Knowledge, attitudes, and practices of the general population about Coronavirus disease 2019 (COVID-19): a systematic review and meta-analysis with policy recommendations

Protocol and registration

We conducted a systematic review of the existing evidence related to KAPs of COVID-19 patients worldwide following Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) statements (Appendix Supplementary file 1).¹⁹ We also registered a protocol for this systematic review in the International Prospective Register of Systematic Reviews.²⁰

Eligibility criteria

We included all studies which met the following inclusion criteria: 1) cross-sectional survey; 2) investigate at least one component of the KAPs model regarding COVID-19 disease worldwide; 3) published or in-press original paper; 4) in English; 5) with a sample representative general population. No restrictions were applied to the setting, time, or quality of the study.

Information sources, search and study selection

We search the PubMed/LitCovid, Scopus, and Web of Sciences for papers in the English language only, up to 1 January 2021. We conducted a search in Google Scholar for retrieving studies that were not cited in the abovementioned databases. In addition, the reference lists of the final included articles were hand-searched. The keywords used in the search were attitude, knowledge, practice, awareness, perception, action, COVID-19, coronavirus disease, SARS-CoV-2, and severe acute respiratory syndrome coronavirus 2. The full search strategy for the PubMed database is provided in Supplementary file 2. When the search was complete, all records were transferred to the Endnote software (V. X8; Clarivate Analytics, Philadelphia, PA) and duplicates were removed. Then, studies based on the title, abstract, and full text were screened by two researchers independently by considering the prespecified eligibility criteria. Disagreements were solved through consultation with a third researcher.

Data collection process and data item

Two researchers independently engaged in the data collection process and extracted data including author, year, journal name, location, study design, data collection tools, sample size, focusing group, mean age or range, gender percent, and result related to KAPs model components. Potential disagreements were solved through consultation with a third researcher.

Quality appraisal

Included studies were critically appraised by two researchers independently. We used the Joanna Briggs Institute checklist developed for cross-sectional studies to appraise the quality of the included studies.²¹ This checklist contains eight simple and clear questions that cover topics such as inclusion criteria for sample; details about study subjects and setting; validity and reliability; criteria for measurement of the condition; confounding variables; and statistical analysis.²² The answer to each questions is yes, no, unclear, and not applicable. Potential discrepancies were resolved by consultation with a third researcher.

Synthesis of results

Descriptive analyses were carried out in most sections and the pooled data reported as a number or percentage for similar data items. We used Microsoft Excel software to design the charts. We categorized the result of each component based on the study by Bdair et al.²³ They categorized each component in two categories as follows: knowledge: (good ≥50) or (poor <50), attitude: (positive ≥50) or (negative <50), and practice: (satisfactory ≥50) or (unsatisfactory <50). The Q-value was applied to discover between-study heterogeneity, and the I² statistic was calculated to assess statistical heterogeneity.²⁴ Based on Cochrane criteria if the heterogeneity was ≥50, we used the random effect model.²⁵ Although there was heterogeneity between the studies above, this was negligible due to differences in settings as well as the use of different questionnaires. However, we used subgroup analysis based on regions to reduce this heterogeneity.²⁶ In addition, a meta-analysis using a random effect model with a 95% confidence interval (CI) was conducted via CMA software (Version 2) based on the percent reported for each component of the KAPs model of the included studies. Publication bias was assessed using Begg's and Egger's tests and visual inspection of the funnel plot.

Additional analysis

We contacted ten experts in the related field including health promotion, public health, health policy, epidemiology, and behavioral science via email and asked for their opinions on how to increase the levels of these components in the community. Comments were translated verbatim and then analyzed using content analysis. The results of this section are presented as policy recommendations.

Study selection

A total of 15,742 records were retrieved from our database search. After removing duplicate, 8270 records were screened by title, abstract, and full text based on eligibility criteria, of which forty-eight studies were included in the final review.^{23 ,} 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64, 65, 66, 67, 68, 69, 70, 71, 72, 73 The PRISMA flow diagram for the complete study selection process is presented in Fig. 1 .

Study characteristics

Forty-two studies encompassing 76,848 participants were included. In addition, 56.53% of the participants were female. The mean age of the participants was 33.7 years. Most studies were from Asia, Africa, and America, (Fig. 2 A). The most important method of data collection was online questionnaires (Fig. 2B). Most studies examined all three components of the KAPs model, but some studies examined two components or one component. More details about the characteristics of included studies are presented in Table 1 .

Quality appraisal

The overall mean quality score of the included studies was 5.70. Of the included studies, 41 studies (85.42%) were scored as good quality (score ≥6), 6 (12.50%) as fair quality (score 3–5), and remaining (2.08%) as low quality (score <3) (Fig. 3 ). The lowest and highest quality scores in the studies were two and six, respectively. None of the studies scored on questions 5 and 6, which were related to identification and deal with confounding variables in the studies (for more details about items see Appendix Supplementary file 3).

Synthesis of results

Among the included studies, 87.5% examined all three components of the KAPs model simultaneously. The most studied component in the studies was the knowledge component with about 100%, followed by attitude and practice with 95.8% and 91.6%, respectively (Table 2 , Fig. 4 ).

Of the studies that examined the knowledge component, 89.5% reported good knowledge, and 10.5% poor knowledge. As well as, of the studies that examined the attitude component, 100% reported a positive attitude. For the practice component, 93.2% reported satisfactory practice, and 6.8% unsatisfactory practice (Table 2, Fig. 5 ).

Meta-analysis

Based on the meta-analysis, the pooled overall score of KAPs components were 78.9 (95% CI: 96.1, 86.2, P = 0.001), 79.8 (95% CI: 80.8, 88.4, P = 0.001), and 74.1 (95% CI: 56.0, 86.5, P = 0.011), respectively. The results of subgroup analysis based on different continents of Africa, America, and Asia were 74.1, 74, and 83.8% for knowledge, 78.7, 63.2, and 85% for attitude, and 59.6, 78.5, and 81.5 for practice components, respectively. The Asia continent had the highest percentage in all three components. The America continent had the lowest percentage in terms of knowledge and attitude, and the Africa continent had the lowest percentage in terms of practice (Table 3 ). Visual inspection of the funnel plot and results of Begg's (0.068) and Egger's test (0.082) did not showed significant evidence of publication bias (Appendix Supplementary file 4).

Policy recommendations

In accordance with experts, the policy recommendations for promoting the KAP components were as follow: holding training courses through virtual mass media; increase the commitment of government officials and policymakers to help conduct training courses; providing appropriate and evidence-based training content to enhance the components of the KAP; designing an integrated international system for measuring cup levels and comparing it between countries.

Strength and weakness

One of the most important strengths of this study was that all stages of the study were conducted with two researchers and in all stages, in cases of disagreement, the third person and consensus were used. In addition, registering the protocol of this study and reviewing and modifying it in the PROSPERO platform is the strength of this study. A large number of the included studies did not report the validity and reliability of the questionnaires. The main reason for this is the rush to publish articles related to coronavirus disease. The included studies were from both high- and low-income countries and therefore generalization of results to all countries should be done with caution. On the other hand, owing to the high speed of publication of articles in this field, some other studies may be published at the time of writing the article and the review process, which has been missed. Of course, owing to the high speed of publishing articles, this limitation is inevitable.

Summary of study findings

We found that about 90% of the samples had good knowledge of COVID-19 (overall score: 78.9%). In addition, 100% of the samples were reported positive attitudes regarding COVID-19 (overall score: 79.8%) and slightly more than 93% of samples performed satisfactory practices (overall score: 74.1%). The level of knowledge, attitudes, and practices is slightly high in the Asia continent. About 90% of the studies used an online questionnaire to collect data, and the most used platforms included Google form, SurveyMonkey, and Qualtrics. The most important social media through which the questionnaires were distributed were Facebook, WhatsApp, and Telegram. The most important sources for learning and staying up to date about COVID-19 mentioned in the studies were television, social media, the internet, radio, and friend and relatives.

Our result showed a high percentage of knowledge, attitudes, and practices in Asian countries. The probable reason for these higher percentages could be related to the factors such as the initial spread of the virus from this continent and the emergency acts that were taken earlier than other continents in this continent.^{78 , 79} However, owing to the lack of studies in developed countries and the change of some factors related to knowledge, attitude, and practice over the past year, the generalizability of these results is low.

The finding of our systematic review demonstrated good knowledge about COVID-19. In most studies, more than 80% of the participants had a good knowledge of issues such as causes, symptoms, ways of transmission, and ways of prevention. In addition, most participants had a high level of knowledge about symptoms such as high fever and dry cough, breathing difficulty and a small number had sufficient knowledge about other symptoms such as chills, headache, muscle pain, sore throat, and loss of taste or smell.^{28 , 33 , 34 , 41 , 49 , 50 , 57 , 63} More than 90% of the participants considered air droplets as a way to spread. This good level of knowledge can be due to widespread information through various means such as public media (television and radio), social media, and government announcements. In addition, preparing several guidelines and reports by WHO, CDC, and local government in times of outbreak and easy access to them have increased the level of information and knowledge of individuals regarding COVID-19.^{28 , 29 , 37 , 45 , 50 , 53 , 60 , 69} On the other hand, factors such as low literacy level, older age, and the presence of the rural population in the samples were among the factors that have reduced the level of knowledge in the studies.^{31 , 35 , 64}

In this review, participants showed a positive attitude regarding COVID-19. Almost all participants believed in the importance of handwashing, disinfecting surfaces, using masks to prevent the spread of infection, resting at home in the event of symptoms, and maintaining social distance and limited contact. Of course, in some cases, there was a negative belief that it could be due to differences in instructions and guidelines by different institutions, such as what was about wearing a face mask at the beginning of the pandemic, and then it was recommended that the whole population should use a mask.^{23 , 34 , 38 , 41 , 80 , 81} Such cases show the importance of integrated guidelines and the focus of decision-making in times of crisis.^{39 ,} 82, 83, 84, 85 Although having a responsible organization can help make better and faster decisions, in such cases, political pressure is exerted by governments that such organizations should put the health of the people at the top and not refuse to make the right decisions due to political pressures.^{11 , 31 ,} 86, 87, 88

In general, the level of practice of the participants in the studies was satisfactory. However, despite the good knowledge and positive attitude of the participants, the level of practice was still sometimes lower than expected. Numerous reasons for unsatisfactory practices have been cited in studies. Lack of availability (for example, masks and disinfectants), imposing financial costs on participants, ambiguity in instructions, not getting used to new conditions such as staying home and wearing a mask, exhaustion from existing conditions, and anxiety and stress of disease were among the causes mentioned in the studies.^{41 , 56 , 73 ,} 89, 90, 91 In this regard, some countries have imposed strict laws and penalties on people who do not follow the guidelines to improve their performance, but in many countries under study, such laws do not exist and have not been applied.^{38 , 50 , 61 , 92 , 93} Another factor that affects the performance of individuals was the presence of decision-makers in public and social media. Seeing a person without a mask at the height of a pandemic hurt a person's good practices.

Given the diversity of settings and questionnaires, the authors of this article recommend that there be a need to design an integrated online system to assess the knowledge, attitudes, and practices of the population about health-related crises. Designing such an integrated system can help better compare countries because integrated items are used for comparison. On the other hand, designing such a system and disseminating its results can accelerate integrated decision-making and improve crisis management. On the other hand, the existence of such an integrated system can lead to an increase in solidarity, which was emphasized by the World Health Organization during the corona pandemic.^{94 , 95}

Acknowledgements

The authors thank the PROSPERO institute for accelerating the review process in the time of Coronavirus. They also thank Birjand University of Medical Sciences for approving our proposal and giving it a code of ethics (IR.BUMS.REC.1399.099).

Ethical approval

None sought.

Funding

Not applicable.

Competing interest

The authors have declared that no competing interests exist.

Author contributions

MA-Z contributed to conception and design. MM, SS, and SH contributed to screen the records, data extraction, and quality appraisal. MA-Z and HA contributed to data analysis. MA-Z contributed to draft manuscript. SH and HA contributed to critical review. All authors approved the final version of the manuscript for publication.

Data availability statement

All relevant data are with the article and the attached supplementary information.