the Research Ethic Committee of Beijing Center for Disease Prevention and Control.

Results

Demographic characteristics

The questionnaires were distributed to13287 Beijing residents, and 11052 of them completed the survey, with a response rate of 83.2 %. The demographic characteristics of the residents is included in Table 1.

Table 1 Health literacy score and Percentage of adequate health literacy by Participant characteristics
Full size table

Reliability and validity of CDHL

The average percentage of correct answer was 61.3 %. The overall discrimination index was 0.428. Most of the items selected were answered with a moderate accuracy rate (average: 61.3 %) and with a good discrimination index (average: 0.428) (Table 2). Cronbach’s alpha was 0.748, indicating acceptable internal consistency for all of the items. Based on confirmatory factor analysis, the items were grouped into three subscales representing knowledge, adoption of preventative measures and behaviors, and health skills (GFI = 0.95, TLI = 0.81, AGFI = 0.94, RMSEA = 0.046). (GFI, goodness-of-fit index; TLI, Tucker-Lewis index; AGFI, adjusted goodness-of-fit index; RMSEA, root mean square error of approximation). For further testing of the dimensionality of CDHL, correlations of the subscales and the Total Score were examined by means of Pearson correlation coefficient. Correlations of the subscales and the Total Score is significant (P < 0.01), and all the three subscales correlate strongly with the Total Score (Table 3).

Table 2 Percentage of correct answer, discrimination index, and item-total correlation of health literacy items
Full size table
Table 3 Inter-Scale Correlations (Pearson’s correlation)
Full size table

ROC analysis

An ROC analysis was used to classify the levels of CDHL. The ROC curve is a graph of sensitivity (y-axis) versus 1 – specificity (x-axis). Maximizing sensitivity corresponds to some large y value on the ROC curve. Maximizing specificity corresponds to a small x value on the ROC curve. Thus a good first choice for a test cutoff value is that value which corresponds to a point on the ROC curve nearest to the upper left corner of the ROC graph. In this study, comparisons by education level indicated a cut-point of 16.5 for differentiating respondents with college education versus no college education (area under the curve =0.69, sensitivity = 0.61, specificity = 0.68). Based on the analysis, participants were classified into two groups: inadequate health literacy (score ≤ 16), and adequate health literacy (score ≥ 17).

CDHL level description

The mean CDHL score for participants was 15.28. The percentage of those who were identified as having adequate CDHL was 41 %. The percentage with adequate CDHL was significantly higher in female (42.4 %) than that in male (39.5 %) (P = 0.002). The percentage also increased with higher education levels. Participants aged between 30 and 39 years recorded the highest percentage of adequate CDHL (49 %), by contrast, the percentage of residents aged >60 was the lowest (29.8 %). Urban residents (48.3 %) were significantly more likely to have adequate health knowledge compared to participants residing in suburban areas (P < 0.001) (Table 1).

Factors associated with overall CDHL

The following factors were found to significantly impact the levels of CDHL according to the multiple logistic regression: gender, age-group, highest level of education, occupation, and self-reported health status and region. Females were more likely to have adequate CDHL (OR = 1.24, P < 0.001). By age group, residents aged 30–39 (OR = 1.39, P < 0.001), 40–49 (OR = 1.26, P = 0.001) and 50–59 (OR = 1.32, P < 0.001) possessed significantly higher level of CDHL when compared to the participant’s aged 18–29. Residents aged more than 60 years old also showed higher CDHL level, however, the difference was not statistical significant. Not surprising, the level of CDHL increased with increases in self-reported health status. Among different occupations, health workers possessed the highest CDHL level (OR = 3.13, P < 0.001), while manual workers have the lowest CDHL level (OR = 0.60, P = 0.005). The CDHL level lowered with the decreasing of self-reported health status. The residents lived in the suburb possessed significantly lower CDHL level than those who lived in the urban areas (Table 4).

Table 4 Multiple logistic regressions for the factors related to CDHL
Full size table

Discussion

Numerous studies examining the issue of health literacy have been conducted since the concept was first introduced [28–31]. In 2007, the National Health Literacy Survey was launched to investigate the levels of health related knowledge and behaviors of Chinese residents [32, 17]. In addition, Xinyin Sun undertook a study measuring the health literacy of infectious respiratory diseases in Beijing [33]. However, most studies conducted to date have failed to cover all three elements of health literacy. We are the first group to explore the issue of health literacy amongst residents of Beijing.

From our results, we found that the proportion of Beijing residents with adequate CDHL (49.0 %) was comparatively higher when compared to the national average of 13.64 % [17]. Rates were particularly low amongst participants: (1) residing in the suburbs; (2) aged 60 years and above; (3) who were mainly engaged in manual labor, and (4) who had completed less than five years of education.

Results from our study showed that gender, age-group, highest level of education, occupation, self-reported health status and region were factors related to the CDHL level significantly. Large-scale surveys of adult health literacy have previously reported associations between at-risk groups and low general literacy skills [34].

People with lower education level demonstrate lower health literacy in comparison with people with higher education level [35, 4, 36]. Low health literacy may be a barrier in access to health information and health care medication use and the prevention of disease [4]. Therefore, low health literacy has been associated with a range of poor health outcomes [37, 38, 35]. Consequently, improving the education level of residents is a key factor in improving CDHL. At the present time, multiple intervention strategies should be applied to residents of different education levels, including timely and accurate delivery of information to the public during a disease outbreak, and additional description of who is at risk, the nature of that risk, and what can be done to avoid exposure and manage illness [39].

The percentage of adequate CDHL of those who were over 60 years old was the lowest of all age subgroups in this study (Table 1). Several factors may affect CDHL levels in those over 60. The first is the decline of memory and verbal fluency, which are strongly associated with health literacy [40]. The second is that lower education levels in those over 60 years old might limit their abilities to obtain knowledge of CDHL. In a cross-sectional survey conducted by Li Wu in China, the proportion of the elderly who had six years or more of schooling was only 19.0 % [41]. The third factor is that the decline of cognition which has an obvious and direct impact on reading comprehension may impose restrictions on acquiring knowledge of communicable disease prevention and control from mass medias [42, 43].

The percentage of adequate CDHL of residents in suburbs was significantly lower than those of residents in urban areas (Table 1). This might be related to their economic status, the allocation of medical resources and the accessibility of CDHL information. Generally, the economic level and education resources in the suburbs were lower and less than in the urban areas in China. It is well established that poverty will directly affect peoples’ ability to make “good” decisions, including those related to their health [44]. Meanwhile less education resources will limit the education level of the local residents. To date, health education and promotion efforts have mainly been launched in urban areas. To change this situation, the publicity of control and prevention knowledge of communicable diseases should be given intensively in the suburbs considering the local epidemic characteristics.

Our study has a number of key strengths. Firstly, in order to improve the practical applicability and tolerance of the subjects, the questionnaire was professionally administered in five to ten minutes. It demonstrated good internal consistency and yielded comparable results. Secondly, the majority of previous investigations have usually focused on assessing the health literacy in limited fields [9, 33], or of certain populations [45, 46]. While our study involved a range of participants from different areas. The research findings were more representative of the overall CDHL in Beijing.

Along with the concept of eHealth literacy has been developed [47], some specific skills to seeking and understanding information of health are increasingly important. However, they may not be familiar to the elders. Consequently, some traditionnal channels, such as television programs, advice from doctors and suggestions from friends remain the main avenues for the elders to improve their health literacy [48], which limits them to acquire health related knowledge as easily as young people do.

Conclusions

Although the method adopted in our study might be different from other studies, it supplied a hint that the CDHL level of residents in Beijing was relatively low compared with those of other cities [49]. In view of the factors that lead to the low CDHL in elder people, measures aiming at improving their CDHL should be carried out. The health and healthcare related information of communicable diseases to the elder adults must be designed beyond vocabulary simplification and specifically to limit the demand on memory and verbal fluency more comprehensively. Meanwhile, public health workers need to consider the possibility of cognitive dysfunction in the elder adults, and adapt their information giving accordingly through multiple medium.