Assessing service availability and readiness to manage cervical cancer in Bangladesh

It was a cross-sectional sub-study of a more extensive parent study called Service Availability and Readiness Assessment (SARA) Survey for NCDs and Disability Service Delivery System in Bangladesh. The study was conducted from December 2017 to June 2018. In this sub-study, variables for assessing the availability and readiness of cancer-related services (mainly focusing on cervical cancer services) were considered.

WHO SARA Standard Tool was used to collect data on the availability of medical equipment related to cancer including their location, functional status, and components of support systems (e.g., logistics, maintenance, and management) [16].

A list of all available public and private health facilities was compiled at the primary, secondary and tertiary levels in Bangladesh using the Management Information System (MIS) and and hospital administration records from the Directorate General of Health Services (DGHS); the National Institute of Population Research and Training (NIPORT) and from the Bangladesh Health Facility Survey (BHFS) 2014 [17, 18]. The list constituted of a total of 19,184 health facilities.

Of these 19,184 health facilities, the parent study of SARA Survey for NCDs and Disability Service Delivery System in Bangladesh selected a sample of 590 health facilities at various levels. These included ten randomly selected tertiary and specialized hospitals for NCDs (i.e., 10%). All secondary level health facilities, that is, 60 district hospitals (DHs) and 61 Mother and Child Welfare Centre (MCWCs), were also included in the sample. Among the primary level health facilities, the sample also included 128 (30% of all) Upazila Health Complexes (UHCs), and 256 Family Welfare Centres (FWCs) and Community Clinics (CCs). Further, we included randomly selected 75 NGOs/ private hospitals.

This sub-study then excluded 267 health facilities from analysis, which included specialized hospitals that did not focus on cervical cancer care (such as those devoted solely to ophthalmology) and facilities where a staff member was not available to complete the survey. The study included a final sample size of 323 health facilities, as depicted in Fig. 1.

The head of each hospital or a management staff member who would have sufficient knowledge of hospital capacity and operations were selected to participate in the survey (Table 1).

Forty-eight interviewers, all Bachelor of Medicine and Bachelor of Surgery (MBBS) graduates, were trained for 2 days on the data collection instrument and data entry into RedCap software using tablets. The interviewers received instructions for uploading the data immediately to the server from each health facility, enabling researchers at the head office to monitor the data collection in real-time. The health facilities included in this study were contacted beforehand and an appointment was set for the interviews. The interviews were conducted in-person and recorded.

Descriptive analyses were performed based on the dataset using the standard core indicators (staff and guideline, equipment, diagnostic facilities, and medicines) and corresponding tracer indicators, as per the reference manual of SARA [19]. There was one variable for assessing the availability and 34 variables for assessing the readiness of cancer related services. Of these 34 variables, four were specific to cervical cancer, which was further analyzed for assessing readiness focusing specifically on services for patients with cervical cancer. Table 2 shows a list of tracer indicators. Data were compiled and analyzed by standard statistical procedures using STATA version 13.

Two outcomes, the “availability” and “readiness” of cancer services, were calculated based upon previous studies that utilized the SARA tool [20, 21]. Service “availability” was assessed by asking if the health facility provided any service for patients with any cancer (yes or no), not specific to cervical cancer. This gave data on the availability of overall cancer services at different levels of health facilities. Service “readiness” was calculated as a composite score using 34 variables and consisted of five stages:

Data was collected on the availability of any type of cancer service readiness tracer indicators at all health facilities.

Tracer indicator scores were calculated for all cancer tracer indicators at each health facility level, using the following formula: number of health facilities with individual tracer indicators present *100/total number of health facilities.

Next, the tracer indicator scores specifically for cervical cancer were extracted (guideline, training, speculum, and acetic acid) which fell under three domains (staff and guideline, equipment, and diagnostic tool). Two of the indicators, guidelines and training, covered the domain of staff and training. Each of the other two indicators covered the domains of equipment and diagnostic tool respectively.

In the next step, the cervical cancer-specific readiness index (CSRI) scores for the cervical cancer-specific domains were calculated for each health facility level. CSRI was done by determining the mean of all tracer indicator scores for each domain (sum of tracer indicator score in each domain/number of tracer indicators in each domain) for each facility level.

The mean cervical cancer-specific readiness index (mean CSRI) score was calculated at each level of health facility by determining the average of the CSRI score of all three domains (sum of readiness index in each domain/number of domains) at each facility level.

Cervical cancer-specific readiness index (CSRI) scores and mean index scores were then compared to a cutoff level of 70%, as evident from a study in Zambia that also used the SARA tool and considered health facilities to be “ready” to manage NCDs when their mean readiness index scores were above the same cutoff score [21]. All analyses were performed using STATA version 13.

A total of 323 health facilities were assessed, where 248 (76.8%) were public and 75 (23.2%) were from the private sector. Facilities included 124 UHCs (primary level) (38.4%), 74 private/NGO hospitals (22.9%), 58 DH (secondary level) (18.0%), 60 MCWC (secondary level) (18.6%), and 7 tertiary level specialized hospitals (2.2%). Most of the NGO and private hospitals are secondary level health facilities. However, only one of the tertiary level specialized hospital was a private hospital. More than one-fourth of facilities were situated in the Dhaka division (Table 3).

Almost three-quarters of the facilities (74%) did not provide any service to cancer patients (Fig. 2). All seven of the tertiary/specialized hospitals provided services for cancer patients, while a few of the UHCs (23.4%) and MCWCs (8.3%) did (Fig. 3). Only about one-third of district hospitals (34.5%) and NGO/private hospitals (31.1%) provided services to cancer patients (Fig. 3).

Frequencies and tracer indicator scores for cancer at each health facility level.

The tracer indicator scores for all 34 cancer tracer indicators at each level of health facility are provided in Table 4.

The tracer indicator scores specific for cervical cancer were 100.0% for guideline, training, and speculum and 85.7% for acetic acid at the tertiary and specialised hospitals. The tracer indicator scores specific for cervical cancer was < 18% for guideline and training, and 62.1% for speculum and acetic acid at the secondary level of health facility, while that at the primary level of health facilities as well as the NGO and private hospitals was < 15 and < 50% respectively.

Cervical cancer-specific readiness index (CSRI) scores at each health facility level: Fig. 4 shows the domain stratified cervical cancer-specific readiness index (CSRI) scores at each health facility level. All CSRIs at the tertiary and specialized hospitals were above the cutoff score of 70% and therefore “ready” to provide cervical cancer services. The CSRI scores of the staff and training domain (guideline and training tracer indicators), as well as the equipment domain (speculum tracer indicator), were 100%, while that of the diagnostic facility domain (acetic acid tracer indicator) was 85.7%. All CSRIs, among all other facilities, were under the cutoff score of 70%. The CSRI scores were much higher for the equipment domain (62.1, 46.7, 46.0, and 32.4%) and the diagnostic facility domain (62.1, 45.0, 42.7, and 18.9%) than for staff and training domain (13.8, 5.0, 9.7, and 8.8%) among the district hospitals, MCWC, UHC, and private hospitals respectively.

Mean cervical cancer-specific readiness index (mean CSRI) scores at each health facility level: Fig. 5 shows the mean cancer-specific readiness index (mean CSRI) scores at different types of health facilities. The tertiary level hospitals have a mean CSRI score of 95.2%, above the cutoff value of 70%. The mean CSRI score of the other levels of health facilities was 46% (secondary level district hospitals), 32.8% (primary level health facilities), 32.2% (secondary level MCWCs), and 20.0% (NGOs and private hospitals), which were all below the cutoff value.

This study’s strengths included health care facilities from rural to urban under public and/ or private support and explored a large representative sample size. On the other hand, since this was a cross-sectional study and the data collected was only valid for the time when the hospital heads were interviewed. Therefore, no trend or change over time could be witnessed in this study. Also, the study could not include all the sample health facilities due to incomplete questionnaire and absence of proper interviewee. While data on the availability of staff and guidelines, equipment for screening, diagnostic facilities and medicines were collected, information on area specific population density was not collected, which was why bed density, facility density, health worker density, and so forth could not be estimated. Further studies to explore the state of HPV vaccine for the targeted population as per national strategy is needed. While working with the SARA tool, the method has some limitations, specially, for cervical cancer services. The SARA tool allowed data collection on the availability of tracer items relevant to VIA, but not on those relevant for other diagnostic methods such as thermocoagulation and cervical conization. Diagnostic and services for cancer are multidisciplinary requiring specific specialization, an issue not adequately focused on the SARA tool to assess the availability and readiness [24, 25]. As such, this study attempted to find the very basics of availability and readiness, while accepting the limitations of SARA methods.