Knowledge and Perception of COVID-19 Pandemic During the First Wave (Feb-May 2020): a Web Based Survey Among Italian Healthcare Workers

Italy was the rst country in Europe to face the coronavirus pandemic. Healthcare workers (HCWs) were at higher risk of contracting COVID-19, because of their closer contact with patients. The present study aimed to analyze HCWs’ knowledge, practices, and attitude towards COVID-19. We set up a cross-sectional survey through SurveyMonkey® and circulated the link in Facebook and Whatsapp closed groups. It was conducted in Italy from the end of February to the rst week of May 2020. The study participants were general practitioners, pediatricians and other health professionals. Data were collected using a well-structured questionnaire including demographic, scope of knowledge, awareness and practice assessment section. and need extensive knowledge and awareness of the disease to take adequate precautionary measures and they are crucial to disseminate good practices.


Introduction
On December 31, 2019, the World Health Organization (WHO) received reports from the Chinese health authorities about the presence of pneumonia cases of unknown cause detected in the city of Wuhan, in the Chinese province of Hubei [1]. Subsequently, the Chinese health authorities identi ed a new Coronavirus, Acute Severe Respiratory Syndrome -Coronavirus 2 (SARS-CoV-2) as responsible for the CoronaVirus Disease 2019 (COVID-19) [2].
The rst two cases of the COVID-19 pandemic in Italy, which tested positive for the SARS-CoV-2 Virus in Rome, were con rmed on January 30 th , both with a travel history to Wuhan, China. On February 21 st , 2020, the Italian National Institute of Health con rmed the rst autochthonous case in Northern Italy (Codogno city-Lombardy region) in a critically ill, hospitalized young man with no travel history to known areas of viral circulation or links to a probable or con rmed COVID-19 case [3][4][5].
This unexpected nding unveiled ongoing transmission in several municipalities in the Lombardy Region [6]. In subsequent days and weeks, case counts and death tolls increased rapidly, rst in Northern Italy, and then in the rest of the country. The Italian government imposed increasingly strict physical distancing measures, starting with the closure of 10 municipalities in the Lodi Province (Lombardy) and one in the Padua Province (Veneto) on the 22 nd of February 2020 [5;7]. This culminated in a national lockdown declared on March 10 th , 2020 and ended on May 3 rd , 2020. COVID-19 is spread by human-to-human transmission through droplets and direct contact; it has an incubation period of 2-14 days [8][9][10][11][12][13].
The clinical presentation of the Sars-CoV-2 infection varies from asymptomatic to very severe pneumonia with acute respiratory distress syndrome, septic shock and multi-organ failure, which can cause death [14][15][16].To date, applying preventive measures to control COVID-19 infection is the most critical intervention.
Healthcare workers (HCWs) are at high risk of infection and they may also contribute to the spread of the disease. Infection prevention and control (IPC) during health care practices when COVID-19 is suspected or con rmed is crucial in order to protect HCWs and fragile patients [17]. The rate of infection reported in HCWs varies across countries [18][19][20][21]. In China, HCWs accounted for 3.8% of all cases, with 14.8% of these having severe/critical disease despite their young age and few comorbidities [22][23]. Other studies however, reported lower illness severity in HCWs, and identi ed PPE use as the main factor associated with decreased infection risk [24].
Measuring scope of knowledge, attitudes and practices in regards to IPC helps to predict HCWs' behaviors in applying preventive and control measures.
Even though the Italian Ministry of Health published guidelines and developed strong initiatives for IPC in healthcare facilities to prevent the spread of the disease, it is crucial to understand if those guidelines were really applied, particularly among HCWs [25].
In order to investigate the knowledge and the practice of HCWs in applying SARS-CoV-2 nonpharmaceutical preventive measures, we conducted a cross-sectional survey at the very beginning of the autochthonous circulation of SARS-CoV-2 in Italy.

Study design
This cross-sectional survey was coordinated by Bambino Gesù Children's Hospital and conducted in Italy from the end of February to the rst week of May 2020.
The survey was set up using SurveyMonkey® and the link to the survey was circulated online through Facebook and Whatsapp closed groups. The study participants were HCWs, including general practitioners, paediatricians, consultants, postgraduate trainees, and other health professionals (nurses, midwives, physiotherapists etc.).
Data was collected using a structured questionnaire that comprised 31 prede ned responses, including the demographic, scope of knowledge (K), awareness (A) and practice (P) assessment sections.
The rst section of the questionnaire consisted of 6 questions regarding demographic details and professional pro le. Section two consisted of 2 questions about the risk perception of HCWs and patients. Section three consisted of 7 questions focusing on the knowledge level of HCWs. Section four had 15 questions regarding attitudes and practices, precautions and procedures to contain the virus.

Sample size
The sample size for the survey was calculated according to the formula adopted in the Raosoft software (http://www.raosoft.com/samplesize.html). Setting the expected proportion of the outcome found in each question of the study at 50% with an accepted margin of error of 5%, we obtained a total sample of 377 individuals, with a con dence level of 95%.

De nitions
Knowledge was de ned as at least an a rmative response to one of the following four questions: -Do you believe that information released by international health authorities regarding the epidemic from COVID-19 in China has been clear enough?
-Do you believe that information disseminated by national and regional health authorities regarding the risks associated with COVID-19 for the Italian population has been su ciently clear?
-Do you think that the de nition of a suspected case of COVID-19 infection is su ciently clear? -Do you think you have been su ciently informed by the national health authorities on how to behave if you are faced with a suspected case of COVID19?
Behavior change was de ned as at least an a rmative response to one of the following three questions: -Since the start of the COVID-19 epidemic, have you changed the way you work? -Since the start of the COVID-19 epidemic, has there been any impact in the organization of visits? -Since the beginning of the COVID-19 epidemic, has there been any impact in your relationship with patients?
Questions on the risk perception for being in contact with COVID-19 for HCWs and their patients were measured using the Likert scale ranging from 0 to 10 (no risk and high risk respectively); the answers were then categorized into ve groups according to the percentile distribution to better show the results in the graph. We considered the pre-lockdown period from February 26 th to March 10 th , 2020 and the lockdown period from March 12 th to May 3 rd , 2020 [26].
Statistical analysis Univariate differences were tested using the Chi-squared test for categorical variables and t test for independent samples for continuous variables. We carried out multivariable ordered logistic regression to investigate the association between the socioeconomic characteristics, the variables investigated in the questionnaire and two outcomes (knowledge and behavior change).
We carried out multiple imputation with chained equations [27] to generate values for missing data points such as sex, age, region and attitude to face the epidemic. All variables included in the models as predictors of outcomes were used to predict missing values [27][28]. Data were assumed to be "missing at random" [27]. Twenty-ve datasets were imputed. Outcomes were not imputed. Data analysis was performed with STATA 13.0 SE (Stata Corporation, College Station, Texas).
At the end of the questionnaire we asked for open comments. We obtained 182 comments. We used the Word Cloud (WC) to visualize the frequencies of keywords. A WC is generated by counting the frequency at which each word appears. WC generation was performed using R packages tm and word cloud. Prepositions like 'for', 'or' and 'in' were excluded from the WC. Combination word concepts such as 'general practitioner', 'front line', 'I would like to', 'mildly symptomatic', 'too much', 'public health', 'Local Health Authority', 'National Health System' and 'Health Protection Agency' were all entered as single words. We showed the words with a frequency greater than or equal to 3.

Results
A total of 958 participants were included. Of these, 320 (33.4%) were general practitioners, 248 (25.9%) were pediatricians and 390 (40.7%) were other health professionals. Most responders lled in the questionnaire during the pre-lockdown period (72.4%). The characteristics of the participants are shown in Table 1. Medical doctors were older than other health professionals, with a mean age of 56.8 versus 38.2 years, respectively. The majority of respondents were female (61.8%). Patients were reportedly more afraid of COVID-19 during the lockdown period (15.8%) than in the prelockdown period (8.4%) (p < 0.001) ( Table 2). Health professionals reported a higher level of perceived risk of contracting COVID-19 from their patients than pediatricians (Fig. 1). According to respondents, the risk of having contact with a patient affected by COVID-19 was far higher in the lockdown period than the pre-lockdown period (19.4% vs 10.2%, p = 0.002) ( Table 2). This risk was higher for general practitioners than for pediatricians and other health professionals, but this difference was not statistically signi cant (Fig. 2).
Respondents reported having changed their clinical practice, more in the lockdown period (81.1%) than in pre-lockdown (46.3%) ( emergency also reported less knowledge (OR 0.13, p < 0.001) ( Table 3). The multivariate model analyzing behavior change showed that general practitioners changed their behavior less than health professionals (OR 0.54, p = 0.008). Participants from Southern Italy changed their behavior less than those from Northern Italy (OR 0.67, p = 0.049). Age positively affected behavior change, and respondents ranging from 46 to 55 years (OR = 1.81, p = 0.024) of age showed more willingness to change their behavior, as did those aged 56-65 (OR = 1.81, p = 0.023) ( Table 3).
Respondents declared to have changed their behavior more during the lockdown period than prelockdown (OR 6.22, p < 0.001). Moreover, those who used the ministry toll-free number to inform themselves, reported the greatest behavior change (OR 2.03, p = 0.001).

Discussion
COVID-19 is a global health problem, especially among HCWs. Italy was the rst European country to face COVID-19, with considerable differences in terms of organization and management strategies throughout the country, resulting in heterogeneous levels of performance across regional health systems.
Healthcare professionals, being in contact with patients, play a crucial role in the transmission of COVID-19; thousands of HCW's, mainly general practitioners, were affected by COVID-19 and died while caring for COVID-19 positive patients [26]. This could be due to several factors: lack of personal protective equipment (PPE), poor knowledge of the virus containment measures, especially in the rst pandemic wave, and heavy workloads. For this reason, we investigated HCWs' knowledge and perceptions of the prevention and control of COVID-19 pandemic.
It is crucial for HCWs to be prepared and to apply all IPC in facing COVID-19 [29][30][31] considering that the prevalence of the infection among HCWs exceeded 10% in Italy [32][33][34] with a consequent loss of capacity for hospitals to respond adequately to pandemic.
Knowledge and perceptions of COVID-19 varied across different categories of HCWs.
Other studies have shown that the majority of HCWs had good knowledge on COVID-19 and showed a positive attitude related to their sense of readiness to confront the disease and implemented good practices towards COVID-19 [30,[35][36][37].
Our study shows that HCWs have a su cient level of knowledge about COVID-19. Moreover, participants frequently reported a change in their behavior in clinical practice during the pandemic. Doctors had a higher level of knowledge and, amongst those, pediatricians were those better informed.
Moreover, doctors declared they had signi cantly modi ed their clinical practice during the pandemic period compared to other health professionals, showing a higher impact of the epidemic on medical doctor's daily routine.
Doctors were also much more con dent in the information coming from the Italian National Health Authorities compared to other health professionals. This indicates that the COVID-19-related updates posted by o cial health authorities had positive implications for improving doctors' knowledge levels.
Obtaining information from institutional sources is crucial for disseminating reliable data about the emerging COVID-19 infection and is essential for HCWs' preparedness and response.
During the pre-lockdown and lockdown periods, all health professionals were informed about the best IPC to be adopted for the containment of COVID-19. Most respondents received their information from institutional channels, while 38,5% had obtained information from other colleagues and 15,8% from social networks.
The ndings of this study suggest a signi cant gap between the amount of information available on COVID-19 and the depth of knowledge among HCWs, particularly regarding disinfection of doctors' o ces and contact surfaces, use of protective glasses and use of alcoholic solution for hand hygiene. All these practices were not extensively applied in the pre-lockdown period, also because they were not routinely used by HCWs before the pandemic; yet their use signi cantly increased during lockdown. This could be due to a low penetration of information and trust in the messages of the health authorities at the beginning of the pandemic, which were perceived only later as important and vital to ght the pandemic.
However, it is important to note that, as expected, the change in IPC measures was mainly driven by those measures less used in the clinical routine (such as use of protective glasses, and use of alcoholic solution for hand hygiene), while well known IPC measures (such as hand washing) increased less.
Other studies, investigating scope of knowledge, attitude and practices, showed that doctors exhibited higher knowledge scores than nurses and paramedics [30]. Knowledge, attitude, and practice of HCWs regarding the use of face masks were found to be inadequate. HCWs showed a positive attitude but moderate-to-poor level of knowledge and practice regarding the use of face masks [35;38-39].
Two interesting results in our study were that the majority of respondents declared to have radically changed their work habits in the lockdown period; moreover, the use of surgical masks among Page 18/24 respondents increased much more in the lockdown period compared to the use of FFP1 or FFP3 masks.
However, this study has some limitations that should be considered. The developed questionnaire was tested among different HCW's in the Bambino Gesù Children's Hospital and open-ended questions were limited to reduce information bias. Moreover, the survey was conducted through the use of an online platform and disseminated through social networks, therefore the type of sampling used may not be representative of all Italian HCWs and could be biased towards respondents with a positive knowledge, attitude and practice.

Conclusions
The majority of respondent HCWs felt to be adequately prepared to deal with pandemic. Healthcare professionals are at high risk of COVID-19 transmission to and from patients and need extensive knowledge and awareness of the disease to take adequate precautionary measures, because they play a crucial role in lowering morbidity and mortality and in disseminating good practices.   Most frequent words reported from participants