01Core principlesThe concepts and mechanisms needed to understand the subject.
The gradient matters. Poor outcomes are not confined to the most deprived group; health often improves stepwise with income, education, housing and power. Universal services therefore remain essential. Proportionate universalism means providing action across the whole gradient with scale and intensity proportionate to disadvantage. A service only for the poorest can miss people just above a threshold, while an identical offer to everyone can favour those best able to respond.
Equality and equity answer different questions. Equality offers the same appointment letter to everyone. Equity asks whether the letter is readable, delivered to a stable address, compatible with work and caring, accessible to disabled people and trusted by the recipient. Some patients need interpretation, longer appointments, outreach, transport support or flexible booking to reach an equivalent opportunity. The goal is fair capability to benefit, not identical process.
Causes operate at several levels. Structural determinants include income distribution, discrimination, legal status and public policy. Living conditions include housing, employment, education, transport, food environment and pollution. Psychosocial stress and exposure shape behaviour and biology. Healthcare contributes through availability, navigation, communication, diagnostic delay and treatment. Explaining inequality as individual lifestyle alone ignores the options and constraints that pattern behaviour.
Current ONS estimates show a substantial deprivation gradient. For England in 2022 to 2024, the slope index inequality in life expectancy at birth was 10.4 years for males and 8.0 years for females; healthy-life-expectancy inequalities were 19.3 and 20.1 years. These population measures do not predict an individual life span. They show a graded distribution requiring both societal and service action.
NHS England’s Core20PLUS5 framework supports action on healthcare inequalities. Core20 refers to the most deprived 20% of the national population by Index of Multiple Deprivation. PLUS denotes locally identified groups experiencing poorer access, experience or outcomes, which may include ethnic minority communities, inclusion health groups, people with learning disability, coastal communities or others defined from data and engagement. The five clinical areas focus accelerated improvement; the framework complements rather than replaces local analysis.
Inclusion health groups, including people experiencing homelessness, substance dependence, vulnerable migration, sex work or contact with the justice system, often face extreme morbidity and fragmented access. Registration, identification, address or immigration misconceptions can block care. Build low-threshold routes, continuity and coordination, and use trauma-informed communication. Do not reduce a person to an inclusion label or assume the same preferences within a group.
Ethnicity data can reveal inequity but are easily misused. Ethnic categories are social and administrative groupings, not simple biological variables. Observed differences may reflect racism, deprivation, occupation, environment, migration history, language, differential treatment or data quality. Use self-ascribed categories where possible, show missingness, avoid treating “BAME” as a uniform group and involve communities in interpretation before attributing cause.
Digital systems can improve convenience and worsen exclusion simultaneously. Online-first booking, app-only results, automated English messages and remote triage impose devices, data, literacy, privacy and confidence requirements. Maintain non-digital routes, reasonable adjustments and human escalation. Monitor who abandons forms, misses remote contacts or receives fewer face-to-face assessments. Digital uptake is a process measure, not proof of equitable access.
Clinical communication needs proportionate support. Use professional interpreters for material decisions; relatives may be conflicted and children should not carry adult clinical interpretation. Offer accessible information and longer or quieter appointments when needed. Ask about preferred contact method and whether messages are safe to leave. Teach-back detects communication failure without testing or blaming the patient. Record adjustments so they persist across encounters.
Measure the pathway, not only the endpoint. Start with the eligible denominator, then offer, successful contact, uptake, quality, treatment, follow-up and outcome. Example: two groups each receive 1,000 screening invitations. Group A completes 700 tests and 95% of positives complete assessment; group B completes 500 and 70% of positives complete assessment. Equal invitations conceal unequal effective access. Stratify by relevant variables and compare absolute as well as relative gaps.
Data can stigmatise or mislead. Small numbers create unstable rates and disclosure risk. Missing ethnicity or housing data are rarely random. Crude comparisons may reflect different age or morbidity structure. Use age standardisation or appropriate modelling when needed, show uncertainty and triangulate quantitative findings with patient experience. Never postpone obvious accessibility fixes solely because a formal statistical threshold was not met.
Action should be co-produced with affected people and frontline staff. Define a specific barrier, select an outcome and balancing measures, test a small change and adapt. Examples include calling non-responders in their preferred language, offering evening cervical screening, creating a walk-in blood-test route, or flagging reasonable adjustments. Check whether extra support diverts capacity or increases stigma, and preserve a route for people outside the targeted category.
At individual level, practise structural competence without turning every consultation into a social questionnaire. Ask the factors likely to change the plan: ability to pay for transport, store medicine, attend fasting tests, read instructions, take time off or contact services. Record and act. At practice level, use searches and multidisciplinary knowledge to identify people missing care, but apply information governance and avoid risk scores that silently reproduce historic under-service.
Key points
- Health inequalities are systematic, avoidable and unfair differences in health, exposure, access, experience or outcome between groups or across a social gradient.
- Equality provides the same resource; equity matches support to different barriers and need; proportionate universalism combines universal provision with greater intensity where disadvantage is greater.
- Measure several stages separately: eligible population, offer, contact, uptake, process quality, treatment, experience and outcome; overall averages can hide loss at one stage.
- Use intersectional assessment because deprivation, ethnicity, disability, sex, age, migration, homelessness, sexuality and other factors can combine rather than act independently.
- Avoid biological or cultural assumptions from group identity; ask the individual about language, trust, costs, work, caring, housing, transport, discrimination and prior healthcare experience.
- NHS England Core20PLUS5 identifies the most deprived 20% nationally, locally defined PLUS groups and five clinical priorities as a framework for focused action.
- Safety-netting is inequitable when it depends on English literacy, digital access, money, transport, stable housing or confidence that the person does not have.
- Co-produce changes with affected people, use small tests, examine unintended consequences and retain universal access while intensifying support for higher-barrier groups.
02Mechanisms and patternsImportant relationships and how to distinguish them.
Repeated non-attendance may reflect inflexible booking, unsafe contact, transport, literacy or distrust rather than lack of interest.
People with greatest need can receive less effective care when services require time, confidence, continuity and resources they are least likely to have.
Multiple social positions and systems interact, so single-category averages can hide a small group with much greater barriers.
Missing ethnicity, disability, housing or communication data can remove affected people from both analysis and targeted support.
App-only, online-first or remote pathways can fail through lack of device, data, privacy, literacy, language or accessible design.
Stereotyping, inaccessible communication and fragmented continuity can delay recognition even after a person reaches the service.
03Interpreting evidenceInformation, measurements and their limitations.
Consider the information, its meaning and its limitations before deciding what follows.
- 01
Pathway equity audit - Why
- Locate where groups are lost between eligibility and outcome.
- Interpretation and limitations
- Measure denominator, offer, contact, uptake, treatment and outcome separately; a similar final count may conceal different need.
- 02
Stratified outcome analysis - Why
- Compare access, experience and outcome across relevant groups.
- Interpretation and limitations
- Use appropriate denominators, absolute gaps and uncertainty; inspect missingness and small-number disclosure before inference.
- 03
Patient and community enquiry - Why
- Understand mechanisms behind a numerical difference.
- Interpretation and limitations
- Use accessible, compensated and trusted engagement; avoid assuming that staff interpretation explains the lived barrier.
- 04
Reasonable-adjustment review - Why
- Check whether recorded needs change the actual care pathway.
- Interpretation and limitations
- An adjustment flag is useful only if booking, consultation, investigation and follow-up teams can see and implement it.
- 05
Intervention balancing measures - Why
- Detect stigma, workload shift or harm caused by a targeted change.
- Interpretation and limitations
- Track wait, continuity, complaints and access for non-targeted patients alongside the intended equity outcome.
04Applied reasoningWorked examples connecting principles to decisions.
01Worked case: unequal diabetes review completionFind the barrier and test proportionate supportPractice data show 72% annual-review completion overall but 41% among patients recorded as needing an interpreter; invitation counts are similar.+
- 1Validate denominator, interpreter coding, contact success, appointment offer, attendance and completed components; stratify without publishing identifiable small cells.
- 2Engage patients and reception, interpreting and clinical staff to test explanations; they identify English-only automated calls, short appointments and inability to request an interpreter online.
- 3The final action is a small co-designed test: preferred-language call booking, pre-booked interpreter and a longer combined review slot, while retaining ordinary routes for everyone.
- 4Verify impact over successive weeks using completed review, patient experience and interpreter availability, plus balancing measures for waiting time and staff workload; adapt before wider rollout.
02Individual consultationMake the clinical plan feasibleA recommended investigation or treatment depends on resources the patient may not have.+
- 1Ask neutrally about transport, cost, work, caring, communication, storage and safe contact where these could alter action.
- 2Agree a clinically sound route using available support, reasonable adjustment or timing rather than documenting “non-compliance”.
- 3Check understanding and ownership, then actively follow up when the person cannot reliably navigate the next step alone.
03Practice dashboardMove beyond an overall averageA quality indicator appears acceptable for the whole registered population.+
- 1Define eligible population and each pathway stage with consistent coding.
- 2Stratify by deprivation and locally relevant PLUS groups, checking missing data and unstable small cells.
- 3Prioritise the largest meaningful avoidable gap, confirm mechanism with affected people and select an actionable measure.
04Digital service changeAssess exclusion before scaleThe practice plans online-first booking or remote monitoring.+
- 1Map device, data, language, literacy, disability, privacy and safeguarding requirements with potential users.
- 2Maintain phone and in-person alternatives and reasonable adjustments, with staff able to override the digital route.
- 3Monitor failed contacts, modality, time to care, experience and outcome by group and modify the design when exclusion appears.
05Checking understandingVerify the reasoning, revisit uncertainties and apply feedback.
- Review deprivation and locally relevant PLUS-group indicators at each important pathway stage, not only overall activity.
- Track completeness and accuracy of ethnicity, disability, language, reasonable-adjustment and contact data with patient consent and governance.
- Measure absolute and relative gaps over time with denominators and uncertainty; a shrinking overall rate can still conceal widening inequality.
- Collect experience in accessible ways, including non-digital options and professional interpretation, and report how feedback changed the service.
- Use balancing measures for delay, workload, continuity, stigma and access among people outside a targeted intervention.
- Review individual harm or missed care for structural contributors and feed specific system changes into practice safety and quality processes.
06Special situationsVariants, exceptions and circumstances that change the usual approach.
Averages hide gradients
A practice can improve overall while the lowest-access group stands still, so distribution belongs beside the headline result.
Same offer can be unequal
An identical invitation demands different time, literacy, transport and trust from different people and may produce unequal opportunity.
Missing data are information
High unknown rates may identify poor trust or workflow and can bias comparisons rather than merely reduce sample size.
Targeting can stigmatise
Extra support should be offered respectfully within universal care and evaluated for labelling, privacy and displacement effects.
Co-production tests assumptions
People experiencing a barrier often identify mechanism and feasible solutions that routine service data cannot reveal.
Clinical plans have social prerequisites
Transport, refrigeration, privacy, money and safe communication can determine whether evidence-based care is possible.
07Common pitfallsFrequent interpretation and management errors.
- 01
Do not interpret a group difference as innate biological or cultural causation without evidence.
- 02
Do not treat repeated non-attendance as informed refusal before investigating access barriers.
- 03
Do not report only an overall average when distribution and pathway stages are available.
- 04
Do not combine diverse ethnic or inclusion groups into one category when important differences are hidden.
- 05
Do not make a digital route mandatory without accessible non-digital alternatives.
- 06
Do not use children as interpreters for material clinical decisions.
- 07
Do not target support using unreliable data without checking false exclusions and stigma.
- 08
Do not measure invitation alone when diagnosis, treatment and outcome can remain unequal.