Preparing for the COVID-19 Pandemic Response in a Country Emerging From an Ebola Epidemic: Assessment of Health Workers' Knowledge, Attitudes and Practices on Coronavirus (COVID-19) in Guinea

Background: study to assess the knowledge, attitudes and practices of medical personnel on the prevention of Covid-2019. This was a cross-sectional study with an analytical aim carried out from 1st to 29th February 2020 among front-line health workers in the health structures of the city of Conakry. Data were collected using the Kobocollect application by administration of a standardized questionnaire. The knowledge and attitude score was categorized in 2: Good if ≥ at the average score and bad if < 50%. Logistic regression models were conducted to identify factors associated with knowledge and attitude. Odds ratios (OR) with their condence intervals were calculated.


Study design:
This was a cross-sectional analytical study conducted between rst and 29th of February 2020 among frontline health personnel in the health structures of the city of Conakry (communal medical centres and the two national hospitals of Conakry).

Framework of the study:
The study was conducted in ve communal medical centres (Coleah, Ratoma, le Flamboyant, Matam and Bernard KOUCHNER), three national hospitals (Donka and Ignace Deen and Sino-Guinean) and three health centres (Maciré, Dixinn and Madina). These hospitals are all located in the city of Conakry, the capital of the Republic of Guinea, which has an estimated population of 1,930,838 [15].
The majority of the health workforce is concentrated in this capital city and all specialized health services are also located here. This unequal distribution of personnel is responsible a signi cant in ux of patients to these facilities, some of whom even come from the for interior of the country.
Despite the impact of the preparedness policies put in place on the continent over the past several years (re ected in particular in the capacity of national laboratories to carry out diagnosis), the Covid-19 epidemic highlights the weakness of coordination among countries in their response and the impact of budget cuts on health systems. Everywhere the epidemic reveals the weaknesses of health systems (limited capacity in intensive care units, lack of health personnel). Underdeveloped countries with known weak health systems will therefore nd it di cult to respond effectively to the predicted epidemic. This is the case in the Republic of Guinea, where the already weak health system was severely impacted by the Ebola epidemic between 2013 and 2014 [12][13][14].
Part of the response to the outbreak of the epidemic involves raising public awareness and training health workers. Thus, to guide health authorities on the need for training and to direct the response towards targeted actions, we organized in February 2020, a few weeks before the noti cation of the rst case of COVID-19, a survey among health personnel to assess knowledge, attitudes and practices to prevent coronavirus. The results of this study will help ll gaps in the training of front-line staff and the preparation of the health system for a possible increase in the number of cases.
The purpose of this study was to assess the knowledge, attitudes and practices of health personnel about COVID-19 and to identify their determinants.
The objectives of this study were to: 1. Describe the knowledge and attitudes of health personnel in the city of Conakry on COVID-19; 2. To identify the factors in uencing the knowledge and attitudes of the health personnel on COVID-19.

Page 5/29
Study Participants : The study focused on front-line health workers in health facilities of Conakry. All health workers present on the day of the survey who gave their verbal consent were eligible for the survey. This is a face-to-face verbal interview with the respondent. All the answers were transferred to the tablets from and transferred to a server at the Infectious Disease Research and Training Center in Guinea.
We sampled for convenience in the selection of health services. This type of sampling was used to select those health services in which staff are considered convenient data sources for this study. Convenience sampling is a type of non-probability sampling in which individuals are sampled simply because they are "convenient" data sources for researchers [16].
The number of health personnel expected in the health facilities in Conakry was 900. The respondent was asked about those present in the services and with a verbal agreement to participate in the study.

Data Collection and Variables
Data were collected by interviewers using a standardized questionnaire with 4 sections as follows: 1-socio-demographic characteristics such as age, gender, health structure, number of years of experience, academic degree, socio-professional category., 2-knowledge about the coronavirus, 3-attitudes to prevention and control of infection, 4-practices and information needs. The Kobocollect application was used to collect the data. The questionnaires were pre-tested for validation purposes. Imperfections in the questionnaire were corrected and the minimum time required to administer the questionnaire was estimated at 20 minutes. Thirty (30) research assistants were mobilized to collect data in the eleven health structures. They were distributed according to the number of staff and the number of services included in the study.
In order to carry out this survey as quickly as possible because of the public health emergency it represented, we adapted the questionnaire used to assess health workers' knowledge of Ebola to Covid-19 [17] and completed it using information provided by the Center for Disease Control and Prevention (CDC) and WHO on the epidemic [3,5].
Knowledge Scoring : Knowledge and attitudes were assessed using structured questions on the virus, symptoms, care, prevention and response activities undertaken by the State. A total of 70 questions to assess knowledge and six (6) attitudinal questions were asked. Each correct answer was scored as 1, while the incorrect answer was scored as 0.
Good knowledge was de ned as a score ≥ at the overall average and poor knowledge was de ned as a score < at the overall average. The same method was used to identify good and bad attitudes towards COVID-19.
To minimize bias, for each respondent the question was administered by a single interviewer and in an isolated o ce.

Data Analysis:
The data collected on kobo collect were extracted and cleaned before being analysed with Stata 14 software. Surveys with a lot of missing data were excluded from the nal analysis. The basic characteristics are presented in proportion, mean and median as appropriate. Standard means and standard deviations and medians were used to describe the quantitative variables. A student or Wilcoxon-Mann-Whitney t-test was used for comparison. The Chi2 or sher exact test was used for the comparison of quality-avoidance variables.
Two logistic regression models were constructed to identify independent factors associated with poor knowledge and COVID-19's poor attitude to prevention.
In our analysis, we retained the variables whose p was ≤ 20% in the bivariate regression analysis. The bottom-up stepwise method was used. The likehood ratio test was used to compare the successive model.
The nal model has been validated by the Link test and the overall t by the Hosmer-Lemeshow goodness-of-t test. We also evaluated the discriminant power of the model with the ROC curve. A discriminant power was judged to be good if the ROC ≥ 80%.

Ethical Considerations:
The study was approved by the Ethics and Scienti c Committee of the Public Health department of the Faculty of Health Sciences and Technology of Gamal Abdel Nasser University, Conakry.

Results:
Between rst and 26th of February 2020, we surveyed 548 frontline health workers in Conakry (capital of the Republic of Guinea) i.e. 60% of the expected sample (Fig. 1).
Nine health workers were excluded from the nal analysis due to missing data because they had not completed the interview with the interviewers More than half of the respondents were under the age of 30 with a median age of 29 years IQ (25-38). The majority of the agents interviewed were women (57.1%) and had less than 5 years' experience in the eld (54.9%) and were nurses in 28% of cases. The majority of the participants were surveyed at Donka National Hospital 130/548 or 23.7%. Nurses were the most represented in the study (28.1%). Most of the health staff reported that the service had not reported any cases of Ebola during the 2013-2016 epidemic) (Table 1).  (Table 2).
Almost all, 546/548 evening 99.6% of the interviewees did not know the emergency number in case of a suspected case and about 50% did not know the incubation period of COVID-19.
Indeed, a large majority (53.8%) of staff reported wanting to consult and prescribe medication for a patient with signs related to COVID-19.
There was a statistical association between gender, age, grade, health structure and level of knowledge. For attitude, the associated variables were age (p = 0.01), health structure (p = 0.01), health status (p = 0.01) and knowledge (p = 0.01).

PCI practices and information sources
Participants' infection prevention and control practices were assessed using a Likert scale (Fig. 1), Study participants reported that they always disinfected surfaces, utensils and work equipment (69.9%) and then washed their hands after a medical procedure (76.6%). They also reported that they always wore boots (26.3%), gloves (63%) and gowns (48.7%) during a medical procedure. Soap and hydro-alcoholic solutions were the most commonly used by staff, with proportions of 55.7% and 42.5% respectively (Table 3).
It should be noted, however, that infection prevention and control (ICP) was not systematically applied by all staff. Thus, preventive gestures were not systematically applied (the rate of application did not even reach 80% for each gesture). (Fig. 2) Sources of information on COVID-19 for health care personnel-19 ( Fig. 2) The media was the main source of information about COVID-19 for health care workers (81%), both online and o ine. The professional and family network was the main source of information for staff, with 65% and 35% respectively. (Fig. 3) Most health personnel (57.5%) had already received training in infection prevention and control and wanted more information on COVID-19 (98%) . Factors associated with poor knowledge of COVID among health care personnel-19 ( Table 5 ) After a logistic regression, the variables associated with poor knowledge of COVID are gender, the structure where the respondent worked, the fact of having trained in ICP in the past, and rank. For the independent determinants of poor attitude, the associated variables are : Knowledge of Ebola cases reported in the department, health facility and grade of health worker.

Discussion
This study assessed the attitudes, knowledge and practices of health personnel in Conakry on COVID-19 and identify factors associated with poor knowledge and attitude of these practitioners. 70.62% of the personnel have a good knowledge of COVID-19 and 57.66% have a good attitude of disease prevention.
The majority of the knowledge acquired on the issue was from the media and internet (81%). The health system did not put in place anything to help the staff to become familiar with the ght because 99.60% did not know the emergency number and did not have a management algorithm (53. 80%°).
Although it was a cross-sectional survey with the possibility of information bias due to the fact that the questionnaires were not self-administered. Some staff would have had the correct answer if they had received the question beforehand. This study has the advantage of giving us an overview of the knowledge and attitudes of health staff regarding COVID-.19 in the pandemic preparedness situation.
Since To ll in the gaps on the issue and to guide decision-makers on the strategy for involving health care personnel, we conducted a knowledge, attitudes and practices survey among 548 health care personnel on COVID-19.
This study shows us that there are still some gains from the previous Ebola epidemic in terms of knowledge, attitudes and infection prevention and control practices.Indeed, it was observed that the majority of health workers have a good knowledge of Covid-19 (70% of the sample). However, the overwhelming majority of health workers, all statuses and structures combined (99.6%) were unaware of the existence of a toll-free number to alert on the presence of a suspected case of coronavirus and only 42.3% knew the incubation period of the virus. In a similar study of Ebola in Guinea, 95.2% of the health workers were knowledgeable and could identify an Ebola suspect [19]. Although this is not the same context, nor the same pathogen, there is an effort on the part of health workers to learn about public health threats. This is re ected in the sources of information consulted: 81% of them obtained their information from the media (Internet, radio). The same source was used by 85% of the surveys in Sierra Leone to obtain information on Ebola [20].
1. 57.1% faced with a patient with signs of COVID-19 said they would consider it but 37% would not. What seems even more worrisome is that 41.2% said they would prescribe a drug according to symptomatology or diagnosis, which raises the question of what happens to patients left to their own devices to pursue their quest for care (with a risk of spreading the virus). Also if 81.8% answered that they would wear PPE (Personal Protective Equipment) to consult them, it remains to be seen what they mean by this when in reality, all that is needed is a mask, gloves and protective glasses. This proportion of staff using personal protective equipment is almost equal to the proportion reported in Nigeria for assessing the PPE of health workers in the Ebola context [17].

2.
60.9% assured that in the event of a case with symptoms of Covid-19 they would put it in isolation and 62.6% announced that they would call the dedicated toll-free number. However, these responses clearly contained a signi cant desirability bias, since all the agents admitted that they were unaware of the existence of the toll-free number that would allow them to do so.
Systematic hand washing (76.6% of respondents) and glove wearing (63%) seemed to be relatively practiced but remain too weak to ensure effective protection for all agents and patients. The wearing of masks (26.1%) is largely insu cient, as is the wearing of gowns (48.7%). Concerning the literature on Haemorrhagic Fevers, the literature shows that the practice of handwashing varies according to whether health workers are practicing in a declared epidemic context or in an alert context. The practice was observed in the rst case, around 80 to 85% in Sierra Leone and Nigeria [17,20], but was very low in the second, as in Iran during the Crimean-Congo fever alert [21]. This issue is crucial for the implementation of infection prevention and control measures among health workers in the context of the COVID epidemic [19] where the feeling of health actors seems to be that if the epidemic is present, it is con ned to a limited fringe of the population (elites and foreigners).
The importance of exercise conditions is clear in the question about the product used to wash hands.
Only 42.5% had a hydroalcoholic gel and 55.7% used only soap and water.
81.6% of health workers reported shaking less than 50 hands a day, which is already too much. It remains to be seen how many avoided all contact. 42.5% of health workers have not received ICP training, and the vast majority of them (98%) would like to receive more information on the Coronavirus. It would also be advisable to identify, through the observation of practices, which types of actors are mainly exposed due to their contact with patients in order to prioritize their training (we can think of actors developing frequent physical contact with patients: care assistants, nurses, but also wardens or reception agents, etc.) but also how the application of ICP impacts on the quality of care produced (in particular by the absence of palpation-type contact).
The agents in the health centres and National Hospitals, a large proportion of whom have less than 5 years of seniority (Health Centre: 47%, HN 38%) compared to the CMCs (15%) had on average a better knowledge of the coronavirus (Health Centre and National Hospital 41.9%, CMC 16.3). This could be explained by the fact that the issue of zoonoses and ICP is now part of the training curriculum for health workers (Faculty of Health Sciences and Techniques and Nursing School), in which they have participated. This therefore calls for priority to be given to upgrading the skills of health professionals with more than ve years' seniority.
There is no strong correlation between level of knowledge of Covid-19 and expression of intention to advise isolation in case of symptoms, systematic mobilization of barrier gestures. This invites us to go beyond the cognitive question of knowledge of the risk to question both its perceptions, in particular the conviction of the actors to be concerned by the risk of the Coronavirus and the factors which do not come under the purely biomedical register which in uences it (risk scales, rumours...), This is not only the political context which gives rise to concern, putting the question of viral risk in the background), but also the conditions for appropriating the knowledge acquired in ordinary practice (conditions of practice, presence of protective equipment or disinfectant products, in ux of patients, leadership role of the medical hierarchy in the establishment). Particular attention should be paid to the impact of discourse on the extraneity of viral risk on the management of national patients. Until now, prevention measures have focused on airports and the gure of the virus carrier has been embodied by the Asian or European traveller. This gure has not been contradicted by the rst proven cases. However, as we enter a phase of local transmission of the virus, it is extremely important to alert health workers to the presence of the virus on national territory and to the fact that it is not limited to a section of the population of non-African origin.
We which allows them to have easier access to information at the central level. On the other hand, they occupy a greater number of administrative positions that allow them to search for information, whereas women, who are mostly con ned to care functions, have fewer opportunities. In the context of the Ebola fever epidemic on the African continent, the lack of time that women could devote to information seeking and their high exposure to risk due to their care functions has been put forward to explain their high lethality [22]. With regard to the pharmacy profession (OR = 9. Although we covered all public structures in the capital, this study has certain limitations: rst of all, private structures were not included, nor were informal spaces for the production of care or the dispensing of medication products. In addition, assessing the level of knowledge about a new virus such as Covid-19, for which knowledge is evolving daily, may seem a challenge. In spite of these limitations, this study is of de nite interest at a time when Guinea, along with the African continent, is organizing its response to the emergence of a new epidemic.

Conclusion:
The epidemic of COVID-19 disease is an "extraordinary" pandemic because of its global scale and the speed of its spread. Despite the collective trauma it is causing worldwide, it is important to recognize the contexts in which it is occurring. While in all countries medical personnel remain at the heart of the response mechanism, their level of knowledge and experience of the epidemic is context speci c. If we observe the legacy of the previous Ebola epidemic among certain Guinean health workers (in terms of knowledge and practices), signi cant gaps remain and are the subject of a demand for training on the part of the actors. Our study has also highlighted the di culties of the appropriation of prevention and protection measures against Covid-19 by health workers for whom the risk appears remote and limited to a transnational or foreign elite.
It would be advisable to associate to this CAP survey an observation of long-term practices allowing in situ observation of actors' practices in order to avoid desirability bias (an observation every day for a week allows actors to "forget" the presence of the observer). This would make it possible to work on the distinction between the o cial norm expressed in the discourse, and the practical rationalities relating to the conditions of practice, professional cultures, and the nature of the interaction between the health worker and the patient (which itself often depends as much on prior inter-knowledge as on the social status of the patient, for example). AT : aparticiper à la conception, la validation des données et la relecture de l'article AC, MSS : Ont participé à la conception, la validation des données et la relecture de l'article ; FLM :A participation à la discussion, la validation des données et la correction de l'article et la traduction de l'articles en Anglais ; BST : Participation à la conception de l'étude, à la supervision de l'enquête, la correction de l'article ;