“There Is No Link Between Resource Allocation and Use of Local Data”: A Qualitative Study of District-Based Health Decision-Making in West Bengal, India

Background: Effective coordination among multiple departments, including data-sharing, is needed for sound decision-making for health services. India has a district planning process involving departments for local resource-allocation based on shared data. This study assesses the decision-making process at the district level, with a focus on the extent of local data-use for resource allocation for maternal and child health. Methods: Direct observations of key decision-making meetings and qualitative interviews with key informants were conducted in two districts in the State of West Bengal, India. Content analysis of the data maintained within the district health system was done to understand the types of data available and sharing mechanisms. This information was triangulated thematically based on WHO health system blocks. Results: There was no structured decision-making process and only limited inter-departmental data-sharing. Data on all 21 issues discussed in the district decision-making meetings observed were available within the information systems. Yet indicators for only nine issues—such as institutional delivery and immunisation services were discussed. Discussions about infrastructure and supplies were not supported by data, and planning targets were not linked to health outcomes. Conclusion: Existing local data is highly under-used for decision-making at the district level. There is strong potential for better interaction between departments and better use of data for priority-setting, planning and follow-up.


Introduction
Decision-making in health systems involves stakeholders reaching consensus on a particular course of action from two or more possible options to address health service challenges. At the district level, local decision-making depends on autonomy over resource allocation and in the planning process [1]. Devolution of administrative powers for the department of health services can result in giving district-level decision-makers a higher degree of autonomy over finances, service distribution, human resources and governance [2]. In contrast, in countries with a centralised decision-making process, local needs might not adequately be reflected in resource allocation [3]. One of the key aspects of a decision-making process is to have a well-structured strategy for problem recognition and building consensus among all stakeholders towards a solution [4,5]. A coordinated process of decision-making The CMOH also supervises the functioning of the National Health Mission's District Programme Management Unit, which assists in preparing plans, an annual budget and in maintaining accounts. CMOHs are supported by three deputy chief medical officers whom each have responsibility for specific areas, such as health administration and tendering; communicable and non-communicable diseases, reproductive and child health services, and the HMIS [26].
In a district, there are several government departments which provide health services indirectly, in addition to the Health Department itself services. For example, the Department of Women and Child Development has a mandate for nutrition for mothers and children, the Department of Rural Development works towards hygiene and sanitation programmes and the formation of Village Health, Sanitation and Nutrition Committees, and the Department of Education is responsible for the health of adolescents.
The District Health Society (DHS) is the primary district-level decision-making forum and facilitates inter-departmental convergence. It is constituted as part of the Health Sector Reform Programme and is the highest policy-making body at the district level, with responsibility for planning and managing all health and family welfare programmes ( Figure 1).  In the Indian context, with the official endorsement of autonomous district planning processes involving multiple departments working towards a common goal, and data available for evidence-based decision-making, critical appraisal of how local decision-making happens is important. This qualitative study assessed the health decision-making process at the district level in West Bengal, the extent to which local data were used for decision-making, planning and resource allocation related to MCH services. This paper adds to the existing literature of gaps and challenges in decision-making processes in LMIC settings, particularly the extent of local data-use for resource allocation for maternal and child health.

Methods
Qualitative data were collected in two districts in the State of West Bengal between June and October 2015: the qualitative approach enabled us to make insights into the decision-making process and the extent of data use decision-making. The data collection methods included observations of decision-making meetings, in-depth interviews with key informants and content analysis of templates from the study districts that are used to collect data for the HMIS (Table 1). We developed an observation checklist to understand how the DHS meetings were conducted; how interaction happened between departments, in terms of data sharing; and how decisions for planning and resource allocation were made. The interview guide included questions on the health administrative process for decision-making, interactions between different government departments for decision-making and planning, the use of data for decision-making and factors deciding resource allocation in a district. The instruments were translated into Bengali and pre-tested in the study area.

Source of Data Sample in Two Districts
Observations District decision-making meetings. 4

In-depth interviews
Respondents from the Health Department 16 Respondents from government departments that provide indirect public health services 6 Respondents from the district administration 2 Collection of data templates, which contribute to HMIS  [27].
Within each district, respondents were selected for semi-structured in-depth interviews from officials involved in the planning process and members of DHS, in consultation with the heads of the health administration and the Chief Medical Officer of Health. The sample included representatives of the health administration and members of the DHS from the public departments, including Health, Women and Child Development, Rural Development and Education, and the district administration, including the District Magistrate. The in-depth interviews were conducted until no significant new responses emerged. The final sample included 24 respondents (Table 1). The interview guide included information on the functioning of the DHS, records of proceedings of DHS meetings, the composition of the DHS, the role and scope of various departments for convergence and improved decision-making, and the maintenance, flow and utilisation of data elements in the district health information system. Open-ended questions were used to capture the information about these aspects. The instrument, developed in English, was translated in Bengali, the local language of the study area. It was pre-tested before finalisation. Trained researchers proficient in English and Bengali, and having knowledge of the local health structure, conducted in-depth interviews. Most of the interviews were conducted in English, and few in Bengali, which were transcribed and translated to English.
Decision-making meetings of the DHS were observed to gain insights into the planning process, the interactions between stakeholders for decision-making and the use of data for planning and resource allocation. Two DHS meeting were observed in each study district. Beyond providing an initial introduction, the research team remained unobtrusive in these meetings.
Given the primary interest to capture the experience of administrative decision-making at the district level, we used a general phenomenological perspective in the development of the semi-structured data-collection guides and analysis [28]. The research team visited different health system levels in the districts to meet the data providers, data managers and programme officers and collected templates of all HMIS and other programme data forms they maintained, both paper-based and electronic to understand the level of compilation.

Data Analysis
We used triangulation of in-depth interviews, observations and content analysis to highlight how local decision-making for health happens at the district level. This triangulation of the three qualitative data collection method allowed us to cross-check and verify findings from each method. The first level of triangulation focused on the existing process, for which data from the in-depth interviews were cross-checked with observation notes from decision-making meetings. Further triangulation was undertaken to cross-check the observation notes from decision-making meetings with the content analysis of the data templates, to show the availability of data in HMIS and whether they were used in decision making.
For all but three of the interviews, respondents gave consent for audio recording. The recordings were transcribed verbatim, and those in Bengali were translated into English. Data were analysed manually by the first author, in collaboration with the other authors, by going through the meeting observation notes and transcripts. Initial a priori codes were identified, to which were added emerging themes from the transcripts. The final analysis included two themes and sub-themes. Theme one highlighted the decision-making structure of health districts and the sub-themes focused on (1) the planning and resource allocation process; and (2) interactions between the health administration and other departments for decision-making. Theme two highlighted the use of data for decision-making. The sub-themes looked at (1) data availability and sharing on maternal, neonatal, and child health between the health and non-health sectors; and (2) the extent of data use for planning in the district decision-making platform.
A database of all the data forms that are maintained at the district level was created using Microsoft Access. Each data form was given a unique number and categorised based on its source, the level the form was maintained, data collected and frequency of reporting. Content analysis of the data type in each form was conducted, according to WHO health system categories of service delivery, contextual factors, medical supplies, workforce, governance and finance, to capture the type of data available for different health system levels, the level of data sharing and the flow. We defined a data element as a recorded event or unit in a data collection form. Methodological details of the content analysis are described elsewhere [23].
Ethical approval for the study was obtained from the London School of Hygiene and Tropical Medicine (reference 6088) and the Health Ministry Screening Committee in India (HMSC/2012/08/HSR).
Written permission from the State Secretary of Health was obtained, and the cooperation of the heads of district health administration was sought before data collection. All respondents provided written consent and where they gave permission, the interviews were digitally recorded. The anonymity of identity and confidentiality of information was maintained during analysis.

Results
The following section describes the decision-making process and use of data for decision-making at a district level.

The Health Decision-Making Process in Districts
Planning and resource allocation process: Decentralised planning is one of the pillars of India's Health Sector Reform Programme that was initiated in 2005, creating a decision making space for district-level staff. Resource allocation is based on a District Health Plan, which each district is expected to prepare annually, but districts in West Bengal develop the plans every three years. District plans are submitted for state approval before being sent on to the national level. In districts, there are regular meetings such as a Public Health meeting on 10th of every month, a Reproductive and Child Health review meetings on 20th of every month, and regular District Task Force (polio) and immunisation meetings to discusses the proposals received from the different health units within the district. The District Programme Management Unit coordinates the planning, troubleshooting and preparation of the District Health Plan and budgeting of various health programmes, combining local needs with the state government's guidelines and priorities.
Respondents mentioned that often that health plans are structured around the State and Central Government's core agenda of health programmes. The shrinking of the decision making space is illustrated by an imbalance between local health needs and resource allocation: despite local health needs often being reflected in the plan documentation, the resource allocation depends on priorities within the State health programmes. Planning remains top-down, and priorities change as and when the Government changes.
"Bottom-up approach should be adopted while making district health plans . . . Suggestions from the community can be considered and discussed . . . But we have to adhere to the priorities set by (the) Government of India and State Government."

(Health Department Representative)
Respondents mentioned that the idea of decentralised planning had been protected to some extent by the DHS, as it manages some flexible untied funds.
"Sometimes, issues related to funds shortage for implementing programmes can be taken care (of) by DHS . . . District specific useful ideas which need funds can be decided at the DHS."

(Health Department Representative)
Interactions between different departments for decision-making: The CMOH has interactions with other departments, such as Education, Women and Child Development and Rural Development, during the meetings. However, their contribution to decision-making is limited because programmes are already planned according to the State Government's health agenda. Respondents from district departments mentioned that although inter-departmental interactions are meant to take place for programme planning and data sharing, there are no discussions regarding financial expenditure. This is due to each department spending funds on their own needs, which have already been approved.
"Interactions with other departments is only need based . . . otherwise, there are no such regular interactions other than DHS forum."

(Health Department Representative)
The respondents emphasised how the DHS, which is supposed to facilitate inter-departmental convergence and to look at district health needs holistically, has not succeeded in bringing all the departments together to make a comprehensive district health plan.
"Our department is not getting much importance in DHS meetings. One representative from our end just attends the meeting and is not aware of DHS functions... and the health department is also not taking the initiative to motivate us . . . Our role is poorly defined." (Other department Representative)

Use of Data for Decision-Making
Data availability and sharing: The content analysis of data available in the study districts showed that there were 94 forms in which health data were being collected, with 6170 data elements. The Health Department maintained 78 data forms, with 3814 data elements, whereas departments providing indirect health services, maintained 16 forms, with 2356 data elements. The Health Department maintains data in both paper-based and electronic formats. Data was collected in paper-based registers at the community level, which was reported to the sub-district (block) level, where data were compiled and then sent to various divisions of the Health Department in the district, such as the immunisation, school health, maternal health, communicable and non-communicable diseases divisions.
In a district, data flowed from level I (community: sub-center) to level II (sub-district: primary and community health centers), from where the compiled forms moved to level III (district: central). The Health Department maintained two online portals to keep track of the indicators needed for health decision-making. Facility performance from sub-center to district hospital was managed through HMIS, and the Maternal and Child Tracking System (MCTS) focused on indicators about reproductive, newborn and child health services.
"HMIS is a structured format with (a) specific set of columns (indicators)... all data coming from different divisions can't be uploaded on HMIS . . . MCTS are specifically reproductive and child health-based data portal, rather than for other public health programmes....."

(Health Department Representative)
The other departments-maintained data in a paper-based format and data sharing with the Health Department could be seen at the sub-district level, where shared indicators included immunisation status and child growth and malnutrition rates. At the district level, there was no policy or process to share data at regular intervals between departments. Although inter-departmental convergence was the mandate of the DHS, no data was shared for joint planning and decision-making.
"Data sharing between Health and the Department of Women and Child is a major challenge. There is no concordance between these two departments. At the district level, data sharing should be mandatory at DHS." (other department Representative)

Extent of Data Use for Planning and Decision-Making
Triangulating the findings of the in-depth interviews with observations of the DHS meetings, revealed that there was limited data sharing, presentation or discussion based on health indicators. Even if such a debate happened, it was mostly on service delivery such as institutional childbirth and immunisation rates and was based on available data. Generally, the issues discussed and decisions at DHS meetings related to allocating funds for infrastructure development and the need for health awareness and training programmes, and were not backed by the available data.
"Yes data is useful for planning (e.g., bed occupancy rate). When (the) Mission Director of NHM (National Health Mission) visited this hospital found bed occupancy rate at 130%. Then we send the proposal of increasing beds in maternity ward from 85 to 120 and was discussed in DHS meeting..."

(Health Department Representative)
Discussions on the construction and renovation of the primary health centers and the requirement for additional beds were based on needs as perceived by the district officials, and not supported by data on infrastructure, despite relevant indicators are available and maintained as part of HMIS. Moreover, there is no further analysis of data on linking infrastructure, human resources and supplies data to health outcome indicators, which is a gap area while making a plan. Similarly, data were not analysed and used to discuss programme monitoring or follow-up of the action plan.
"We are not monitoring our programmes on the basis of our own data... we are not utilising the data... in fact, we are not benefiting from the large volume of data that we are collecting."

(Health Department Representative)
"Data is very much useful while preparing (the) district health plan. Data supports us every time, but it is also true that due to lack of time and inadequate manpower... it is not utilised. Data is a fascinating tool if we use it properly."

(Health Department Representative)
There was no utilisation or sharing of data while taking decisions to allocate funds to programmes that were not specified in the District Health Plan. DHS planning is top-down because it follows State and central Government agendas, and local health demands, evident from the data, were not linked to resource allocation.
"There is no such link between funding and data; in my personal opinion, funding is particular (predefined state guideline) and never linked with data."

(other department Representative)
We also triangulated the direct observation with the content analysis of information systems that were available in a district (Table 2), to look at the feasibility of using data for planning, decision-making and resource allocation. MCH-related issues that were discussed in DHS meetings included service delivery issues, health outcomes and infrastructure and supplies. In the four observed DHS meetings, 21 issues were discussed, and action plans were developed, yet decisions on only nine of these issues were based on data. In contrast, our content analysis showing that data for 20 of the 21 issues were maintained and collected at the district level as part of HMIS and MCTS. In addition, the Department of Women and Child Development also maintains data related to malnutrition, which is available at the district level, but not shared with Department of Health, and thus not used during the decision-making meetings.

Discussion
This study shows that districts have a decision-making structure, including representatives of all relevant government departments, and resources available to support local plans. However, a well-defined decision-making process is lacking, and there are limited interactions between departments for formal data sharing. Triangulating the observation data with that of the content analysis showed that less than half of 21 issues discussed in the district decision-making meetings were based on data, despite relevant HMIS data being available.
Aligning with local priorities is limited by various factors, including the limited use of health data for making plans and decision, lack of coordination among decision-makers and lack of autonomy [1,29]. Decentralised planning is a cornerstone of the Health Sector Reform Programme in India, and aims to empower local district health administrators to make plans based on local health needs. Studies have shown that within a decentralised structure, administrators can make locally relevant plans which can benefit their communities [2,29,30]. This study adds to a body of literature on decision making space: district-level managers are not able to make an efficient allocation of resources as per local health needs [10]. Still, with some degree of flexibility in terms of the resources available to the district administration, local needs can be addressed. At district level in India, there are many other departments apart from the Health Department that provide public health services and work towards achieving common health outcomes in terms of improvement of service coverage for maternal, neonatal and child health care [22]. To have a holistic decision-making process, the DHS provides space for formal interaction between the multiple players. If this formal interaction could be extended to include data sharing across departments, it would help to align the resources available for community health needs and avoid duplication of effort. Hence, an effective process can contribute to developing a holistic health plan at the district level. In similar settings, district staffs have been reported to have inadequate training and understanding of HMIS, and this is a lack of institutionalisation of HMIS, resulting in a lack of integration into everyday activities [16,22,30].
We found that although data are available across district level departments, they were not analysed and used for decision-making. This wastes resources, if data that has been collected is not utilised. There are further ethical implications, as the study has shown a loss of accountability opportunities, inefficiency in the utilisation of the health system resources and ineffective availability of services, hampering fairness and equity [30,31]. This is similar in other LMICs settings where utilisation of local health data for planning is often sub-optimal [30][31][32]. Studies have shown that timeliness and appropriateness in data availability often result in non-utilisation [31]. Moreover, health-care decision-makers have to deal with a large volume of evidence and may or may not have the capacity to use this evidence in decision-making [33]. HMIS provides opportunities for data-based decision-making, particularly within a decentralised health system where local data can help in setting district health priorities and planning, resource allocation and introducing new services or improving existing service delivery as per the needs of the local population [34][35][36].
This study highlights how the existing decision-making forum at the district level could be strengthened through well-defined structured processes, better coordination between different departments and formal sharing of data to develop a holistic health plan. The study result shows the need for contextualised guidelines and job aids, which can enhance data utility alongside formalising interaction between departments. This system-strengthening initiative could support the local administration in evidence-based decision-making, planning and resource allocation based on local health priorities.
This study has limitations in terms of geographical representation, having been conducted in two districts of one state in India, although care was taken to select typical districts with respect to Indian health systems. The study focused on decision-making for maternal, newborn and child health, which may differ from decision-making for non-communicable and infectious diseases. In addition, interactions for joint planning, resource and data sharing between the Government and the private sector were not captured, because the private sector is not represented in the DHS. Observer bias may also have been a factor when recording the proceedings of the meetings, but care was taken to match observations with the minutes of the meeting to reduce the likelihood of bias. Moreover, during the time frame of the study, only four meetings were conducted in the districts, which limited our sample for observations. Lastly, while conducting the content analysis of data available at the district level, the study did not assess the quality of the data that are maintained as part of the HMIS.

Conclusions
With a decentralised administration and financial autonomy, and the availability of local MCH data, existing forums at the district level in India have strong potential for playing a central role in evidence-based decision-making for planning and resource allocation. The study further highlights the need for a well-defined decision-making process, which could be achieved through capacity-building using sound guidelines and job aids that can facilitate collaboration and data use among stakeholders. This type of package could provide a platform for district health administrations to undertake situation analysis, engage stakeholders, prioritise, develop action plans and follow-up the action plans based on local data.