A Visualized and Scientometric Analysis of Health Literacy Research

Purpose The health literacy concerned by numerous policy and advocacy organizations is springing up rapidly and has evolved substantially over the past few decades. During the period of COVID-19, people who are in the absence of effective treatment and limited availability of vaccination need a higher health literacy to protective themselves. In order to provide a summary of the health literacy research, a visualized and Scientometric analysis is applied in this study. Methods Based on a scientific review of 3,670 data on health literacy from Web of Science, this research uses Citespace software to systematically and objectively describe health literacy to determine the knowledge evolution structure between articles and investigate research trends. Results The results show that the annual outputs of publications have grown rapidly since 2003. The USA and Australia make larger contributions compared with other countries in terms of quantity of publications and worldwide collaboration relationship between them. The “Soc Sci Med,” “J Health Commun.” “Patient Educ Couns” and “J Gen Intern Med” pay more attention to health literacy research. The leading authors with influence and authority are Wolf MS, Osborne RH, and Paasche-Orlow MK. Health literacy research in this field focuses on care, knowledge, and education, and so on. An emerging trend of health literacy with Covid-19, online tools, nursing, and obesity will spread across the globe. Conclusion Compared with simply reviewing existing articles, the major contribution in this study is a comprehensive review of yearly scientific output, journals, countries, institutions, contributors, highly cited papers, and keywords for health literacy research. The review also provides valuable and seminal guidelines for interested researchers on health literacy research.


INTRODUCTION
Along with changing demographics, significant advancement in science and technology, improvement of individuals' health beliefs as well as the complex demands of health in modern society, individuals and counties pay more attention to the health knowledge and skill to maintain or improve quality of life. Many countries or organizations, such as the EU (1, 2), American (3), and Australia (4,5) have enacted policies to promote health literacy. Amid the COVID-19 pandemic, dealing with an overabundance of information rife with misinformation and hoaxes requires adequate health literacy to improve public health worldwide (6). However, there are still challenges with low health literacy to embed health literacy principles into effective self-management of diseases. On these grounds, health literacy is a crucial factor associated with health care and has greatly attracted scholars' interest. Many definitions for health literacy have been proposed from different perspectives. For example, health literacy is the skill or ability to retrieve, understand, analyze and act on this information as defined by the American Medical Association Foundation (7). Nutbeam divided the evolving concept of health literacy into two different views, respectively, as a clinical "risk, " or a personal "asset" (8). Based on the above definitions, several relevant studies on health literacy from different disciplines have been carried out to enhance their ability to recognize disorders and seek effective treatments (6,(9)(10)(11)(12)(13)(14)(15)(16)(17)(18).
However, it is difficult for researchers to understand the current status and development trends from numerous qualitative and quantitative researches from a single perspective. For example, some studies concentrated on the assessment of health literacy (19)(20)(21). Some scholars investigate the relationship between health literacy and health behavior (6,12,15,16,22). In a word, a comprehensive, structured overview of health literacy is of great necessity. The literature reviews for a specific topic of health literacy are performed in terms of region difference [e.g. Malaysia (23), China (24), and Singapore (25)], age difference [e.g., young adults (26), older adults (27), and children (28)] or gender difference [e.g., women (24), men (29)] and so on. An exploratory bibliometric analysis of health literacy covering the period 1997-2007 is conducted by Bankson, but only including analyzing the characteristics of literature amount and core journals (30). Therefore, it is necessary to summarize previous research results and update the emerging trend and hotspots in health literacy research.
Given the incomplete and out-of-date literature research, this paper will use Scientometric and visualized analysis to demonstrate knowledge structures and developments in the field of health literacy from 2003 to 2021 with the help of Citespace.
The manuscript is organized as follows. In the next section, this study describes the materials and methods used to acquire the bibliometric analysis data. The following section contains the detailed results involving visualization and analysis on the distribution of scientific output, main source journals, countries, institutions, contributors, highly cited articles, and keywords. The final section includes the main findings, limitations, and future research.

Material Collection
The literature data in the areas of health literacy collected from the SCIE and SSCI citation index database in the Web of Science Core Collection are cover more scientific, comprehensive, accessible, and authoritative journals papers than other databases. Therefore, the data used in this study come from the index of Science Citation Index Expanded (SCI-EXPANDED), Social Sciences Citation Index (SSCI) of WOS Core Collection databases. A rapid upward growth pattern of published articles on health literacy after 2003 can be explained by the publicities undertaken by the Institute of Medicine, the American Medical Association, and the Medical Library Association (MLA) around this time (30,31). Therefore, this study downloaded health literacy-related multidisciplinary studies since 2003. The time period of these publications is from 2003 to 2021 for an integral vision of the research development status and the evolution experienced during the last two decades by the research field. In the final data set, document type only retains the journal paper. Based on the retrieval strategy described above, data containing the terms "health literacy" in the title are collected over a period of "all years" and 3,670 papers are available (as of 16 August 2021).

Methods
There are many Scientometric tools developed to identify and visualize hotspots, evolution, and emerging trends within a specific field, such as Citespace, VOS viewer, Bibliometrics, Sci2, SciMAT, SATI, and Bibexcel. Among these Scientometric tools, Citespace is regarded as the most popular software to perform data analysis in terms of data acquisition, data processing, visualization, and interpretation (32). First, Citespace has an advantage on the academic search of the Web of Science (WOS) core or other databases compare with Bibliometrix, Sci2, SciMAT, SATI, and Bibexcel. Second, some key functions and important metrics in CiteSpace can be used for more convenient and accurate data processing. It is beneficial to detect the research frontier in a particular field using the function of burst detection supported by Citespace to identify features with high intensity in a limited time (33). Third, a variety of visualization methods, especially in burst detection, time zone views, time distance measured by warm and cold colors make visual map a better readable and easier interpretation. Last but not the least, Citespace is widely used in Consumer Privacy Research (34), online learning (35), hospitality Research (36), intercultural competence (37), and other disciplines. Given the powerful function and advantages of Citespace, this study applies it to health literacy research.

Visualization and Analysis on Yearly Quantitative Distribution of Scientific Output
The number of published papers over time plots the annual changes in the attention paid by international experts and scholars to health literacy. Figure 1 indicates the total annual number of statistical results for 3,670 literature data related to health literacy in the WOS core collection, published from 2003 to 2021. A total of 3,670 publications on health literacy consists of articles and reviews, with an average annual publication of 193 papers. Based on the line chart of the number of annual publications, it can be seen that there is a continuous macro upward development trend and scholars' interests. In a word, the  The visualization of the journal co-citation network is plotted as shown in Figure 2; Table 1 below illustrates a list of the top 20 journals in the field of health literacy. The centrality and cited frequency of the literature published by a journal are the key indicators to evaluate the influence and importance of the core journal in a research field. The "Soc Sci Med, " "J Health Commun, " "Patient Educ Couns" and "J Gen Intern Med" play important roles in the evolution of health literacy according to a statistical analysis of centrality. The purple rings around the outer rim of some nodes of these academic journals indicate they have a closer interrelationship with other journals. From the perspective of the cited Frequency, the most highly cited journal related to  health literacy research is Gen Intern Med" with the frequency of 2,042. The "Patient Educ Couns" "J Health Commun" and "BMC Public Health" in Figure 2 are three highly cited journals that corresponded with the relatively large node. Among them, the other journals with high cited frequency such as "Jama-J Am Med Assoc, " "Health Promot Int, " "Soc Sci Med" and "Ann Intern Med" are also of great value for this area. In terms of the dual criteria of impact factor and JCR category, the most influential journals publishing health literacy research are the Lancet (IF

Visualization and Analysis on Main Countries and Their Cooperation
According to the cooperation network of analysis results between various countries, the publications on health literacy research come from 87 countries and territories, and the top−10 most influential countries and territories with the largest number of published papers are shown in Figure 3; Table 2. There is no doubt that the biggest node in the USA, which indicates the USA makes the largest contributions to health literacy research and ranks first in the number of publications with the frequency of 1,618. In terms of the centrality of published papers at each node, the USA also keep first place among the degree of communication between the connected countries/regions of publications from 2003 to 2021. Therefore, the USA is the most active and influential research country in the research field of health literacy research. Meanwhile, Australia, People R China, Canada, England, Germany, Netherlands, Taiwan (China), South Korea, Japan, and other countries or territories make decreasing contributions to health literacy research. The number of publications in Australia has published over 400 publications, ranking second, followed by People R China, with 180 papers published. Canada, England, Germany, and Netherlands published more than 100 but <170 papers. Taiwan (China), South Korea, Japan, and other countries are <100 papers. Most of the productive research countries are from North America and Europe. In a word, there is an actual uneven geographical distribution and the cooperative relationship with other countries of the number of articles published by each country or territory.

Visualization and Analysis on Main Institutes and Their Cooperation
The leading research institutes of health literacy research are presented in Figure 4; Table 3. In Figure 4, there are 394 institutions of publications. The thickness and quantity of the ties indicate the collaboration degree among institutions. In terms of Figure 4, the relationship in the cooperation between the key institutions focuses on the cooperation between domestic institutions, such as the strong tie between Univ N Carolina and Univ Illinois, the connection of Northwestern Univ and Emory Univ and linkages in the relationship between Univ Melbourne and Deakin Univ. The thicker the connecting line, the higher degree of link strength between the different institutions, and vice versa. Table 3 presents the top 10 institutions in the field of health literacy research. The top-10 core academic institutions in the collaboration network graph conclude seven American institutes and three Australian institutes. In terms of the node size and publications, Univ N Carolina (72) and Northwestern Univ (70) are the largest number of academic outputs among these institutes, followed by Univ Melbourne (54), Univ Sydney (54), Vanderbilt Univ (47), Univ Calif San Francisco (41), Emory Univ (41). The remaining institutes with more than 30 and <40 papers are expected to continue outstanding contributions to the field of health literacy research, such as Boston Univ (39), Univ Texas Austin (38), and Deakin Univ (36). In addition, Emory Univ, Vanderbilt Univ, and Northwestern Univ have a burst value over 6. In terms of the burst of publications in Table 3, which means that these universities have strong power in this field over a short period.

Visualization and Analysis on Main Contributors and Their Cooperation
The author collaboration networks of main contributors are plotted as Figure 5, with 699 authors as well as 1,162 collaboration ties described. Besides, the top 10 most influential authors from the output of CiteSpace are listed in Table 4.
Wolf MS ranks on the top of 10 highly cited scholars in the field of health literacy research. He comes from Northwestern University. His main research contributions are the relationship between health literacy and health management, health status, health outcomes, health risk behaviors, health disparities, health literacy skills, and health beliefs. The older people are the main data source collected in his health literacy research. Osborne RH, from Deakin University, ranks second. Prof. Osborne RH tends to develop health literacy scale tools, such as health literacy measure tool, health literacy assessment tool, and the health literacy management scale, to evaluate health outcomes, investigate health literacy needs, health behavior. In addition, he also pays more attention to the effect of electronic health literacy on health outcomes of socially disadvantaged groups, such as older adults, ethnic minority groups, low-income groups, low-literacy groups, and rural communities (31). Paasche-Orlow MK, working at Boston Univ, focuses on how to improve limited health literacy to reduce health disparities. He argues that the application of information technology in health literacy can reduce disparities by using intelligent, convenient, and actionable systems which provide information, advice, and behavioral support to low-literacy populations at the place and time it is needed (32). The other influential scholars on health literacy are Sorensen K, Kripalani S, Rowlands G, Schulz P J, Lee HY, Schillinger D, Rothman RL. These researchers make further contributions on definitions and conceptual frameworks of health literacy, Health Literacy Measurement, and health behavior, and so on.

Visualization and Analysis on the Highly Cited Articles
The advantage of highly cited papers for bibliometric analysis is very obvious, that is, the highly cited papers are regarded as one of the most important symbols to evaluate the influence of the general development trend and research frontiers (33). Based on the data, 3,674 papers, this study use Price's law to identify highly cited papers. M is computed to identify the lowest cited frequency of highly cited papers. The formula is as follows: Where n max is citation frequency of highly cited papers. By employing the formula, publications that were cited beyond 32 times were generally referred to as highly cited. Finally, a total of 637 highly cited articles is available.
The top 10 highly cited articles, including title, year, authors, journal, and citation, are shown in Figure 6; Table 5. Based on the abstract of these articles, the scope of these most influential papers in the field of health literacy focuses on low health literacy, health outcomes, the concept of health literacy, and eHealth literacy. Specifically, 3 out of 10 articles are review articles. has the highest citation frequency with 1,824 and suggests it to be the most important article. The number of another three articles also has been cited over 1,000 times, which highlights its importance in the field of health literacy.

Visualization and Analysis on Keywords Co-occurrence
The core and focus of research topics in a specific knowledge domain can be generalized by keywords. The keywords cooccurrence of health literacy are clearly illustrated in Figure 7 with Citespace and the top 20 keywords with high Frequency, centrality, and burst strength in health literacy are presented in Table 6. Every node in Figure 7 represents a keyword. The frequency occurrence of keywords can be measured by the size of the node, which is an important indicator to judge the importance of keywords. Otherwise, the keyword with a high centrality also is a symbol of the current research hotspots and emerging research trends. According to Figure 7; Table 6, the node of "health literacy" ranks first of all the keywords with the highest count of 1,955 and centrality of 0.23. Regardless of the keyword "health literacy, " the other keywords such as "Care, " "knowledge, " "Outcm (Outcome), " "Education, " "Communication, " "Adult, " "Impact, " "Association"  and "Behavior" at the Frequency range from 745 to 293. Further, the keywords such as "Risk, " "Intervention." "Breast cancer, " "Care, " "Adult, " "knowledge, " "Education, " "Association" and "Behavior" have a larger centrality compared with others, which indicated that they play a more active role in connectivity in the keywords network. Meanwhile, this study can judge the popularity of certain keywords in a short time according to burst terms. It could be seen that "literacy, " "Medicare enrollee, " "knowledge, " "hospital admission" and "readability" were the top five research frontiers of health literacy as shown in Table 6. Based on burst strength, the research frontiers of health literacy research can be concluded into the following three categories. First of all, a functional level of health literacy offered by Nutbeam remains a research hotspot (34). Reading ability is an important component of all basic skills of literacy. Beker et al. argued that there is a higher risk of hospital admission for individuals with inadequate or marginal functional health literacy (35). Secondly, data sources for many papers were mainly from Medicare enrollees or managed care enrollees aged 65 or older. Elderly individuals have worse knowledge and self-management skills which is suitable to investigate the relationship between low health literacy and health outcomes (36). Thirdly, the elder with a higher prevalence of chronic diseases, such as asthma, diabetes, congestive heart failure, hypertension, has a heavy burden of medical knowledge due to frequent use of health care services (37).
In addition, the time zone map shows the evolution of the prevalent frontiers of the core research topic over an approximately 20-year period as shown in Figure 8.

Main Finding
This study conducts a visualized and Scientometric analysis of health literacy research and the main findings are discussed in detail as follows: First, this study has been a scientific review of 3,670 publications and shows a significantly strong increasing trend of health literacy research during the period of 2003 to 2020. This is consistent with the prior result (30). In addition, this study forecast the publication amount will reach a higher level in 2021 because of the more extensive attention in the academic field.
Regarding the core journals, journal co-citation analysis identified the top 20 influential journals in the health literacy domain. It is clear that health literacy research attracts more and more attention because high-quality journals, such as Lancet As far as the main strength of health literacy research was concerned, the two most outstanding countries are USA and Australia in this area. Based on the publication amount of each institute, the top 10 institutions in the field of health literacy research all come from USA and Australia. In addition, better cross-national co-operations should be strengthened in the research process compared with strong cooperation between domestic institutions, according to the result of visualized and Scientometric analysis.
Through in-depth analysis of contributors, individual scholars with the most published articles are Wolf MS from Northwestern University, Osborne RH from Deakin University, and Paasche-Orlow MK from Boston Univ, who made significant contributions in this field.
Our results showed that the highly cited articles highlighted and synthesized some important issues of theoretical knowledge and empirical studies in the field of health literacy, including the relationship between low health literacy and health outcomes, the evolving concept of health literacy, the measurement of health literacy as well as patients' health literacy skills on accessing and effectively applying information technology (38).
Apart from the above discovery, this study use three indicators, namely keyword frequency, keyword centrality, and keyword burst, to investigate the research hotspots of health literacy. Some high-frequency keywords, such as health literacy, care, knowledge, outcome and education, can better show the current focus of the theme of change and evolution. The high-centrality keywords indicate the related terms, involving health literacy, risk, intervention, breast cancer, and care et al., are at the core of health literacy research. Based on the statistics of keywords with high bursts, this research find that kinds of enrollee, literacy, knowledge, and chronic disease are important directions in health literacy research. Linking this result with previous literature, a possible explanation is that Medicare managed care enrollees are mainly 65 or older with inadequate or marginal health literacy and chronic disease, who have urgent demand to increase health knowledge (35,39).
Last but not least, many scholars pay more attention to health literacy (12,(40)(41)(42), COVID-19 (6,16,17,(42)(43)(44), nursing (12,45) and obesity (46,47), just as time zone map has shown. For example, some scholars find that whether information technology in health care has the potential to improve the self-management level of patients with chronic diseases is associated with intended users' literacy (48,49). The massive open online courses (MOOCs) are beneficial to promote digital health literacy (50). During the COVID-19 pandemic, higher health literacy and eHealth literacy is beneficial for the public to prevent disease diffusion and promote both personal and organizational security (51). For the nursing setting, low health literacy harms health outcomes and the costs of the healthcare system (52). Many studies investigate the relationship between health literacy and obesity, arguing that populations with higher health literacy can increase people's control over their weight (46).

Limitation and Future Research
Inevitably, a few inherent limitations of this visualized and Scientometric study are listed as follows. First, the datasets this research use were published in English and other articles are excluded for reasons such as the usage of Korean. Therefore, this may affect the results of the analysis to some extent due to the lack of top journals from non-English speaking countries and the limit by the given data sources. Second, this study may underestimate the influence of the newly published articles. In future research, how to use other analysis tools is important to discover the potential value of the newly published articles. Third, this  study only integrates the data of Web of Science and ignores the data of PubMed. Because PubMed data is open access, it may be more advantageous to collect PubMed data and may highlight different pictures. Therefore, it is recommended that future scholars collect data from various other sources, especially PubMed, to further analyze this concept in more detail.

DATA AVAILABILITY STATEMENT
The original contributions presented in the study are included in the article/supplementary material, further inquiries can be directed to the corresponding author.