Study design and setting

This cross-sectional study aimed to assess the HRQOL of the COVID-19 patients following their one-month illness experience since diagnosis by real-time reverse transcriptase-polymerase chain reaction (RT-PCR) test for SARS-COV2. We considered the one-month duration of illness experience of COVID-19 patients after RT-PCR test positive and used the CDC-HRQOL questionnaire to assess their HRQOL. Both recovered (who became RT-PCR negative within 30 days) and non-recovered (who remained RT-PCR positive after 30 days since diagnosis) patients who had completed one month of illness after diagnosis by RT-PCR test were eligible for the current study. We conducted the study at the National Institute of Preventive and Social Medicine (NIPSOM) of Bangladesh from June to November 2020. NIPSOM is the apex public health institute in the country, holding the central laboratory designated by the Ministry of Health and Family Warfare of the Bangladesh government for COVID-19 diagnosis.

Study population

The study population included all adult COVID-19 patients diagnosed by RT-PCR assay at the central laboratory of NIPSOM. We enrolled the patients of both sexes, aged >18 years and who completed the one-month duration of illness after being diagnosed by RT-PCR test. The study excluded those patients who had no contact number or wrong contact number (Telephone/Cell); who didn’t respond to a phone call on three separate occasions at three different times on the 31^st day after being RT-PCR positive; who were unwilling to participate; and patients who had incomplete interviews. We collected data through telephone interviews, and there was no scope to investigate whether the clinical characteristics of the individuals excluded from the study were similar or different from the individuals who were included. We excluded some individuals from the study who did not respond to our telephone calls and whose telephones were inactive. As a result, we couldn’t collect data on the severity and couldn’t be concerned about omitting any patient with particular severity. We retrieved the COVID-19 patients using the records of the central laboratory of NIPSOM.

Sample size and sampling

All the COVID-19 patients diagnosed by the NIPSOM laboratory in July 2020 (from 1st to 31st July) before the data collection period were selected purposively for data collection. The central laboratory of NIPSOM diagnosed a total of 1342 COVID-19 patients in July 2020. Out of them, 1204 patients were enrolled in the study following the selection criteria. The selection criteria of the study included: (i) Completion of the one-month duration of illness after being RT-PCR positive; (ii) Adults aged >18 years; and (iii) Capable of giving informed consent. So, the final sample size of the study was 1204 who were enrolled following the purposive sampling technique. According to Health Information System (HIS), Directorate General Health Services (DGHS), Bangladesh, the month of July saw a total of 75,955 confirmed COVID-19 patients, out of which 53,062 were male while 22,893 were female [13]. We used these data as a reference for assessing representativeness.

Data collection

For data collection, we recruited 15 data enumerators and three supervisors, and all of them were physicians and had a Master of Public Health (MPH) degree as an additional qualification. We organized a three-day training program on pretesting, data collection instruments, and the technique of telephone interviews to make the data enumerators and supervisors well-skilled. For collecting data through a telephone interview, the study enrolled those patients who completed 30 days after being diagnosed with COVID-19 by RT-PCR. Each patient gave informed consent and participated voluntarily in the study. The data enumerator made an appointment with each patient on the 30th day to conduct a telephone interview on the 31st day. Before taking the consent, each patient was informed about the objectives, study procedure, risks, and benefits of participation in the study. To minimize the variability among the responses obtained by different interviewers, we trained the interviewers extensively on telephone interviews and conducted shadow interviews randomly among the responses obtained by different interviewers.

A semi-structured questionnaire was used for collecting data through telephone interviews. The questionnaire comprised variables on sociodemographic characteristics (age, sex, marital status, education, occupation, place of residence, type of family, number of family members, monthly family income), health-related quality of life, symptoms of COVID-19, and comorbidities. A checklist was used to collect laboratory data of the COVID-19 patients by reviewing their medical records obtained from the central laboratory of NIPSOM. Both data collection instruments were pretested on the COVID-19 patients diagnosed in June 2020, and accordingly, necessary corrections and modifications were performed for its finalization.

To ensure the validity of the data, each telephone interview was recorded using a digital recorder along with filling up the questionnaire on paper. To ensure the quality, consistency, and relevancy of the data, we explained data collection instruments and procedures in detail to the data enumerators and supervisors. Detailed discussion was also done on each question in the questionnaire to enable them to collect data consistently following the same approach and sequence. The supervisors supervised data enumerators during data collection, randomly selected some participants, and interviewed them to verify their responses. Moreover, the investigators examined the variability among the responses obtained by different interviewers. To ensure the quality of data and precision of data analysis, we performed double entry of data. To minimize the recall bias, data on symptoms and other relevant attributes were retrieved by reviewing the medical records of the patients, which were filled up by the trained physician during the diagnosis process. We also posed probing questions during data collection through telephone interviews to verify the response to minimize recall bias.

Psychometric properties of the questionnaire

HRQOL was assessed via CDC’s Core and Optional HRQOL Modules. We used thirteen out of fourteen items from CDC HRQOL-14 for measuring the quality of life of COVID-19 patients. We excluded question 2 of the activity limitation module from our study because it is applicable for the estimation of the overall HRQOL of an individual considering all the health problems through future follow-up. But the present study conducted the assessment of the quality-of-life changes caused only by COVID-19at a specific time (one month after diagnosis). Among 13 items, four items are part of the Core HRQOL module, and the remaining nine questions are part of a 10-item Optional Module which measures more detailed aspects of HRQOL. Question 1 assesses self-rated general health with five responses ranging from excellent to poor. For statistical analysis, responses were grouped into two categories considering the grouping done in another study conducted in Pennsylvania [14]. These categories were named as: i) ‘good health’, which included ‘excellent, very good, or good’ responses and ii) ‘not so good health’, which included ‘fair or poor’ responses. In the remaining three items of the Healthy Days Core Module (duration of physical illness, duration of mental illness, and duration of usual activity limitation) scores ranged from 0 to 30 days, with higher scores indicating poor perceived health. Unhealthy days were an estimate of the overall number of days during the previous 30 days when the respondents felt that either their physical or mental health was not good. To obtain an estimate of the overall unhealthy days of a patient, responses to questions 2 (duration of physically unhealthy days) and question 3 (duration of mentally unhealthy days) were added together, with a logical maximum of 30 unhealthy days. Eight additional items (presence of a health-related activity limitation, needing help with personal care, needing help with routine care, and five questions of healthy days’ symptoms module) were treated as binary (yes/no) variables. The remainder of the HRQOL items were treated as continuous variables, with a minimum score of 0 to a maximum score of 30. The summary index of unhealthy days and healthy days was measured by the standard procedure described along with the questionnaire [15].

Though the CDC HRQOL-14 measures were initially developed for use in the U.S. English-speaking population, the concepts assessed by the measures are believed to be universal and are therefore capable of being adapted for use in other cultures and languages [11]. As the tool was not used previously in the Bangladeshi population, it was first translated into Bangla and then back translated into English. Followed by pre-testing was done to ensure the validity and reliability of the data collection instrument. Analysis revealed Cronbach’s alpha reliabilities for the three modules of the CDC HRQOL-14; healthy days’ core module, activity limitations module, and healthy days’ symptoms module as 0.86, 0.81, and 0.76, respectively.

Statistical analysis

Data were analyzed using SPSS STATISTICS (Version 25.0, IBM Statistical Product and Service Solutions, Armonk, NY, USA). A Chi-square Goodness-of-Fit test was used to assess the representativeness of the sample. The normality of the variables was also tested with the Shapiro-Wilk test/Kolmogorov-Smirnov tests of Normality. Continuous data were portrayed in the form of mean and standard deviation. Categorical data were depicted as frequency and percentages. Descriptive statistics estimated mean, standard deviation, and frequency while inferential statistics included chi-square test, independent sample t-test, and F-test. We used the chi-square test to compare the significant difference between two categorical variables. The t-test was used to find a significant difference in the means of two quantitative variables, and the F-test was used to find a significant difference in the means of three or more quantitative variables. Significant F-test results were further analyzed by Post-Hoc test. To identify the confounders, we developed an adjusted logistic model using all the significant categorical variables and calculated the adjusted odds ratio (AOR) by logistic regression analysis. A p-value <0.05 was considered significant with 95% confidence interval (CI). All the statistical tests were two-sided and performed at a significance level of α = 0.05.

Ethical considerations

We obtained ethical approval from the Institutional Review Board (IRB) of the National Institute of Preventive and Social Medicine (NIPSOM), Mohakhali, Dhaka-1212, Bangladesh (Ref. No. NIPSOM/IRB/2020/5, Date: 17/05/2020). Participation of the COVID-19 patients was voluntary, and they had the freedom to withdraw their consent at any stage of the study. Before data collection, informed verbal consent was obtained from each participant over the telephone after describing the study procedure and ethical issues. With the permission of each participant, the informed consent was also recorded using an audio recorder. During data collection, the confidentiality of data and privacy of the patients were maintained strictly. We used the collected data anonymously only for the current study.

Background characteristics

In the present study, around three-fourths of the COVID-19 patients were males, and the rest were females. This finding indicates that males were at a higher risk of being infected by COVID-19 than females. In this respect, other studies conducted in China [17] and Vietnam [8] revealed reverse findings, where females were affected more than males. It may be due to the divergence of the socio-cultural context of Bangladesh, which is different from those countries. Based on the local social and cultural frame, males are more involved in outdoor activities and more exposed to social gatherings than females. So, the chance of exposure and being affected is higher in males than females. Around half of the COVID-19 patients were young adults (age group 30–49 years), and only 11.0% of the patients were elderly (aged 60–90 years). The demographic profile of Bangladesh depicts that 40.07% of the population belongs to the age group of 25–54 years, and only 6.42% belongs to the age group >60 years [1]. In this regard, we can claim that the age distribution of the patients is in alignment with the national age distribution.

An almost equal number of the COVID-19 patients reported from the urban (50.2%) and rural (49.8%) communities. The study of Islam MZ et al., [1] identified a different finding in Bangladesh where reported patients were more in number from the urban communities. Though both the studies were carried out in the same country, a different result may be due to differences in the study periods. The former study was conducted earlier compared to the current one. At the beginning of the COVID-19 pandemic, limited diagnostic facilities were available in the urban setting of Bangladesh. Moreover, rural people of the country were less aware of the COVID-19 diagnosis. As a result, the former study found more patients reported from the urban setting than in the current study.

In respect of occupation, nearly half of the patients were service holders. In this regard, the study conducted by Islam MZ et al., [1] also identified most of the patients as service holders, and it was relatively lower (32.5%) than in the current study. It is evident that a significant portion of the population of Bangladesh is engaged in various occupational activities, and they must move frequently within the community, which exposes them to disease infection. As a result, the service holders are more vulnerable to getting an infection of COVID-19 than other occupational groups. We can mention that during the early stage of the pandemic, there was a restriction on population movement due to lockdown, shutdown, quarantine, and panic about the disease. It could be a valid reason for comparatively fewer service holders in the former than in the current study.

Symptoms and comorbidities of COVID-19 patients

Due to poor diagnostic facilities and limited RT-PCR testing capacity, asymptomatic individuals in a developing nation like Bangladesh were less likely to be investigated for disease screening. Moreover, due to ignorance and superstition, a significant proportion of asymptomatic patients didn’t participate in the investigation procedure of the disease. Lack of contact tracing and community-based screening initiatives in a setting with limited resources may also be responsible for the detection of fewer asymptomatic patients. The study found that 11.4% of asymptomatic patients attended the health facilities to exclude COVID-19 followed by exposure history, travel history, and treatment of diverse ailments. The present study portrayed that more than one-third (35.5%) of the COVID-19 patients had different comorbidities like hypertension, diabetes mellitus, ischemic heart disease, chronic lung, kidney, and liver diseases. The former study by Islam MZ et al., [1] detected nearly the same proportion (33.9%) of COVID-19 patients with similar comorbidities. Due to the higher risk of worse morbidity and mortality consequences of COVID-19 among the comorbid patients, they were referred to different health facilities to exclude COVID-19.

Different items of HRQOL of the COVID-19 patients and associated factors

We used laboratory records to retrieve the data from those COVID-19 patients who completed their 30 days’ duration of illness after being RT-PCR positive. We obtained information on patients’ general health condition, mental and physical illness, health-related activity limitations, needing help with personal and routine care, and healthy days’ symptoms module. We evaluated and analyzed all items separately.

At the end of 30 days after diagnosis by RT-PCR test, the general health condition was ‘good’ in 70.1% of patients, and it was ‘not so good’ in 29.8% of patients. ‘Not so good’ health condition was significantly associated with female gender (AOR = 1.565, CI = 1.01–2.42), joint family (AOR = 1.324, CI = 1.00–1.75), having a symptom of COVID-19 (AOR = 32.871, CI = 8.06–134.0), and comorbidity (AOR = 1.700, CI = 1.26–2.29). ‘Not so good’ health condition also increased with increasing age. The symptoms of COVID-19 aggravate the morbidity and illness severity of the patients, and in turn, interfere with their general health condition. It is a fact that females and elderly patients conserve relatively lower body immunity to protect against the illness progression and its clinical severity. As a result, their duration of suffering lasts longer than the males and younger adults.

The estimated mean duration of physical illness was around ten days and eight days for mental illness. The mean duration of the usual activity limitation was around seven days and it was significantly higher in the age group 50–59 years. The presence of comorbidity poses an incremental effect on the worse progression of the disease and impairs the health condition of the patients. In the present study, patients’ need for help with routine needs was also significantly higher among urban patients having a symptom and comorbidity. The duration of limited health-related activity was significantly higher in males and patients having symptoms and comorbidity. The duration of healthy days was significantly lower among patients with symptoms and comorbidities. Usual activity limitation and limited health-related activity were significantly higher among patients having symptoms. Coexisting clinical symptoms and comorbidities deteriorate the physical and mental well-being of the patients, which decreases their ability to perform daily activities.

The mean duration of feeling pain was significantly higher in patients having symptoms and comorbidity. The mean duration of feeling sad, blue, or depressed was higher in the ever-married patients and those having symptoms and comorbidity. A study conducted in China revealed a higher risk of pain/discomfort and anxiety/depression among older people having a chronic disease and lower income [17]. On the contrary, the mean duration of feeling worried was significantly higher in females, ever-married patients, and those having symptoms. The mean duration of feeling very healthy was lower in the older adults (50–59 years) and those having symptoms and comorbidity. The present study revealed a significantly higher duration of symptoms in females, ever married, and patients with clinical manifestation and comorbidity. These findings entice the attention of the health policymakers and healthcare managers to invent prioritized medical measures for those high-risk COVID-19 patients to improve their HRQOL and general health condition.

The current study concerns a limitation in collecting some relevant information like severity of COVID-19 infection, treatment options, presence of any other short-term illness during the one month, etc. We could not identify and manage the heterogeneity in the COVID-19 severity among the sample because the severity could not be measured through telephone interviews and reviewing medical records having scarcity of relevant information. It is a fact that any short-term disease could impair the HRQOL during the 30 days of illness period and could confound the effect of COVID-19. We asked the patients about any short-term illness, but they gave no significant response. We did not perform any clinical examinations and investigations of the patients due to their sporadic distribution throughout the country. Telephone interviews and inadequate information in the medical records were also a few reasons behind it. This cross-sectional study design only provided data on the HRQOL of COVID-19 patients after 30 days of being diagnosed, and no comparison was possible due to the lack of a control group and any previous baseline data. So, we could not determine whether the HRQOL of the patients was stabilized or returned to baseline after 30 days. We only assessed the HRQOL of the COVID-19 patients for one month since diagnosis (whether it was stabilized/returned to baseline or not) following the specific guidelines of the CDC HRQOL-14 questionnaire. Although we considered a shorter period of illness to prevent recall bias, measuring the effects of COVID-19 on the HRQOL of the patients couldn’t be impartial. The study considered clinical symptoms, treatment, comorbidities, and socioeconomic features as crucial determinants of HRQOL. The sample size was adequate for exploring the patients’ experiences and providing substantial evidence and directions for policy-making and future analytical studies.

Advantages of CDC HRQOL-14 questionnaire

Firstly, there is a policy value of estimating the burden of disease or disability in days, months, or years because it provides concrete measures that can be understood and used by legislators, policymakers, and healthcare managers to devise effective interventions in the preventive and clinical management of COVID-19. Secondly, the healthy days’ measures permit a wide range of options for statistical analysis because of their continuous, cardinal, and bounded (Range = 0 to 30) mathematical properties. Moreover, the quantitative approach of CDC HRQOL over other tools offers ample scope to measure the duration of physical and mental illness. Accordingly, it contributes to developing a vibrant clinical guideline for the effective management of diverse communicable and non-communicable diseases. The core CDC HRQOL-14 measures are the briefest validated set of generic HRQOL measures with a clear and explicit definition of HRQOL, and a transparent summary measure. The CDC HRQOL-14 measure is the most cost-effective and less burdensome among all HRQOL measures [15]. The CDC HRQOL-14 questionnaire is designed in such a manner that it can estimate the overall number of days during the previous 30 days when the respondents felt that either their physical or mental health was not good.

The CDC HRQOL-14 instrument can calculate the HRQOL of the patients for 30 days, whether it has returned to baseline/stabilized or not [11]. So, the instrument was appropriate for the present study to assess HRQOL of COVID-19 patients comprising physical and mental health for 30 days after diagnosis. One of the additional advantages of the CDC HRQOL-14 questionnaire is that the whole interview could be completed over the telephone, a data collection technique that is profoundly beneficial during this COVID-19 pandemic [12].