SHARE Knowledge Hub

Wellbeing Strategy and Programmes

A workplace wellbeing strategy is a governed plan to prevent harm from work, improve conditions that support health and connect people with proportionate help. It uses worker evidence and risk assessment to set priorities, resource changes and test whether controls work across shifts and groups. Individual programmes can add support, but they cannot compensate for harmful workload, unsafe behaviour or inaccessible work.

Updated 2026-11-11GB and NI

Understanding Wellbeing Strategy and Programmes

A workplace wellbeing strategy is the governed plan by which an organisation prevents work-related harm, creates healthy and inclusive work, supports people with health needs and evaluates whether its actions improve outcomes without creating new disadvantage. Programmes are the coordinated interventions used to deliver that strategy. Campaigns, apps, health checks, challenges, webinars and benefits are optional components, not the strategy itself. The foundation remains compliance with health and safety, equality, employment, consultation and data-protection duties.

Why this matters at work

Common failures include a benefits catalogue described as strategy, awareness used instead of risk control, mandatory disclosure, health scorecards for managers, unsupported ROI, public competitions, apps procured before need is defined, employee resilience blamed for overload and a launch celebrated without capacity or evaluation. Each failure can create harm while appearing active.

Workers exposed to traumatic material, violence, death, serious injury or repeated human distress need an occupational-risk approach. Identify foreseeable exposures, reduce them where possible, prepare people realistically, provide supervision and recovery, manage cumulative load and establish specialist support. Selection for resilience should not replace safer staffing or exposure controls.

How organisations manage this in practice

Prevention should begin with the design and management of work. Remove avoidable hazards and demands, substitute safer processes, redesign staffing and tasks, improve control and role clarity, provide resources and supervision, manage relationships and change, and make the physical environment suitable. Individual education, coaching and coping support sit later in the hierarchy because they rely more heavily on the person while leaving the source in place.

HSE's Management Standards cover demands, control, support, relationships, role and change. They help identify organisational causes of work-related stress and provide a suitable risk-assessment approach when correctly applied. Use surveys, focus groups, operational data and worker consultation to identify gaps, develop solutions, assign action and review. The HSE Indicator Tool should not be the only data source.

Give the strategy a senior accountable owner and connect it to risk assessment, consultation, occupational health and operational decision making. Workers and representatives should help identify priorities, develop controls and assess whether changes work. Keep formal consultation obligations distinct from optional wellbeing events. An employee survey is one source of information; it is not evidence that identified risks have been controlled or that the employer can transfer responsibility to a committee. Use a clear action process with resources, due dates and escalation for unresolved issues. Review incident, absence and work-design information carefully, respecting the limits of each measure and protecting confidentiality. Report whether controls were implemented and effective, with any limitations and unmet needs. Revisit the strategy after significant change, new evidence or poor outcomes. Make it safe to raise concerns and show what was done in response. The practical test is whether leaders change the conditions they control and maintain credible support, rather than producing a policy statement or an annual calendar alone.

A programme decision needs an evidence trail

A programme should start with a work need, not a purchase. Delivery, benefits, harms and uncertainty inform a continuing owned decision.

Employee talking privately with a wellbeing adviser in a UK office. Photorealistic illustration.
Illustrative UK workplace scene. Read the practical checks below for responsibilities, evidence and limits. Download image · Download smaller image
  1. Work need

    Use consultation and work evidence, such as an identified reception rota gap, without treating individuals as the source of the hazard.

  2. Change and deliver

    Resource a relief plan and check that it reaches the intended shifts and workers; participation is delivery evidence, not proof of benefit.

  3. Evaluate benefit and harms

    Consider mixed results, privacy, exclusion, unintended burden and uncertainty without attributing causation to absence or productivity alone.

  4. Decide and return

    The programme owner continues, redesigns, stops or considers scaling with a fresh context check. Return the decision and evidence to needs and priorities.

Read the complete task-specific control examples

ERICPD is a planning cross-map: Discipline verifies and sustains every control. ESEAP compares effectiveness. A label does not authorise the task; retain the owner, evidence and stop or change decision.

Wellbeing Strategy and Programmes · ERICPD

Eliminate
Remove a programme practice that creates avoidable harmful demand, such as mandatory activity outside paid time or contradictory productivity targets.
Reduce
Replace an individual-only resilience offer with a portfolio that includes staffing, recovery, fair work arrangements and voluntary support.
Isolate
For a defined exposure, implement a feasible physical or technical change such as a quieter workstation or controlled routine contact.
Control
Set achievable workload priorities, paid access to support, accountable managers, consultation and confidential referral routes.
PPE
PPE does not control organisational overload or poor relationships.
Discipline
Review work-risk outcomes, unintended effects and equity of access, and close actions with field evidence.

Wellbeing Strategy and Programmes · ESEAP

Elimination
Remove a programme practice that creates avoidable harmful demand, such as mandatory activity outside paid time or contradictory productivity targets.
Substitution
Replace an individual-only resilience offer with a portfolio that includes staffing, recovery, fair work arrangements and voluntary support.
Engineering controls
For a defined exposure, implement a feasible physical or technical change such as a quieter workstation or controlled routine contact.
Administrative controls
Set achievable workload priorities, paid access to support, accountable managers, consultation and confidential referral routes.
PPE
PPE does not control organisational overload or poor relationships.

What is a workplace wellbeing strategy?

It is a governed plan for preventing harm, improving healthy work and providing proportionate support, with objectives, resources, owners, measures, evaluation and stop decisions.

Wellbeing includes how people function, connect, cope, recover, belong and find purpose at work. A strategy links these outcomes to work design and management rather than treating wellbeing as an individual lifestyle project.

Context. Required question: Which work factors and groups need action?; Evidence: Mixed needs evidence.

Plan. Required question: What will change, for whom and why?; Evidence: Objectives and theory of change.

Delivery. Required question: Who owns resources and controls?; Evidence: Action plan and governance.

Learning. Required question: Did work and outcomes improve without new harm?; Evidence: Evaluation and leadership decision.

Is a programme calendar a wellbeing strategy?

No. Apps, talks, exercise classes, EAPs, campaigns and awareness days may form part of support, but they do not replace prevention, work-risk control, adjustments or healthy management.

A crowded calendar can coexist with impossible workload, bullying, poor staffing or exclusion. Judge the strategy by whether work factors change and intended outcomes improve, not by the number of activities purchased.

Wellbeing app. Possible role: Optional individual support.; Cannot prove: That workload or culture is safe.

Awareness campaign. Possible role: Language and signposting.; Cannot prove: That people can raise concerns without detriment.

EAP. Possible role: Confidential support route.; Cannot prove: That organisational risk is controlled.

Fitness challenge. Possible role: Optional activity for some workers.; Cannot prove: Improved population health or inclusion.

What principles are non-negotiable?

Prevent harm at source, consult workers, protect choice and privacy, include every work pattern, use evidence honestly, assign authority and review effects including harm.

No initiative may be used to shift responsibility for harmful work onto individuals, compel health disclosure, penalise non-participation or claim success from attendance alone.

Prevention first. In practice: Change demands, control and support.; Reject: Resilience as substitute for control.

Voluntary support. In practice: No detriment for declining.; Reject: Participation linked to appraisal.

Minimum data. In practice: Purpose-limited access and retention.; Reject: Named health dashboard for managers.

Learning. In practice: Act on positive and adverse findings.; Reject: Campaign continued for appearance.

How can a weak strategy allow harm to develop?

A strategy fails when it counts activities while leaving excessive demand, hostile contact or unsafe work unchanged. Repeated pressure can disturb sleep and concentration and contribute to work-related stress; in a safety-critical task, fatigue or poor communication can also contribute to physical injury. Identify the exposure and change the work, rather than labelling a person as the hazard.

Use worker accounts, incidents and work observations to trace source, route, frequency and affected group. A private offer may help one person, but it does not remove an overloaded rota or an abusive public interaction. Acute threats need an immediate task and emergency response; persistent pressure needs assessed and resourced work redesign.HSE Management StandardsHSENI Management Standards

Exposure route, outcome and action
Work source and routePossible harm and early signalControl or stop decision
Unmanageable demand and missed breaksRepeated pressure can disrupt sleep and concentration; fatigue and errors can build and may contribute to stress-related ill health.Reduce demand or add competent cover; hold a safety-critical task when attention or staffing cannot be assured.
Hostile public contact, bullying or retaliationThreats may cause acute distress or injury; repeated exposure can contribute to anxiety or trauma and deter reporting.Separate the threat, use trained security and fair complaint routes; call emergency services for immediate danger.
Lone or night work with unreliable contactA person in crisis or a hazardous task error may go unrecognised; delay can worsen harm.Test contact and handover, stop affected work when contact fails and confirm accepted emergency help.
Programme pressure or identifiable health reportingCoercion and privacy breach can cause distress, exclusion and delayed help-seeking.Pause the activity, restrict access, protect the worker and correct the data and choice design before restart.

What is the strategy lifecycle?

Move from context and need through consultation, priorities, objectives, controls and programme selection, due diligence, implementation, monitoring, evaluation, corrective action and leadership review.

Prevention-first strategy lifecycle

Prevention-first strategy lifecycle. Readable native learning diagram with labelled relationships and an accessible text equivalent.
Learning point: Prevention-first strategy lifecycle. Use the named evidence and owner for the actual task; this illustration does not authorise work or change the current review status.
Read the complete visual as text
  1. 1. DEFINE NEED AND CONSULT: Strategy owner uses mixed work and worker evidence, including night, remote, lone, contractor and disabled groups. Confirm GB or NI scope and consult representatives before decisions.
  2. 2. PRIORITISE AND DEFINE OUTCOMES: Remove or redesign harmful work first: workload, staffing, role, behaviour, change and support. Assign owners, resources and measurable objectives; optional programmes add support.
  3. 3. CHECK PROGRAMMES: Compare work change, existing routes, purchase and no purchase. Verify evidence, equality, privacy, clinical boundaries, handover, evaluation and exit before readiness approval.
  4. 4. IMPLEMENT AND MONITOR: Resource managers and delivery, verify access and accepted referral routes, monitor work controls, process, outcomes and adverse effects.
  5. 5. EVALUATE AND DECIDE: Leadership receives evidence, uncertainty and worker feedback. Continue, redesign, scale through new approval, or stop with people supported and data controlled.

Each gate can return the proposal for redesign or stop it. Starting with a preferred product reverses the process and can hide the problem that actually needs control.

Understand. Gate question: What work problem and affected groups are evidenced?; Output: Needs and risk statement.

Design. Gate question: Will the proposed route address causes and be inclusive?; Output: Objectives and control plan.

Approve. Gate question: Are evidence, resources, privacy and ownership adequate?; Output: Signed readiness record.

Operate. Gate question: Is delivery safe and as intended?; Output: Monitoring and corrective action.

Review. Gate question: Continue, redesign, scale or stop?; Output: Leadership decision and feedback.

How should wellbeing governance be structured?

Use a senior sponsor, accountable strategy owner, cross-functional steering group, worker representatives, programme owners and independent challenge with written terms and decision rights.

The group needs authority to change work and budgets, not only coordinate communications. Conflicts of interest, vendor influence and approval thresholds should be visible.

Senior sponsor. Authority: Resolve blocked resources and risk.; Accountability: Overall strategy decision.

Steering group. Authority: Prioritise, challenge and recommend.; Accountability: Minutes, actions and conflicts.

Programme owner. Authority: Deliver, monitor and pause within scope.; Accountability: Controls and evidence.

Independent reviewer. Authority: Challenge claims and evaluation.; Accountability: Recorded opinion and limitations.

When and how should workers be consulted?

Consult affected workers and representatives before priorities, procurement, measures and delivery are fixed, then explain what was heard, decided and changed.

Include night, remote, part-time, temporary, agency, contractor and disabled workers. Provide information, time, accessible routes and protection from retaliation. A survey sent after purchase is not consultation.

Needs. Consult about: Work factors, barriers and existing support.; Close the loop: Publish protected themes.

Design. Consult about: Objectives, access and unintended effects.; Close the loop: Show accepted and rejected proposals.

Pilot. Consult about: Usability, safety and fairness.; Close the loop: Correct before scale.

Review. Consult about: Actual effects and next decision.; Close the loop: Continue, redesign or stop with reasons.

How is a defensible wellbeing needs assessment made?

Triangulate work-risk assessments, protected worker feedback, absence and turnover patterns, incidents, complaints, change plans, equality evidence and service data without treating any one source as the answer.

Survey scores are indicators, not diagnoses or proof of cause. Check who is missing, subgroup re-identification risk, data quality, seasonal effects, selection bias and whether trends reflect work or external change.

Risk assessment and observation. Useful for: Work factors and exposed roles.; Limitation: May miss unreported experience.

Survey or focus group. Useful for: Perception and possible hot spots.; Limitation: Selection, wording and anonymity limits.

Administrative trend. Useful for: Where to investigate.; Limitation: Does not prove cause or programme impact.

Complaints and incidents. Useful for: Serious barriers and control failure.; Limitation: Access and under-reporting vary.

How does work design sit inside the strategy?

Assess tasks, staffing, pace, hours, shifts, remote and lone work, public contact, travel, role clarity, change, support, equipment and environment, then control foreseeable harm through ordinary management.

Work can cause, contribute to, trigger or reveal poor wellbeing. The response should examine job and organisational factors as well as individual needs and should not default to blame.

Demand and staffing. Possible control: Priorities, capacity, breaks and recovery.; Evidence of effect: Actual workload and missed-break trend.

Control and autonomy. Possible control: Decision latitude and escalation.; Evidence of effect: Worker-reported control and exceptions.

Support and relationships. Possible control: Supervision, resources and behaviour control.; Evidence of effect: Access, action and complaint themes.

Change and role. Possible control: Consultation, clarity and transition support.; Evidence of effect: Questions resolved and controls reviewed.

How should the six Management Standards be used?

Use demands, control, support, relationships, role and change as a structure for work-stress assessment, worker discussion, action planning and review in the correct GB or Northern Ireland route.

The six areas are not a wellbeing score or clinical diagnostic tool. Use mixed evidence and consultation to identify underlying causes and test controls.

AreaAskExample action
DemandsAre workload, patterns and environment manageable?Rebalance volume and staffing.
ControlCan people influence how work is done?Increase discretion within safe limits.
SupportAre resources and supervision available?Provide competent cover and escalation.
RelationshipsIs unacceptable behaviour controlled?Standards, reporting and fair action.
RoleAre responsibilities compatible and clear?Resolve conflict and ambiguity.
ChangeAre people informed and involved?Consult before and during transition.

How should priorities be chosen?

Prioritise material work risks, serious inequality, repeated concerns, exposed groups, feasible controls and strategic changes using transparent criteria agreed with workers.

Popularity is not the same as need. A visible low-risk campaign must not displace action on harmful demand, violence, bullying, job insecurity or inaccessible work.

Risk and severity. Question: What harm is foreseeable and how serious?; Record: Risk evidence and uncertainty.

Reach and equity. Question: Who is affected or excluded?; Record: Population and access analysis.

Control opportunity. Question: Can work be changed now?; Record: Owner, resource and due date.

Evidence gap. Question: What must be tested before scale?; Record: Pilot and evaluation question.

How can the hierarchy of controls prioritise changes to harmful working conditions in a wellbeing programme?

Assess the work demands and remove or redesign avoidable sources of work-related stress before relying on individual support.

HSE’s risk-assessment route considers demands, control, support, relationships, role and change. Assign a responsible owner and review whether the work change helps.

ERICPD (Eliminate, Reduce including substitution, Isolate, Control, personal protective equipment (PPE), Discipline) is a SHARE decision prompt cross-mapped to HSE’s five-level hierarchy. Choose stronger feasible controls first, verify them in the actual task and record why a stronger option cannot be used. Discipline means sustaining controls through maintenance, supervision and learning. Apply this proportionately to the actual work risk; body size, distress or a health result is not itself a control hierarchy. Organisational rules are administrative controls, and support services do not replace prevention.

A new support portfolio does not replace harmful work

Replacing one support service with another does not itself substitute the source of risk. For example, replacing an on-call rota that repeatedly interrupts recovery with an assessed staffed cover arrangement changes the harmful demand; verify that workload and risk have not simply transferred to another group. Keep optional support separate from this work-control decision.

What should wellbeing objectives look like?

Set specific, measurable, time-bound objectives for work conditions, access, control implementation and learning, without turning worker feelings or health into performance quotas.

An objective should state the affected group, baseline, intended change, measure, owner, decision date and limits. Avoid targets that pressure people to report higher wellbeing or lower absence.

Improve wellbeing by 20 percent. Better objective: Close agreed late-shift workload controls by Q2 and test effect with staff.; Protection: No individual score target.

Increase app uptake. Better objective: Offer accessible optional support and assess whether intended users find it useful.; Protection: No participation pressure.

Reduce absence. Better objective: Identify and control work factors in roles with repeated stress signals.; Protection: No causality claim or attendance coercion.

Who owns which decisions?

Name the owner for strategy, work controls, clinical or OH advice, privacy, equality, procurement, safeguarding, evaluation, corrective action and final continuation or stop.

Managers need authority and escalation routes. Peer supporters and champions listen and signpost but do not diagnose, keep case files, investigate or own risk alone.

DecisionOwnerBoundary
Work-risk controlsOperations and line managementOwn task, workload, staffing, hours and practical controls; do not transfer this duty to a wellbeing programme.
Clinical or occupational-health adviceQualified clinician or competent occupational-health professionalAssess and advise within competence and consent; managers receive only the work information they need.
Privacy and worker dataPrivacy lead and accountable controllerApprove purpose, access, retention, processors and incident route before collection.
Programme hold, release or stopSenior sponsor and named programme ownerResource the decision, retain specialist approvals and act on evaluation or harm signals.

What resources must the strategy provide?

Budget time, competent people, manager capacity, worker participation, accessible delivery, adjustment funding, data governance, evaluation and cover for action on work.

Training without time or authority to change workload is not implementation. Resource plans should include maintenance, supervision, contract management and exit costs, not only purchase price.

Management capacity. Minimum provision: Time, authority and cover.; Failure signal: Actions repeatedly deferred.

Competence. Minimum provision: OH, safety, privacy, equality and evaluation advice.; Failure signal: Decisions outside capability.

Worker participation. Minimum provision: Paid time and accessible routes.; Failure signal: Only office-day voices heard.

Lifecycle funding. Minimum provision: Implementation, evaluation and exit.; Failure signal: Pilot cannot be reviewed or stopped cleanly.

How does the strategy transfer across sectors?

The same prevention principle produces different decisions in different work: reduce a production demand, re-plan a construction interface, protect hospitality staff from abuse or reschedule an unsafe delivery route. Each setting needs its actual risk assessment, funded owner and check across shifts and contractors.

A programme can provide optional support alongside these controls, but high uptake does not prove that work risk has fallen. Check near misses, worker experience and the planned control; stop unsafe work immediately where cover, communication or protection is missing.

Sector task, harm route and strongest feasible control
Task and peopleHarm route and outcomeControl, owner and stop check
Manufacturing — Operators on a short-staffed lineMissed relief and high pace can cause fatigue, stress and attention errors around machinery.Operations reduces output or adds competent relief; verify breaks and hold hazardous running without safe staffing. HSE work-stress control. HSENI work-stress control.
Construction — Trades facing a late sequence changeUnclear roles and time pressure can produce poor coordination, stress and serious task error.Principal contractor re-plans interfaces and confirms supervision; stop a lift or high-risk task while roles are unclear. HSE work-stress control. HSENI work-stress control.
Hospitality — Reception staff facing an abusive guestRepeated hostility can cause distress; an acute threat can cause physical injury.Duty manager removes exposure and uses trained security and safe cover; call 999 for immediate danger. HSE work-stress control. HSENI work-stress control.
Logistics — Lone driver with an impossible routeTime pressure and isolation can cause fatigue and unsafe driving, with collision injury possible.Dispatch re-plans demand, breaks and contact; stop driving when safe function or contact cannot be assured. HSE work-stress control. HSENI work-stress control.

How should a wellbeing programme portfolio be balanced?

Map each initiative to a documented need and tier: organisation-wide prevention is the foundation, followed by optional individual support and targeted support where justified.

Show dependencies and gaps. Several individual apps do not become an organisational strategy, and one targeted service should not be expected to repair unsafe work design.

TierExamplesPortfolio test
Organisation-wide preventionWork design, staffing, workload, control, support and consultationFoundation is linked to assessed work risks and owned corrective action.
Optional individual supportInformation, employee assistance, peer signposting or learningChoice is real, access is equitable and the offer does not replace work controls.
Targeted specialist supportOccupational health, referral, rehabilitation or return-to-work supportEligibility, consent, competence, accepted handover and follow-up are clear.

How should a programme be selected?

Start with the defined need and objective, compare no-purchase and work-design options, assess evidence and fit, consult users, review risks and approve only with an evaluation and exit plan.

Free or discounted products still require due diligence. A prestigious logo, testimonial, award or supplier case study is not evidence that the programme will work for this workforce.

GateQuestionOutcome
NeedWhat problem and population are defined?Reject solution looking for a problem.
AlternativesCan work or an existing route solve it better?Prefer effective source control.
FitDoes evidence match the objective and group?Pilot, redesign or reject.
ReadinessAre data, access, referral and resources controlled?Approve only when conditions are met.

How should evidence and vendor claims be judged?

Check intervention, comparison, population, setting, outcomes, duration, effect size, harms, independence and uncertainty, then state only claims the evidence supports.

Do not promise reduced absence, productivity, health improvement or return on investment from weak association, testimonials or a vendor calculator. Record conflicts and whether evidence is peer reviewed, independent and transferable.

Claim checkAskDecision
Population and settingDoes it match roles, shifts and risk?Limit scope or test locally.
OutcomeIs it meaningful and measured validly?Reject proxy-only claims.
Comparison and attributionWhat else could explain change?Use cautious contribution language.
Harms and conflictsWere adverse effects and funding reported?Require disclosure and monitoring.

What must procurement check before contract?

Confirm scope, evidence, accessibility, safeguarding and clinical boundaries, data flow, security, sub-processors, service levels, incident response, change control, audit rights, exit, deletion and ownership.

Programme proposal: need, options and evidence

Programme proposal: need, options and evidence. Readable native learning diagram with labelled relationships and an accessible text equivalent.
Learning point: Programme proposal: need, options and evidence. Use the named evidence and owner for the actual task; this illustration does not authorise work or change the current review status.
Read the complete visual as text
  1. Start: A worker group, risk assessment, survey, leader or supplier proposes an intervention.
  2. Question 1: Are need and affected population clearly evidenced? Decision owner: Programme owner.
  3. YES: Check immediate risk and existing controls. Owner: Programme owner. CONTINUE → QUESTION 2.
  4. NO / UNKNOWN — HOLD: Do not procure; gather proportionate mixed evidence and consult. Owner: Programme owner. HOLD → REPEAT QUESTION 1 AFTER CORRECTION.
  5. Question 2: Would changing work or improving an existing route better address the need? Decision owner: Operations / programme owner.
  6. YES: Assign work control, resources and review. Purchase may only be additional and justified; continue all proposal gates if still needed. Owner: Operations / programme owner. CONTINUE → QUESTION 3.
  7. NO: Define objective, theory of change and success, harm and stop measures; compare credible options including no purchase. Owner: Operations / programme owner. CONTINUE → QUESTION 3.
  8. UNKNOWN: Consult and compare work changes and existing routes before committing a programme. Owner: Operations / programme owner. HOLD → REPEAT QUESTION 2 AFTER CORRECTION.
  9. Question 3: Does independent evidence match intervention, population, setting and outcome? Decision owner: Evidence / programme reviewer.
  10. YES: Record evidence and remaining limitations; continue all readiness gates. Owner: Evidence / programme reviewer. CONTINUE → Access, privacy and delivery readiness.
  11. NO: Check whether the claim or risk itself is unsupported or unsafe. Owner: Evidence / programme reviewer. CONTINUE → QUESTION 4.
  12. UNKNOWN: Treat effectiveness as uncertain and assess unsupported claims or material risks. Owner: Evidence / programme reviewer. CONTINUE → QUESTION 4.
  13. Question 4: Unsupported diagnostic/clinical/outcome claim or unresolved safety or privacy risk? Decision owner: Clinical / privacy / programme owner.
  14. YES: Hold or reject. Correct the claim or control; do not pilot unresolved safety or privacy issues. Owner: Clinical / privacy / programme owner. HOLD → REPEAT QUESTION 4 AFTER CORRECTION.
  15. NO: Record the weak-effectiveness learning question. A bounded pilot may be considered only after every remaining gate passes. Owner: Clinical / privacy / programme owner. CONTINUE → Access, privacy and delivery readiness.
  16. UNKNOWN: Hold until the relevant competent reviewer resolves the claim or risk. Owner: Clinical / privacy / programme owner. HOLD → REPEAT QUESTION 4 AFTER CORRECTION.
  17. Outcome: A supported proposal or bounded learning question proceeds to every remaining gate; work controls continue.
  18. Stop and escalation boundary: Weak effectiveness evidence does not release a safety, privacy or access issue.

Access, privacy and delivery readiness

Access, privacy and delivery readiness. Readable native learning diagram with labelled relationships and an accessible text equivalent.
Learning point: Access, privacy and delivery readiness. Use the named evidence and owner for the actual task; this illustration does not authorise work or change the current review status.
Read the complete visual as text
  1. Start: A proposal remains under consideration after need/options/evidence review.
  2. Question 1: Have all intended groups and representatives been consulted and access, equality, choice and alternatives checked? Decision owner: Programme / equality owner.
  3. YES: Verify clinical, OH, urgent, safeguarding and complaint boundaries with accepted handovers. Owner: Programme / equality owner. CONTINUE → QUESTION 2.
  4. NO / UNKNOWN — HOLD: Redesign before any pilot: include day/night, remote, part-time, agency, contractor and disabled workers. Owner: Programme / equality owner. HOLD → REPEAT QUESTION 1 AFTER CORRECTION.
  5. Question 2: Are purpose, lawful processing and the complete data flow controlled? Decision owner: Privacy and security leads.
  6. YES: Verify every processor/sub-processor, access, retention, incident response and DPIA decision. Review AI, profiling or automated decisions specifically. Owner: Privacy and security leads. CONTINUE → QUESTION 3.
  7. NO / UNKNOWN — HOLD: Hold or reject collection. Resolve purpose, Article 6 basis, applicable Article 9 condition, security and all vendor/data roles. Owner: Privacy and security leads. HOLD → REPEAT QUESTION 2 AFTER CORRECTION.
  8. Question 3: Do contract, change, deletion, exit, resources, competence and stop authority have accepted owners? Decision owner: Procurement / clinical / operations owners.
  9. YES: Test readiness with actual users, access, deletion and referral handover. Owner: Procurement / clinical / operations owners. CONTINUE → QUESTION 4.
  10. NO / UNKNOWN — HOLD: Hold or redesign; a free product and a paid product need the same evidence. Owner: Procurement / clinical / operations owners. HOLD → REPEAT QUESTION 3 AFTER CORRECTION.
  11. Question 4: Have actual user, access, deletion and handover tests passed with all gate issues closed? Decision owner: Accountable owner.
  12. YES: Submit complete gate evidence and residual effectiveness uncertainty for explicit decision. Owner: Accountable owner. OUTCOME → RECORD AND HAND OVER.
  13. NO / UNKNOWN — HOLD: No pilot: correct failed tests, unknown sub-processors, excluded shifts or unaccepted clinical handover. Owner: Accountable owner. HOLD → REPEAT QUESTION 4 AFTER CORRECTION.
  14. Outcome: Complete gate evidence proceeds to accountable decision; a failed gate remains held.
  15. Stop and escalation boundary: Individual mood/activity data is not a line-manager dashboard; a consent checkbox alone is not all legal grounds.

Procurement should involve intended users and the specialists who will own the risks. Contract language must not convert optional participation into monitoring or permit reuse of worker data for unrelated product development.

Due-diligence areaMinimum evidenceHold trigger
Need and alternativesDefined need, population, work-design and no-purchase optionsNo clear objective or affected workforce.
Evidence and claimsIndependent evidence, limits and proportionate evaluation planDiagnostic, safety or outcome claim exceeds evidence.
Safety and referralCompetence, service hours, escalation and tested handoverUrgent or safeguarding route is absent or untested.
Data and securityPurpose and flow map, access, processors, retention, deletion and incident routeUnknown data flow, sub-processor or lawful route.
Equality and accessIntended-user testing, adjustments and a practical alternativeMaterial group is excluded or participation is coerced.
Contract and exitService levels, change, audit, continuity, export and deletion evidenceOwnership, supplier change or exit cannot be controlled.

How should vendor data flows be controlled?

Map every data item from collection to deletion, including inferred health status, devices, integrations, analytics, profiling, AI, hosting, support access, sub-processors and onward sharing.

The employer remains responsible for its purposes and cannot outsource accountability. Review vendor term, model, hosting or ownership changes before they take effect.

Data-flow pointDecisionEvidence
CollectionIs each field necessary and expected?Field-purpose register.
Use and inferenceIs health or risk inferred or scored?Logic, condition and safeguard.
Access and sharingWho can see identifiable or small-group data?Role matrix and processor list.
ExitCan data be exported and securely deleted?Tested return and deletion evidence.

What privacy rules apply to wellbeing data?

Define a necessary purpose, Article 6 lawful basis and Article 9 condition where health data is processed, then minimise data and access, explain use, set retention and security and complete a DPIA where required. Hold collection or procurement while purpose, route, sub-processors, access, inference, retention or exit is unresolved.

Voluntary programme use does not make employment consent an automatic lawful basis. The named privacy owner releases processing only after the actual flow, necessity, proportionality, transparency, access, small-group risk, incident response, deletion and contract evidence are approved; small groups may be identifiable without names.

Purpose and legal route. Control: Document basis, condition and necessity.; Failure to avoid: Collect now and decide later.

Choice. Control: Real alternative and no detriment.; Failure to avoid: App consent tied to employment pressure.

Aggregation. Control: Suppress or combine small cells.; Failure to avoid: Manager infers individual health.

Access and retention. Control: Need-to-know roles and deletion schedule.; Failure to avoid: Indefinite shadow health record.

How should equality and access be designed?

Assess barriers for disabled and neurodivergent workers, pregnancy and menopause, religion, language, digital access, caring, income, location, shift and contract status, then provide feasible alternatives.

Offer paid-time access where appropriate and do not make a smartphone, wearable, gym membership, disclosure or fixed daytime attendance the only route. Consultation should test dignity and practical usability.

Shift or location. Design response: Equivalent timing and remote or offline route.; Evidence: Access by group and worker feedback.

Disability or neurodivergence. Design response: Accessible format and reasonable adjustment.; Evidence: User testing and adjustment review.

Digital or financial. Design response: Employer equipment or non-digital alternative.; Evidence: No personal cost or device requirement.

Language or literacy. Design response: Plain language and interpretation route.; Evidence: Comprehension testing.

What does genuinely voluntary participation mean?

Workers can decline, pause or choose an alternative without loss, scrutiny or effect on appraisal, attendance, capability, promotion, insurance, reward, scheduling or discipline.

Do not classify non-participation as disengagement or health risk. Make any mandatory safety or training element explicit, justified and separate from personal wellbeing activity or data.

Invitation. Acceptable: Clear purpose, choice and alternatives.; Not acceptable: Manager pressure or repeated chasing.

Incentive. Acceptable: Modest and non-discriminatory if used.; Not acceptable: Financial penalty or intrusive condition.

Data. Acceptable: Aggregate implementation insight.; Not acceptable: Line-manager view of personal activity.

Withdrawal. Acceptable: Simple and without detriment.; Not acceptable: Loss of access to unrelated benefit.

What are the clinical and occupational-health boundaries?

Publish role boundaries so managers and champions can address work and signpost support without diagnosing, treating or making clinical fitness decisions.

Managers own immediate work controls and agreed follow-up. A competent clinician or occupational-health professional owns assessment and clinical or functional advice within consent and scope. The programme owner must provide supervision, an alternative contact and a tested handover route; pause any feature that leaves an unqualified person as the sole responder.

What must different roles learn and demonstrate?

Awareness, work-risk control, programme approval, clinical support and audit are different competences. Test the decision each role makes, identify who assesses it and keep proportionate evidence. Training cannot substitute for authority, staffing or a functioning control.

Use scenarios based on a blocked break, unsafe public contact, privacy breach and urgent disclosure. Reassess after a changed role, service route, failed control or new process. Do not imply that all roles need the same certificate or fixed renewal interval.

Learning, evidence and reassessment by role
RoleLearning and practical demonstrationAssessor, record and trigger
Worker or nearby colleagueRecognise harmful work, report and use a current emergency route without diagnosing.Supervisor checks a scenario and reporting route; record briefing, revisit after a task or route change.
Frontline manager or contractor coordinatorPlan safe work, listen privately, hold unsafe duty, consult and hand over to a named owner.Senior manager observes a scenario and a redacted action file; reassess after failed control or incident.
Wellbeing lead or programme purchaserDefine need, test evidence and alternatives, equality, data, referral and exit plan.Governance lead checks a decision file and stop gate; reassess after vendor or purpose change.
OH, clinical or safeguarding specialistWork within professional scope and give only necessary functional or urgent advice.Professional lead checks scope and accepted handover; use the relevant competence process.
Senior leader or designerFund source controls, approve priorities and challenge unsupported outcome claims.Board sponsor checks action closure and worker feedback; reassess after structural change.
Safety adviser, privacy specialist or auditorTrace law, hazard, data route, control evidence and adverse effects.Competent lead samples records and real shifts; repeat after findings or changed law.

How should a strategy and programme be communicated?

State the need, objective, eligibility, voluntary or mandatory elements, data use, confidentiality limits, support and complaint routes, accessibility, evaluation and stop process in plain language.

Do not use fear, moral judgement, exaggerated benefits, before-and-after body or mood claims, or language that makes wellbeing a test of commitment. Keep support separate from promotional pressure.

Launch. Must include: Why, for whom, choice and owner.; Avoid: Guaranteed benefit.

Privacy. Must include: Data, recipients, retention and rights.; Avoid: Vague anonymous claim.

Support. Must include: Clinical, OH and urgent boundaries.; Avoid: App as emergency service.

Learning. Must include: What will be measured and how used.; Avoid: Hidden performance monitoring.

What must be ready before launch?

Do not launch until work controls, consultation, evidence, equality and privacy review, referral routes, competence, access, resources, monitoring, incident response, evaluation, stop authority and exit arrangements are verified.

Test the actual worker journey, including night shift, remote access, accessibility, privacy notices, referral handover and vendor support. Named approvals must be recorded.

Access. Test: Every intended group can use or choose an alternative.; Release evidence: Scenario and user test.

Safety. Test: Referral, safeguarding and urgent routes work.; Release evidence: Accepted handover drill.

Data. Test: Notices, roles, flow and deletion match reality.; Release evidence: DPIA or assessment and contract.

Operations. Test: Owner, capacity, complaints and stop route are live.; Release evidence: Signed checklist and issues closed.

How should pilots and programme changes be controlled?

Weak effectiveness evidence may justify only consideration of a bounded learning pilot, not pilot approval. Reject unsupported diagnostic claims and hold unresolved safety, privacy, data, equality or access issues. Every proposal must pass consultation, equality and accessibility, voluntary choice, clinical and safeguarding, data, vendor, contract, resources, competence, handover and readiness gates before a named owner approves a pilot.

Release evidence must include representative consultation including night and contract workers, an equality and access test, clinical and safeguarding boundaries, a complete data map and lawful-route record, DPIA decision, known sub-processors, security and contract controls, deletion and exit tests, resources, competence, an accepted referral-handover drill, monitoring, evaluation and stop authority. Weak effectiveness evidence alone does not authorise a pilot: it may support only a bounded, independently appraised learning question after every readiness gate passes. Unknown sub-processors, excluded night workers or unresolved safety, privacy, equality or access safeguards prohibit pilot approval. Scaling requires new evidence and approval.

Evidence and claim. Hold or stop: Reject unsupported diagnostic claims; treat weak effectiveness evidence only as a learning question.; Release evidence and owner: Independent evidence appraisal, bounded uncertainty and approved learning objective; evaluation owner signs.

Consultation, equality and choice. Hold or stop: Redesign if night, disabled, part-time, agency or contract workers are excluded, or choice is not voluntary.; Release evidence and owner: Representative consultation, accessibility and equality tests and feasible alternatives; equality and workforce owners sign.

Clinical, data, vendor and contract. Hold or stop: Hold or reject for unsupported clinical use, unsafe referral, unknown sub-processors or uncontrolled processing, change, deletion or exit.; Release evidence and owner: Approved boundaries, data map, lawful route, DPIA decision, sub-processors, security, contract, deletion and exit; clinical, privacy, security and procurement owners sign.

Resources, readiness and decision. Hold or stop: Do not start without resources, competence, accepted handover, monitoring, evaluation, stop authority or named approval.; Release evidence and owner: Readiness test and referral drill passed; accountable programme owner approves a time limit, baseline, review date and stop criteria. Scale needs new approval.

What should routine monitoring cover?

Monitor resources, delivery fidelity, reach and equitable access, work-control completion, support and complaint performance, data quality, outcomes and adverse effects against defined thresholds.

What evidence can a programme decision use?

What evidence can a programme decision use?. Readable native learning diagram with labelled relationships and an accessible text equivalent.
Learning point: What evidence can a programme decision use?. Use the named evidence and owner for the actual task; this illustration does not authorise work or change the current review status.
Read the complete visual as text
  1. INPUTS AND PROCESS: Budget, trained capacity and manager time show resources. Reach, wait, dropout and delivery fidelity show implementation; participation is not proof of benefit.
  2. LEADING WORK CONTROLS: Workload actions and worker-verified closure test prevention. Conversation counts or EAP uptake cannot replace actual control checks.
  3. OUTCOMES: Use a baseline, theory of change and objective-relevant change with qualitative experience and equitable access. Absence or productivity alone does not prove causation.
  4. ADVERSE EFFECTS: Check pressure, stigma, exclusion, privacy loss and displaced controls. Averages can hide group harm; suppress identifiable small cells.
  5. DECISION AND UNCERTAINTY: Triangulate alternative explanations. State limitations; do not use wellbeing scores as diagnosis or fitness thresholds. Leadership investigates, corrects, holds or stops.

Use the least intrusive data that can answer the question. Monitoring tells you what is happening; it does not by itself establish why or prove programme impact.Magenta Book evaluation guidance

Inputs. Example: Budget, trained capacity and manager time.; Decision use: Can delivery be sustained?

Process. Example: Access, wait, drop-out and control closure.; Decision use: Is implementation working fairly?

Leading control. Example: Workload or change actions verified.; Decision use: Are causes being addressed?

Outcome and harm. Example: Relevant change and unintended effect.; Decision use: Continue, investigate or stop.

How should a programme be evaluated?

Plan evaluation before launch using a theory of change, baseline, questions, process and outcome measures, qualitative evidence, comparison or alternative explanations where feasible, and decision dates.

Ask what worked, for whom, in which context, why, at what cost and with what unintended effects. Use proportionate expertise and distinguish contribution from causation.

Evaluation elementQuestionOutput
Theory of changeHow should activity lead to outcome?Assumptions and failure routes.
ProcessWas delivery implemented and received as intended?Reach, fidelity and context.
Impact or outcomeWhat changed and what else may explain it?Cautious effect assessment.
Value and decisionWere benefits, costs and harms acceptable?Continue, redesign, scale or stop.

How should wellbeing metrics be interpreted?

Treat participation, survey scores, absence, turnover, productivity and service use as partial signals with definitions, denominators, uncertainty and context, never as individual diagnosis or fitness thresholds.

High uptake can reflect pressure or incentives; low uptake can reflect good health, poor access, distrust or irrelevance. Falling absence may reflect presenteeism or policy change. Triangulate before acting.

MetricCan indicateCannot establish alone
ParticipationReach and implementation pattern.Health benefit or commitment.
Wellbeing scoreReported population trend.Diagnosis, safety or individual cause.
Absence or turnoverPattern needing investigation.Programme impact or worker motivation.
ProductivityOperational change in context.Wellbeing causation.

Evaluation: average improvement with a coverage failure

A 60-person service compares eight baseline weeks with eight pilot weeks. Recorded missed-break shifts fall from 24 of 240 staffed shifts to 12 of 240 after extra relief cover, while a demand reduction could also explain part of the change. Only 48 day workers received the new cover; 12 night workers were excluded and report continued missed breaks. The evaluation owner records the denominator, coverage and alternative explanation, redesigns night cover and repeats the review before considering expansion. Attendance or the overall average alone cannot justify a causal wellbeing claim.

What adverse and unintended effects should be watched?

Look for coercion, stigma, distress, exclusion, retaliation, unsafe delay, workload added by participation, privacy loss, false reassurance, displacement of controls and inequitable benefit.

Monitoring signal: contain, correct and verify

2 linked parts. Follow the question numbers and complete the applicable route.

Part 1 of 2 · Questions 1–3

Monitoring signal: contain, correct and verify. Readable native learning diagram with labelled relationships and an accessible text equivalent.
Continue · Part 2 of 2 · Questions 4–5Monitoring signal: contain, correct and verify. Readable native learning diagram with labelled relationships and an accessible text equivalent.
Learning point: Monitoring signal: contain, correct and verify. Use the named evidence and owner for the actual task; this illustration does not authorise work or change the current review status.
Read the complete visual as text
  1. Start: Feedback, complaint, incident, vendor change or evaluation suggests a failure.
  2. Question 1: Immediate danger, serious harm, coercion, discrimination, privacy/security breach, clinical/safeguarding failure or uncontrolled vendor change? Decision owner: Programme owner / competent response lead.
  3. YES: Contain now; stop affected processing/activity; protect and support people; use emergency, safeguarding, privacy, security or formal route; notify senior owner. Owner: Programme owner / competent response lead. CONTINUE → QUESTION 2.
  4. NO: Validate signal without exposing individuals. Owner: Programme owner / competent response lead. CONTINUE → QUESTION 2.
  5. UNKNOWN: Restrict affected activity while competent owners assess potential material harm. Owner: Programme owner / competent response lead. HOLD → REPEAT QUESTION 1 AFTER CORRECTION.
  6. Question 2: Are data quality, denominator, baseline, missing groups and alternative explanations understood? Decision owner: Evaluation / privacy owner.
  7. YES: Classify implementation, work-control, outcome, adverse-effect and vendor/data issues; more than one can apply. Owner: Evaluation / privacy owner. CONTINUE → QUESTION 3.
  8. NO / UNKNOWN — HOLD: Gather protected evidence; protect small groups and do not let a vendor control essential assurance evidence. Owner: Evaluation / privacy owner. HOLD → REPEAT QUESTION 2 AFTER CORRECTION.
  9. Question 3: Is corrective action owned, resourced and consulted with affected workers? Decision owner: Accountable owner.
  10. YES: Determine whether operation can continue safely during correction. Owner: Accountable owner. CONTINUE → QUESTION 4.
  11. NO / UNKNOWN — HOLD: Escalate blocked authority or resource; action remains open. Owner: Accountable owner. HOLD → REPEAT QUESTION 3 AFTER CORRECTION.
  12. Question 4: Can affected activity safely continue while corrected? Decision owner: Competent programme owner.
  13. YES: Apply correction and enhanced monitoring. Owner: Competent programme owner. CONTINUE → QUESTION 5.
  14. NO: Hold programme or affected feature; support users and correct before restart. Owner: Competent programme owner. HOLD → REPEAT QUESTION 4 AFTER CORRECTION.
  15. UNKNOWN: Hold affected activity until safety and data risks are resolved. Owner: Competent programme owner. HOLD → REPEAT QUESTION 4 AFTER CORRECTION.
  16. Question 5: Are effectiveness, access, privacy and unintended effects verified, with residual risk accepted? Decision owner: Accountable owner / reviewers.
  17. YES: Prepare leadership findings, limitations, worker response and recommendation. Owner: Accountable owner / reviewers. CONTINUE → Leadership programme decision.
  18. NO: Escalate resources, redesign or terminate. Owner: Accountable owner / reviewers. HOLD → REPEAT QUESTION 5 AFTER CORRECTION.
  19. UNKNOWN: Keep action open and obtain actual evidence; do not use an optimistic dashboard as closure. Owner: Accountable owner / reviewers. HOLD → REPEAT QUESTION 5 AFTER CORRECTION.
  20. Outcome: Verified correction proceeds to explicit leadership decision; unresolved harm remains contained.

Make adverse reporting easy and independent of the programme owner or vendor. Protect complainants and small groups. A positive average must not hide serious harm to a subgroup.

Coercion or retaliation. Immediate response: Stop pressure and protect worker.; Review: Management and employment route.

Privacy or re-identification. Immediate response: Contain access and notify privacy route.; Review: Data, contract and reporting duty.

Clinical or safeguarding gap. Immediate response: Use competent route and hold affected pathway.; Review: Boundary and escalation design.

Work controls displaced. Immediate response: Restore risk action.; Review: Portfolio and leadership decision.

How should corrective action work?

Contain harm, support affected people, preserve necessary evidence, notify the correct lead, investigate system causes, consult, redesign or terminate, assign actions and verify effectiveness before release.

Do not close an issue because a vendor apologised or attendance remains high. The action owner must have authority and a due date, with escalation for blocked resources.

Contain. Action: Pause affected use, sharing or promotion.; Evidence: Time and scope of hold.

Investigate. Action: Examine work, process, data and governance causes.; Evidence: Protected factual record.

Correct. Action: Change control, contract, design or owner.; Evidence: Action and resource.

Verify. Action: Test that risk and harm are reduced.; Evidence: Release or termination decision.

When should a programme be held, stopped or released?

Hold or stop for serious safety, privacy, coercion, discrimination, unsupported diagnostic or outcome claim, unknown data flow or sub-processor, uncontrolled vendor change, failed handover, excluded workforce group, displaced work controls or inability to evaluate material risk.

Only the named accountable owner may release the affected programme or feature after competent specialists verify the cause, implemented and tested controls, worker consultation, equality and access, lawful data and contract routes, accepted clinical or safeguarding handover, residual risk, evaluation readiness and recorded approvals.

DecisionTriggerAuthority
HoldUncertainty can expose people while checked.Programme owner or any named stop role.
Stop or terminateMaterial harm, unlawful route or failed control.Senior accountable owner with specialist advice.
RedesignObjective remains valid but route is weak.Steering group and affected workers.
ReleaseConditions closed and verification passed.Named approval body.

How should crisis and safeguarding boundaries be built in?

Before launch, name the emergency, crisis, safeguarding and complaint routes, their hours, acceptance point and fallback when a route cannot be reached.

Workers and peer supporters must know how to obtain immediate help without becoming investigators or sole responders. The programme owner should test handover, safe contact, minimum information sharing, work cover and follow-up. Use the applicable current national and local route: the cited domestic-abuse guidance applies to England and reserved matters in Wales, with Welsh devolved guidance needed for other matters; the cited crisis page is specific to Northern Ireland. Its employer-support discussion is relevant guidance, not a universal statutory employer duty.

How should contractors, agency workers and supply chains be included?

Host and employer should agree which work risks, consultation, programme access, data, referral, incident and corrective-action duties each controls before work starts.

Do not exclude people exposed to the same work factors or duplicate incompatible assessments. Workers need one clear route even where contractual responsibilities are shared.

Work design. Agree: Who controls task, staffing, hours and site.; Evidence: Coordinated risk action.

Access. Agree: Eligibility, alternatives and paid time.; Evidence: Worker communication.

Data. Agree: Controller, processor and sharing roles.; Evidence: Transparent data map.

Escalation. Agree: Urgent, complaint and stop contacts.; Evidence: Tested route and owner.

What should leaders review?

Review context, risk, objectives, resources, worker voice, control completion, portfolio fit, access, privacy, outcomes, harms, complaints, vendor performance and assurance, then make explicit decisions.

Accountable decision and evaluation

Accountable decision and evaluation. Readable native learning diagram with labelled relationships and an accessible text equivalent.
Learning point: Accountable decision and evaluation. Use the named evidence and owner for the actual task; this illustration does not authorise work or change the current review status.
Read the complete visual as text
  1. Start: Every readiness gate is closed and the accountable owner has the evidence.
  2. Question 1: Is residual effectiveness uncertainty material but bounded? Decision owner: Accountable programme owner.
  3. YES: Approve only a time-limited learning pilot with baseline, support, decision date and stop criteria. Owner: Accountable programme owner. CONTINUE → QUESTION 2.
  4. NO: Record approval with conditions, named owners and evaluation arrangements. Owner: Accountable programme owner. CONTINUE → QUESTION 2.
  5. UNKNOWN: Do not approve; define and bound the uncertainty with competent evidence. Owner: Accountable programme owner. HOLD → REPEAT QUESTION 1 AFTER CORRECTION.
  6. Question 2: Is delivery being evaluated for process, outcomes, equity, adverse effects, cost and alternatives? Decision owner: Evaluation / programme owner.
  7. YES: Assess credible benefit and safeguard effectiveness. Owner: Evaluation / programme owner. CONTINUE → QUESTION 3.
  8. NO / UNKNOWN — HOLD: Hold expansion; correct the evaluation and protect participants. Owner: Evaluation / programme owner. HOLD → REPEAT QUESTION 2 AFTER CORRECTION.
  9. Question 3: Are benefits credible and safeguards effective? Decision owner: Leadership / workers / competent reviewers.
  10. YES: Continue within conditions; any scale needs a new context and approval check. Owner: Leadership / workers / competent reviewers. OUTCOME → RECORD AND HAND OVER.
  11. PARTLY: Redesign and retest through due diligence; do not automatically renew. Owner: Leadership / workers / competent reviewers. HOLD → REPEAT QUESTION 3 AFTER CORRECTION.
  12. NO / MATERIAL HARM: Stop affected service; execute support, contract exit, data return/deletion and lessons. Owner: Leadership / workers / competent reviewers. STOP / REFER → APPROPRIATE RESPONSE.
  13. UNKNOWN: State uncertainty and obtain evidence; hold scale and any unbounded activity. Owner: Leadership / workers / competent reviewers. HOLD → REPEAT QUESTION 3 AFTER CORRECTION.
  14. Outcome: Continue with next review, redesigned bounded programme, newly approved scale or supported exit.
  15. Stop and escalation boundary: Vendor testimonials, participation or a before/after score alone do not establish effectiveness.

Leadership programme decision

Leadership programme decision. Readable native learning diagram with labelled relationships and an accessible text equivalent.
Learning point: Leadership programme decision. Use the named evidence and owner for the actual task; this illustration does not authorise work or change the current review status.
Read the complete visual as text
  1. CONTINUE: Accountable owner records conditions, next review and communication to workers. Preserve limitations and monitor adverse effects.
  2. SCALE: Recheck new context, affected groups, data/vendor arrangements and capacity. Obtain new approval before expanding.
  3. REDESIGN: Return to due diligence and bounded pilot readiness; retest failed access, handover or data controls.
  4. STOP: Support people, execute contract exit and data return/deletion, verify closure and return lessons to the strategy.

Do not accept activity counts as assurance. Challenge whether the organisation changed harmful work, whether evidence supports claims and whether unresolved issues need resources or stop action.

Are causes controlled?. Evidence: Risk actions and worker verification.; Decision: Fund, escalate or redesign.

Is delivery fair and safe?. Evidence: Access, adverse and complaint evidence.; Decision: Correct, hold or stop.

Are outcomes credible?. Evidence: Evaluation and limitations.; Decision: Continue, test or withdraw claim.

Is governance effective?. Evidence: Overdue actions, conflicts and assurance.; Decision: Change ownership or capacity.

How does the strategy improve over time?

Repeat needs evidence, consult, review policies and controls, test changes, learn from adverse effects and update objectives, routes and the portfolio through a documented Plan, Do, Check, Act cycle.

Close the loop to workers and preserve what did not work as organisational learning. A yearly relaunch with the same unresolved work risks is not continual improvement.

Plan. Action: Refresh context, risk and priorities.; Evidence: Updated need and objectives.

Do. Action: Implement controls and approved programmes.; Evidence: Resources and delivery record.

Check. Action: Monitor, evaluate and seek worker feedback.; Evidence: Findings and uncertainty.

Act. Action: Correct, scale, redesign or stop.; Evidence: Leadership decision and verified action.

What assurance should be retained?

Keep proportionate evidence of decisions, consultation, risk and equality review, data governance, procurement, competence, implementation, monitoring, evaluation, corrective action and approval.

Assurance should show how conclusions were reached and limitations handled, without retaining unnecessary personal health narratives. Independent review should be strongest where risk, novelty or claim size is greatest.

Decision record. Owner: Steering group or accountable owner.; Retention focus: Rationale, conditions and conflicts.

Risk and privacy. Owner: Relevant competent leads.; Retention focus: Current controls and residual risk.

Evaluation. Owner: Evaluation owner with independent challenge.; Retention focus: Methods, findings and limitations.

Worker feedback. Owner: Consultation owner.; Retention focus: Themes, response and closure without exposure.

What is the Great Britain legal framework?

In Great Britain, general health and safety, risk assessment, consultation, equality and data-protection duties may apply to work risks and programme processing; there is no single statute that approves a wellbeing strategy or product.

Use the Health and Safety at Work etc Act 1974, Management Regulations 1999, the correct consultation regulations, Equality Act 2010 and UK data-protection framework within their scope. Obtain advice for the facts.

FrameworkRelevanceBoundary
HSWA 1974: Health and Safety at Work etc Act 1974.General health, safety and welfare duties.GB framework.
Management Regulations 1999: Management of Health and Safety at Work Regulations 1999.Risk assessment and arrangements.GB instrument.
Consultation regulations: Safety Representatives and Safety Committees Regulations 1977; Health and Safety Consultation with Employees Regulations 1996.Union and non-union routes.Use correct workforce route.
Equality Act 2010: Equality Act 2010.Discrimination and adjustments.GB, not NI disability framework.
UK GDPR and DPA 2018: Data Protection Act 2018; Personal and health data.Document basis, condition and safeguards.

What is the Northern Ireland legal framework?

Northern Ireland has separate health and safety, management, consultation and disability-discrimination instruments, while UK data protection applies across the United Kingdom.

Use the HSW NI Order 1978, Management Regulations NI 2000, the correct NI consultation regulations and Disability Discrimination Act 1995. Do not substitute GB citations or the Equality Act 2010.

FrameworkRelevanceBoundary
HSW NI Order 1978: Health and Safety at Work Northern Ireland Order 1978.General health, safety and welfare duties.NI only.
Management Regulations NI 2000: Management of Health and Safety at Work Regulations Northern Ireland 2000.Risk assessment and arrangements.NI instrument.
NI consultation regulations: Safety Representatives and Safety Committees Regulations Northern Ireland 1979; Health and Safety Consultation with Employees Regulations Northern Ireland 1996.Union and non-union routes.Use correct NI route.
Disability Discrimination Act 1995: Disability Discrimination Act 1995.NI disability and adjustment framework.Not Equality Act 2010.
UK GDPR and DPA 2018: Data Protection Act 2018; Personal and health data.UK-wide with local service routes.

How should management arrangements be checked?

Assign responsibility, assess the actual risk, implement controls, monitor whether they work and review after change or failure.

HSE’s Plan, Do, Check, Act guidance supports management review. It does not replace topic-specific law, technical assessment or competent sign-off.

Which warning signs require immediate action?

Act on coercion, retaliation, unsafe work, serious distress, privacy or security breach, discriminatory exclusion, unsupported clinical claims, failed referral, unknown data flow and vendor change outside approval.

The purpose is to trigger the correct control and authority, not to diagnose a person or protect programme reputation.

Immediate danger. Immediate action: Use emergency or safeguarding route.; Owner: Trained responder.

Privacy or security breach. Immediate action: Contain and notify privacy/security route.; Owner: Controller and incident owner.

Participation pressure. Immediate action: Stop pressure and protect affected workers.; Owner: Senior and employment owner.

Unsupported claim or adverse effect. Immediate action: Withdraw claim or hold programme.; Owner: Programme and assurance owner.

Controls displaced by campaign. Immediate action: Restore work-risk action.; Owner: Operations and safety owner.

What do the acronyms and key terms mean?

Use terms that separate wellbeing, psychosocial risk, work-related stress, intervention, monitoring, evaluation, health data, adjustment and clinical care.

Acronyms and terms
Term or lettersExpansion and meaning in this topic
HSE and HSENIHealth and Safety Executive for Great Britain; Health and Safety Executive for Northern Ireland.
OH, HR and EAPOccupational health; human resources; employee assistance programme. They have different advice, employment and support roles.
GB and NIGreat Britain and Northern Ireland: related but distinct safety and equality legal routes.
Wellbeing strategyGoverned plan for preventing harm, improving work and providing proportionate support.
Programme or interventionActivity intended to create a defined change, requiring evidence and evaluation.
Psychosocial hazardWork factor that can cause psychological or physical harm, never a label for a person.
Theory of changeTestable explanation of how resources and activities should produce outcomes and for whom.
Validated measureMeasure tested for a specific purpose and group; use it only within those limits.
ERICPDEliminate, Reduce including substitution, Isolate, Control, personal protective equipment, Discipline.A SHARE decision prompt; sustain effective controls and learning.
PPEPersonal protective equipment.Use for assessed residual exposure; it does not replace stronger feasible controls.

Wellbeing strategy. Meaning: Governed plan for prevention, healthy work and support.; Boundary: Not a calendar of benefits.

Programme or intervention. Meaning: Activity intended to create defined change.; Boundary: Needs evidence, controls and evaluation.

Psychosocial hazard. Meaning: Work factor with potential for psychological or physical harm.; Boundary: Not a label for a person.

Theory of change. Meaning: How inputs and activities are expected to produce outcomes.; Boundary: Assumptions must be tested.

Monitoring. Meaning: Routine evidence about delivery and signals.; Boundary: Does not establish causation alone.

Evaluation. Meaning: Systematic assessment of design, delivery and outcomes.; Boundary: Must state uncertainty and alternatives.

Special-category data. Meaning: Personal data concerning physical or mental health.; Boundary: Needs a lawful basis, condition and safeguards.

Reasonable adjustment. Meaning: Change intended to remove or reduce disability disadvantage.; Boundary: Legal framework differs in GB and NI.