DISTRICT
The Field Guide is the concept library inside DISTRICT. It explains the 39 implementation science ideas the simulation uses, from context assessment and missed opportunities for vaccination to contribution analysis and Learning in Action, with sources for each. Examples refer to Kwara North, the fictional district in the simulation.
Context assessment examines the social, geographic, cultural, political and health system factors that influence implementation.
Why it matters. Interventions that perform well in one setting may perform poorly in another if contextual differences are not recognised and addressed.
In immunization. Context assessment helps identify why children are missed, which barriers are most important, and which implementation strategies are most likely to succeed in a specific community.
Equity in immunization means coverage gains reach the people with the greatest unmet need, including those hardest to serve.
Why it matters. Aggregate coverage gains can mask widening inequity. A programme that improves the headline figure while leaving its most excluded communities further behind is not meeting its equity mandate.
In immunization. Disaggregating coverage data by ward, wealth quintile, and population group reveals which communities are being left behind and allows programmes to direct resources accordingly.
Data is there to guide action. Different indicators reveal different implementation challenges.
Why it matters. DTP1 reflects access. Dropout rates reflect retention. MOV rates reflect service quality. When administrative and survey data disagree, check data quality.
In immunization. Programme teams use these indicators to identify implementation failures, prioritise corrective actions, and monitor improvement over time.
CFIR (Damschroder et al., 2009; updated 2022) organises the determinants of implementation success into five domains: Innovation (characteristics of the intervention), Outer Setting (external context and structural barriers), Inner Setting (the implementing organisation), Individuals (knowledge, beliefs, and motivations of implementers and recipients), and Implementation Process (how implementation is planned, executed, and evaluated).
Why it matters. CFIR prevents the common error of treating all context problems as the same type, which leads to applying the same strategy to barriers that differ in kind. Each CFIR domain points to a different category of solution.
In immunization. In Kwara North: Kpako reflects an Outer and Inner Setting barrier (seasonal roads, no nearby service point). Tudun Wada reflects an Individuals barrier (fears fed by rumours). Sabon Gida reflects an Outer and Inner Setting barrier (mobility, no tracking across facilities). Rafin Zuwo reflects an Inner Setting and Implementation Process barrier (stretched staff, no screening at triage).
Source: Damschroder LJ et al. (2009). Implementation Science 4:50. Updated: Damschroder LJ et al. (2022). Implementation Science 17:75.
Context has several dimensions. It varies across communities and shifts within each community over time.
Why it matters. Structural context (geography, poverty, migration routes) is relatively stable and shapes initial programme design. Dynamic context (seasonal cycles, economic rhythms, political transitions) changes over time and can render a well-designed programme ineffective if monitoring does not detect the shift.
In immunization. The Kpako attendance drop in the second half of every month illustrates dynamic context: the same community behaves differently at different points in the economic cycle. Monitoring should track programme performance and also changes in the context where implementation takes place.
Collaborative implementation treats communities and stakeholders as co-designers rather than targets. In co-diagnosis, community members, frontline workers and programme staff identify barriers and design solutions together.
Why it matters. Community members often know things about local conditions that programme managers do not. Co-designed strategies are more likely to fit local conditions and to be accepted by the people they serve.
In immunization. Co-diagnosis tools include community mapping, participatory rapid appraisals, village health committee consultations, and CHW-led household surveys. IA2030 Strategic Priority 2 aims for all people and communities to value, actively support and seek out immunization services.
A Missed Opportunity for Vaccination (MOV) occurs when an eligible individual contacts a health service but does not receive a vaccine for which they are eligible and have no contraindication. Missed opportunities can happen at several kinds of contact: immunization contact (eligible but not fully vaccinated), sick-child contact (status not checked), other preventive service contact (growth monitoring, ANC), and community outreach contact (present but not vaccinated due to stock-out or provider absence).
Why it matters. MOVs are distinct from zero-dose: a zero-dose child has no DTP contact; a MOV child has had contact but was not vaccinated. They require different responses. Children who default on the schedule may also experience a MOV at a later contact, but defaulters and MOVs are separate concepts.
In immunization. MOV reduction strategies include same-day vaccination at every contact, vaccination screening at triage, and stock-out prevention protocols. The primary cause is usually a failure in the service delivery system. Caregiver refusal is seldom the main reason.
Source: World Health Organization (2017). Methodology for the assessment of missed opportunities for vaccination. Geneva: WHO. Ogbuanu IU, Li AJ et al. (2019). PLoS One 14(1):e0210648.
A theory of change is an explicit, testable account of how and why a set of activities is expected to produce outcomes. It maps the results chain (inputs → activities → outputs → outcomes → impact) and states the assumptions linking each step.
Why it matters. When a programme underperforms, one or more underlying assumptions may not hold in practice, even when planned activities are implemented as intended. Making assumptions explicit allows them to be monitored and tested rather than discovered through failure.
In immunization. Every implementation framework serves the theory of change: CFIR diagnoses which assumptions are at risk, ERIC selects strategies to protect them, and RE-AIM tests whether the causal chain actually held. Set out the programme theory before you select interventions.
A zero-dose child has not received a first dose of a diphtheria, tetanus and pertussis-containing vaccine. Each one marks a place where the health system did not reach a family: no service close enough, a visit that went badly, a session at the wrong time, or a decision at home that nobody helped the family make.
Why it matters. Labels such as “refuser” or “non-compliant” place the problem on the caregiver and stop the inquiry. Treating the gap as a system failure keeps attention on the things a programme can change.
In immunization. In DISTRICT every ward’s zero-dose children trace back to a system cause: distance in Kpako, unanswered fears in Tudun Wada, lost records in Sabon Gida and missed opportunities in Rafin Zuwo. The Zero-Dose Learning Agenda (ZDLA) works from the same starting point.
The iceberg separates what you can see (low coverage, a rumour) from what sits below it: the events and patterns that produced it, and the norms and structures that keep it in place. The 5 Whys asks “why?” repeatedly until the answer is something a team can act on. Taiichi Ohno described the 5 Whys in Toyota Production System (1988).
Why it matters. A response aimed at the surface label rarely lasts. Teams that keep asking why find causes they can change, and the answer often points to the health service itself.
In immunization. In Tudun Wada the label was “rumours”. Below it: babies had fever after vaccination, nobody explained it or followed up, and husbands then withheld permission. The response that follows adds side-effect counselling and a follow-up visit to the work with trusted leaders.
WHO and UNICEF’s Human-centred design for tailoring immunization programmes (2022) maps a caregiver’s path in six stages: Knowledge, awareness & belief; Intent; Preparation, cost & effort; Point of service; Experience of care; After service.
Why it matters. A family can drop out at any stage. Mapping the journey shows where a barrier sits, so the response is aimed at that point and not at awareness by default.
In immunization. Tudun Wada’s problem began at “After service”: a fever with no one to ask. That experience then shaped “Intent” for the next child. Kpako’s problem sits at “Preparation, cost & effort”.
Gender intentionality means asking, at every step, how gender shapes who decides, who travels, who pays and who is heard, and designing with that in mind. The ZDLA treated it as a thread through every phase of its work.
Why it matters. In many households the person who brings the child is not the person who decides. Programmes that speak only to mothers can miss the decision entirely.
In immunization. In Tudun Wada mothers need their husbands’ agreement, so fathers get their own sessions. Nigeria’s PHCUOR implementation guidelines (NPHCDA, 2018) state that “at least one executive position in the WDC should be occupied by women”. Women on Ward Development Committees are a route to hearing mothers directly.
An implementation strategy is the action used to support delivery of an effective intervention. The vaccine is the intervention; mobile outreach, community mobilisation, and same-day vaccination are implementation strategies.
Why it matters. Without appropriate strategies, even highly effective vaccines fail to reach the populations that need them most. Match each strategy to the barrier you found; do not apply one approach everywhere.
In immunization. In zero-dose programmes, strategies differ by community. Geographic barriers require changed delivery infrastructure. Trust barriers require community engagement. Mobility barriers require records that follow the child.
Source: Powell BJ et al. (2015). Implementation Science 10:21.
Feasibility is the degree to which an implementation strategy can be carried out within the constraints of a given setting: resources, skills, infrastructure, and operational capacity.
Why it matters. A strategy that is technically correct but operationally impossible wastes design time and creates false expectations. Feasibility belongs in implementation design from the start.
In immunization. Before selecting a strategy, assess whether it can realistically be executed: Are supplies available? Do staff have the required skills? Can transport and logistics support the model?
Adaptation means modifying an intervention to fit local context while preserving core components. Fidelity means delivering it as designed. You need both.
Why it matters. Core components are causally linked to outcomes and should not be changed. Delivery elements, such as session timing, language, and location, can be adapted without undermining effectiveness. The FRAME tool (Wiltsey Stirman et al., 2019) provides a structured way to document what was changed, why, and by whom.
In immunization. Changing session timing for a pastoralist community is an adaptation. Changing the antigen is not. Undocumented adaptations accumulate into programme drift: a gradual erosion of effectiveness that may not show in coverage data until outcomes decline.
Source: Wiltsey Stirman S et al. (2019). Implementation Science 14(1):58.
The ERIC compilation (Powell et al., 2015) catalogues 73 discrete implementation strategies, which Waltz et al. (2015) grouped into nine clusters, including Engage consumers, Change infrastructure, Adapt and tailor to context, Train and educate stakeholders, and Utilise financial strategies.
Why it matters. ERIC provides a common language for selecting, reporting, and comparing implementation strategies across settings. It is used as a planning menu, a diagnostic tool when programmes are failing, and a reporting standard for implementation research.
In immunization. In zero-dose programmes, Engage consumers is a frequently under-used cluster. Programmes often invest in supply-side infrastructure while under-investing in the community trust and demand-building strategies needed to reach the hardest-to-reach.
Source: Powell BJ et al. (2015). Implementation Science 10:21. Waltz TJ et al. (2015). Implementation Science 10:109.
COM-B (Michie, van Stralen & West, 2011) states that any behaviour requires three conditions simultaneously: Capability (knowledge and physical capacity), Opportunity (physical access, affordability, and social norms), and Motivation (beliefs, trust, and habit).
Why it matters. COM-B is the demand-side complement to CFIR. Where CFIR diagnoses implementation determinants in the health system, COM-B diagnoses the behaviour of the people the system is trying to reach. Matching the intervention to the specific deficit (capability, opportunity, or motivation) prevents common errors, such as running an education campaign where the barrier is distance.
In immunization. Applied to Kwara North: Kpako and Sabon Gida face physical opportunity barriers (distance, mobility). Rafin Zuwo’s barrier sits in the service itself. Tudun Wada faces a motivation and social opportunity barrier (rumours, declining trust). Each requires a different response.
Source: Michie S, van Stralen MM, West R (2011). Implementation Science 6:42.
Co-design brings the people affected into shaping the response. A “How might we…?” question turns a root cause into an open design prompt; IDEO.org’s Field Guide to Human-Centered Design (2015) uses this method. Ideas are then sorted by likely impact and effort.
Why it matters. Solutions designed without the people who will use them often fail on details nobody in the office could see.
In immunization. Mission 2 starts from “How might we help mothers in Tudun Wada through the night after vaccination?” and sorts five ideas by impact and effort before choosing which to try first.
Fidelity measures whether an intervention is delivered as designed: the right components, the right dose, to the right population, by staff with the right skills.
Why it matters. With low fidelity, when a programme underperforms you cannot tell whether the intervention does not work or was never delivered. Fidelity monitoring separates design failures from delivery failures. Each calls for a different response.
In immunization. Common fidelity failures include stock-outs at outreach sites, cancelled sessions due to transport, and incomplete schedules. Monitoring fidelity in real time allows corrective action before coverage is affected.
Reach measures what proportion of the intended target population actually received the intervention.
Why it matters. High fidelity does not guarantee high reach. A programme can run well-delivered sessions that consistently miss the communities that need them most. Reach and fidelity are independently measurable and require separate monitoring responses.
In immunization. Low reach with high fidelity suggests the delivery model does not fit the target population (a strategy problem). High reach with low fidelity suggests delivery quality needs strengthening (a training or supply problem).
Effective monitoring tracks both implementation (what was done) and outcomes (what changed) in near-real time, and routes findings to decision-makers quickly enough to support corrective action.
Why it matters. Monitoring is useful only when it leads to timely action. Reports no one reads, or data that arrive after the decision, change nothing.
In immunization. In district immunization, routine monitoring draws on DHIS2 dashboards, CHW field reports, supportive supervision registers, and community feedback. Each source captures a different aspect of programme performance.
Supportive supervision differs from punitive inspection. It uses routine data and direct observation to identify problems, solve them jointly with health workers, and agree on specific corrective actions with timelines.
Why it matters. Supervision that assigns blame rather than solving problems reduces health worker motivation and drives under-reporting. Supportive supervision improves both implementation quality and the reliability of the data used to manage programmes.
In immunization. In district immunization, effective supervisory visits integrate DHIS2 data review, CHW field report analysis, direct service observation, and agreed action plans. Repeated visits of this kind build a continuous improvement cycle.
Every implementation decision has second-order effects. Before choosing a strategy, ask two questions: will it solve the immediate problem, and what might it disrupt elsewhere in the system?
Why it matters. Three common failure modes: resource displacement (surging resources to one site weakens another); campaign-routine tension (intensive pushes can undermine routine systems and create campaign-dependence); and incentive distortion (coverage incentives can drive over-reporting or neglect of harder-to-reach populations).
In immunization. Plan for second-order effects from the start: backfill displaced staff, build surges into routine systems, and watch for unintended consequences.
An after action review asks what was planned, what happened, why, and what to change. WHO’s Guidance for after action review (2019) is written for reviewing the response to a public health event, and notes that AAR also works “as a routine management tool for continuous learning and improvements”.
Why it matters. Short, regular reviews catch problems while they can still be fixed. A change log records each adjustment so the team later knows what it actually implemented.
In immunization. In Mission 3 the review rhythm pairs a user advisory group with after-action reviews and a change log, and holds separate groups for mothers and fathers where mothers would not speak freely in a mixed group.
Effectiveness asks whether outcomes improved for those reached. Equity outcomes ask whether the most deprived groups benefited proportionally or more than better-served populations.
Why it matters. IA2030 tracks zero-dose children and coverage in the lowest-performing districts because headline coverage gains can conceal equity failures. A programme that raises aggregate coverage while leaving the most excluded communities behind is not meeting its equity mandate.
In immunization. Report both the aggregate change and the change for the hardest-to-reach subgroup. A widening gap between the two is a programme failure, even when the headline number improves.
RE-AIM (Glasgow et al., 1999) evaluates real-world implementation across five dimensions: Reach (who received the intervention?), Effectiveness (what impact did it have?), Adoption (did the intended sites and staff take it up?), Implementation (was it delivered as designed?), and Maintenance (can gains continue?).
Why it matters. No single metric can substitute for evaluation across all five dimensions. High Effectiveness with low Reach is not success. High Reach with poor Implementation produces waste.
In immunization. Applied to Kwara North: Rafin Zuwo showed strong Reach and Effectiveness. Transport breakdowns in Tudun Wada were an Implementation (fidelity) failure. Kpako showed low Reach despite sustained effort, which points to an Implementation and Appropriateness problem.
Source: Glasgow RE et al. (1999). American Journal of Public Health 89(9):1322–1327.
Implementation outcomes (Proctor et al., 2011) are distinct from programme outcomes. Programme outcomes (coverage, dropout, disease reduction) measure what happened to the target population. Implementation outcomes measure what happened to the programme itself.
Why it matters. The eight implementation outcomes are: Acceptability, Adoption, Appropriateness, Feasibility, Fidelity, Penetration, Sustainability, and Cost. They explain why programmes succeed in one setting and struggle in another, often before coverage data reveals a problem.
In immunization. When penetration is low despite sustained effort (as in Kpako), investigate Acceptability and Appropriateness first, ahead of effort or sustainability. Implementation outcomes direct the diagnostic question.
Source: Proctor E et al. (2011). Administration and Policy in Mental Health 38(2):65–76.
When an outcome improves, rigorous evaluators ask how they know the intervention caused the change. Rival explanations include: secular trends (improvement was happening anyway), concurrent interventions (other programmes contributed), data artefacts (recording improved while vaccination did not), and regression to the mean (poorly performing areas improve naturally).
Why it matters. District programmes rarely have randomised controls. Instead, evaluators build a plausibility argument: evidence that makes a causal contribution credible. The strongest forms are dose-response (largest gains where the intervention was most intensive) and specificity (the mechanism matches exactly what was changed).
In immunization. Good evaluation presents the most credible explanation the available evidence supports, without claiming certainty. Give funders that explanation instead of a naive “coverage rose, therefore we succeeded”.
Contribution analysis, developed by John Mayne (ILAC Brief 16, 2008; Evaluation, 2012), builds a reasoned case that a programme contributed to a change. It sets out the expected chain of results, checks each link against evidence, and weighs other explanations.
Why it matters. Most district programmes have no comparison group. Contribution analysis gives a disciplined way to say how confident you are, and why, without overclaiming.
In immunization. The ZDLA rated its contribution claims as strong, moderate or low confidence. In Mission 4 you weigh whether Rafin Zuwo’s fall in missed opportunities links to same-day vaccination and triage screening.
The Immunization Delivery Cost Catalogue (IDCC) Codebook (2024) separates financial cost, the money actually spent, from economic cost, which adds the value of resources used without new spending, such as health worker time and donated items.
Why it matters. A strategy that looks cheap in the budget may rely on staff time that is taken from other work. Decision makers need both figures to judge whether it can be sustained.
In immunization. Mission 4 asks how to treat the time the Village Health Worker, the facility in-charge and supervisors spent on the Kpako outreach post, which the ₦360,000 did not count.
Sustainability is the ability to maintain programme outcomes after initial funding, external support, or project activities decline.
Why it matters. Many programmes achieve short-term success but fail to sustain it. Sustaining results depends on political will, financing systems, workforce capacity, and community ownership.
In immunization. Ask what would continue and what would stop if external support ended tomorrow. The answer often reveals sustainability risks long before coverage indicators begin to decline.
Stakeholder engagement means finding the people and organisations that influence or are affected by a programme, and building working relationships with them.
Why it matters. Different people hold different levers. Political leaders control budgets. Community leaders shape trust. Health managers influence delivery quality. Each requires a different engagement approach and responds to different evidence framings.
In immunization. Present the same programme result in terms that matter to each decision-maker: a finance official, an elected leader or a community representative. The evidence stays the same; only the framing changes.
The DRIVE framework (Documenting and Reporting Implementation Strategies of Vaccination Efforts) proposes eight elements for documenting immunization implementation strategies: (1) name, (2) implementation problem, (3) context, (4) strategy definition, (5) mechanism of action, (6) specifications and results, (7) facilitators and barriers, and (8) additional information.
Why it matters. Voltage drop during scale-out is often a documentation failure: the original strategy was sound but poorly recorded. A district that records only its coverage outcome, without the mechanism, context, adaptations, and barriers, leaves no transferable learning. The next district starts from scratch.
In immunization. Document every implementation strategy you use in enough detail that someone in another district, facing the same barrier profile, could understand what you did and why it worked. DRIVE is a proposed framework (Adamu et al., 2025) that has not yet been validated; it is introduced here as a practitioner discipline.
Source: Adamu AA et al. (2025). Discover Public Health 22:34. https://doi.org/10.1186/s12982-025-00417-9
Scale-up expands an intervention within similar contexts. Scale-out adapts it for different populations or settings. The distinction matters because scale-up assumes the original context is similar enough to replicate, while scale-out requires deliberate re-contextualisation.
Why it matters. Assessing contextual fit should be part of planning for replication and scale. Replicating a trust-barrier strategy in a community where the primary barrier is geographic distance misallocates resources and produces no results. Effectiveness can also diminish during expansion if implementation quality, contextual fit, or support systems are not maintained.
In immunization. WHO/ExpandNet scaling-up guidance, the Intervention Scalability Assessment Tool (ISAT; Milat et al., 2020), and the FRAME documentation tool provide structured support for scaling decisions. Between them they cover scalability and context assessment, fidelity, and the documentation of adaptations.
Embedding learning means the cycle of asking, testing and adjusting keeps running after a project ends: inside supervision, review meetings and the people who hold those roles.
Why it matters. The ZDLA found these practices last when they are built into meetings and structures that already exist and local health officials lead them. Teams named three supports: skills for health system staff, an environment that rewards adapting plans, and technical support that hands over to local staff.
In immunization. Mission 5 asks you to protect Kwara North’s gains by placing reviews, change logs and community feedback inside monthly review meetings, supervision visits and community platforms, and by costing what it takes each month to keep them going.
Studying implementation means collecting evidence on purpose, in the places where a programme runs, about why it works or fails there. Routine data show what happened and where; the reasons sit in the local context.
Why it matters. Every mission in DISTRICT turned on a local reason the dashboard could not show: a narrow denominator, a monthly attendance rhythm, a service point families rarely used. Decisions that ignore those reasons repeat the same mistakes.
In immunization. Six steps: start from a decision; name the outcome, determinant or strategy you are studying; pick a design that fits routine work (process evaluation, rapid qualitative methods, routine data over time); choose sites and people on purpose; check existing data, then add a little; collect ethically and feed back fast. Mataki is the companion field tool for this work.
Sources: Peters DH, Tran NT, Adam T, eds. (2013). Implementation research in health: a practical guide. WHO/AHPSR. Moore GF et al. (2015). BMJ 350:h1258. Ghaffar A et al. (2017). Bull World Health Organ 95(2):87.
The Acceptability of Intervention Measure (AIM), Intervention Appropriateness Measure (IAM) and Feasibility of Intervention Measure (FIM) each ask four short questions on a five-point scale from completely disagree to completely agree. They are free to use.
Why it matters. Routine settings need measures that are short enough to use during a supervision visit and still useful. Glasgow and Riley call these pragmatic measures.
In immunization. When penetration is low despite delivery effort, as in Kpako, ask the health workers delivering the service to complete the AIM, IAM and FIM, and set the scores beside what caregivers tell you in interviews.
Sources: Weiner BJ et al. (2017). Implementation Science 12:108. Glasgow RE, Riley WT (2013). American Journal of Preventive Medicine 45(2):237–243.
WHO’s Data Quality Review toolkit looks at four dimensions: completeness and timeliness, internal consistency of reported data, external consistency with other sources such as surveys, and external comparison of population data (denominators).
Why it matters. Routine data are the most frequent record a district has. Their weaknesses are known and checkable. Checking them first tells you how far to trust each figure.
In immunization. Mission 1’s 78% came from a target population of 6,500 against a district estimate of 7,440: an external comparison of population data. When a register entry looks wrong, check it against tally sheets and stock records and report it as a limitation.
Source: World Health Organization (2017). Data quality review: a toolkit for facility data quality assessment. Geneva: WHO.
Learning in Action is a cycle for improving immunization in a specific place, based on the cycle the ZDLA calls learn-by-doing (LxD). It has three phases: identify problems and root causes; brainstorm and co-design; implement and adapt. Gender intentionality runs through all three.
Why it matters. Evidence collected in the place where a programme runs answers questions that national averages cannot: why children are missed here and what changes that.
In immunization. The methods module walks the cycle and shows where each phase lives in Mataki: Assess for barriers and root causes, Plan for strategies and expected mechanisms, Track and Review for adapting.