Readiness of public health facilities to provide quality maternal and newborn care across the state of Bihar, India: A cross-sectional study of district hospitals and primary health centres

Introduction: Poor access to quality healthcare is one of the most important reasons of high maternal and neonatal mortality in India, particularly in poorer states like Bihar. India has implemented initiatives to promote institutional maternal deliveries. It is important to ensure that health facilities are adequately equipped and staffed to provide quality care for mothers and newborns. Methods: We conducted a cross-sectional study of 190 primary health centres (PHCs) and 36 district hospitals (DHs) across all districts in Bihar to assess the readiness of facilities to provide quality maternal and neonatal care. Infrastructure, equipment and supplies and staffing were assessed using the WHO service availability and readiness assessment and Indian public health standard guidelines. Additionally, we used household survey data to assess the quality of care reported by mothers delivering at study facilities. Results: PHCs and DHs were found to have 61% and 67% of the mandated structural components to provide maternal and neonatal care, on average, respectively. DHs were, on average, slightly better equipped in terms of infrastructure, equipment and supplies by comparison to PHCs. DHs were found to be inadequately prepared to provide neonatal care. Lack of recommended handwashing stations and bins at both DHs and PHCs suggested low levels of hygiene. Only half of the essential drugs were available in both DHs and PHCs. While no association was revealed between structural capacity and patient-reported quality of care, adequacy of staffing was positively associated with the quality of care in DHs. Conclusion: Examining all DHs and a representative sample of PHCs in Bihar, this study revealed the gaps in structural components that need to be filled to provide quality care to mothers and newborns. Access to quality care is essential if progress in reducing maternal and neonatal mortality is to be achieved in this high-burden state.

We conducted a cross-sectional study of health facilities in Bihar during July 2016 to October 2016. Facility surveys were conducted in block (subdistrict) and district level government-run public health facilities. This study also uses household maternal and child health survey data collected during October 2016 to December 2016 by CARE India. There are 36 district hospitals in Bihar, all of which were invited to participate in the facility survey. There are 534 blocks (subdistricts) in Bihar, 190 of which were sampled for the facility survey. The number of blocks vary widely per district. Hence, blocks were sampled proportionally according to the total number of blocks per district. The selected sample had blocks ranging from one to nine per district with a median of six blocks. Each block contains one block PHC, all of which (from the 190 sampled blocks) were included in the facility survey. Household survey data were collected using five different questionnaires for mothers who had a child belonging to the following five age groups: (i) 0 to 2, (ii) 3 to 5, (iii) 6 to 8, (iv) 9 to 11 and (v) 12 to 23 months old. A mixed sampling methodology of population based-estimation and lot quality assurance sampling (LQAS) (a small sample survey design based on binomial distribution) was used.23 The sampling ‘lots’ in this survey were the blocks/subdistricts. All 534 blocks in 38 districts were included in the study data collection. The number of anganwadi centres (AWCs, village level institutions providing basic healthcare services) sampled from each block was determined using proportional allocation, however if this resulted in a sample of less than 19 AWCs, then 19 AWCs were sampled in order to meet a minimum sample threshold per block. The sampled AWC were selected within each block using simple random sampling. Five households per AWC were selected, with one each from mother of following five age groups: (a) 0 to 2, (b) 3 to 5, (c) 6 to 8, (d) 9 to 11 and (e) 12 to 23. In total, 15 667 AWCs were selected ranging from 19 to 123 per block. Within each sampled AWC catchment area, households were identified through systematic sampling.23 Briefly, an index household was chosen within each AWC catchment area using a random number table. Starting with the index household, data collectors visited every fifth household looking for eligible mothers. This approach aimed to obtain a wide distribution of households (minimising the effect of clustering), while remaining feasible and practical for data collection purposes. The pilot phase of the study did not observe any significant differences in household characteristics when alternative sample intervals of 10th, 15th and 20th households were selected. The data collectors continued moving in a circular manner, following the ‘right-hand rule’, until five eligible households had been interviewed per AWC catchment area, one household for each age group questionnaire. To reduce the recall bias, data on quality care presented in the analysis were restricted to mothers with children aged between 0 to 2 months. Of the mothers who also delivered at the DHs or PHCs that were covered in the facility survey (ranging from 1 to 17 mother per facility) were included in this analysis. Data were collected using a standardised structured survey tool designed based on the Service Availability and Readiness Assessment tool developed by the WHO and the United States Agency for International Development.24 The tool was modified for the Indian context using the Indian Public Health Standards (IPHS) guidelines.25 26 To evaluate the structural capacity of the facility, the availability and condition of infrastructure, equipment and supplies in different departments, including the labour room, newborn care corner, immunisation room, laboratory, operation theatre, drug store and data operation were assessed. Information on infrastructure and equipment was collected through interviews with the facility-in-charge and staff nurse as well as through direct observation. The pharmacist or drug store-in-charge was interviewed, and the responses were validated through the drug register to collect information on supplies availability. The medical officer in charge (MOIC) at the PHCs and hospital manager at the DHs were also interviewed to obtain information on the number of health personnel employed at the facilities and the number of personnel that were sanctioned (number of staff expected to be employed) to the facilities for each of the health cadres, including medical officers (MOs), staff nurses, auxiliary nurse midwife (ANM), laboratory technicians and pharmacists. This information was also cross-checked with the facility registers. Availability of 30 services related to family planning, safe delivery, antenatal care and neonatal and child care was assessed and the reasons for unavailability were asked from the MOIC in PHCs and the hospital manager in DHs. Three pilot tests were conducted in the facilities outside the study sample to refine the survey tool and to train the enumeration team. The survey was conducted by 60 enumerators over the 4 month period. Enumerators all had prior experience in conducting facility surveys and received further training over 10 days on using the study tool and conducting this survey. Periodic data checks for completeness and outliers were conducted by a data management team in Patna, Bihar. Where information was missing due to absenteeism or lack of time provided by the respondent, a second visit to those facilities was organised. One-to-one interviews were conducted with consenting and eligible mothers by trained data collectors, using a standardised questionnaire and following standard operating procedures. Information collected from mothers and of interest to this study included the household characteristics, the place of delivery and care received at the place of delivery. Patients were not involved in the study. Data analysis was conducted using Stata V.13 (Stata Corporation, USA). The current status of the facilities was assessed on three broad parameters, namely, the structural capacity, staffing and the quality of care provided at the facilities. The structural capacity of the facilities was assessed by computing readiness scores of 0 to 1 for infrastructure, equipment and supplies. ‘Infrastructure readiness’ included the availability as well as the condition of different components, wherever applicable. For equipment, ‘readiness’ implied the availability as well as functionality of the equipment and for supplies, readiness was defined by availability.24 Infrastructure readiness of the facilities included nine broad components (such as power, water, transport, handwashing stations) at the PHCs.24 An additional three components (availability of different rooms, computer and internet and blood bank) were assessed for DH infrastructure score (details of components are listed in online supplementary table S1). bmjopen-2018-028370supp001.pdf The equipment readiness of the facilities was assessed by scoring the availability and functionality of 48 essential (according to IPHS guidelines) maternal and newborn health equipment (items listed in online supplementary table S2). A score of 1 was assigned if the equipment was observed to be available and in a functional state. In case of unavailability or available but not functional equipment, a score of 0 was assigned. Similarly, supplies readiness was assessed by considering the availability of 76 essential maternal and child health drugs that were expected at the facilities as per the IPHS guidelines and contextualised based on the requirements in Bihar (listed in online supplementary table S3). The mean across the three components of infrastructure, equipment and supplies was computed to generate a score for structural capacity ranging from 0 to 1 per facility. The mean across facilities was computed to get an overall score for structural capacity. Detailed methods of scoring have been provided in the online supplementary data. We assessed the availability of human resources by computing the ratio of filled to sanctioned positions, as reported by the MOIC and the hospital manager or equivalent authority in charge in the PHCs and DHs, for each health cadre in each facility. The ratio of total filled to total sanctioned positions for permanent staff, combining all cadres, was computed to generate an overall staffing index for each facility. The availability of health staff was also compared with the essential requirements mandated by IPHS guidelines. In PHCs, we considered staff requirement based on the monthly delivery load of more than 20, as provided by the IPHS guidelines.25 In DHs, the staff requirement based on the bed strength were rounded down to compare with the mandated guidelines.26 For instance, for DHs with less than or equal to 200 beds, we considered the staff requirements for 100 beds as defined by IPHS guidelines. For ANMs, the IPHS requirement of 0.45 staff per bed was considered. (online supplementary table S4). The relationship between availability of services (that were unavailable in at least 10% of the PHCs and DHs) and structural capacity and staffing index was explored by assessing the pairwise correlation coefficients between the indices at the facility. Our primary aim was to describe the structural readiness of facilities to provide essential maternal and newborn services. We also conducted analyses of household survey data to explore the quality of care at facilities as reported by women who both participated in the household survey and delivered at study facilities. Each mother was asked 11 questions during the household survey pertaining to the treatment and care that they and their newborns received during delivery. Each question was assigned a score of 0 (not performed/don’t know) or 1 (performed). Household survey data was merged with facility data by matching the names of facilities where mothers delivered with the facility names collected during facility assessment survey. A quality of care index for each PHC and DH was generated by taking the average score of the 11 questions for all those household survey participants who delivered within the facility. All data were assessed at the facility level. The relationship between structural capacity, staffing and quality of care indexes were visually explored using scatter plots and trend lines as part of this exploratory analysis. Ethical approval was granted by the Indian Institutional Review Board. At each facility, the purpose of the study was explained and informed consent was obtained from the MOIC and the hospital manager or equivalent authority in charge in the PHCs and DHs, respectively. For the household survey, ethics approval was obtained from Ashirwad Ethics Committee, Ashirwad Hospital and Research Centre, Ulhasnagar, India, and informed consent was taken from the mothers.

• kaur, japneet
• franzen, samuel r.p.
• newton-lewis, tom
• murphy, georgina a.v.