Knowledge of COVID-19 and Health Literacy Among Patients Seeking Care in an Emergency Department

ways To examine

average COVID-19 knowledge score of 7.78. The BHLS and COVID-19 knowledge scores were positively correlated in both bivariate and multivariate analyses. Education and income levels were statistically signi cant in the multivariate regression.

Conclusions
To better prevent further increases in COVID-19 transmission, community-based interventions can be more cost-effective when targeting sociodemographic groups that have lower general health literacy. In particular, individuals of low educational and with low income levels should be prioritized. Background COVID-19 has changed Americans' health behavior in a dramatic way. New scienti c discoveries, along with an abundance of information and misinformation have created confusion in public understanding of COVID-19. For example, the CDC did not encourage wearing face masks at the start of the pandemic but reversed that decision in April by stating masks do have some level of bene t and should be worn by the public. 1,2 It is well established that a lower level of health literacy is associated with worse health outcomes. [3][4][5] One particularly important role of health literacy is to improve patients' ability to understand the prevention, treatment, and prognosis of infectious diseases which leads to more effective personal prevention strategies and coping mechanisms. A study from 2016 showed that a low health literacy with infectious diseases, such as tuberculosis, malaria and in uenza, was associated with decreased protective behaviors and understanding of antibiotic usage. 6 The same study also demonstrated that patients with a lower health literacy score were less likely to be currently vaccinated and less likely to receive future vaccine as compared to patients with a higher health literacy score. Additionally, in a study focused on outpatient antibiotics prescribed in emergency departments (EDs) , patients with low health literacy had a decreased number of prescriptions lled at 3 days. 7 As the US continues to experience a daily increase of COVID-19 cases, there is an urgent need to identify ways to improve individuals' knowledge of COVID-19 to achieve effective prevention. If a better COVID-19 knowledge is associated with an overall higher level of health literacy, then community interventions to prevent COVID-19 transmission among low health literacy individuals could be much more cost-effective than targeting the general population. This is particularly relevant to communities with limited resources.
Because the current understanding of COVID-19 is very limited, there are no studies that have addressed this important question. The primary objective of the current study was to examine whether better knowledge of COVID-19 was associated with patients' overall health literacy among patients seeking care in an ED. The secondary objective was to demonstrate whether patients' primary source of COVID-19 information was associated with their COVID-19 knowledge.

Methods
A convenience sample was collected in an urban ED of an annual average volume of 80k. Approval from the local IRB was obtained prior to the start of the study. The duration of the study was a 6-week period from 05/25/2020 to 07/05/2020. Written consents were obtained. A total of 252 participants completed the study, exceeding the minimal sample size of 220 calculated by the initial power analysis using a beta of 99%. A commercial online platform, Qualtrics XM , was used for data collection from a survey. The inclusion criteria were 1) currently seeking ED care; 2) 18-89 years of age; and 3) Glasgow Coma Scale (GCS) of 15. Patients aged 17 years or younger, psychiatric patients and trauma activations were excluded. Patients in ED rooms who met the inclusion criteria were briefed with a verbal description of the study and the study yer was given. Participants could choose between completing the survey online on their phone or computer at any time during the study period and completing the survey in the room while waiting for care to be completed with the assistance of a data collector who helped the patient utilize a tablet device.
Patients' knowledge of COVID-19 was examined by answers to whether they agreed to a mix of 10 correct and incorrect statements about COVID-19 that included basic epidemiology, prevention, diagnosis, treatment and prognosis. The total number of correct answers was used as the COVID-19 knowledge score.
2. You can get COVID-19 through contact with an infected person.
3. Face masks, and frequent hand washing or use of hand sanitizer can prevent getting COVID-19 4. You don't have COVID-19 if you don't have any symptoms. 5. Fever, cough, and shortness breath are the most likely symptoms of COVID-19.
6. People with chronic health problems, such as high blood pressure, diabetes, heart disease and kidney disease, are more likely to get COVID-19 and die.
8. COVID-19 vaccine is available now in the US.
9. More than half of people who had COVID-19 died.
10. COVID-19 is more deadly among young persons and children.
Health literacy was measured by the Brief Health Literacy Screen (BHLS) that has been demonstrated to have good concordance with several commonly used instruments measuring health literacy, REALM, Sand TOFHLA. 8-10 BHLS has been tested in a wide variety of clinical settings including EDs, and can be self-administered. [11][12][13] The BHLS uses three questions with each question scored on a 5-point Likert scale: always (1), often (2), sometimes (3), rarely (4) and never (5). The summation represents the nal BHLS score.
1. How often do you have problems learning about your medical condition because of di culty understanding written information?
2. How con dent are you lling out medical forms by yourself?
3. How often do you have someone help you read hospital materials?
Because there were diverse sources of information about COVID-19, participants were asked about their primary source of COVID-19 information. The replies were grouped into 4 categories: social media and social circle of family and friends, internet sites and searches, 3 rd -party reports (TVs, radios and newspapers), and scienti c sources (CDC and professional journals).
Patients' sociodemographic information was also collected: age, gender, race, education level, employment status, household income and primary language spoken at home. The frequency of ED visits in 2019 and the frequency of outpatient clinic visits in 2019 were collected to gauge participants' familiarity and interaction with the healthcare system at the baseline prior to the COVID-19 pandemic, whether they had a primary care provider (PCP), and whether they had any comorbidities (hypertension, hyperlipidemia, diabetes, heart diseases, chronic kidney disease, stroke and liver diseases).
Statistic software Stata (StataCorp, College Station, TX) was used for analyses. Distributions of the COVID-19 knowledge and BHLS scores were rst examined. Descriptive statistics of the primary source of COVID-19 information, participants' sociodemographics and familiarity and interaction with the healthcare system were reported. Bivariate analysis between the COVID-19 knowledge score and each independent variable was conducted. A multivariate regression was conducted to examine the independent associations of BHLS and the primary source of COVID-19 information and the COVID-19 knowledge score, respectively, controlling for the confounding of patients' sociodemographics and familiarity and interaction with the healthcare system. The results from the multivariate analysis are shown in Table 2. Even after controlling for the confounding variables, BHLS was still found to be statistically signi cant (p=0.02). However, the primary source of COVID-19 information was no longer signi cant. In contrast to the results from bivariate analyses, participants' employment status, race and whether patients had a primary care provider were no longer statistically signi cant after controlling for confounding variables. Compared to patients who did not graduate high school, those with an education level higher than high school had a signi cantly higher COVID-19 knowledge score by 1.61 (p<0.01). No signi cant difference was found between high school graduates and those who did not graduate high school. Lastly, lower income was associated with a lower COVID-19 knowledge score by 0.67 (p=0.02).

Discussion
There are several limitations in the current study. First, the convenience sample consists of only ED patients. The results may not be representative of the general population who do not seek ED care. A second concern is that despite the pertinence and accuracy of BHLS, further research using other measurements of health literacy is warranted. Third, this is a pilot study based on a relatively small sample size, although the sample size was larger than the minimum obtained from the power analysis. Certain trends, like the relationship between the primary language spoken at home, or race with COVID-19 knowledge, could have shown to be statistically signi cant if a larger sample had been obtained. Lastly, the data collected were from a single urban tertiary hospital in a city with about 1/3 of the population being Hispanic. The generalizability of the conclusions to other areas of the US could be limited.
This study found that even after controlling for patients' characteristics, patients with better general health literacy had better knowledge of basic epidemiology, prevention, diagnosis, treatment, and prognosis of COVID-19. This is in line with ndings from studies of other infectious diseases that patients with lower health literacy had worse knowledge of antibiotics, decreased immunization rate and health screenings. 4,6,14 Reading comprehension, education level, English pro ciency, and cultural differences were identi ed to be common additional contributing factor to health literacy. 15 Despite these similarities, the ways the public acquire the knowledge of COVID-19 are distinctly different from those for other infectious diseases in the past, as COVID-19 is presented in a completely different epidemiological, cultural, technological, and even political context. The public have been exposed to some aspects of COVID-19 daily, as information and misinformation about COVID-19 permeate all media sources. Additionally, the drastic change in society with a near-nationwide shutdown and all individual lives being disrupted to some level created a uni ed front combating the transmission of COVID-19, making the efforts more focused than those for other infectious diseases in recent US history.
Consequently, it is not surprising to nd that overall, the participants of the current study had a good knowledge of COVID-19.
Furthermore, even after controlling for confounding variables, this study found that ED patients with low income and those with lower education had a lower level of COVID-19 knowledge. Similar ndings were demonstrated in an outpatient setting. In a recent study of adult outpatient clinic patients with at least 1 chronic condition, researchers found that blacks, the poor and those with low health literacy were less worried about COVID-19, less likely to believe that they would become infected and felt less prepared for an outbreak. 16 It is very concerning that recent studies, however limited in number, identi ed that the same subpopulations had higher mortality and hospitalization rates of COVID-19. 17,18 This is consistent with prior studies that demonstrated lower education and income resulted in poor outcomes in other diseases. 15,[19][20][21] It is likely that the worse health outcomes of these disadvantaged subpopulations are the result of the lack of effective COVID-19 prevention and coping strategies stemming from the inadequate COVID-19 knowledge. This is further complicated by the already existing barriers to medical care for these patients. Therefore, identifying ways to improve the COVID-19 knowledge within these subpopulations should become an integral part of any community-based interventions to better prepare the public for the pandemic and decrease the health disparities.
Various interventions have been shown to be effective for other diseases in the past, such as simpli ed wording during media presentations, numerical charts, addition of images and increased funding to media sources. [22][23][24]   BHLS Score (N=252)