Study design and settings
This was a cross-sectional study of primary healthcare facilities conducted between May and October 2021. Data was collected on the availability of a set of items that are required for NCD services at the primary healthcare facilities including guidelines and staff, basic equipment, diagnostic facility, and essential medicines. The readiness of the healthcare facilities was assessed for the following four major NCDs: cancer, CRIs, CVDs, and DM. The specific NCDs were defined according to the diagnosis of the healthcare providers of the respective healthcare facilities. Although the National Cancer Control Strategy and Plan of Action 2009-2015, focused on the prevention and management of commonly prevalent cancer types (e.g., breast, colorectal, esophagus, lung, cervix, lips and oral cavity, etc.), the provision of all major type of cancers are not available in the primary healthcare level in Bangladesh (32, 33). Cervical cancer has been gradually increasing over the past years and now is the second most prevalent among the women in Bangladesh and subsequently gained greater public health response (34). The National Cervical Cancer Control Program (2017-22) guideline expanded the provision of cervical cancer services at the primary healthcare level (35). Considering the scope of services provision, prevalence, health burden, and public health response, this study addressed only cervical cancer readiness (35, 36).
Study sample and sampling technique
The sample size was calculated using the following formula: (Z2*P*d2)/(V2*P) provided by the Monitoring and Evaluation to ASsess and Use REsults Evaluation (MEASURE) as a sampling manual for the facility surveys (37). The anticipated proportion of the healthcare facilities, with the attribute of interest P = 50%, design effect d = 1.2, and the relative variance (V2) as the square of the relative error taken as 20%, as used by a previous study (38). This calculation yielded the minimum required sample size of 115 healthcare facilities. Anticipating a 10% non-response rate, we surveyed 126 healthcare facilities. The sample was selected by a multi-stage stratified random sampling technique. Bangladesh is divided into eight administrative divisions (39). Each division is further divided into several districts, and each district consists of several sub-districts locally known as upazila. The public healthcare facilities are established under administrative units and operate in the following three levels: primary, secondary, and tertiary (40). The primary care facilities include the Upazila Health Complex (UHC) at the headquarters of a sub-district, union sub-center (USC)/union health center (UHC), family welfare center (FWC) at the union level (hereafter referred to as ‘ULF’ to mean all healthcare facilities at the union level), and community clinic (CC) at the ward level (41). Along with the public facilities, private and NGO-operated (hereafter referred to as ‘private facilities’) health facilities function within the sub-districts. The total number of healthcare facilities and related information was collected from the ‘Facility Registry’ database of the Directorate General of Health Services (42). This study covered 126 healthcare facilities from the following administrative districts of Bangladesh: Cumilla, Jhenaidah, Rajshahi, and Sylhet. Using an electronic structured questionnaire (REDCap), the facility head or management staff member was face-to-face interviewed to collect data.
Inclusion/exclusion criteria
The health facilities were included based on the following criteria: (1) facilities located in sub-districts level, and (2) facilities providing an NCD-related service (prevention or management). The facilities were excluded based on the following criteria: (1) did not meet the inclusion criteria; (2) facilities that had less than six months of service; (3) facilities that provided specialized services at the sub-district level (i.e., a tuberculosis clinic); (4) healthcare facilities that had been temporarily established to address the emergency residents (i.e., camp hospitals).
Data collection team and training
Eight interviewers with a Bachelor of Medicine, Surgery, or Anthropology were involved in the data collection process. Before the interviews, one week of training covering the topic of data collection instruments, filling in the electronic questionnaire, and obtaining information from medical records and data entry into the RedCap software was provided (43). Additionally, the training focused on the contents of the questionnaire, building rapports, moderating interviews and discussions, taking notes, approaching and inviting interview questions, the organization and function of the primary healthcare system, NCD service delivery package, effective communication with the facility head, and scheduling interviews that enhanced efficiency in the quality data collection. The questionnaire had the following three modules: facility identification, general service availability, and disease-specific readiness. The questionnaire was developed in plain English and then translated into Bengali (the local language). The Bengali version was again translated into English to check the consistency of meaning between versions. The pretest for the questionnaire was conducted, and the necessary feedback was accommodated in the final version. The interview was conducted in Bengali.
Participants’ consent
Beginning of the interview, the data collectors informed the participants (e.g. the facility head/ management staff) about the purpose of the study. Then the Explanatory Statement was provided by the data collectors to the participants and allowed them to read and ask questions. Upon their agreement to participate, participants were required to read and sign a consent form. The consent form explained the purpose of the study, the freedom to participate, and how participants' information would be used while maintaining their individual/facility information confidential.
Quality assurance of data collection
To ensure the quality of data collection, the procedure was monitored by the first author throughout the survey. A random consistency check for approximately 5% of the interviewed questionnaires was done by the investigators. A regular group discussion and follow-up meeting were conducted with data collection teams to discuss and share the experience, challenges, and overcoming strategies in conducting interviews. Supportive supervision was provided to the data collectors as required.
Data collection instrument
Data was collected using a facility survey tool that was developed based on the WHO’s service availability and readiness assessment (SARA) manual (WHO-SARA) (44). The WHO-SARA tool was designed to generate a set of indicators to determine whether the facilities meet the standardized requirements of general or specific services with reliable quality. This tool offers indicators to monitor and assess several standardized items, including human resources, basic equipment, supplies and technologies, and essential medicine, which are required for general and NCD-specific service delivery in the healthcare sector. The WHO-SARA is considered a reliable methodology and is widely used to evaluate the readiness of healthcare facilities. The survey questionnaire was designed based on the WHO-SARA methodology, with a slight modification according to the standard set by the Bangladesh Ministry of Health context.
Outcome measures
The outcome variable in this study is the ‘readiness’ of primary healthcare facilities in terms of four specific NCDs. The readiness variable was rated as an index grouped into four domains as proposed in the WHO SARA methodology: (i) guidelines and staff, (ii) basic equipment, (iii) diagnostic facility, and (iv) essential medicine. Each of these domains has multiple indicators which were measured in nominal scales. In the first domain, there are two indicators: the availability of guidelines, and trained staff for every four NCDs, which was categorized as ‘yes’ for facilities with guidelines, and at least one trained staff for each specific NCD and ‘no’ otherwise. In both the second and third domain, there are 15 basic equipment and diagnostic items and each of them was categorized as ‘yes’ for the existing facilities, and ‘no’ otherwise. In the fourth domain, 28 essential medicines (a list from Bangladesh health ministry) were categorized as ‘yes’ with facilities reporting the availability of each specific medicine, and ‘no’ otherwise.
Statistical analysis
Based on the WHO-SARA manual, a descriptive analysis was conducted to define a set of tracer items/readiness indicators for NCDs. The service readiness was assessed into the following four domains: staff and guidelines, equipment and supplies, diagnostic facility, and essential medicine. Based on the SARA tool, the ‘availability’ and ‘readiness’ were categorized as ‘yes’ for facilities with individual items for each specific NCD and ‘no’ otherwise. In each domain, the scores for the tracer items were calculated and expressed as percentage points (0%–100%). The mean availability service readiness was assessed in the following three levels: (i) determining the tracer score items for four major NCDs at each facility level (number of facilities with tracer items*100/the total number of facilities); (ii) calculating the readiness index (RI) of the facility based on the four domains (the mean score of tracer items in each domain); (iii) determining the overall readiness score based on the facility types (the average domain indices for all four domains). The indices were displayed for each of the following facility types: UHCs, ULFs, CCs, and private facilities. These indices were compared to an agreed cut-off threshold of 70%, which means that a facility index below the 70% cut-off was considered not ready to manage NCDs at that level. The summarized scores of the facility-level data are presented as means and standard deviations. The results were summarized and expressed as frequency and percentage. All statistical analysis was conducted using SPSS version 22.
Data storage and management
During the data collection period, the data was saved in the secure REDCap web-based application hosted at Monash University. The application was only accessible by the research team. When the data collection was completed, the data was exported to the IBM SPSS statistical package and saved in the secure faculty-allocated network storage (Monash (S:) drive). Facilities’ identifiers, such as names and addresses, were removed from the main database, saved in a separate secure electronic folder, and not used for the data analysis. Only the research team had access to the electronic databases.