Measuring Outcomes in Spanish Long-Term Care: Data, Evidence and System Accountability

A long-term care system can know exactly how many people receive a service and still know too little about whether their lives are improving. An older person may receive every scheduled home-help visit yet become increasingly isolated. A residential centre may meet staffing and accreditation requirements while residents have little control over their routines. A region may reduce administrative waiting while people continue waiting for the service identified in their Programa Individual de Atención. Activity is important evidence, but it is not the same as outcome.

This distinction is central to the next stage of analysis within the Spain Ageing, Long-Term Care & Community Support Knowledge Hub. Spain’s Sistema para la Autonomía y Atención a la Dependencia (SAAD) produces national and regional information on dependency recognition, services, economic benefits and system operation. Autonomous Communities also generate substantial administrative, provider, inspection and service information. The strategic question is how that evidence can be connected to what happens to people.

Measurement matters because Spain’s long-term care system is both rights-based and decentralised. National legislation establishes common foundations while Autonomous Communities administer much of the system, develop service networks and exercise major responsibilities for quality and implementation. Without comparable evidence, it becomes difficult to distinguish legitimate regional variation from avoidable inequality, effective innovation from activity growth, or increased expenditure from improved outcomes.

The stronger opportunity is not to create one enormous score for Spanish long-term care. It is to build an evidence architecture in which administrative data, quality indicators, lived experience, workforce information and person-centred outcomes answer different but connected questions. Accountability becomes stronger when decision-makers can see not simply what the system delivered, but what that delivery enabled.

Spain already measures a great deal, but much of it describes the system rather than the outcome

The SAAD depends on information. Dependency applications need to be processed, recognised grades recorded, PIAs established and services or economic benefits activated. National information is required to understand how the system is operating across Spain and to support intergovernmental oversight and financing.

These administrative datasets are indispensable. They can show how many people are recognised as dependent, what types of services or benefits are used, how activity changes over time and how Autonomous Communities differ across important dimensions.

But administrative data primarily answer questions about the operation of the programme.

They do not automatically show whether somebody feels safer, has retained autonomy, participates in community life, experiences continuity of care or believes their support reflects their preferences. Nor do they reveal automatically whether a family carer is becoming unable to sustain an arrangement, whether home-help timing makes support useful or whether a residential service provides a genuinely person-centred life.

The distinction is fundamental. The number of people receiving teleassistance is an activity measure. Whether teleassistance reduces avoidable crisis, increases confidence or enables somebody to remain safely connected at home is an outcome question.

Spain therefore needs both. A credible measurement system starts with reliable administrative information and then connects it to quality data, KPIs and performance metrics capable of showing what service delivery means in practice.

Three different questions should not be collapsed into one measure

Long-term care measurement becomes clearer when it distinguishes three levels of evidence.

• System activity: who receives support, what type, when, where and at what level of expenditure.
• Service quality: whether support is safe, reliable, adequately staffed, person-centred and delivered according to applicable requirements.
• Person and population outcomes: whether people maintain or improve autonomy, quality of life, social participation, stability, dignity and other outcomes that matter to them.

These levels influence one another but are not interchangeable.

A high-performing administrative system can process applications quickly while the resulting services lack capacity. A well-run provider can deliver excellent support within a region where access remains unequal. A person’s health may deteriorate despite skilled care because their underlying condition is progressive, meaning that stability rather than improvement represents a strong outcome.

That last point is particularly important. Long-term care cannot be evaluated using a simplistic improvement model.

For somebody recovering after an acute episode, increased independence may be appropriate. For a person with advanced dementia or a progressive neurological condition, the meaningful outcome may be comfort, continuity, participation, reduced distress or maintaining abilities for as long as possible. For another person, effective support may enable them to make decisions and remain connected even while physical dependency increases.

Outcomes therefore require interpretation against the person’s circumstances and goals.

Person-centred outcomes begin with the PIA but should not end there

The Programa Individual de Atención is a critical point in the SAAD because it determines the appropriate service or benefit modalities following recognised dependency, within the applicable regional framework and with participation from the person.

That creates an important opportunity for outcome measurement.

If the PIA records only what service will be provided, later review can establish whether the service occurred. If it also captures what the support is intended to enable, review can examine whether the intervention remains appropriate.

For example, two people may both receive home help but seek different outcomes. One wants to continue preparing part of their own meals after a stroke. Another needs reliable personal care that allows them to remain at home with dignity despite increasing frailty. The same service category therefore has different measures of success.

This connects with outcomes-focused and goal-led support. A person-centred outcome should be specific enough to guide support but flexible enough to recognise changing need.

It should also avoid making the person responsible for outcomes outside their control. A goal such as “remain independent” is too broad if inaccessible housing, insufficient support or a progressive health condition make the outcome unrealistic. Measurement should identify the contribution of care, not use outcomes to blame people for deterioration.

The most useful review question is often not “did the person improve?” but “did the support achieve the purpose agreed with the person, given their changing circumstances?”

Scenario: the same home-help hours produce different outcomes

Isabel and Carmen are fictional older residents in two Spanish municipalities. Both have Grade II dependency and each receives a similar volume of home-help support. At regional level, the administrative data show two active services and broadly comparable delivered hours.

For Isabel, the service arrives reliably at agreed times. Workers know her routines and support her with personal care while encouraging her to continue preparing breakfast independently. She remains able to attend a local activity twice each week and says the support gives her confidence.

Carmen receives the same number of recorded hours, but visits change frequently. Morning assistance sometimes arrives after she would normally be dressed and eating. Different workers complete tasks quickly but have little continuity. Carmen has stopped attending community activities because she cannot predict when support will arrive.

Activity data make the cases look similar. Outcome evidence shows that they are not.

A mature measurement approach would retain delivered hours but add indicators such as continuity, visit reliability, participation, complaints, changes in function and the person’s own experience. It would also interpret those indicators rather than automatically converting them into one score.

The Quality Dashboard Builder can help organisations structure this combination of activity, quality and outcome evidence. It is not a Spanish SAAD instrument, but the underlying principle applies directly: the strongest dashboard shows whether operational delivery is producing the outcomes it was intended to support.

Regional comparison is necessary, but crude league tables can mislead

Spain’s 17 Autonomous Communities operate within a common national dependency framework but differ in population structure, geography, historic service infrastructure, workforce markets, budgets and service mix.

Comparison is therefore necessary for accountability. Without it, persistent territorial differences can be dismissed as local arrangements even where they affect practical access or outcomes.

But comparison must be intelligent.

A region with an older population may legitimately have higher levels of dependency support. A sparsely populated territory may face higher home-care travel costs. One Autonomous Community may use more formal services while another has a different balance between services and economic benefits. A headline expenditure or service-use rate tells only part of the story.

Meaningful comparison therefore needs context. Population age, recognised dependency, grades of need, rurality, service mix, workforce supply and other factors can influence what the numbers mean.

This does not mean every difference can be explained away. It means accountability should ask whether variation is defensible.

Regional measurement should help distinguish three possibilities: differences driven by legitimate context; differences driven by policy choice; and differences that indicate weaker access, capacity or implementation. That is a stronger approach than ranking regions from first to seventeenth on one indicator.

Data quality determines whether comparison is trustworthy

Comparable evidence depends on common definitions, consistent recording and timely information. Digitisation alone does not guarantee any of these.

If one service records a missed visit only when no worker attends at all, while another also records visits so late that they no longer fulfil their purpose, headline missed-visit rates are not directly comparable. If regions use different operational definitions for particular indicators, national averages can conceal methodological variation.

Data quality therefore requires governance over meaning as well as technology.

Important measures need clear definitions, ownership, validation and review. Decision-makers should know how current the information is and what is excluded. Material changes in recording practice should be distinguishable from genuine changes in performance.

This is where the discipline of data quality, metrics and performance dashboards matters. A dashboard that allows rapid comparison can amplify errors just as efficiently as it can expose good practice.

The strongest evidence systems make uncertainty visible. They do not present provisional, incomplete or inconsistently defined data with false precision.

Workforce information belongs inside outcome measurement

Long-term care outcomes cannot be interpreted separately from workforce capacity.

Continuity, competence and availability influence what people experience directly. A service can appear fully operational because every required shift has technically been covered while relying heavily on unfamiliar workers, overtime or unstable staffing arrangements. Those conditions may affect relationships, observation of changing need and person-centred practice before they produce a formal incident.

Workforce data therefore provide leading indicators as well as operational information.

Vacancies, turnover, sickness, continuity, agency or temporary staffing, qualifications, supervision and geographic distribution can all help explain emerging quality patterns. They should not automatically be treated as outcomes in themselves, but they can reveal whether services have the capability required to sustain them.

This is particularly relevant as Spain develops more community-based care. Workforce demand becomes geographically dispersed, and home-support capacity is shaped by travel, scheduling and local labour markets.

The connection with workforce assurance is therefore strong. If a region reports growing community provision without showing whether the workforce is stable enough to deliver it, the measurement architecture remains incomplete.

Outcome measurement should ask not only what happened, but what operational conditions made the result likely.

Scenario: residential quality looks strong until resident experience is included

A fictional residential centre in Catalonia reports favourable conventional indicators. Staffing levels meet applicable requirements, medication incidents are low, inspections have identified no serious concerns and relatives submit relatively few formal complaints.

On paper, the service appears stable.

As part of a broader person-centred review, the centre collects structured feedback from residents using communication methods appropriate to individual needs. The findings reveal a different issue. Residents feel safe and generally like the staff, but many say their days are repetitive, mealtimes are too fixed and opportunities to leave the centre are limited.

No major compliance failure has occurred. The evidence nevertheless identifies a quality gap.

The centre responds by reviewing routines, staff deployment and community participation rather than treating low complaint numbers as proof that residents are satisfied. Individual goals are incorporated more clearly into support planning, and managers begin measuring participation and choice alongside incidents and staffing.

This demonstrates why service-user feedback and co-production belong within assurance rather than sitting beside it as optional engagement activity.

The outcome is not that every resident suddenly becomes more active. Some prefer quieter routines. The improvement is that individual preference becomes visible and the service can distinguish chosen routine from institutional routine.

Measurement becomes more person-centred when it helps the organisation understand what life actually feels like, not simply whether formal controls are functioning.

Family evidence is important, but it is not a substitute for the person’s voice

Families often hold essential information about long-term care. They may notice deterioration, missed support, communication problems or changes in behaviour that are not obvious from formal records. They can also show whether the practical burden of an arrangement is increasing.

That evidence should be taken seriously.

But family experience and service-user experience are not always identical. A relative may prioritise safety while the person values autonomy. A family member may favour residential care because community support creates anxiety, while the person wants to remain at home. In other cases, a relative may understate strain because they believe providing care is their responsibility.

A measurement system therefore needs to preserve different perspectives rather than merging them into one satisfaction figure.

Family-carer sustainability can itself be a meaningful outcome dimension where informal care is central to the arrangement. The relevant evidence may include whether the carer can continue willingly and safely, whether respite is available, whether employment or health is being affected and whether formal services reduce rather than simply reorganise unpaid work.

This is particularly important in Spain because family care remains a substantial component of long-term support. A system that appears financially efficient because relatives absorb unmet demand may be measuring public activity while overlooking the full care burden.

Waiting should be measured as a pathway, not a single number

Timeliness is another area where measurement can become misleading if stages are collapsed.

A person may wait for dependency assessment, recognition, development of the PIA or activation of the resulting service. Each stage reflects a different operational process and may require a different response.

An administration that reduces assessment delay has achieved something useful. But if approved services cannot start because workforce or capacity is unavailable, the person may still experience a long period without the intended support.

This is why waiting-list evidence should connect administrative flow with practical access.

Relevant indicators can include time at different stages, changes in need while waiting, interim support, whether the person receives the preferred service and what happens when the approved option is unavailable.

The purpose is not to create ever more indicators. It is to avoid a performance measure that improves on paper while pressure is displaced into another part of the pathway.

This principle also strengthens quality monitoring systems. Measures should follow the service journey sufficiently far to detect whether apparently successful process improvement has changed the actual outcome.

Technology can make outcomes more visible, but it can also create measurement excess

Digital care records, scheduling platforms, teleassistance, sensors and regional information systems can produce large volumes of data. This creates opportunities for earlier identification of deterioration and more timely performance intelligence.

It also creates a risk that measurable behaviour is treated as more important simply because it is easy to collect.

For example, an advanced teleassistance service may record calls, alerts, response times and device status automatically. Those indicators are useful. They do not necessarily reveal whether the person feels more secure, whether avoidable emergencies have reduced or whether the technology is intrusive.

Digital monitoring therefore needs to distinguish operational telemetry from human outcomes.

The same applies to home-care scheduling. Arrival times can be measured precisely, but punctuality alone does not describe the relationship, quality of support or whether the visit enables the person’s goals.

Technology should help connect evidence rather than overwhelm decision-makers with every available data point.

A strong digital measurement strategy identifies a limited number of questions first and then collects the data needed to answer them.

Scenario: teleassistance is performing technically but not achieving its preventive purpose

A fictional Autonomous Community expands advanced teleassistance for older people living alone. The implementation dashboard looks impressive. Device installation targets are exceeded, response times are fast and technical uptime is high.

After one year, however, regional analysts examine a broader set of evidence. They find that a subgroup of people generate repeated non-emergency contacts linked to loneliness, anxiety and difficulty managing daily routines. The service responds efficiently each time, but the underlying pattern rarely changes.

The technology is functioning exactly as designed. The outcome architecture is incomplete.

Instead of treating frequent contact as merely increased activity, the region develops a pathway through which repeated patterns can prompt appropriate social-service or health review, with the person’s involvement. The objective is not to discourage people from contacting teleassistance but to recognise when repeated contacts represent an unmet need that another service should address.

Performance measurement then changes. The administration still monitors response time and technical reliability, but it also examines recurring-contact patterns, successful escalation, subsequent support and the person’s experience.

The lesson is broader than teleassistance. Digital services often produce excellent technical metrics. Long-term care accountability requires another step: whether the technology altered the person’s situation in the way the service was intended to do.

Outcome evidence should influence purchasing and service design

Autonomous Communities and local entities, depending on the service and regional arrangements, may directly provide services or purchase them from external organisations. Contracts and other service arrangements therefore create an important route through which measurement influences delivery.

If specifications focus overwhelmingly on volume, providers will naturally organise around volume. Hours, places, response times and staffing remain important, but they should be connected to the purpose of the service.

For home help, that may include continuity, independence and reliability. For day services, participation, maintenance of function and family sustainability may matter. For residential care, autonomy, relationships, safety and quality of life should be visible. For personal assistance, the degree of control exercised by the person is fundamental.

Outcome-based thinking does not require all payment to be tied directly to outcomes. That can create perverse incentives, particularly where providers support people with progressive or complex needs.

The stronger approach is to make outcomes part of accountability: service arrangements define what evidence should be gathered, organisations explain performance in context and recurring gaps influence future design.

Measurement should create better conversations about value rather than simplistic financial rewards for favourable numbers.

National accountability requires comparability without erasing regional responsibility

Spain’s decentralised system creates a distinctive accountability challenge. National government needs sufficient information to understand whether the common framework established by Law 39/2006 is producing equitable and effective access across the country. Autonomous Communities remain responsible for many of the operational decisions that shape actual delivery.

A national dataset therefore cannot replace regional governance, while regional reporting cannot make national comparability unnecessary.

The strongest architecture is layered.

Nationally, common definitions and core indicators can show whether the SAAD is functioning consistently enough to protect the underlying right to dependency support. The Territorial Council provides an important mechanism through which common criteria and system information can support shared oversight.

Regionally, administrations need substantially more detailed evidence about service capacity, workforce, provider performance, geography, quality and outcomes. Local and provider information should identify operational patterns before they become regional averages.

The governance task is then to connect the levels.

If a national indicator identifies persistent variation, responsibility should not stop at publication. The evidence should prompt analysis of causes. If one region identifies a successful approach, wider learning can examine which mechanisms are transferable without assuming every Autonomous Community should adopt an identical model.

This reflects the wider discipline of continuous improvement: data become valuable when they lead to inquiry, action and subsequent evaluation.

Accountability needs an explanation of causation, not merely performance colour coding

Dashboards frequently simplify performance into red, amber and green categories. The visual clarity can be useful, but long-term care rarely has simple causes.

A fall in delivered home-help hours might reflect workforce shortages, provider withdrawal, bad weather, inaccurate recording or deliberate adjustment after reassessment. Rising residential admissions could indicate inadequate community capacity, demographic change, improved access for people previously waiting or changing preferences.

Governance therefore needs interpretation.

A strong performance process asks several questions when a material variation appears:

• Is the underlying data reliable and comparable?
• Which population or geography is affected?
• What operational factors plausibly explain the change?
• Does the variation affect rights, safety, autonomy or access?
• Who has authority to respond?
• How will leaders know whether the intervention worked?

The Governance Maturity Assessment can help organisations structure these relationships between evidence, responsibility, escalation and action. It is not designed to replicate Spanish public-administration structures, but its underlying question is directly relevant: does performance information reach the level capable of changing the cause?

Accountability is weak when numbers are published but responsibility for responding remains unclear.

Scenario: a regional average hides a local failure in continuity

A fictional Autonomous Community reports stable home-care performance. Delivered hours remain close to plan, complaints have not increased materially and the proportion of missed visits appears low.

When data are segmented geographically, one inland province shows much higher staff turnover and more frequent worker changes. People are still receiving most of their scheduled visits, which is why the regional headline remains stable.

Qualitative feedback explains the consequence. Older people with dementia describe greater anxiety when unfamiliar workers arrive. Families report repeatedly explaining routines to new staff. Managers observe that subtle deterioration is being identified later because continuity is weak.

The region now sees that the problem is not primarily total care volume. It is the stability of the local workforce.

Action therefore focuses on recruitment, travel patterns, employment conditions and continuity rather than purchasing additional headline hours. Outcome monitoring then tests whether worker continuity improves and whether complaints, missed deterioration and family concerns reduce.

This example demonstrates why regional averages require disaggregation and why workforce evidence belongs inside quality governance.

It also shows the danger of treating every poor indicator as a provider-performance issue. The underlying cause may sit partly in regional labour markets, geography or service-purchasing arrangements. Accountability requires evidence to reach the level where those structural decisions can be changed.

Measuring independence requires care around interpretation

Independence is an important objective of the SAAD, but it can be measured badly.

A reduction in formal support is not automatically evidence of increased independence. It may reflect improved function, but it may also mean that family members are providing more unpaid care or that the person is receiving less support than they need.

Similarly, increased support is not automatically a negative outcome. For somebody whose condition has progressed, more assistance may preserve autonomy by allowing them to remain at home, communicate choices or continue participating in ordinary life.

Outcome measurement must therefore understand independence as control and capability rather than simply fewer care hours.

This is where recording and evidencing person-centred care becomes important. The evidence should explain how support relates to the individual’s priorities and changing situation.

Long-term care often succeeds by compensating for declining function in a way that preserves the person’s life. A person can become more physically dependent while retaining greater autonomy because the right support is available.

Measures need to be sophisticated enough to recognise that distinction.

Outcome frameworks must include people who communicate differently

Person-reported outcomes can become exclusionary if the only accepted evidence is a conventional survey.

People with dementia, intellectual disability, sensory impairment, acquired communication difficulties or other needs may require different methods to express preferences and experience. Some people communicate through behaviour, familiar routines, supported conversation or observation alongside direct communication.

Families and workers may contribute evidence, but their interpretation should not automatically replace the person’s perspective.

Accessible outcome measurement therefore requires flexible methods. Questions may need simplified language, visual formats, supported communication or repeated observation rather than a single interview.

This is not an optional methodological refinement. If people with the most complex needs are systematically absent from outcome datasets, a system can appear increasingly person-centred while measuring only people who find standard feedback mechanisms easiest to use.

Equity in measurement means making the evidence process accessible as well as making the resulting services accessible.

Outcome measurement should support learning rather than defensive reporting

Performance systems can create unhealthy behaviour when organisations believe that every unfavourable outcome will be interpreted as failure.

Long-term care involves uncertainty. Falls can occur despite appropriate support. A family arrangement can break down unexpectedly. A person may choose an option that involves greater risk. Progressive illness can change outcomes even where care is good.

If organisations are punished simply for reporting unfavourable events, measurement becomes less reliable because the incentive shifts towards presentation rather than learning.

A stronger accountability model distinguishes avoidable weakness from legitimate complexity.

Providers and public administrations should be expected to understand their data, identify recurring patterns, investigate material variation and demonstrate improvement. They should not be expected to produce permanently perfect numbers.

This is the difference between assurance and performance theatre.

Outcome evidence is most valuable when it supports transparent inquiry: what happened, why, whether it reflects a pattern, what can be changed and what evidence will show whether that change worked.

Spain’s next measurement challenge is connecting datasets around the same policy questions

Spain does not need every long-term care dataset to become one system before measurement can improve. The more immediate opportunity is to connect evidence conceptually around shared questions.

If policymakers want to know whether community care is becoming more sustainable, they may need to examine home-help capacity, personal assistance, teleassistance, workforce, residential admissions, family burden, waiting and person-reported experience together.

If the question is whether regional inequality is narrowing, national administrative data need to be interpreted alongside service intensity, waiting, population characteristics, workforce and practical access.

If the question is whether deinstitutionalisation is improving lives, counting residential places is insufficient. Evidence should examine housing, autonomy, community participation, support continuity and whether institutional practices reappear in smaller settings.

The analytical discipline is to start with the policy question rather than with the dataset that happens to be easiest to access.

Digital transformation can eventually make these connections faster. It should not determine what Spain chooses to value.

What Spain’s experience offers internationally

Spain’s measurement challenge is shaped by its own institutional architecture: a national rights framework, decentralised administration through the Autonomous Communities, diverse provider arrangements and substantial family involvement. Other countries will organise long-term care differently.

The transferable lessons lie in the evidence design rather than the institutions.

First, administrative activity should not be mistaken for outcomes. Counting services is necessary, but it does not reveal what those services enable.

Second, decentralised systems require both common comparability and local interpretation. National standards can make territorial variation visible without assuming all regions should look identical.

Third, workforce, waiting, family burden and quality data should be interpreted alongside person-centred outcomes. They often explain why similar service volumes produce different experiences.

Fourth, measurement should recognise progressive conditions and complex needs. Improvement is not the only legitimate outcome; stability, dignity, reduced distress and maintained participation may be equally important.

Finally, accountability needs a response mechanism. Publishing data is not enough if persistent variation does not prompt investigation, action and learning.

Other systems can adapt these principles without copying the SAAD. The underlying lesson is that an evidence system should illuminate the relationship between public resources, service delivery and human outcomes rather than allow any one of those dimensions to stand in for the others.

Conclusion

Spain has already developed substantial information infrastructure around the SAAD. The next challenge is to make that evidence more useful for understanding whether the system is achieving the outcomes that justify its existence: autonomy, dignity, timely support, safety, participation and sustainable care around people whose needs may change substantially over time.

That requires more than additional indicators. Administrative data need to remain reliable and comparable, but they should be connected with service quality, workforce stability, waiting, family experience and the voices of people receiving support. Regional comparison should expose unjustified variation without reducing the Autonomous Communities to crude league tables. Digital tools should improve visibility without allowing what is easiest to count to become what matters most.

The strongest accountability model is therefore layered. National evidence should show whether common rights are being realised across Spain. Regional evidence should explain service capacity, quality and territorial variation. Providers and local services should understand what happens in everyday delivery. Person-centred evidence should test whether those systems produce lives with greater control and security.

Measurement becomes meaningful when it changes decisions. If recurring evidence leads to better workforce planning, redesigned pathways, improved service availability or stronger support around individual goals, data have become part of care rather than merely administration. For Spain, that is the central opportunity: to move from knowing how much the SAAD does towards understanding, with increasing precision, what that activity achieves for the people whose lives depend on it.