Theoretical Framework
Our theoretical framework integrates socio-ecological systems thinking with a practical implementation science framework and tools from health improvement science. Exploring the scale-up of the ‘24/7 BHU’ initiative from a systems perspective implies looking carefully at what happens when the initiative is rolled out across the province. It means exploring the system elements that interplay with the intervention components and the processes that ensue - particularly those that amplify or dampen intended mechanisms - and the potentially varied trajectories and outcomes across local settings. Such a framework draws attention to proximal, intermediate and distal processes and to both the ‘hardware’ and the ‘software’ of the health system. So, while mapping the core health system ‘building blocks’ 19 is essential, so too is understanding the shifts that ensue in the power, roles, relationships, understandings, and motivations of key actors within the system as the scale-up proceeds. Furthermore, socio-cultural values and norms are key system elements to consider.20,21 Health services are social institutions, culturally embedded and politically contingent, replicating the norms and ideologies of wider society.22 Importantly, much prior work in Pakistan underscores the impact of the gender order on the health ecosystem, shaping the understandings and behaviours of managers, practitioners, patients and family members.23,24 Similarly, socioeconomic and caste (zaat) hierarchies are relevant, yet rarely addressed in health services research and policy development in Pakistan. A systems approach also recognises that interventions and initiatives are not fixed over time and place, but rather change occurs both because of deliberate policy direction and via interpretation and ‘mutation’ at local level.
Systems thinking is challenging and the various elements and processes within the health ecosystem that are relevant to understanding the scale up of the ‘24/7 BHU’ initiative are potentially overwhelming. We have therefore selected an implementation tool – the revised version of Promoting Action on Research Implementation in Health Services (PARiHS) - to support the engagement of government stakeholders in the analytical process. PARiHS can be a useful tool, providing a structure within which to make sense of the complexity of implementation, determine critical factors and processes affecting success, and identify actionable strategies for improvement.25 PARiHS identifies three interacting elements that shape implementation. Evidence [E] refers to "codified and non-codified sources of knowledge" that key stakeholders have; Context [C] refers to characteristics of the environment in which the initiative is implemented and Facilitation [F] refers to the ways in which deliberate support may promote implementation (e.g. helping people to shift attitudes, skills and ways of working). A helpful operational guide to PARiHS has been produced which sets out a comprehensive framework of prompting questions through which to interrogate the unfolding nature of an initiative.25 The tool therefore provides a useful starting point and has been found to be intuitively appealing to policy-makers25 who may naturally view any new initiative as separate from the context into which it is introduced and focus their attention on the resources and actions they view as within their own remit. We will use the PARiHS framework to consider the complex, dynamic and iterative interactions between the three elements (E, C and F). We will begin to map their inter-relations and identify the ways in which they shape, and are shaped, by the wider system. In line with system perspectives, this mapping process will involve a temporal element, exploring the scale up ‘journey’ over time. In so doing, we will be mindful of the possibility that the ‘24/7 BHU’ initiative may change the boundaries of the health system (e.g. by setting up relationships with new community actors) and may produce unanticipated effects (such as the displacement of activities). We will also consider that factors that support or undermine roll out in the early days (e.g. fit between the initiative and provider role understandings) may differ from those that are sustaining the initiative over time. Similarly, objectives and outcomes of the initiative may need to be revised prospectively.25
In the later stages of the project, we will combine our systems thinking and the PARiHS framework with tools from health improvement science - PDSA cycles – to plan, pilot and evaluate modifications to the ‘24/7 BHU’ initiative. PDSA is a pragmatic, structured learning approach to testing changes in complex systems that recognises the influence of local settings and unpredictability of outcomes.26, 27 This approach puts policy-makers and practitioners at the centre of the improvement process.28
Detailed project description
The project consists of two overlapping work packages (WP). WP1 will explicate programme theory, assess performance, describe how the initiative plays out in different settings, and identify system processes influencing scale-up. WP2 will take this integrated knowledge to develop and pilot improvement approaches. An integrated knowledge translation approach29 will engage key policy and practice stakeholders throughout the project (Figure 1).
Work Package 1
A mix of quantitative and qualitative research methods will explicate the theory of the ‘24/7 BHU’ initiative, examine its operationalisation across varied local settings and develop a system model of the processes that are shaping scale-up. WP1 will consist of three modules. Data will be generated at BHU, district and directorate levels. Directorate level data will be collected from Provincial Department of Health in Lahore. District and BHU-level data generation will focus on those 19 Punjab districts that have implemented the 24/7 initiative for two years or more and within which at least 20 BHUs have been upgraded (i.e. those where it can be considered that the initiative is underway).
WP1-Module 1: Programme theory explication and initial system modelling
Addresses research question:
Q1: What is the programme theory underlying the ‘24/7 BHU’ initiative? What assumptions and understandings of the wider system are embedded in the theory?
We draw on Rossi’s definition of programme theory as a “construction of a plausible and sensible model of how a program is supposed to work” and “provides causal links between the operation of the program and its intended effects”(p134).30 This module will explicate, in all its complexity, the elements and processes of the ‘24/7 BHU’ initiative. It will articulate the design of the scale-up, the processes, inputs, outputs, outcomes and costs of the initiatives. In addition, it will begin to map out the wider socio-ecological system – to situate the deliberate activity of the initiative within a wider set of elements and processes. We will start to consider the complex dynamics of the initiative as it is rolled out, for example, being open to the possibility of simultaneous pathways to impact or alternative causal strands as the initiative plays out in different local settings. We will also look for evidence of reinforcing and dampening feedback loops.31
Data will be generated through a document review and interviews, both informed by the conceptual framework described above. Documents to be reviewed include: grey literature related to the ‘24/7 BHU’ initiative; Punjab Health Department policy and planning documents; PSPU monitoring reports; and minutes of provincial and district-level meetings. As aspirational documents, policy and planning documents are a key source of ‘deliberate and conscious statements of policies and strategies at particular points in time and can at the very least be regarded as public avowals of commitment to certain objectives and even values’ (p261).32 A semi-structured template will be designed and employed systematically to extract content across the documents.
To further refine the programme theory, a series of theory-driven in-depth key informant interviews, individually and in groups of 2-4, will be conducted with 20-25 provincial and district level policymakers and programme managers. Respondents may include Director-General Health, Directors of the ‘24/7 BHU’ implementing unit, the Integrated Reproductive, Maternal, Newborn & Child Health and Nutrition (IRMNCHN) program, and the Policy and Strategic Planning Unit. Key personnel working for international organizations and donors directly associated with the ‘24/7 BHU’ initiative, specifically World Bank and the Foreign, Commonwealth and Development Office (FCDO), will also be interviewed. As the analysis proceeds it will identify a small number of respondents who are particularly engaged in the process for follow-up interviews to refine and consolidate the analysis. Interviews will engage respondents in critical interrogation of the initiative and its performance, drawing on the PARiHS framework where helpful and ensuring sufficient openness for respondents to refine, refute and re-define the emergent theory and system model.
Data management and analysis: Documents and interview transcripts (recorded subject to respondent consent) will be uploaded to the computerised textual analysis package Quirkos.33 Analysis will involve an iterative process of theory gleaning, theory refining and theory consolidation. First, interpretative analysis of the documentary sources will be used to develop a draft formulation of: programme inputs, outputs and outcomes; causal processes linking these elements; and any contextual factors identified as potential moderators of these relationships. This analysis, supported by the PARiHS framework, prior evidence on the Pakistani maternal healthcare system, and investigators’ extensive contextual experience, will inform development of the interview guide. Following completion of the interviews, analysis of transcripts will first focus on extracting segments that speak directly to elements of the draft programme theory, re-drawing, confirming or refining it, in some way. Subsequently, we will map evidence on inter-relationships between programme components, wider influences, feedback loops and unanticipated effects, thereby beginning to extend the theory towards an early system model. We will adopt a coding system that maps interview segments to system elements and processes and allows an audit trail of the iterative process of building explanations and refining the model. Since these initial interviews will also include more open-ended, respondent-led discussion around broader issues of maternal health policy and service delivery in Pakistan, we will complement the analysis described above, with an inductive, thematic analysis using line-by-line code and retrieve approaches. This analysis will seek to uncover ‘softer’ system characteristics including latent meanings, incentives, relationships and power structures. The product of Module 1 will be visual representations and textual descriptions of the early system model. We will explore various ways of engaging policy stakeholders with the emerging understanding including casual loop diagrams, figures employing the PARiHS E, C and F categorisation, vignettes and textual summaries. The analysis will highlight gaps and embedded assumptions within the initiative that require exploration on the ground. Inconsistencies, if any, between understandings of the ‘24/7 BHU’ initiative held by different stakeholders within the system will also be noted.
WP1 - Module 2: Assessing patterns of performance
Addresses research questions:
Q2: To what extent are the key performance indicators of the ‘24/7 BHU’ initiative being met in Punjab? How does performance vary across settings and between sub-groups of women?
Q3: How does the ‘24/7 BHU’ initiative work in the dynamics of different settings? What system processes are supporting or undermining successful scale up?
We adopt an efficient design for WP1 module 2 combining analysis of data from the Punjab Health Department’s routine monitoring system alongside the collection of new data.
Analysis of Punjab Health Department routine data
The Punjab Health Department monitors the ‘24/7 BHU’ initiative through data routinely collected and reported by the Policy and Strategic Planning Unit, a monitoring unit operating at arms-length from the implementing unit (IRMNCH). This monitoring arrangement – as an audit-and-feedback system34 is an important element of the ‘24/7 BHU’ initiative that will form part of the programme theory and be examined in the detail as described below in WP2-module 3. In addition, it offers a useful means by which dimensions of performance of ‘24/7 BHU’ can be assessed. Punjab Health Department and PSPU have granted access to the monitoring data to our research team and this will enable a series of analyses that extend the PSPU’s routine reporting and provide additional insight into the progress of the scale-up. Precise details of the analyses and the insights these can offer will be agreed via early stakeholder meetings and a workshop in which the programme theory will be discussed (see WP2 below). However, early discussions have clarified available indicators and identified some key areas for investigation. Analyses will examine (i) data completeness, consistency and plausibility, and (ii) variation in performance across BHUs and across districts.
Currently, the key performance indicator for the ‘24/7 BHU’ initiative is facility functionality around-the-clock. Since January 2017, births are notified by SMS text message in real-time by midwives to the local Monitoring and Evaluation (M&E) person, the M&E manager and the District Public Health Officer. These data will allow examination of the distribution of reported birth timing. Exploration of 24-hour birth patterns will provide an indication of whether deliveries are being reported as taking place at BHUs after dark (taking into consideration seasonality). Evidence indicates that natural births do not happen evenly distributed around the clock, with 38% of births naturally occurring within the night-time hours of midnight to 8am.35 We will also examine birth volumes in comparison with expected numbers based on demographics and prior survey findings on proportion of facility-based births. The Health Department collects further information relevant to the availability of skilled birth attendants via the ‘Facility Analysis Tool’ administered by district supervisory officers on a monthly basis. Data on presence, phone contact and place of residence of skilled staff will be analysed.
Indicators of facility quality are also important, and receiving increased attention by the Health Department as the programme is scaled-up. The routine system currently collects indicators of BHU functionality including: stock availability (e.g. blood Hb strips and metres; essential medicines); equipment (e.g. thermometers; electronic scale); documentation (e.g. referral records) and facility ‘outlook’ & maintenance (e.g. signage; toilets; drinking water; cleanliness).36
Descriptive analyses will be performed at BHU, district and province levels. We will work closely with the Director Technical PSPU (Government of Punjab) to define the analyses and guide data handling, presentation and interpretation of results Trends over time will be examined. A particular interest will be to examine the degree of variation in key performance indicators within and between districts. Bivariate and multi-variable analyses will be performed to test key hypotheses regarding factors operating at BHU and district level that support (or undermine) performance. For example, it might be hypothesised that better physical BHU condition will be positively associated with personnel regular attendance. A detailed analysis plan will be informed by the programme theory and system modelling undertaken with government stakeholders in Module 1.
Recognising the newness of the programme and the challenging objective of providing round-the-clock high quality care (with significant demand- and supply-side obstacles), accurate reporting of birth timing, and other indicators of programme functioning, is likely to be affected by both administrative error and some degree of fraud. We will therefore undertake verification exercises (see below). ‘Margin of error’ statements will then be incorporated into the results based on the Punjab DH monitoring system data. Data will be analysed using Stata 17.37
Verification and supplementation of Punjab Health Department data via new data collection at District and BHU level
The nature and objectives of the ‘24/7 BHU’ initiative dictates a three-level cluster sample in which a sample of districts is first drawn from the whole province, followed by a sample of BHUs from within each district, and finally a sample of women (births) is drawn from within each BHU. The ultimate outcomes relate to women’s access to good quality care when needed, but important determinants of these outcomes may operate at different levels within the system. In drawing our sample, there is a trade-off to be made between the number of districts and the number of BHUs. The sample could consist of fewer districts and more BHUs per district, or more districts with fewer BHUs within each, with concomitant implications for the power to detect associations between BHU-level and district-level factors and programme performance indicators. Findings from our analyses of the Health Department routine data, and our prior systems modelling work, will therefore guide our sampling approach. In particular, working closely with Health Department stakeholders, we will identify key factors at BHU and district level that are believed to be (i) influential, (ii) variable and (ii) potentially modifiable. This assessment will inform the appropriate balance of districts and BHUs in the sample. Maintaining this design flexibility is important both to ensure that the emerging understanding of the system is taken on board and also to promote ownership of the project and its findings among government stakeholders.
At this stage we have therefore determined the size of the woman (birth) level sample that is desirable. Taking the proportion of births that receive ‘good’ care (assessed via achievement of a threshold on a composite measure) as the outcome, sample size calculations indicate that a sample of 1,500 women(births) is sufficient to detect a small standardised difference between sub-groups of women defined by key characteristics (such as educational level) with 90% power [a series of calculations were run assuming varied splits across comparator groups, an α of 0.05 and a design effect of 1.5].38 Similarly, this sample size will be adequate to detect important differences in the proportion of deliveries occurring at night (as an indicator of round-the-clock access to care). We have currently budgeted for data collection for this number of births (as well as other data as described below) for a geographically dispersed sample over the north and south of the province (recognising important variation in socioeconomic development), and assuming data collection for a minimum of 30 women (births) per BHU. This would then allow data collection from up to 50 BHUs; if the analysis of routine data suggests BHU variation is of key importance, then we will ensure this maximum number of BHUs is included, but possibly spread over fewer districts; if district variation is more important, then collecting data from more districts will be prioritised with fewer BHUs per district.
Verification of Health Department data: We will collect data on (1) timing of deliveries and (2) facility readiness, in order to assess accuracy of routinely reported data, a priority for the Health Department.
(1) Reported timing of deliveries:
The Health Department, specifically PSPU, have agreed to implement a process whereby SMS messages reporting deliveries will be copied to the research team. A fieldworker will then verify the birth timing. Verification will happen in one of three ways:
(i) The fieldworker may already be at the facility observing the labour/delivery as part of our supplementary data collection (see below);
(ii) During day-time hours the fieldworker will attend the facility as soon as possible;
(iii) During night-time hours, the fieldworker will attend the facility the following morning and, if necessary, make a home visit (women often return home very promptly but patient address and phone number are routinely collected by BHUs.)
The field-team will verify a minimum of 30 birth reports per BHU, verifying each one in sequence from the initial entry into each BHU fieldwork period. Verification will involve recording the time of the birth and a set of questions designed to (1) verify who delivered the child; and (2) record key indicators of quality of care and (3) patient satisfaction. We are aware that our field-team’s presence may prompt staff to adjust their reporting of births via SMS. For instance, the accuracy of timing may improve due to increased attention to detail and/or reduction in any prior fraudulent reporting of night-time births. Analysis will therefore compare the observed pattern with historical data for the BHUs and districts. It is unlikely that fieldworker presence would prompt a major change in the provision and uptake of night-time services within the short period of data collection in each locality. The verification process will therefore allow an assessment of both reporting accuracy and the proportion of births taking place at night. This information will inform interpretation of the routine monitoring data discussed above.
2) Facility readiness:
Structured observations, review of documentation and short staff interviews will be used to collect information on key indicators (identified in conjunction with Health Department stakeholders as useful ‘sentinel’ markers from within the full range of indicators currently reported by PSPU). Indicators will cover aspects of physical infra-structure (e.g. labour room privacy; water supplies); stocks and equipment; and functioning (e.g. referral records), and also skill-mix and numbers of staff. As above, analysis will involve comparison with the routine data for the same BHUs and districts.
Supplementation of Health Department data: Two important additional dimensions will be examined, (1) work processes and (2) users’ experiences of care.
1) Work processes:
Informed by the programme theory and systems modelling, recognised standards and assessment tools39-41 and the PARiHS framework25, we will develop structured data collection tools to capture key indicators at district and BHU level via structured observation and short staff interviews. At district level, the tool will focus on key aspects of governance and management: including human resource allocation; M&E system functioning; financial reporting; record keeping; communications with and support to BHUs. At BHU-level, the work processes tool will include indicators of: personnel management (job descriptions; supervision arrangements; night-time security); referral systems (referral protocols; transport system; efficient communication and cooperation between actors in the referral chains); supply chains (forecasting; functional chain of supply); monitoring and record keeping.
In addition structured clinical observations of patient-provider interactions will be undertaken for a minimum 30 antenatal care and 30 labour and delivery interactions (from entry to exit) for each BHU. In addition, structured clinical observations of patient-provider interactions will be undertaken for a minimum of 30 antenatal care and 30 labour and delivery interactions (from entry to exit) for each BHU.
2) User perspectives:
Exit interviews will be carried out with at least 30 women in each BHU. As far as possible, these will be the same women whose labour/birth experience has been observed. Women will most often be followed up at home a day or two post-delivery. Interviews will not be longer than half an hour and will ask for respondents’ perspectives on: staff availability and contact time; staff-patient communication; privacy; information; choice; respect; dignity; and emotional support. Exit interviews will also take the opportunity to ask brief questions about choice of delivery place and whether the woman would have attended the facility after dark, and it not, why not. Socio-demographic data will also be recorded. Data will be collected by clinically trained fieldworkers working in pairs. They will be trained and assessed on WHO clinical standards, use of assessment tools and objective scoring to foster quality.
Profiling BHUs
‘Thumbnail’ BHU sketches of each BHU will be developed. A template will be prepared for this purpose informed by the programme theory and system modelling and this will be populated via a series of brief semi-structured interviews with personnel, observational activity during the course of site visits, and de-briefing with data collectors. A particular objective will be to identify local level innovations, ‘work arounds’ and ‘interpretations’ of the initiative that appear to be working well. In addition, local people and providers will be engaged in discussions of how things might work better to elicit “change ideas” for later consideration in WP2. These sketches will be combined with quantitative summary measures for BHUs to produce overall BHU profiles, which will be examined qualitatively to determine whether BHU ‘types’ can be identified (reflecting particular expressions of system functioning).
Data management and analysis: Quantitative data will be analysed using Stata 17.37 For each domain of interest, assessment criteria will be scored, summed and rated using accepted external standards (e.g. ‘non-functional’ / ‘partially functional’ / ‘fully functional’ referral system). Indicators will be reported for each BHU and distributions of BHU ratings will be reported by district and overall for the province. We will also describe the profile of women using BHU facilities in the different districts and overall. Bivariate statistics and multi-variable, multi-level modelling will then be used to explore factors associated with better/worse performance on key indicators at district, BHU and woman level.
WP1 - Module 3: Understanding system processes in detail
Addresses research questions:
Q3: How does the ‘24/7 BHU’ initiative work in the dynamics of different settings? What system processes are supporting or undermining successful scale up?
Q4: How does the ‘24/7 BHU’ programme theory relate to system influences on scale-up? What are the key areas of misalignment?
Building on the insights gained via Module 2, this module aims to generate a deeper understanding of the system processes that ensue, and how the scale up of the ‘24/7 BHU’ initiative unfolds, in different settings. An Institutional Ethnography (IE) will be conducted in 4-5 contrasting BHUs selected using the profiles generated above, and including positive deviants, in three districts, their district headquarters offices and the provincial health directorate in Lahore. Increasingly recognized as a powerful tool in implementation research and quality improvement, IE is a qualitative method that examines everyday processes, relationships, and interactions at work to identify broader ‘ruling relations’ within the organization and wider socio-ecological system.42,43 Data will be generated using a range of formal and informal observations, guided interviews and group discussions. Data generation will be informed, but not constrained, by our prior system modelling work and the PARiHS framework. We will thereby aim for an effective balance between ‘bottom-up’ inductive data generation that excels at revealing the taken-for-granted and unexpected, and a more structured approach that tests hypotheses and guards against the inadvertent oversight of key system elements or processes.
Data management and analysis: Data generation and analysis will be concurrent and will be guided by an ongoing dialogue with, and refinement of, the programme theory and system modelling. Qualitative data will be collected using audio recorders and detailed note-taking. A database of the transcribed interviews, focus groups, and observation notes will be created in Quirkos. Interpretive accuracy will be assessed by peer debriefing within the research team and other colleagues and respondent validation as an on-going activity.44 To make the analysis process manageable, within-case analysis will occur first. We will prepare operational memos for each BHU and district. These will follow a loose structure guided by earlier module findings and document key aspects of ‘what is/has been going on’ in each site. Next, holistic thematic memos will draw out more interpretive claims about ‘how the system is/has been functioning’ in each BHU/district setting and will include excerpts of empirical material in support of claims (e.g. direct quotations; observation notes). Cross-case analysis will then be initiated through the systematic comparing across memos in a series of team analysis workshops to identify patterns and relationships in the data (e.g. processes that amplify or dampen aspects of the initiative). A combination of data tabulation, diagramming and narrative techniques will be used to organise the data into manageable evidence ‘nuggets’. We envisage an analytical dialogue between the empirical data and the emergent system model. Hypothesised causal processes identified in the system model via earlier modules will be tested against ethnographic data and either confirmed, refined or refuted. Plus, empirical data will be mapped onto the system model, so that, where necessary, new elements and relationships will be added to the model.
Work package 2
Addresses research questions:
Q4: How does the ‘24/7 BHU’ programme theory relate to systems influences on scale-up? What are the key areas of misalignment?
Q5: How can system elements and processes be manoeuvred to improve implementation of the initiative overall and for sub-groups of women who are poorly served?
Q6: What key system elements and processes may need to be monitored and addressed to sustain the initiative going forward?
The goal of WP2 is to identify feasible and acceptable modifications to the ‘24/7BHU’ initiative that can be tested to assess their promise in supporting improved implementation in a wider range of settings across Punjab. Ongoing dialogue with senior Health Directorate and PSPU staff confirm commitment to innovation and a desire to develop capacity in improvement science techniques.
WP2 - Module 1: Integrating understanding and identifying plausible improvements
An initial one-day project launch workshop will be held in month 2. This event will engage key policy and practice stakeholders, members of the Policy and Programming Research Stakeholder Group (PPRSG), in the project aims and design and aim to engender a sense of ownership of the project. Ongoing one-to-one and small group meetings will subsequently take place regularly during WP1 to share progress with key policy-makers.
In month 24, a second deliberative workshop will bring together members of PPRSG with district health officers of the sampled districts to discuss findings from WP1 Modules 1 and 2 and to generate a shared understanding of the emerging system model and areas of uncertainty. Priorities for investigation during Module 3 will be identified. This workshop will also introduce the PDSA cycle approach to improvement and invite early discussion on possible “change ideas”. A key objective will be gain consensus that the PDSA process will generate learning but cannot promise achievement of desired outcomes; stakeholders must accept the possibility that improvement goals cannot be achieved under current constraints, and that new issues may arise.27 After this workshop, regular meetings and webinars will maintain stakeholder engagement in the project activities.
In month 38, a two-day deliberative workshop will bring together PPRSG and district health officers to discuss the integrated findings from across the whole of WP1. Data will be presented via evidence summaries using a mix of formats (visual, textual, vignettes etc.) that will be co-produced with key stakeholders prior to the workshop to ensure clarity and relevance. These will summarise: patterns and variations in implementation status; gaps between programme theory and operational realities; processes at work in successful facilities (including local innovations and ‘work arounds’); and “change ideas” proposed by different actors in the system. A refined system model will also be shared (visually in causal loop diagrams and in narrative form) that integrates these evidence elements and engages stakeholders in the ‘systems thinking’ approach. We will also provide workshop participants with stimulus by presenting a range of improvement approaches that have been tried in other contexts.45 An informal and participatory environment will encourage open debate and dialogue. As advocated by Reed and Card,27 the workshop process will promote clear understanding and framing of the system challenges. Workshop participants will then work collectively to identify and elaborate potential ways to enhance system enablers and/or circumvent obstacles. The focus will be on achieving standardized function rather than on following a single form of provision (e.g. round-the-clock care could vary in form with some BHUs having midwives in situ 24 hours while others operate an ‘on-call’ system). This will produce a list of candidate “change ideas”. Finally, workshop participants will be engaged in a discussion to establish prioritization criteria that will be used to select a small number of “change ideas” to be taken forward. Workshop discussions will be carefully documented and an accessible report prepared immediately afterwards.
WP2 - Module 2: Plan, pilot and document modifications
The next step will involve team members and senior managers in IRMNCHN in (i) identifying a district and 2-3 BHUs within which a Plan-Do-Study-Act26 approach is to be undertaken; and (ii) selecting priority “change ideas” to be piloted. Action will then move to the district and BHU level. PDSA cycles will be led by local managers and practitioners at the district and/or BHU level, with hands-on support from the research team. The steps are:
(a) Establish the local team, roles and a shared understanding of the PDSA approach;
(b) Develop a ‘proof of concept’ piloting plan defining the elements of the 24/7 BHU initiative to be adjusted, articulating the ‘logic’ of the modifications and establishing success criteria;
(c) Define the scope and duration of the pilot (anticipate quick development cycles, 6-8 weeks);
(d) Develop a clear work plan for successful implementation of the modification(s);
(e) Identify and minimize any risks to the pilot;
(f) Determine a suitable monitoring system (likely to be both quantitative and qualitative);
(g) Implement the strategy;
(h) Evaluate effects quantitatively and qualitatively against the success criteria;
(i) Collectively discuss implications and agree on next steps (abandon, modify, expand);
(j) Repeat cycle with new or refined “change idea”
In addition to the monitoring and evaluation embedded within the PDSA cycles, team members will document the process as it unfolds through observation and informal interviews with the key actors in order to capture key learning. Products from WP2 will be co-produced by the research team and the policy/practice partners and will include:
- Documentation of each PDSA cycle undertaken describing the “change idea”, the ensuing system changes and outcomes observed, learning and follow on actions. We will experiment with visual, text-based and short video capture as ways of sharing learning across the programme.
- A reflective account of the PDSA process as a whole as it was undertaken and experienced across the sites and recommendations for future use of this approach by the government and PSPU.
- A longer-term improvement strategy capturing “change ideas” generated but not piloted during this project as well as system aspects warranting further attention as the ‘24/7 BHU’ initiative continues.