This cross-sectional survey used both web and paper-based questionnaires to obtain data from HCWs in government, private, quasi-government, and Christian Health Association of Ghana (CHAG) health facilities located in four regions of the country. The online survey via Google forms provided a convenient approach to obtain data from a cross-section of health personnel while minimizing face-to-face contacts. This data collection strategy was used in similar situations elsewhere . Paper-type questionnaires supplemented the online survey and were self-administered to study participants who could not access the online survey due to phone and technological barriers. Using the paper questionnaires also facilitated ease of recruiting other categories of HCWs such as laundry staff, cleaners, laboratory technologists, and sonographers whose nature of work prevented them from participating in the study.
Four regions (Greater Accra, Ashanti, Central and Western) were purposively selected because they were the epicenters with the highest incidence of the coronavirus infections. For instance, as of 8th April 2021, Greater Accra Region (GAR) had recorded 50,241 cases; Ashanti 15,379; Western 5,717 and Central 3,294 .
Characteristics Of Participants And Sample Size
Eligible respondents included both clinical and non-clinical HCWs recruited from different health facilities. Clinical staff should be providing direct patient care in any department/unit of the hospital such as physicians, general registered nurses, midwives, pharmacists, laboratory technologists, and nurse aides or health assistants. On the other hand, non-clinical staff were personnel who did not directly provide patient care, example, administrative support staff. At the time of the study, they should not have been infected with COVID-19. The sample size was calculated using Cochran’s formula N = z2 * p(1-p)/d2 assuming a response rate of 50%, 95% confidence interval (CI), z of 1.96, and 5% margin of error. A further 10% was added to counteract any errors in completing the questionnaires, resulting in a final estimated sample size of 414.
Non-probability sampling techniques were used to recruit potential participants - purposive, convenience and ‘chain referral’ techniques. The aim was to complement and facilitate ease of access to data collection. For instance, the ‘chain referral’ sampling strategy was purposed at reaching several other HCWs indirectly in the specified regions through peer and social networks, and groups. We anticipated that given the double burden of work (including shift system schedules) at this time, they may not be easily accessible to participate in the study, hence, these multiple sampling procedures.
We designed and pre-validated a questionnaire based on HCWs background characteristics, and compliance with IPC practices which included: hand hygiene, PPE use (face masks), social distancing, and disinfection practices at the workplace. These four main IPC practices are the most basic yet importantly observed IPC measures observed by all categories of health workers in health facilities globally. The questionnaire comprised of 19 items with two sections: section A focused on participants background details such as age, sex, marital status, number of living children, highest educational level (completed), religious affiliation, category/type of HCW, number of years since being employed, type of health facility, and region of workplace. Section B consisted of eight items on compliance with hand hygiene practices, wearing of PPE (face masks), social distancing, and disinfection practices at the workplace (Table 2). The internal consistency with Cronbach’s alpha coefficient on the eight-item compliance measures was 0.80.
The outcome variable was perceived risk of COVID-19. It was measured as a dichotomous variable: 1 = Yes, and 0 = No. The explanatory or predictor variables were the level of compliance constructs assessed by eight items. These items focused on hand hygiene, wearing of face masks, social distancing, and disinfection practices. These items were assessed on a three-point scale from Not compliant at all = 0, Sometimes compliant = 1 and Always compliant = 2. The control variables were age, sex, marital status, highest educational level, religious affiliation, category/type of HCW, number of years since being employed, type of health facility, and region of workplace. Questions on the frequency of COVID-19 tests, and ever tested for COVID-19 were included.
This study is part of a larger study on HCWs knowledge, attitude and perceived vulnerability to COVID-19 and the likelihood of contracting COVID-19. It was approved by the University of Ghana Ethics Committee for the Humanities (ECH016/20–21) and the Ghana Health Service Ethics Review Committee (GHS-ERC 012/08/20). All ethical principles and considerations in the study were in accordance with the Declaration of Helsinki. Informed consent (mostly written) was obtained from all the study participants for the paper-based questionnaires while a statement of informed consent was included in the Google form prior to starting the questionnaire.
Data analysis involved univariate, bivariate, and multivariate analysis. Univariate analysis included simple descriptive statistics using frequencies to describe respondents background characteristics. Bivariate analysis was performed with chi-square tests to determine the association between the independent and dependent variables. Multivariate analysis was performed to identify the factors associated with perceived risk of COVID-19 and to also examine the effect of all the study variables on HCWs perceived risk of COVID-19. The type of multivariate analysis used was binary logistic regression. All analysis were performed in STATA version 12.