Research design, setting and study population
A cross-sectional study was conducted in two tertiary care hospitals of Kathmandu: Tribhuvan University Teaching Hospital (TUTH) and Manmohan Cardiothoracic Vascular and Transplant Center (MCVTC). All patients with chronic diseases conditions were involved in the study who attended the outpatient department of different departments of TUTH and MCVTC for their routine treatment and follow-up between the period of 22 August 2021 and 19 December 2021.
Study population, sampling method and sample size
The study enrolled patients age 18 years and above who visited the selected hospitals for treatment and follow-up of chronic disease. Those patients with serious health complications including advanced liver failure, advanced respiratory problems, advanced heart failure, hearing impairment, and advanced dementia or mental disorders were excluded.
Study participants were selected systematically, recruiting every 10th patient from the patient register. Sample size was assumed at 50% prevalence of KAP, since there was no published data that showed knowledge, attitude, and practice related to COVID-19 among chronic disease patients in Nepal, and calculated using Cochran’s formula, n = Z2pq/d2. The estimated sample size was 384, but after adjusting for possible non-response rate of 10%, the final sample size was 422. Fifteen chronic disease patients were interviewed per day.
Research instrument and pre-testing
The research tool included structured questionnaires related to socio-demographic information of participants, their clinical characteristics, and specific information about KAP on COVID-19. The tool had 13 knowledge related items (clinical presentation, route of transmission, prevention and control of the outbreak) scored as true, false or not sure [18,19]. In attitude-related questions, six statements were adopted from the literature assessing participants’ attitudes toward COVID-19, using a five-point Likert scale, with scores ranging from 1 (Strongly disagree) to 5 (Strongly agree). Total scores ranged from 6 to 30, where high score indicated positive attitude [20]. For practice-related questions, total eight questions were used from a previous study, allowing participants to choose between two options ‘Yes’ and ‘No’[21]. All the tools were developed by the researchers with the help of extensive literature review and consultation with experts. The pre-testing was done in 15 chronic disease patients of Bir Hospital located in Kathmandu. Based on pre-testing, minor changes were made in the wording and sequence of the questions. The questionnaires were developed in both English and Local (Nepali) languages.
Data collection procedure
Informed written consent was taken before initiating the data collection from respondents. Data were collected by face-to-face interviews with participants and the range of duration was12 to 15 minutes. Research assistants (medical undergraduates) were involved in the interview at the study sites. The research assistants were personally briefed and trained by the first author beforehand. Confidentiality was maintained by keeping the response safely and was only used for this research purpose. Anonymity of respondents was maintained by assigning code numbers instead of participants’ names. Considering the COVID- 19 pandemic situation, the patients’ safety was assured by maintaining physical distance (at least 2 meters) and wearing a face mask and gloves.
Data analysis
Data was entered in the EpiData version 3.1, then transported to the Statistical Package for Social Sciences (SPSS IBM) version 21.0 for analysis. At first, we performed a descriptive analysis which included calculation of frequency, percentage, mean and median for presentation of socio-demographic and KAP-related variables. The Chi-square test was used to assess differences in categorical variables. Variables with p value of ≤ 0.2 were retained in the multivariable logistic regression to determine potential factors associated with the outcome variable; thus, adjusted odd ratios were calculated. We also checked multi-collinearity between predictor variables. For final result of logistic regression, p value was set statistically significant at <0.05. Additionally, we examined the effect of several predictor variables (age, sex, occupation, marital status, family income, family support, study site, duration of chronic disease, and health insurance) on COVID-19 knowledge; the effect of occupation, education, family support, study site, and duration of chronic illness on COVID19 attitude; and the effect of age, ethnicity, occupation, marital status, family income, family support, duration of chronic illness, and health insurance on the participant’s practice. The level of significance was considered at 5% with p< 0.05 and 95% confidence interval to determine the strength of association between independent variables and outcome variables.