Assessment of risk perception and risk communication regarding COVID-19 among healthcare providers: An explanatory sequential mixed-method study in Bangladesh

Background

The role of health professionals is crucial during an outbreak such as in the current coronavirus disease 2019 (COVID-19) pandemic to maintain population health and provide assurance in retaining healthcare system order. Hence, a clear understanding of how healthcare providers are being communicated with about the risk and how they perceive the risk is essential for emergency preparedness and crisis management¹ during public health emergencies.

When risk is the anticipation of a catastrophe, perception applies to the mental processes through which a person deals with the disastrous event². Studies of risk perception examine the judgments people make when they are asked to characterize and evaluate any hazardous situation³. Empirical studies show that perception and acceptance of risk have their roots embedded in social and cultural factors⁴ and research evaluating risk perception often offers important pointers concerning the selection of dimensions that need to be focused on for risk management. On the other hand, risk communication is multi-directional communication and engagement with the population at risk, so that they can make informed decisions to protect themselves⁵. In any health emergency, risk communication is directed to share information essential for saving lives, preserving health, and minimizing harm through changing perception and behaviour⁵. Communicating risk with healthcare providers is important as it might influence their understanding of the risk, willingness to serve at the frontline and enhance their preventive practices in times of need.

A recent qualitative study in China reported that healthcare providers experienced several challenges while working in COVID-19 wards that include heavy workloads, exhaustion from wearing protective gear, fears of being infected, and a sense of powerlessness while fulfilling their professional responsibilities for patients’ wellbeing⁶. During the severe acute respiratory syndrome (SARS) epidemic in Japan, a study revealed a high level of risk perception among healthcare providers and hence, emphasized on planning and implementing institutional measures during any health emergency⁷. Bangladesh reported its first Severe Acute Respiratory Syndrome Coronavirus-2 (SARS-CoV-2) positive case on 8 March 2020, around three months after the first reported case in China. Yet, experiences of Bangladeshi healthcare providers during the COVID-19 pandemic remain mostly unexplored. With a steep rise in new COVID-19 cases in Bangladesh, many healthcare providers have already been infected. Under the circumstances, a clear understanding of different perspectives of disease risk and prevention is needed in order to develop effective prevention strategies⁸. A detailed understanding of risk perception is also essential for effective risk communication and risk management⁹. This study has been designed to examine the communication made with healthcare providers and their perceptions regarding the risks related to COVID-19.

Research questions

Study design

This will be a two-phased explanatory sequential mixed-method study. According to the study design, a quantitative cross-sectional study will be conducted in the first phase to evaluate the general pattern and quantifiable measures of the research objectives. In the second phase, qualitative data will be collected for in-depth understanding and exploration of participants’ experiences and insights. Data from both quantitative and qualitative phases will then be triangulated to illustrate the answer to each research question.

Study settings and study population

Registered physicians and nurses working at different tertiary care hospitals in Dhaka, the capital of Bangladesh, will be invited. More than half of the COVID-19 patients in Bangladesh are concentrated in Dhaka¹⁰. Here, some healthcare facilities have been dedicated for the treatment of COVID-19 patients while others are open to all patients. Using a lottery method, six facilities will be selected as study sites from the list of public tertiary healthcare facilities in Dhaka.

Participant recruitment and data collection

First phase: Quantitative data collection. The sample size has been calculated by 4pq/L2. Considering the perception of 50% (p), q as 1− p, 5% allowable error (L), 95% confidence interval, and 10% non-response rate, the calculated sample size is 440 participants. Thus, recruiting at least 440 participants will be adequate for this study.

During the ongoing COVID-19 pandemic, healthcare providers in tertiary care hospitals in Dhaka are working in shifts following a duty roster. For this purpose, physicians and nurses working for each hospital have been divided into several groups. Usually, one group is working continuously for a pre-fixed period and then going into quarantine as another group replaces them. Considering this context, randomly selected physicians and nurses serving at the selected hospitals or departments of hospitals during the pre-fixed data collection period will be approached.

The following selection criteria will be applied for this study.

Inclusion criteria

Exclusion criteria

A list of physicians and nurses will be prepared from the selected hospitals, applying the selection criteria for collecting quantitative data, and a self-administered questionnaire will be distributed among them. A group of data collectors will be trained on the questionnaire beforehand. Data collectors will be made available at the hospitals throughout the data collection period for any clarification regarding the questionnaire.

Second phase: Qualitative data collection. Qualitative data will be collected through in-depth interviews (IDIs) and document review. A strategic sampling strategy with gender balance will be followed for qualitative data collection. Primarily, ten physicians and ten nurses working at the sampled hospitals will be selected through purposive and snowball sampling for interview. Qualitative data will be collected until data saturation.

At first, the selected potential participants will be sent informed consent forms and a permission letter from the corresponding hospital administration asking for their participation in this study. After obtaining their written approval on consent forms, interviews will officially proceed. Secluded places within hospital premises or adjacent to hospitals as per the convenience of the participant will be the preferred interview locations. Maintaining proper physical distancing and other personal protective measures, IDIs will be conducted and digitally recorded. In addition, verbal and non-verbal expressions of the participants will be recorded by note taking. All audio-taped interviews will be transcribed verbatim immediately after each interview.

Documents mentioned by the participants during the interview that need further exploration to accomplish research objectives will also be reviewed.

The data collection plan for collecting qualitative data is shown in Table 1.

Data collection tool

First phase: Quantitative data collection tool. For quantitative data, a structured questionnaire will be constructed encompassing three aspects of risk perception - cognitive, affective, and psychometric. Cognitive risk perception will be assessed by asking the participants to rate their perceived susceptibility to and perceived severity of COVID-19 using a Likert scale. “Standard questionnaire on risk perception of an infectious disease outbreak” developed by the Municipal Public Health Service Rotterdam-Rijnmond¹¹ and constructs of the health belief model (HBM)¹² will be followed to set up the questionnaire on cognitive risk perception. The affective dimension of risk perception will be evaluated through fear, anxiety, trust, and general concerns about COVID-19. To evaluate fear, the Fear of COVID-19 Scale¹³, a well-validated tool will be used. A validated Bengali version of this tool is also available¹⁴. Permission has already been obtained for using this tool for this study. For psychometric risk perception, the psychometric paradigm suggested by Slovic et al. will be used, which focuses on the qualitative dimensions of the perception on COVID-19 such as perceived voluntariness, catastrophic ability, controllability, severity, personal impact, and novelty¹⁵. A German risk perception survey questionnaire¹⁶ will also be followed to construct the questionnaire for evaluating psychometric paradigm. Questions to evaluate risk communication will be developed based on literature review and supported by the mental theory of risk communication¹. Infection prevention and control (IPC) practices will be assessed based on the IPC guideline provided by WHO¹⁷ for the healthcare providers. The questionnaire will be pre-tested prior to the data collection among healthcare workers at a primary healthcare facility in Dhaka.

Second phase: Qualitative data collection tool. A semi-structured guide for IDIs will be prepared, focusing on issues mentioned by the participants in the first phase of the study that need additional explanations. Pilot interviews will be conducted to test the questions in the semi-structured guide and necessary modifications will be made before starting the formal interviews, as recommended by Magnusson and Maracek¹⁸.

Documents shared by the respondents during IDIs will be reviewed. Participants will be asked for two types of documents that are relevant to the study objectives and will reflect their experiences on their risk perception, risk communication and preventive practices: public documents such as office notices, training manuals, guidelines, or protocols; and private documents such as personal notes or logs. For example, participants will be asked to share their experiences about the methods of risk communication during IDIs. If they mention any documents while describing their experience, these will be sought and reviewed.

The methodological matrix for the study is presented in Table 2.

Outcome variables

The outcome variables to be assessed in this study are presented in Table 3.

Statistical analysis plan

First phase: Quantitative data analysis. Descriptive analysis will be performed for socio-demographic and other professional characteristics. For continuous variables, mean (standard deviation, SD), median, maximum, and minimum will be calculated. Normality assumption will be made by Shapiro-Wilk test and a p-value of less than 5 percent will be considered as an asymmetric distribution. For categorical variables, rate, percentage, and proportion will be calculated. Perceived risk will be assessed using means with SD and comparisons will be made using the student T-test and analysis of variance (ANOVA) test based on different characteristics such as age, sex, profession, living status, type of healthcare facility, etc.

Multiple linear regression analyses will be performed to determine how much variability of risk perception is predicted by mediums, influencers, and content of risk communication, trust, fear, and anxiety. In addition, the association of risk perception and risk prevention practices will be explored through regression analysis. The role of HBM constructs to explain the healthcare providers’ compliance with IPC guidelines will also be analysed through regression analysis. A p < 0.05 will be used as the level of significance. A window-based statistical software package, preferably SPSS-23, will be used for analysis.

Second phase: Qualitative data analysis. Qualitative interviews will be transcribed verbatim immediately after interview and will be checked by two researchers via thorough listening of the interview recordings. Data analysis will start immediately after completion of the first transcript while interviews will still be ongoing. A strategic plan will be developed for analysing interviews and documents, based on the generic coding method proposed by Alase¹⁹ and suggestions given by Creswell²⁰. Firstly, a qualitative codebook will be developed based on literature review on the research topics. This codebook will contain a list of potential codes with definitions, examples, and instructions on usage. These codes will provide preliminary guidance on coding process and will be changed based on the information learned in the process of data analysis. Secondly, researchers will read interview transcripts and documents several times, organise responses into block of sentences or statements, condense them into meaningful chunky statements, and list repeatedly expressed words or phrases by the participants. Thus, the codebook will be furnished and applied to all interviews and documents. Thirdly, re-reading of and listening to all the documents and interviews will be done and chunky statements will be condensed into fewer non-repetitive non-overlapping statements and encapsulated to produce the central meaning or meaning units of the interviews and documents. Meaning units will then be grouped into sub-categories and then categories. Consequently, themes will emerge that can answer research questions.

After analysing and evaluating quantitative and qualitative data separately, triangulation or combination of both data types will be done. Data will be compared and converged following Creswell’s guidance²⁰ to increase data validity, reduce potential bias, minimize limitation, and thus, generate in-depth knowledge on research topics.

Validity/quality assurance strategy

Creswell²⁰ put emphasis on establishing validity of the scores and findings from both quantitative and qualitative measures in any mixed method study. With a view to ensuring the accuracy of the overall study findings, some measures are planned to be executed. A well-calculated and adequate sample size will be deployed in both phases of the study. Findings of the quantitative phase will be carefully analysed to find out potential areas that will need further in-depth explanation and will be included in the qualitative data collection tool. Samples will be drawn from the same population for each phase of the study to validate the outcomes. Two different survey interviewer manuals will be prepared in the local language of Bangladesh for the two phases of data collection. A training session will be organized where an adequate number of data collectors will be trained to introduce themselves, explain the purpose of the study, obtain informed consent, administer the data collection tool, preserve confidentiality, and recognize possible negative reactions and respond properly. Fieldwork activities of data collectors will be monitored and supervised regularly to ensure the validity of data. Every transcript and document will be revised thoroughly by two separate researchers to ensure authenticity and credibility. Codes will be cross-checked by different researchers. Consensus on each meaning unit and study finding will be made by all researchers.

Ethical statement

Ethical approval for this protocol has been obtained from the Institutional Review Board of Bangabandhu Sheikh Mujib Medical University at its 199th meeting (Memo number- BSMMU/2020/6040). All physical data, transcripts and documents will be coded and stored in locked cabinets to secure participants’ information. Only research personnel will be allowed to access the data. The collected information will be used for research purpose only. Several techniques have been adopted to minimize social, physical, and legal risk during the data collection process. Participants will have the right to withdraw from the research at any time. Each of the participants will be given a special identification number for safeguarding confidentiality and protecting anonymity. An informed consent form will be developed containing detailed information about the aim and objectives of the study, the procedure of the study, benefits and risks of participation and the identity of the principal investigator. Informed written consent to participate in the study will be sought from every respondent in both phases of the study.

Dissemination

Study findings will be disseminated through an online dissemination seminar. In addition, articles will be written and published in international peer reviewed journals and the data set will be shared in the Mendeley Data repository.

Study status

Data collection in the first phase of this study has been conducted from 17 to 30 May 2020. The second phase of data collection was completed in August. Now, we are undertaking data analysis and report writing.