Topic 7.6 — Workforce Equality, Diversity & Inclusion

In one sentence: Building a digital, data and technology workforce that fairly reflects the people it serves — and where everyone can belong, contribute and progress — is not an optional value statement but a statutory duty, a clinical-safety concern, and a precondition for building services that work equitably for the whole population.

Why this matters in health and care#

Health and care is delivered by people, for people. Who builds your digital services shapes what those services assume, who they work for, and who they quietly exclude. A team that is homogeneous in background, disability status, gender, ethnicity or lived experience will tend to build products that work well for people like them and badly for everyone else — and it will do so without noticing, because the gaps are invisible from the inside. In a sector where a poorly designed service can mean a missed diagnosis, an inaccessible appointment, or a family unable to reach care, this is a safety and equity issue, not merely a matter of fairness to staff.

The stakes are also legal and public. NHS and social-care organizations are public authorities bound by the Equality Act 2010 and the Public Sector Equality Duty, and are held to account through the NHS Workforce Race Equality Standard (WRES) and Workforce Disability Equality Standard (WDES). The evidence gathered under these standards is stark and persistent: staff from ethnic minority backgrounds and disabled staff report worse experiences of recruitment, progression, bullying and disciplinary processes than their white and non-disabled colleagues. A workforce that does not feel it belongs cannot give its best, and the well-documented link between staff experience and patient experience means that inequity among staff becomes inequity for patients.

This topic is about the workforce — your colleagues, recruits and leaders — and is deliberately distinct from the user-facing concerns of Topic 1.9 — Digital Inclusion & Accessibility. The two are joined: an inclusive workforce is one of your strongest defenses against inequitable services. Treat this alongside Topic 7.0 — Team Structures, Roles & Responsibilities, Topic 7.3 — Workforce Strategy, and Topic 7.5 — People & Organizational Development.

Core concepts#

Equality, diversity, inclusion and belonging. These four words are often bundled but differ. Equality is fair treatment and the absence of unlawful discrimination. Diversity is the range of difference present in a workforce — visible and invisible. Inclusion is whether that difference is actually welcomed, heard and able to shape decisions. Belonging is the felt experience of being accepted as yourself. You can be diverse without being inclusive (people present but ignored), and inclusive-sounding without belonging (people masking who they are to fit in). Aim for all four.

The Equality Act 2010, protected characteristics and the Public Sector Equality Duty. The Equality Act 2010 is the UK's consolidated anti-discrimination law. It protects nine protected characteristics: age, disability, gender reassignment, marriage and civil partnership, pregnancy and maternity, race, religion or belief, sex, and sexual orientation. The Public Sector Equality Duty (PSED, section 149) requires public bodies to have due regard to eliminating discrimination, advancing equality of opportunity, and fostering good relations — in employment as well as in services.

WRES, WDES and the pay gaps. The Workforce Race Equality Standard (WRES) and Workforce Disability Equality Standard (WDES) are mandated NHS mechanisms that require organizations to report on a fixed set of indicators — such as likelihood of being appointed from shortlisting, entry into formal disciplinary processes, and experience of harassment — broken down by ethnicity and disability. Gender and ethnicity pay-gap reporting (see gender pay gap) exposes the average earnings difference between groups, which is largely a story about who holds the senior, better-paid roles rather than unequal pay for the same job.

Intersectionality. Intersectionality is the recognition that characteristics combine: a disabled woman of color does not experience "race issues" and "gender issues" and "disability issues" as separable streams, but as a single, compounded experience. Single-axis monitoring (race or gender or disability) can hide the very groups facing the sharpest disadvantage.

Unconscious bias and its limits. Implicit stereotypes are the automatic, unexamined associations that shape decisions such as who gets shortlisted or stretched. Awareness matters, but the evidence that one-off "unconscious bias training" changes behavior is weak — awareness without structural change tends to fade. Bias is better designed out of processes than trained out of individuals.

Bias in HR and AI hiring tools. Algorithmic bias is the tendency of automated systems to reproduce and amplify the inequities in their training data. CV-screening, video-interview scoring and "culture fit" algorithms have repeatedly been shown to disadvantage women, disabled applicants and minority candidates, while wearing a veneer of objectivity. Automating a biased process makes it faster and harder to challenge, not fairer.

Reasonable adjustments and the social model of disability. A reasonable adjustment (reasonable accommodation) is a change to a job, workplace or process that removes a disadvantage a disabled person would otherwise face — often free or low-cost. It flows from the social model of disability: people are disabled by barriers in the environment, not solely by their condition, so the fix is usually to remove the barrier.

Psychological safety and allyship. Psychological safety — the shared belief that you can speak up without humiliation or punishment — is the mechanism through which inclusion becomes real: without it, difference stays silent. Allyship is the active use of one's own advantage to support colleagues from marginalized groups; done well it shifts effort off those already carrying it, done badly it is performative.

Best practices#

1. Treat EDI as a delivery and safety issue, not an HR side-quest. Frame inclusion in the language your teams already use: better products, fewer blind spots, safer services. Homogeneous teams ship the biases they cannot see, from pulse oximeters calibrated on light skin to forms that assume a stable address; diverse, included teams catch these earlier. Owning EDI as an engineering and product concern — not delegating it to a standalone committee — is what turns intent into changed practice.

2. Redesign recruitment to remove bias, don't just train recruiters. Anonymize applications, use structured interviews with pre-agreed questions and scoring, and always have a diverse, trained panel. Write job descriptions that describe the outcome rather than a narrow pedigree, and cut "nice to have" requirements that screen out capable non-traditional candidates. These structural changes reduce the influence of implicit stereotypes far more reliably than asking people to be less biased.

3. Scrutinize any AI or automated hiring tool before and during use. If a supplier offers CV-screening, gamified assessment or video-interview scoring, demand evidence of validation against adverse impact across protected characteristics, and keep a human accountable for every decision. Under the Public Sector Equality Duty you cannot outsource your duty to an algorithm. Treat these tools as you would any clinical algorithm — subject to bias testing, transparency and the governance in Topic 8.4 — AI & Emerging Technology Governance.

4. Fix progression and retention, not just the front door. Diverse hiring is undone if people cannot progress or choose to leave. Analyse your own WRES and WDES data: where in the pipeline do ethnic minority and disabled staff fall away, and why? Invest in sponsorship, stretch opportunities, transparent promotion criteria, and reverse mentoring — and act on exit-interview themes rather than filing them.

5. Make reasonable adjustments fast, funded and normal. Establish a clear, low-friction route to workplace adjustments — ideally a passport that travels with the person between roles — so no one has to re-justify their needs each time they move. Most adjustments are cheap or free; the real barrier is usually delay, stigma or a manager who does not know the process. Applying the social model of disability, assume the barrier is fixable and start from "how", not "whether".

6. Build psychological safety and real allyship, not just networks. Staff networks for disabled, LGBTQ+, and ethnic minority colleagues are valuable but should not carry the whole burden of change. Give them sponsorship, time and a route to influence decisions, and expect the majority — especially leaders — to practice allyship by using their position to remove barriers. The test of psychological safety is whether a junior colleague would challenge a senior one's assumption in a design review; build the conditions where they would.

7. Set EDI objectives into leadership accountability. What leaders are measured on is what changes. Put specific, evidenced EDI objectives into senior appraisals and board reporting, tied to your WRES/WDES action plans and pay-gap trends. Boards should see the same disaggregated data quarterly that they see for finance and performance, and be expected to explain movement.

8. Measure honestly and resist tokenism. Collect and use disaggregated data — including intersectional cuts where numbers allow — but pair the numbers with lived experience gathered safely. A single senior appointment from an under-represented group is not a strategy, and celebrating it as one can demoralize the people still facing the barriers. Measure the system (shortlisting rates, disciplinary rates, adjustment turnaround, staff-survey belonging scores), not the anecdote.

9. Close the loop from workforce inclusion to equitable services. Explicitly connect your inclusive-workforce work to equitable outcomes for patients and citizens (Topic 1.9). Involve staff with relevant lived experience in research, design and testing — with proper support and recognition, not as unpaid representatives. The purpose is not only fairness to staff but services that do not fail the people least like the average developer.

Questions to discuss with your team#

1. Where in our recruitment, progression and retention pipeline are people from under-represented groups actually falling away — and what in our process, not in them, is causing it? This question forces a move from good intentions to evidence, which is where most EDI conversations quietly stall. In a digital, data and technology function the under-representation is often severe — these roles skew heavily male and, at senior levels, white — so the temptation is to blame a "pipeline problem" in wider society and stop there. The honest work is to look at your own data: relative likelihood of appointment from shortlisting, who enters formal capability or disciplinary processes, who is promoted, and who resigns and why. WRES and WDES give you a national frame for this, but the value is in interrogating your local numbers and asking which specific steps — the wording of adverts, the composition of panels, the informal "tap on the shoulder" that decides who gets stretch work — produce the drop-off. A good answer names concrete stages and process changes with owners, resists the comfort of externalizing the cause, and commits to re-measuring rather than assuming the fix worked.

2. When we adopt tools and processes that promise to make hiring or performance "objective" — including AI — how do we know they are not just automating our existing bias? Digital teams are drawn to automation and data, which makes them unusually exposed to the seductive idea that an algorithm removes human bias from hiring. In reality, systems trained on historic decisions learn historic inequities, and a video-interview scorer or CV-ranker can disadvantage disabled applicants, women or minority candidates while appearing neutral and being harder to challenge than a human panel. The tension to debate is speed and consistency versus fairness and accountability: automation can standardize, but standardizing a biased baseline entrenches it. Consider whether your suppliers can evidence adverse-impact testing, whether a named human remains accountable for each decision, and whether a rejected candidate could meaningfully contest the outcome. Under the Public Sector Equality Duty the duty cannot be delegated to a vendor, so a good answer treats these tools with the same scrutiny as a clinical algorithm — bias-tested, transparent, and always with a human in the loop.

3. Is it genuinely safe for the least powerful person in the room to say "this design assumes something that will fail people like me" — and how would we know? Inclusion only changes products when difference can be voiced, and voice depends on psychological safety, which is unevenly distributed and easy to overestimate from a position of comfort. In practice the person best placed to spot that a service will exclude carers, disabled users or a particular community is often the most junior or the only person from that background in the team — precisely the person who pays the highest price for speaking up and being wrong. The debate should surface how your team actually responds when someone raises an equity concern that slows delivery down: is it welcomed as a caught defect, or treated as a distraction from the sprint? Look for evidence rather than reassurance — when did a challenge like this last happen, who raised it, and what changed as a result? A strong, honest answer distinguishes the team's stated openness from its observed behavior, acknowledges that belonging is not felt equally by everyone present, and identifies what would have to be true for the quietest voice to be heard.

4. When a colleague needs a workplace adjustment, does our process make it fast, funded and dignified — or does the burden of proof and delay fall on them? Reasonable adjustments are where a stated commitment to disability inclusion meets daily reality, and digital teams have both an advantage and a blind spot here. The advantage is that many adjustments in a technology role — assistive software, flexible hours, remote or asynchronous working, changes to meeting formats — are cheap, quick and already within the team's gift; the blind spot is assuming that because the tooling exists, the barrier is solved. The tension to examine is speed and dignity versus process and scrutiny: an adjustment delayed for weeks of assessment, or one that must be re-justified every time someone changes team, is a barrier in its own right, however well-intentioned the paperwork. Consider whether a manager actually knows how to action a request without escalation, whether an adjustments passport travels with the person, and who pays — the individual's energy and disclosure, or a properly funded central budget. Applying the social model of disability, an honest answer starts from "how do we remove this barrier" rather than "should we", names a turnaround time it is prepared to be held to, and treats stigma and delay, not cost, as the real obstacles.

5. Whose disadvantage does our data hide because we only ever look at one characteristic at a time? Most monitoring counts race, sex and disability on separate axes, which is exactly how the groups facing the sharpest, compounded disadvantage stay invisible — a disabled woman of color is not the average of three separate gaps, and single-axis dashboards will never show her. In a digital, data and technology function the numbers in any one intersectional cell are often small, which is both a genuine statistical difficulty and a convenient reason never to look. The trade-off to debate is rigor and anonymity versus visibility: cutting data too finely risks identifying individuals and breaching confidence, yet refusing to cut it at all guarantees you miss the people the standards exist to protect. Pair the quantitative limits with lived experience gathered safely, so that where the cell is too small to report, the story is not simply lost. A strong answer commits to examining intersectional patterns wherever numbers responsibly allow, is candid about where they cannot, and resists treating a headline WRES or WDES figure as the whole picture when it may be averaging away the worst-affected.

6. Are we asking under-represented colleagues and their networks to fix inclusion in their spare time, while the people with the most power stay comfortable spectators? Staff networks and individual advocates do vital work, but there is a recurring injustice in expecting those who carry the disadvantage to also carry the unpaid labor of remedying it, often on top of a full delivery role. The question surfaces who actually does the work of change: who runs the network, who sits on the inclusion panels, who is asked to speak on the panel or review the design for accessibility — and whether that effort is recognized, resourced and reflected in workload, or simply assumed. The tension is between valuable grassroots energy and a leadership that quietly outsources its own accountability to it; a network with sponsorship, time and real influence is empowerment, the same network with none of those is exploitation. Genuine allyship means the majority — especially senior leaders — using their position to remove barriers rather than delegating the effort downward. An honest answer looks at where the burden currently sits, whether lived-experience contribution to research and design is paid and credited or extracted for free, and what leaders are personally doing rather than sponsoring others to do.

In practice: a health & care example#

An integrated care system (ICS) is scaling up a central digital, data and technology directorate of around two hundred people, pulling together staff from several trusts and new external hires. Early data is uncomfortable: the directorate is 80% male, its senior band is almost entirely white, and its inherited WRES and WDES figures show ethnic minority and disabled staff far more likely to enter formal processes and far less likely to be promoted. A "culture fit" screening tool, brought over from a partner trust, is quietly ranking out candidates with career breaks and non-linear paths.

The chief digital information officer refuses to treat this as an HR problem. The directorate rewrites its job descriptions around outcomes, drops inflated degree and "10 years' experience" filters, anonymizes first-sift applications, and requires diverse, trained interview panels using structured scoring. The culture-fit algorithm is suspended pending adverse-impact evidence the supplier cannot produce, and is replaced with a structured work-sample task. A reasonable-adjustments passport is introduced so disabled staff moving between the constituent trusts never have to re-justify their needs.

Progress is measured honestly. The board sees disaggregated shortlisting, progression, disciplinary and adjustment-turnaround data every quarter, alongside belonging scores from the staff survey, and each senior leader carries a specific EDI objective. The disability and ethnic-minority staff networks get sponsorship, time and a standing slot in design reviews, so lived experience shapes services rather than decorating a report. Within a year the shortlisting gap narrows and adjustment turnaround falls from weeks to days; the progression gap barely moves, and the directorate says so plainly rather than claiming a win.

Sector lenses#

Startup#

A small digital-health startup has no EDI function, and the founders' first dozen hires set a culture that is extremely hard to change later — homogeneity compounds through referral hiring from similar networks. The upside is agility: it is far cheaper to build inclusive recruitment habits, a no-fuss adjustments process and a speak-up culture into ten people than to retrofit them into a thousand. The acute risks are a monoculture that builds a product blind to whole user groups, and a lone under-represented hire who carries an invisible burden and leaves. The pragmatic move is to make structured, bias-aware hiring and explicit psychological safety non-negotiable from the first hire, and to widen sourcing beyond the founders' own networks before the pattern sets.

Small business#

An established small provider — a GP practice, a community pharmacy, a care home or domiciliary-care agency, a single-site clinic, or a small health-tech supplier past the pre-revenue stage — has real patients and staff but no HR or EDI specialist, and inclusion is whatever the owner or practice manager makes time for between operational fires. It is still bound by the Equality Act 2010 and, if it delivers NHS-commissioned services, may be pulled into WRES/WDES expectations, yet must meet the same duties as a large trust without the data teams, central adjustment budgets or trained panels to do so. The practical risks are recruitment run entirely through word of mouth in a narrow local network, a reasonable adjustment stalling because no one knows the process, and a single manager's goodwill standing in for policy. The proportionate response is to adopt a handful of cheap, high-leverage habits — a structured interview with agreed questions, a job advert focused on outcomes, a known route to fund and action adjustments, and honest use of Access to Work — rather than a formal program it cannot staff.

Enterprise#

A large NHS trust, ICS or health-tech firm has formal EDI functions, staff networks and mandated WRES/WDES reporting, but fights scale, inertia and inconsistency — inclusion that is real in one team is absent in the next, and the gap between published policy and daily experience is the central problem. The enterprise's advantage is data and leverage: it can analyse pipelines at scale, run pay-gap analysis, fund adjustments centrally, and hold leaders accountable through appraisal and board reporting. The characteristic failure is treating reporting as the goal — producing a polished WRES action plan that changes nothing on the ground. The task is to convert statutory reporting into operational change, with disaggregated data governed as seriously as finance and named ownership for every gap.

Government#

A national body — NHS England, the Department of Health and Social Care, a devolved government or a large local authority — sets the standards that everyone else implements, holding the levers of the Public Sector Equality Duty, WRES, WDES and mandatory pay-gap reporting. Its power is framework and funding rather than direct line management, so its impact depends on adoption it cannot simply mandate, and its own workforce is under the same public scrutiny it applies to others. Accountability runs to ministers, Parliament and the public, which makes honesty about slow progress politically hard but essential. The recurring risk is that national EDI mechanisms become compliance rituals — data returned, boxes ticked — without changing the lived experience of staff, so government work must be evaluated for real-world impact and must model the inclusion it requires of the system.

Common failure modes#

• EDI as a poster, not a process. Values on the wall and a network group, but recruitment, progression and adjustments unchanged. Nothing measurable moves.
• Unconscious-bias training as a fix. A one-off e-learning module treated as the intervention; awareness rises briefly, behavior and structures do not change.
• Automating bias. Adopting AI or "culture fit" screening that entrenches historic inequity behind a claim of objectivity, with no adverse-impact testing or human accountability.
• Tokenism and the single success story. Celebrating one senior appointment as proof of progress while the pipeline behind it is unchanged — which demoralizes rather than reassures.
• Reporting without action. A compliant WRES/WDES return that documents disparities year after year without closing them; the data becomes an alibi.
• Overloading the marginalized. Expecting under-represented staff and their networks to fix inclusion in unpaid extra time, while leaders and the majority stay passive.
• Single-axis blindness. Monitoring race, sex and disability separately and missing the intersectional groups facing the sharpest, compounded disadvantage.

Maturity model#

Checklist#

• Job descriptions describe outcomes, not narrow pedigree; inflated filters removed.
• Recruitment uses anonymized sift, structured scoring, and trained diverse panels.
• Any AI or automated hiring tool is evidenced for adverse impact, with a human accountable for every decision.
• WRES, WDES and gender/ethnicity pay-gap data are reviewed by the board, disaggregated, with action owners.
• Progression and exit data are analysed by characteristic, and pipeline drop-off points are being addressed.
• A fast, funded reasonable-adjustments process (ideally a passport) is in place and known to managers.
• Staff networks have sponsorship, time and a real route to influence decisions.
• Senior leaders carry specific, evidenced EDI objectives in their appraisals.
• Lived-experience input into research, design and testing is resourced and recognized, not extracted unpaid.
• Belonging and psychological-safety measures are tracked alongside representation, and reported honestly.
• Intersectional cuts are examined where numbers allow, not just single-axis monitoring.
• EDI is owned as a delivery and safety concern, linked explicitly to equitable services (Topic 1.9).

Key sources#

• Equality Act 2010 and the Public Sector Equality Duty — legislation.gov.uk and the Equality and Human Rights Commission
• NHS Workforce Race Equality Standard (WRES) — NHS England annual reports and technical guidance
• NHS Workforce Disability Equality Standard (WDES) — NHS England annual reports and technical guidance
• NHS Equality, Diversity and Inclusion Improvement Plan and the NHS People Promise — NHS England
• NHS Staff Survey — for belonging, bullying, discrimination and engagement measures
• Gender pay gap and (voluntary) ethnicity pay gap reporting guidance — GOV.UK
• Reasonable adjustments and the Workplace Adjustments/Access to Work schemes — GOV.UK and Business Disability Forum
• Behavioural Insights Team / Government Equalities Office reviews of the evidence on unconscious bias training
• The Kotter, Edmondson and West bodies of work on psychological safety and inclusive, compassionate leadership

References#

1. Equality Act 2010 — Wikipedia — https://en.wikipedia.org/wiki/Equality_Act_2010
2. Diversity (business) — Wikipedia — https://en.wikipedia.org/wiki/Diversity_(business)
3. Belongingness — Wikipedia — https://en.wikipedia.org/wiki/Belongingness
4. Gender pay gap — Wikipedia — https://en.wikipedia.org/wiki/Gender_pay_gap
5. Intersectionality — Wikipedia — https://en.wikipedia.org/wiki/Intersectionality
6. Implicit stereotype — Wikipedia — https://en.wikipedia.org/wiki/Implicit_stereotype
7. Algorithmic bias — Wikipedia — https://en.wikipedia.org/wiki/Algorithmic_bias
8. Reasonable accommodation — Wikipedia — https://en.wikipedia.org/wiki/Reasonable_accommodation
9. Social model of disability — Wikipedia — https://en.wikipedia.org/wiki/Social_model_of_disability
10. Psychological safety — Wikipedia — https://en.wikipedia.org/wiki/Psychological_safety
11. Allyship — Wikipedia — https://en.wikipedia.org/wiki/Allyship
12. Equality Act 2010 — legislation.gov.uk — https://www.legislation.gov.uk/ukpga/2010/15/contents
13. Public Sector Equality Duty guidance — Equality and Human Rights Commission — https://www.equalityhumanrights.com/en/advice-and-guidance/public-sector-equality-duty
14. NHS Workforce Race Equality Standard (WRES) — NHS England — https://www.england.nhs.uk/about/equality/equality-hub/workforce-equality-data-standards/equality-standard/
15. NHS Workforce Disability Equality Standard (WDES) — NHS England — https://www.england.nhs.uk/about/equality/equality-hub/workforce-equality-data-standards/workforce-disability-equality-standard/
16. NHS Equality, Diversity and Inclusion Improvement Plan — NHS England — https://www.england.nhs.uk/long-read/nhs-equality-diversity-and-inclusion-improvement-plan/
17. Gender pay gap reporting: guidance for employers — GOV.UK — https://www.gov.uk/guidance/gender-pay-gap-reporting-overview
18. Unconscious bias training: an assessment of the evidence for effectiveness — Government Equalities Office / Behavioural Insights Team — https://www.gov.uk/government/publications/unconscious-bias-training-an-assessment-of-the-evidence-for-effectiveness

See also Topic 1.9 — Digital Inclusion & Accessibility, Topic 7.0 — Team Structures, Roles & Responsibilities, Topic 7.3 — Workforce Strategy, and Topic 7.5 — People & Organisational Development.