Design:
The data gathered via a cross-sectional phone survey involving 213 persons who received COVID-19 care in July 2020 from any COVID-19-specific hospital designated by the Government of Bangladesh (GoB) were subjected to psychometric analyses to validate the existing ROP-Scale (8).
Participants:
The study participants were recruited from both public and private COVID-19 designated hospitals across Bangladesh and were identified through the database of COVID-19 patients provided by the Aspire to Innovate (A2i) Programme, Information and Communication Technology (ICT) Division of the GoB that contains patient’s name, disease outcome, date of admission and discharge, hospital name, and contact information. We sampled potential respondents from the database and extracted the contact information for the survey which was conducted over the phone due to the pandemic.
Only adult patients (aged ≥18 years) who were treated in one of the COVID-19 dedicated hospitals in July 2020 whose correct phone number was available in the database were eligible for participation. Those that did not consent to participate or could not be reached after calling three times at different time points were excluded from the sample.
Content validity:
Item clarity and content validity were established using a modified Delphi technique. The first author, who is a health policy and systems researcher with expertise in psychometrics, served as the facilitator. He convened the panel of experts (all of whom are the article co-authors) comprising of two public health physicians, a communication expert, and a biostatistician experienced in psychometrics, and provided them with the existing ROP-Scale, along with the guidelines for refining its 34 items through three iterations. The aim was to ensure that the scale items are (1) comprehendible over the phone, (2) applicable to the COVID-19 inpatient hospital context, and (3) appropriate for a questionnaire survey (the original ROP-Scale items were designed for structured observation and the five response categories were anchored in an outpatient consultation scenario), and that (4) the whole questionnaire is short enough not to discourage voluntary participation.
Sample size and sampling techniques:
The sample size was determined based on the 10:1 “n to p ratio” (9, 10), where ‘n’ is the minimum sample size, and ‘p’ is the number of items. Since the initial tool consisted of 20 items, the required sample size was 200. Anticipating a 20% non-response rate, as the survey was to be conducted over the phone and the patients may be weak and reluctant to participate after recovering from COVID-19, we approached 250 patients and obtained 213 valid responses. We used the RAND function in MS Excel (MS Office Professional Plus 2016) to randomly select the 250 patients based on simple random sampling technique.
Data collection instruments and procedures:
The survey questionnaire consisted of two parts: (1) Socio-demographic characteristics (age, gender, education, occupation, current residence, religion, marital status, and monthly income), and (2) ROP-Scale items. From the database, we also extracted the location and type of hospitals (public or private) in which the patients received care, duration of hospital stay, and disease severity. The ROP-Scale section was further subdivided into seven sub-sections, namely (1) Beginning part, (2) History taking, (3) Examination, (4) Prescribing, (5) Explaining, (6) Leaving, and (7) Throughout consultation. Respondents were also asked to rate their level of overall satisfaction with the services received from the doctors in the hospital. The 20-item draft tool was pilot-tested on five non-sampled patients and the language was improved for intelligibility based on their feedback. The original ROP-Scale items and the draft COVID-19 ROP-Scale items used for data collection in this research are shown in Table 1.
Table 1: ROP-Scale’s original items and draft items for this study
Name of Sub-scales or Domains
|
Definition
|
Items in Original ROP-Scale
|
Items Retained in the Draft COVID-19 ROP-Scale
|
Friendliness
|
How a physician communicates with a patient
|
- Asking patient's name
- Engaging in social talks
- Asking about patient's family
- Friendliness
- Giving courage and reassurance
- Sense of humour
|
- Engaging in social talks
- Friendliness
- Giving courage and reassurance
- Sense of humour
|
Respecting
|
How a physician explicitly shows respect to a patient
|
- Greetings by physician
- Showing respect explicitly
- Listening to patient's complaints completely
- Listening to patient's complaints attentively
- Examining the patient with care
- Encouraging patient to ask questions
- Listening attentively to patient's questions
- Closing salutation by physician
- Non-verbal communication by physician
- Compassionately touching the patient by physician
|
- Greetings by physician
- Showing respect explicitly
- Listening to patient's complaints attentively
- Examining the patient with care
- Encouraging patient to ask questions
|
Informing and guiding
|
How a physician empowers a patient
|
- Suggestions on disease prevention and health promotion in general
- Facilitating follow-up
- Quantity of issues explained and the quality of explanation
- Quantity of issues explained
- Asking patient if s/he understood the explanation
- Explaining the cause of disease to the patient
- Explaining the diagnosis of disease to the patient
- Explaining the prognosis of disease to the patient
- Explaining the treatment to the patient
- Explaining the preventive aspects to the patient
|
- Facilitating follow-up
- Explaining the cause of disease to the patient
- Explaining the diagnosis of disease to the patient
- Explaining the prognosis of disease to the patient
- Explaining the treatment to the patient
- Explaining the preventive aspects to the patient
|
Gaining trust
|
How a physician may gain trust of the patients, or refrains from doing something that may breach trust of the patients
|
- Earning trust of patients
- Service oriented, not business-like attitude
- Not using jargon
- Not being involved in illegal activities
|
- Service oriented, not business-like attitude
- Not being involved in illegal activities
|
Financial sensitivity
|
Understanding financial need of the patients and providing support if needed, going beyond the consultation
|
- Considering the socio-economic status of the patient
- Trying to understand the socio-economic status of the patient
- Informing the cost of treatment
- Providing financial assistance if needed
|
- Trying to understand the socio-economic status of the patient
- Informing the cost of treatment
- Providing financial assistance if needed
|
A team of masters-level students at the Communication Department of a Bangladeshi university served as data collectors. When contacting each of the respondents randomly chosen from the A2i database, they provided a brief explanation of the study objectives and procedures, as well as voluntary nature of their participation, and ensured them of the anonymity and confidentiality of the information gathered through the survey. After obtaining verbal informed consent from the eligible respondents, the data collectors conducted the interview over the phone, which took 20−30 minutes to complete. The respondents answered the 20 responsiveness questions using a 10-point scale, where 1 indicates negativity and 10 indicates positivity. Thus, a higher score corresponds to greater ROP.
Data management and analysis:
All data analyses, including data management, cleaning, missing value imputation, calculation of descriptive statistics, and psychometric analysis were conducted using Stata 16. We performed an exploratory factor analysis (EFA) to measure the factor structure and psychometric properties of the initial COVID-19 ROP-Scale comprising of 20 items. As only 53 and 36 respondents answered the question related to examining patients (item 6) and offering financial assistance if needed (item 9), respectively, these two items were excluded from further analysis. The remaining 18 items were subjected to correlation matrix analysis, which revealed that item 10 (Explaining the cause of the disease to the patient) and item 11 (Explaining the disease diagnosis to the patient) were highly correlated (correlation coefficient value >0.95). Therefore, to avoid the multicollinearity issue, item 10 was excluded, and EFA was performed again on the remaining 17 items. Since some of the values were missing, the maximum likelihood approach with expectation-maximisation (EM) algorithm was used to estimate the covariance matrix (11). The factormat command in Stata was used to obtain the factor solution. Orthogonal varimax option was chosen for factor rotation (12) and Kaiser’s criteria (eigenvalue > 1 rule) were adopted to determine the number of factors to be retained in the model (13). As only items with factor loadings ≥0.4 were retained (14), items 3 (Friendliness of the physician), 5 (Listening to patient's complaints attentively), and 15 (Encouraging patient to ask questions) were excluded. Before EFA, we checked the suitability of data for factor analysis by conducting Bartlett’s test and Kaiser-Meyer-Olkin (KMO) test (15).
Cronbach's alpha was used to assess the internal consistency reliability of all the items and all four retained sub-scales or domains. The corrected item-total correlation was also calculated. Factor 4, with its two items—18 (Service oriented, not business-like attitude) and 19 (Not being involved in illegal activities)—were eliminated at this stage, resulting in 12 items. Concurrent validity of this final version of COVID-19 ROP-Scale was assessed by examining Pearson’s correlation between the COVID-19 ROP-Scale score (i.e., the summed score of all item scores) and overall patient satisfaction, under the assumption that responsiveness would be positively correlated with satisfaction (16, 17). Figure 1 shows the COVID-19 ROP-Scale validation process.