CASE STUDY · PUBLIC HEALTH · UX RESEARCH
Reframing a Public Health Dashboard from Clarity to Accountability
A two-phase qualitative study revealing why “clear” dashboards still failed public-health teams, and defining design principles that make surveillance evidence trustworthy, explainable, and reusable in public health decision workflows.
SESSIONS
20 (n=7 + n=13)
MY ROLE
Lead User Researcher,
worked with an external delivery partner and domain experts
ORGANISATION
University of Sheffield
METHODS
Semi-structured interviews |
Prototype think-aloud sessions / design probe | Thematic analysis
Exec Summary
Five things to know before reading further
01
Grounded the problem in real workflows
Framed a two-phase study with epidemiologists to understand how WBE is interpreted, validated, and reused across public health roles in real decision settings.
02
Reframed WBE as a validation signal (Discovery)
WBE was rarely used as a standalone trigger. Its value depended on how easily it could be compared with other indicators, understood in context, and judged as reliable for action.
03
Used a prototype to reveal hidden requirements (Alpha)
A rapid prototype exposed how users interpret, validate, and act on WBE in practice — surfacing needs that interviews alone could not uncover.
04
Showed that clarity was not the main problem
Breakdowns came from missing context, unclear accountability, and weak support for validation and reuse, rather than not simply chart clarity.
05
Shifted the design goal from clarity to accountability
The design direction moved from “make the dashboard clearer” to “make outputs trustworthy, defensible, and reusable across reporting and decision workflows.”
Problem + User Complexity
It is a foundational research to define how a future WBE dashboard should support trust, interpretation, and action in public health workflows
The problem wasn't simply design a dashboard that showcases WBE data — it is making WBE signal trustworthy inside real decision workflows
An emerging signal without clear operational workflow
WBE showed promise for disease monitoring, but it was not yet clear how the data could support real public health decisions in practice.
Future design needed more than visual clarity
Existing dashboard examples showed data effectively, but did not address the UK-specific workflow, accountability, accessibility, and reporting needs that a deployable public-sector dashboard would have to support.
Complex, multi-role decision environment
The study identified three user groups within a shared decision pipeline: public health practitioners, local decision-makers, and data analysts, each requiring different views, levels of detail, and validation support.
Research need to uncover unspoken user requirements
Stakeholders could not fully specify these needs upfront. The study therefore had to uncover how different roles actually interpret, validate, and act on WBE data in practice, and what a future service would need to support.
HOW I FRAME THE STRATEGY
The external delivery partner led recruitment, prototype design, and running sessions across discovery and alpha. My role was to work with epidemiologists to build domain understanding, frame the research questions, and synthesise findings into design-relevant insights.
x
The strategy focused first on understanding how WBE fit into complex public health workflows (Phase I), then on using a prototype to test how people actually interpreted and trusted it in realistic decision contexts (Phase II).
Approach
Two phases: map real workflows first, then probe deeper requirements through a rapidly designed prototype
Phase I (Discovery) · FOUNDATION
Semi-structured interviews
Established baseline workflows, pain points, and expectations before introducing any design. Participants walked through how they access WBE outputs and where the process breaks down.
n = 7 · 45–60 min each
Phase II (Alpha) · DESIGN PROBE
Prototype-anchored think-aloud sessions
Scoped from Phase I findings, a rapid prototype was used as an alpha-style probe with scenario-based tasks. Participants imagined using the dashboard in their role and thought aloud, revealing gaps in interpretation, responsibility, validation, and reuse not surfaced in interviews alone.
n=13 · 45–60 min each
Data Synthesis
Hybrid thematic analysis: structured around research aims, refined through emergent role-specific practices
Deductive codes were aligned to the core research questions around current practices, barriers, and design opportunities.
Inductive coding captured emergent behaviours, role-specific practices, and unanticipated patterns.
I synthesised the data into actionable insights and design implications, using Dovetail to rapidly cluster evidence, generate initial tags, and maintain a structured research repository across both phases.
Key Findings
Four insights that shaped the design direction
1
WBE dashboards works best as a concordance-checking tool
Public health practitioners rarely used WBE in isolation. Instead, they compared it with established indicators and treated divergence as something to investigate. The dashboard’s role was therefore not to act as a standalone alert, but to support rapid, defensible cross-signal interpretation.
"We use it (WBE) as a signal... it does not typically trigger standalone action."
2
“Clear” charts fail when they are not self-describing
Participants stalled even on visually polished views when legends, unit definitions, time windows, update rhythm, or normalisation methods were not persistently visible. Under time pressure, clarity depended on embedded context, not aesthetic simplicity alone.
"[It is] not initially clear what the [map] is showing… [it] needs explanation."
3
Accountability is a spatial concept
Wastewater (WBE) catchments rarely aligned with the administrative or service boundaries teams were accountable for. This made it difficult to judge who the data applied to, who should act on it, and how it should travel through decision-making structures.
"Spatial misfit… creates ambiguity about who the data actually represents."
4
Trust depends on governance and reproducibility, not usability polish alone
Users were more likely to trust WBE outputs when figures were traceable, citable, and reproducible. Confidence depended on provenance, method transparency, update rhythm, and revision visibility, not just whether the interface looked polished.
"[I think the design] requests demographic overlays and visual mapping to aid interpretation."
Design Recommendations + Decisions
Four insights that shaped the design direction
Role-based entry points
Different user groups needed different levels of context, detail, and actionability. A future dashboard should support public health practitioners in validating and interrogating data, while also giving decision-makers outputs that are easier to interpret and reuse in briefings and operational discussions.
GDS Standard 2 — Solve a whole problem for users
Self-describing views
Core interpretive information (legends, units, time windows, update rhythm, and provenance) should remain persistently visible.
This supports interpretation under time pressure and ensures outputs remain understandable when reused beyond the interface, including for non-specialist users.
GDS Standard 2 — Solve a whole problem for users
Jurisdiction crosswalk + overlays
Wastewater (WBE) catchments should be translated into administrative and service boundaries, making accountability explicit and supporting alignment with existing public-sector reporting structures.
GDS Standard 5 — Make sure everyone can use the service
WCAG 2.2 AA
Reusable evidence outputs
The dashboard should support exportable figures with embedded provenance, versioning, and contextual metadata. Outputs must remain interpretable outside the dashboard, supporting reporting, auditability, and accessibility across roles.
GDS Standard 10 — Define what success looks like and publish performance data
WCAG 2.2 AA
Impact + Decision Framing Shift
What actually changed
Optimise for visual clarity and on-screen WBE chart polish
Add more widgets and data views
Pasted screenshots into briefings
Single map view for all users
→
→
→
→
Design for accountable, defensible reuse under scrutiny
Role-based entry points that match governance context
Versioned figure citations with embedded provenance
Jurisdiction crosswalk so ownership is legible
As formative research, the main impact was not a final dashboard release but a requirements reframing: from visual clarity as the primary design goal to accountability, reproducibility, accessibility, and jurisdiction-awareness as first-class requirements for future delivery.
The study also clarified what success would need to look like in a live service: increased confidence in interpretation, reduced ambiguity around responsibility, and more reliable reuse of outputs in reporting and decision workflows.
NEXT RESEARCH QUESTIONS
How should a WBE dashboard support auditable interpretation, escalation, and communication across multi-stakeholder decision workflows?
RQ 1
How can dashboards balance specialist analytical depth with accessible, self-describing outputs that support interpretation and reuse by non-specialist decision-makers?
RQ2
The insight successfully reframed the dashboard from a visualisation tool to a governed public-service interface, where accessibility, traceability, and accountability are core requirements.