Antenatal Care Services and Preparedness of Health Facilities in Bangladesh: Evidence From Two Nationally Representative Surveys

Objective To assess whether health facilities in Bangladesh are prepared to provide ANC services and to investigate facility characteristics that are linked to preparedness. Methods This cross-sectional analysis used publicly available data from two waves of Bangladesh Health Facilities Survey (BHFS) that was conducted in 2014 and 2017 using stratied random sample of facilities. 1508 and 1506 health facilities were included as study samples from the 2014 and 2017 BHFS respectively. The outcome variable ‘ANC services preparedness’ was calculated as an index score using a group of tracer indicators. Multivariable logistic regression models were used to identify the important correlates of ANC service preparedness The percentage of facilities providing ANC services has increased slightly from 97.4% in 2014 to 98.8% in 2017. Union level public facilities were less likely to be prepared for providing ANC service than district level public facilities in 2014 (Odds ratio (OR): 0.20, Condence interval (CI): 0.08-0.50, P-value=<0.001). Similar results were also found in 2017 (OR: 0.14, CI: 0.06-0.33, P-Value=<0.001). The facilities with high basic amenities score were more likely to be prepared for providing ANC than the facilities with low basic amenities score in 2014 (OR: 2.52, CI: 1.02-6.12, P-Value=0.04) and in 2017 (OR: 1.79, CI: 1.08-2.99, P-Value=0.02). The overall offer ANC poor. ndings not only health but also in produce a and a baby at the end of pregnancy.


Page 3/16
reproductive aged women deaths [13]. The third sustainable development goal (SDG) established a global objective, implying that Bangladesh must reduce MMR to less than70 per 100,00 live births [8]. It is estimated that 90% of maternal deaths and pregnancy associated diseases can be prevented by appropriate medical care at various steps of childbearing di culty [14] and good ANC singly can lessen 20% of maternal mortality. [15] An effective ANC program requires quali ed health care providers in a functional health center with referral facilities as well as adequate supplies and diagnostic capabilities [4]. Therefore, health facilities should be well prepared with regard to trained health professionals, medications, supplies, equipment, amenities, and infrastructure to meet this demand. [16] Several studies in Bangladesh have identi ed the factors responsible for frequency of ANC visit. [15,[17][18][19] For example, mothers level of education, place of residence, administrative division, media access, birth order were the common signi cant predictors of frequency of ANC visit in the published studies. Moreover, other studies have attempted to assess the utilization of ANC service in Bangladesh were limited to rural areas [6, 20,21]). Furthermore, some studies identi ed the level and trends of frequency of ANC visits [22][23][24]. However, the preparedness to provide ANC services by health facilities has never been studied in Bangladesh.
Health facility preparedness is a vital aspect that indicates a facility's commitment to con rm cumulative availability of items needed to offer a speci c service [25]. Assessing health facilities preparedness for ANC service is critical for not only health planning but also decision making in order to deliver a healthy mother and a baby at the end of pregnancy, which ultimately helps in reducing maternal mortality through improving maternal health. A sound understanding of the factors in uencing facility readiness to provide ANC service is important to assist maternal and newborn health system progress. Therefore, the present study sought to evaluate the preparedness of health facilities to provide ANC services in Bangladesh. We also explored facility characteristics associated with the preparedness using two nationally representative surveys.

Study population and setting
This study used publicly available cross-sectional data from two waves of Bangladesh Health Facilities Survey (BHFS) that was

Selection of study samples
Among all facilities that offer ANC services were included as nal study samples. From 2014 BHFS, 1508 facilities were included as study sample and 1506 facilities were taken as nal study sample from 2017 BHFS.

Outcome variable
The outcome variable 'ANC services preparedness' is de ned as the readiness or willingness of the facility to provide ANC services.
The WHO has determined a group of tracer indicators requires to be available for a health facility to be regarded as prepared to provide quality ANC services [25]. In this study, a somewhat smaller con ned and Bangladesh speci c-proper version of the WHO-suggested ANC service preparedness measure was constructed. The outcome variable was calculated as an index score using WHO de nition of service preparedness index where the greater scores imply that the facilities have greater service preparedness (Table 1) [28]. For example, if a health facility has a score of 100, it means that the facility is completely prepared to offer ANC service, or if a facility has a score of 50, the facility has 50% preparedness to offer the service. We created dichotomized variable by cutting off it in the middle.
The facilities who scored 50% or higher were categorized as prepared and not prepared which scored less than 50%. This cut-off point was utilized similar in prior studies. [29][30][31]  The following tracer items such as a regular electricity, improved water source, visual and auditory privacy during consultations, client latrine, communication equipment (landline/mobile phone), and computer with internet access were included in the basic amenities score. Moreover, waste bin, sharps box, general disinfectant, syringes/needles, sterile disposable gloves, hand hygiene, running water, soap/hand disinfectant were included in the infection prevention score. For each facility, the score is equal to the sum of the availabilities (i.e. value = 1) of all the tracer items, divided by the total number of items. The score variables are continuous, and the distribution of scores was classi ed into two equal parts. The facility was regarded as "low" score if mentioned to have at least a half (score of ≤ 50%) and greater than a half (> 50%) considered as "high".

Data analysis
The proportion of ANC service preparedness between the categories of various potential factors was compared using chi-square test.
Multivariable survey logistic regression model was used to identify the health facility characteristics associated with ANC service preparedness. The variable managing authority was not included in the nal model due to very few frequencies in one category. Also, duty schedule and quality assurance activities were not included in the nal model because of extensive amount of missing values. Multicollinearity was assessed and any high correlation between the potential factors was not found. All data management and analyses were conducted using Stata 13 (StataCorp, College Station, TX, USA). To account for the complex survey design, we weighted all our analyses using the weight option in Stata with the sampling weights provided in the dataset. For modeling exercise, we used 'svy' command of Stata to account the survey design, primary sampling unit and cluster.

Univariate analysis
The percentage of facilities providing ANC service has slightly increased from 97.4% in 2014 to 98.8% in 2017. The availability of tracer indicators for ANC service preparedness is shown in Fig Table 1).

Multivariable analysis
The results of the logistic regression model for ANC service preparedness are shown in Table 4. Union level public facilities were less likely to be prepared for providing ANC service than district level public facilities in 2014 (Odds ratio (OR): 0.20, Con dence interval (CI): 0.08-0.50, P-value = < 0.001). Similar results were also found in 2017 (OR: 0.14, CI: 0.06-0.33, P-Value = < 0.001). The facilities with high basic amenities score were more likely to be prepared for providing ANC than the facilities with low basic amenities score in 2014 (OR: 2.52, CI: 1.02-6.12, P-Value = 0.04) and in 2017 (OR: 1.79, CI: 1.08-2.99, P-Value = 0.02). The facilities that did not maintain Individual client cards or records for ANC clients were less likely to be prepared for providing ANC service than the counterparts in 2014

Discussion
In this study, we assessed the preparedness of health facilities including its characteristics associated with the preparedness to provide ANC services in Bangladesh. Since 2014, there have been only small changes in the number of facilities that offer ANC services. The preparedness of offering ANC was evaluated using ve domains of the service. Among the domains, shortage of ANC guidelines, staff training, and laboratory diagnostic capacity were elds of weakness. In line with a prior study [17], the overall ANC service preparedness score was observed to be low. Moreover, we found that facility type, basic amenities score, and individual client cards or records for ANC clients were signi cantly associated with ANC service preparedness in both the 2014 and 2017 BHFS.
A step to evaluate the quality of ANC service is assuring that all facilities have availability of ANC service guidelines for the health workers which would assist to con rm that protocols are followed [32]. However, in this study, such guidelines were lacking in most of the private facilities. Moreover, this study demonstrated that extensive training on any of the ANC topics: ANC screening, counseling, complications of pregnancy and their management, nutritional assessment of pregnant woman, and prevention of mother to-child transmission of HIV was rare across private facilities. These topics are usually contained in-service training given to government facility staffs who is employed in maternal and newborn care [33]. The safe motherhood program in Bangladesh should also set up a system to comprise private facility staff in same training in the future.
In Bangladesh, many mothers suffer from anemia, and it is estimated that about 20% maternal deaths are because of preeclampsia or eclampsia [27]. To diagnose these conditions, hemoglobin and urine protein testing are essential part of ANC which were missing in the most of union-level public facilities and community clinics, as well as some district and upazila-level public facilities ( Table 2).
Absence of diagnostic tests such as hemoglobin and urine protein test can result in tardy diagnosis or non-diagnosis of pregnancy complexity like pre-eclampsia/eclampsia [32]. Thus, it is necessary for all health facilities with preference given to the union-level facilities and CCs to have the capability to perform hemoglobin and urine protein test. Moreover, the ministry of health should give nancial and technical support to the facilities in setting up laboratory diagnostic services.
Our analysis identi ed facility type as a signi cant factor for ANC service preparedness in both the survey years where district and upazila level public facilities are more likely to be prepared for offering the service compared with all other type of facilities. Because the availability of all the tracer indicators for ANC service preparedness was better in district and upazila public health ( Table 2). The suboptimal preparedness found in other public facilities might be because of abstruse policy on how to assign funds in these facilities that may cause shortages and discriminations in the allocation of medical supplies [34]. The fortifying of the union level as well as community clinic facilities which are generally situated in the rural areas and private facilities is greatly required to address the shortage of tracer indicators of ANC service preparedness.
In this study, the basic amenities score was a signi cant factor for ANC service preparedness. This nding is important as the availability of basic amenities (regular electricity, improved water source, client latrine, communication equipment and computer with internet access etc.) is crucial to client contentment with health services provided at a facility [27]. Facilities having high basic amenities score were 2.52 times and 1.79 times more likely to be prepared in 2014 and 2017 respectively for offering ANC service compared to facilities having low basic amenities score. It may be because if a health facility has high basic amenities, the facility is well decorated and therefore are more likely to enhance the availability of services that may end in high preparedness of service.
The ANC card is an essential source of health information which gives every pregnant woman with an individual record of her medical as well as obstetric history over time. The woman is suggested to bring the card with her, whither she move, and to show the card every time she visits any health facility [35]. Our analyses showed that the facilities that did not maintain individual client cards or records for ANC clients were less likely to be prepared for providing ANC service than the counterparts. This may be since facilities did not maintain ANC card or records for clients are less likely to ameliorate the availability of ANC services that are essential aspect to evaluate the readiness of the facility to provide ANC services.
Consistent with the prior studies [16,36], the present study found signi cant regional variation on ANC service preparedness in 2017.
In 2014, almost all divisions were less likely but in 2017 all other divisions were to more likely to be prepared for providing ANC service than Dhaka division. The observed geographical variations in offering ANC services may be due to disparities in health seeking behavior, service availability and quality [15]. Further studies may be conducted to examine the underlying causes behind geographical variation of ANC service among the health facilities in Bangladesh. Nevertheless, we did not observe any signi cant difference by facilities location in accord with a prior study. [16] Visual and auditory privacy is crucial during consultations with health professional since it permits clients to report their problems in detail without reluctance [27]. In this study, we found signi cant association between su cient privacy for ANC exam with ANC service preparedness in 2017 but not in 2014. Now a days, women are more aware about their right of privacy and con dentiality and performs physical exams during their ANC visit in a facility where privacy could be protected. Moreover, we found that facilities having no su cient privacy during ANC exam have 0.42 times lower chance of being prepared for offering ANC service than the facilities having the su cient privacy in 2014. It is recommended that women's privacy and con dentiality in examination and counselling during every ANC visit should be protected to the reasonable extent at health facilities in Bangladesh.
External supervision not only forti es a health system but also gives chance to health workers providing quality services and ameliorating performance [37]. The ndings of this study indicate that facilities that received external supervision were more likely to be prepared to offer ANC service compared with facilities did not receive external supervision prior to 6 months of the survey in 2014.
Therefore, supervision visits are necessary to entail observation of the ANC client examination/counseling at health facilities to con rm that standard guidelines are maintained. [32] We found a signi cant association between visual aid for client education and ANC service preparedness in 2014 but not in 2017, where facilities having visual aid for client education were more likely to be prepared for providing ANC service. Visual aids at Bangladeshi health facilities require to be made available and it should also be checked regularly.
It is crucial that a health facility have infection control equipment and supplies suitable to the services delivered [7]. Our ndings gave evidence that infection prevention score as well as injection safety precaution guideline are signi cant factors for ANC service preparedness in 2014 where facilities having high infection prevention score and having infection precaution guideline have higher odds of being prepared for offering the service than their counterparts.

Strengths and limitations
This study has several strengths. First, to our knowledge this study is the rst of this nature in Bangladesh giving key insight into the preparedness of health facilities for offering ANC service. This study used nationally representative samples of health facilities in Bangladesh in which our ndings reveals essential information about the factors responsible for ANC service preparedness. Second, SPA data are collected using a complex sampling strategy, hence the estimates of this study were corrected for using cluster effect and sample weights. Finally, we conducted a comparative study using the 2014 and 2017 BHFS, hence the changes that has been made in ANC service preparedness among health facilities of Bangladesh can be observed.
Results from this study should be taken into account in the context of some limitations. First, data were collected at a particular point in time and hence, this study is incapable of inferring causality. Longitudinal research is required to better understand the factors associated with ANC service preparedness. Second, this study did not catch provider-level data that would give more idea about the preparedness of care from the provider's perspective [16]. Third, our analysis concentrated on health facility preparedness, which is an essential issues but not an assurance of offering quality ANC services [38]. Finally, although SPAs of other countries give information on other measures of quality based on observations of ANC consultations as well as client exit interviews, the BHFSs only give information on service availability and readiness. Future research is needed to assess quality of ANC service at health facilities in Bangladesh.

Conclusion
Although the percentage of facilities providing ANC services has increased slightly from 2014 to 2017, many Bangladeshi health facilities lack the ANC guidelines, staff training, and laboratory diagnostic capacity. The overall preparedness score to offer ANC service was also poor. There was a disparity in overall preparedness of ANC service provision by facility type. To improve the ANC

Declarations
Ethics approval and consent to participate: NA, as data used in this study is publicly available.
Consent for publication-NA/Not Applicable.
Availability of data and materials: All data is publicly available in dhsprogram website (www.dhsprogram.com) Con ict of interest: There are no potential con icts ( nancial, professional, or personal) to disclose by any of the authors.
Funding: The authors did not receive any speci c grant from funding agencies in the public, commercial, or not-for-pro t sectors.
Authors The corresponding author accepts responsibility as guarantor.