In this study, a cross-sectional online survey was conducted on the public platform created by Jiangsu Provincial People’s Government and managed by the Jiangsu Provincial Center for Disease Control and Prevention (CDC). The platform included Jiangsu health codes, through which every citizen had the obligation to fill in health information during the COVID-19 outbreak; otherwise, they were not permitted to enter/exit public places and their workplaces [16]. The platform included approximately 30,000,000 participants. The questionnaire used in this study has been embedded on the platform since February 15, 2020, participation was anonymous and voluntary, and participants had one chance to fill in their information. This study was approved by the Ethics Committee of the Jiangsu Provincial CDC.

Data were collected using an online questionnaire through WeChat (Multimedia Appendix 1). The questionnaire was created according to the national guideline for the diagnosis and treatment of COVID-19 and revised via expert evaluation. It includes items on personal information; protection knowledge, skills, and behaviors related to COVID-19; access to information; and current information needs. The Cronbach alpha coefficient and the Kaiser-Meyer-Olkin value for the behavior section was 0.723 and 0.838, respectively, indicating that the research data were relatively true and reliable.

Personal information included demographic data, such as sex, age, educational level, occupation, and place of residence. The knowledge section was composed of 7 single-choice questions and 3 true-or-false questions, which were scored 3 points each, including initial symptoms, distribution of death cases, transmission routes, conditions for killing viruses, mask selection, medical observation days, fever temperature, new coronavirus infection after influenza vaccination, selection of disinfection products, and infectiousness of asymptomatic individuals. The skill section consisted of 9 single-choice questions, scored 3 points each, including cough or sneeze treatment, home quarantine measures, measures from outside to inside (ie, measures implemented by individuals when they go back home from public places or workplaces), mask use, return notice (ie, matters that individuals should pay attention to when they come to or return to Jiangsu Province from other places), washing hands correctly, precautions, quarantine during travel, and attention to household alcohol disinfection. The behavior section comprised 11 scale questions, scored 0-3 points each, including no partying, wearing masks, wearing gloves, washing hands, no contact with live poultry, daily ventilation, weekly disinfection, distinction between the common cold and COVID-19, correct identification of epidemic information, workplace precautions, and community precautions. The highest possible score for each of these sections is 30, 27, and 33 points, respectively. The total score was calculated as follows:

Three methods were used to ensure data quality. Questionnaires filled out before 12 AM on February 15, 2020, were excluded, as the questionnaire was still in testing and was not officially published. Incomplete questionnaires were also excluded. Finally, questionnaires with irrelevant answers or obvious errors were excluded.

Frequencies, proportions, arithmetic means, and standard deviations were used to present the data. The chi-square test, the independent samples t test (two-tailed), and a one-way analysis of variance were conducted, as appropriate. A multivariate linear regression analysis was performed to identify the factors associated with the knowledge score, skill score, behavior score, and total score for COVID-19. Further, a binary logistic regression analysis was used to explain the selection differences under different characteristics for key items. Unstandardized regression coefficients (β) and odds ratios and their 95% CIs were used to explain associations between variables. The questionnaire data were exported to Microsoft Excel 2016 (Microsoft Corp) for data processing and analysis in combination with SPSS 26.0 (IBM Corp). P values of <.05 were considered statistically significant.

In view of the impact of epidemic changes on public perceptions, the time-varying reproduction number (R_t) over a 7-day moving average and its 95% credible interval were estimated in R version 4.0.0 (R Foundation for Statistical Computing) using the method developed by Thompson et al [17], and the serial interval derived from a previous epidemiological survey [18], in combination with the officially published epidemic data of Jiangsu Province, mainland China, and the entire world. Thereafter, the maximal information coefficient [19,20] was applied to test for correlations among the daily participation number, average protection score, number of confirmed cases, and R_t.

In total, 52,066 participants were included in the investigation of the status of protection against COVID-19 from February 15 to April 21, 2020, after excluding 344 respondents (including 47 test accounts, 228 with incomplete answers, and 69 with irrelevant answers) (Table 1). Of these, there were 30,212 (58.03%) men, and the male-to-female sex ratio was 1.38:1. The study population mostly comprised those aged 31-40 years (19,131/52,066, 36.74%), followed by those aged 21-30 years (14,226/52,066, 27.32%) and 41-50 years (9885/52,066, 18.99%). In terms of educational level, the proportion of “junior college and bachelor’s degree” was the largest, at 40.89% (21,291/52,066), and the proportion of “master’s degree and above” was the smallest, at only 4.69% (2441/52,066). For the occupational classifications, enterprises (18,187/52,066, 34.93%) and business and service industries (5906/52,066, 11.34%) accounted for a relatively large proportion. One-third of the participants lived in rural areas, while the other two-thirds lived in urban areas. The number of participants involved in the investigation each day is shown in Figure 1, with 3 peaks, which generally corresponded to the time for resuming work and study in batches in Jiangsu Province.

For the total protection score, 65.91% of participants (34,315/52,066) had scores over 90 points. The univariate analysis showed that there were significant differences in the knowledge score, skill score, behavior score, and total score for sex, age, educational level, occupation, and place of residence (all Ps<.001, except for sex and behavior score [P=.009]; Table 1).

In this study, access to personal protection knowledge on COVID-19 and information needs were also investigated. Television shows, government websites, and news outlets (46,145/52,066, 88.63%), as well as the government’s WeChat public account (45,657/52,066, 87.69%), were the main sources for acquiring personal protection knowledge. The sources for participants of different sexes and places of residence were similar. Participants of all ages rarely obtained protection information from microblogs or via communication among family, relatives, and friends. Those with an educational level of high school or less obtained information mostly from television shows, government websites, and news outlets, while those with an educational level of junior college or above obtained information from the government’s WeChat public account. Participants with occupations categorized under government agency, institution, and enterprise, as well as medical practitioners, obtained information more often from the government’s WeChat public account. Participants believed that the government’s media and WeChat public accounts (48,307/52,066, 92.78%) and authoritative medical experts (46,062/52,066, 88.47%) were the most reliable information sources (Table S4 in Multimedia Appendix 2).

The current information needs of the participants included the latest epidemic developments (46,729/52,066, 89.75%), disease treatment progress (42,181/52,066, 81.01%), and daily protection knowledge (41,451/52,066, 79.61%). Participants of different sexes had large differences in information needs in terms of disease treatment progress, prevention and control status in epidemic areas, and social dynamics; the differences in the other aspects were smaller. Participants of different ages, especially those aged 21-60 years, were very eager to understand the latest epidemic developments. The information needs of those with an educational level of junior high school or below concerned the latest epidemic developments and daily protection knowledge; conversely, the information needs of those with an educational level of high school or above were the latest epidemic developments and disease treatment progress. All participants paid less attention to material supply and social dynamics. Those with occupations categorized under government agency and institution, enterprise, business and service industry, freelancers, medical practitioners, and those who were unemployed had higher needs for epidemic developments and disease treatment progress, while farmers, students, and retirees had higher needs for epidemic developments and protection knowledge. Conversely, those living in rural areas were more interested in obtaining information on epidemic developments and daily protection knowledge than those living in urban areas (Table S5 in Multimedia Appendix 2).

We analyzed the R_t trends and attempted to determine associations among the daily participation number, average protection score, number of confirmed cases, and R_t values during the investigation. Owing to the implementation of strict control measures [21], the R_t in Jiangsu Province declined below 1, close to 0, and the R_t in mainland China also dropped significantly (Figures 3A and B). However, the number of confirmed cases worldwide has been increasing, with the global R_t showing a trend of first rising and then declining and maintaining a value around 1 (Figure 3C). The correlation analysis revealed that the daily participation number and average protection score were significantly associated with the number of confirmed cases and R_t in Jiangsu Province, mainland China, and the entire world (maximal information coefficient >0.70, range: 0.76-1.00; Figure 4).

Emerging infectious diseases are usually unpredictable with a lack of effective vaccines and drug treatments, and have direct or indirect negative impacts on economic development, social stability, and the public’s quality of life [22,23]. During the prior outbreaks of SARS and MERS, investigations were conducted to understand the public’s knowledge, attitudes, and practices [24-29]. Similarly, this study has shown that there was a relatively strong relationship between epidemic development and public perceptions.

A total of 52,066 participants were included in the study, of whom 65.91% had a total protection score above 90 points, indicating that the protection knowledge and skills were well understood and the actual action ability was strong. Unfortunately, there were still deficiencies in knowledge, skills, and actual behaviors in 34.09% of participants, and, hence, precise health education measures need to be provided.

For the knowledge section, the correct answer rates for initial symptoms, transmission routes, and selection of disinfection products were less than 80%; specifically, the rate for initial symptoms was only 62.01% (32,286/52,066), which was less than that of hospital visitors [14]. This may be because the outbreak time of COVID-19 overlaps with that of the common cold, flu, and other diseases, and these are also respiratory diseases, which have certain similarities in clinical manifestations. Similarly, owing to limited knowledge on COVID-19, which is currently an emerging infectious disease, individuals will tend to choose transmission routes and disinfection products that they consider reasonable. It is suggested that relevant departments should release authoritative health information on COVID-19 in a timely manner and strengthen the promotion and education of daily protection knowledge to meet the public’s needs.

For the skill section, the correct rate for home quarantine measures was relatively low; 23.10% (12,029/52,066) of the respondents were unaware that they cannot participate in family dinners. Yet, it is necessary to ensure that dishes and chopsticks are used and sterilized separately to avoid cross-infection. This may be attributed to the fact that although most quarantined individuals have no symptoms or mild symptoms, there is still a probability of presymptomatic transmission [30], which prompts relevant departments to provide health tips on epidemic prevention and control and consider such families as key monitoring objects.

For the actual behavior section, most individuals could reduce risk behaviors and take necessary protective measures, similar with previous study findings [11,13,31]. However, the proportion of those able to disinfect their homes weekly and distinguish the initial symptoms between the common cold and COVID-19 was lower than that of those presenting other behaviors, as in a previous study [14]. This may be related to the limited knowledge on COVID-19 and the lack of self-protection ability, indicating that public health information literacy needs to be improved. In the regression analysis, those with high educational levels were relatively unable to disinfect their homes weekly and clearly distinguish between the common cold and COVID-19. The reason may be that these individuals usually pay more attention to personal protection and that they are more cautious in answering the questions on the difference between the common cold and COVID-19.

Sex, age, educational level, occupation, and place of residence affected the total protection score for COVID-19 at different degrees. Women (average score of 90.58, SD 8.66) tended to have higher total protection scores than men, which is similar to the findings of previous investigations [13,31]. A large difference between men and women was observed in the skill scores—23.88 (SD 4.14) for men and 24.27 (SD 3.84) for women. Conversely, the total protection scores of those aged 21-60 years tended to be higher than the scores of those aged ≤20 years, which may be attributed to the current situation of resuming work and study; thus, these individuals need to actively obtain information on protection knowledge for COVID-19 and improve their self-protection ability. The total score was influenced by the educational level; the total score of those with higher educational levels tended to be higher than that of those with lower educational levels. Relevant studies also showed that individuals with higher educational levels were more willing to accept new knowledge and skills and adopt healthier practices [14,32,33]. In comparison with the unemployed, all participants, except the retirees and especially those with occupations categorized under government agency and institution, enterprise, business and service industry, medical practitioners, and students were more likely to have higher total scores, which may be attributed to their professional characteristics. Those living in urban areas had higher total protection scores than those living in rural areas. This may be attributed to the insufficient basic medical resources and relatively weak primary public health prevention strategies in rural areas. Further, the results may be associated with the relatively limited access to the internet and online health information resources [13]. Therefore, it is necessary to focus on the dissemination of health education knowledge on COVID-19 for men, those aged ≤20 years, those with an educational level of junior high school or less, those who are unemployed, and those living in rural areas.

This study showed that the government, television shows, and news outlets were the main sources for protection knowledge, which accounted for a higher proportion than that in a US sample [12]. Information released by the government and authoritative medical experts was also considered as reliable information, which is different from that reported in a previous study [31]. Moreover, this study emphasizes the need to continue to publicize the latest epidemic developments, disease treatment progress, daily protection knowledge, and other information on COVID-19.

There were several limitations in this study. First, the questionnaire used was designed based on a literature review and was used for the investigation after being revised via expert consultations. With the absence of a rigorous design process, the reliability of information may decline. Second, since the online surveys were conducted through WeChat and only in Jiangsu Province, the research samples are biased and limited. Given the differences in the resumption of work and study across different regions, our conclusions may change when expanding the study population. Lastly, the questionnaire only considered the influencing factors of the total protection score for COVID-19 from an individual perspective, without considering the influence of macro factors, such as government policies and society.

A high proportion of study participants had good protection knowledge and skills related to COVID-19. The factors influencing the total protection score for COVID-19 included sex, age, educational level, occupation, and place of residence. The study results suggest that relevant government departments need to update accurate information on COVID-19 in a timely manner, such as the latest epidemic developments and disease treatment progress, via official media and new media channels, and continue to promote daily protection knowledge on COVID-19. When resuming work and study, relevant departments need to apply different health education measures and conduct extensive and in-depth health education and promotion activities to guide the public, especially men, younger individuals, individuals with low educational levels, the unemployed, and individuals living in rural areas, to adopt positive and healthy behaviors. Doing so will ultimately reduce the negative impact of COVID-19.