Data, Evidence and Outcomes in Lithuanian Long-Term Care: From Activity to Accountability

A municipality can know how many people received social care at home last year. A healthcare provider can know how many home nursing visits it delivered. A ministry can monitor expenditure, staffing and service capacity. Yet none of those figures, on their own, answers the question that matters most to an older person or disabled person: did the support enable me to live safely, independently and with a reasonable quality of life?

This distinction is becoming increasingly important for Lithuania. Long-term care is developing across systems that have historically been divided between healthcare responsibilities led by the Ministry of Health, social protection responsibilities led by the Ministry of Social Security and Labour, services organised through municipalities, compulsory health insurance financing and different forms of state and municipal funding. The wider structure and reform direction are explored throughout the Lithuania Ageing, Long-Term Care & Community Support Knowledge Hub.

Lithuania now has more reason than ever to understand what those different parts of the system achieve together. Population ageing is increasing long-term-care demand. Community and home-based provision is becoming more important. Integrated long-term care is developing, while formal care capacity remains constrained and municipalities operate in very different demographic and geographic circumstances. Recent international data also indicate that more than one in five Lithuanians aged 65 and over received long-term-care services in 2023 under the relevant comparative definition, while the formal long-term-care workforce remains small relative to the older population.

The next stage of maturity is therefore not simply collecting more information. It is connecting information to decisions. Lithuania needs evidence that can distinguish activity from access, access from quality and quality from outcomes. That is the foundation of meaningful accountability.

Lithuania's data challenge reflects the structure of long-term care itself

Long-term care in Lithuania does not operate as one administratively unified service. Healthcare and social services have different legal, financial and organisational foundations. Health-related long-term care can be financed through the compulsory health insurance system, while social services are organised through municipalities and supported through municipal budgets, state transfers and personal contributions under the applicable arrangements.

That division inevitably affects information.

A health service may record diagnoses, nursing interventions, clinical activity and reimbursement information. A municipality needs information about social-service eligibility, support needs, provision and expenditure. A residential social-care institution has its own care, staffing and quality information. Cash benefits and individual assistance arrangements create further administrative data.

Each dataset can be legitimate and useful while still providing an incomplete view of the person.

Historically, this fragmentation has been one of the structural challenges identified in Lithuania's long-term-care system. Reform towards greater integration is therefore partly an information challenge. Coordinated services are difficult to operate if professionals cannot obtain the information they need, and integrated policy is difficult to evaluate if national decision-makers cannot connect activity across the health and social domains.

This makes interoperability and system integration more than a technical ambition. It concerns whether different parts of the system can develop a sufficiently coherent understanding of need, service use and outcomes to make better decisions.

Activity data remain necessary, but they cannot define success

Long-term-care systems need basic operational information. Lithuania needs to know how many people receive particular services, where they live, how much provision costs, how long people wait and whether capacity is increasing or contracting.

Activity measures can identify pressure. A rise in applications for home support can signal changing demand. Increasing utilisation of home nursing can affect workforce planning. Residential occupancy can inform capacity decisions. Differences between municipalities can trigger questions about access.

The problem arises when activity becomes a proxy for outcome.

Delivering 20,000 additional hours of home support demonstrates that more activity occurred. It does not establish whether people became more independent, whether family-carer pressure reduced, whether preventable deterioration was avoided or whether people received support early enough.

Similarly, a reduction in residential admissions could reflect successful community support, but it could also reflect insufficient residential capacity or unmet need. A low utilisation rate may indicate a healthier population, or it may reveal barriers to access.

Evidence therefore needs interpretation.

For Lithuania, a mature long-term-care evidence framework would distinguish at least four dimensions:

• need — who requires support and what type of need exists;
• access — who receives services, how quickly and with what geographic or financial barriers;
• quality — whether services are safe, reliable, person-centred and appropriately delivered; and
• outcomes — what changes, stabilises or is prevented because support was provided.

Those dimensions should connect rather than operate as separate reporting exercises.

National averages can conceal municipal reality

Lithuania's 60 municipalities differ considerably. Population structure, settlement patterns, fiscal capacity, provider availability and workforce supply affect what services can be organised locally. Vilnius, Kaunas or Klaipėda therefore cannot be assumed to represent the operating conditions of a smaller or more rural municipality.

National data can show the direction of travel while concealing variation underneath it.

This matters because geographic inequality is itself an outcome of system design. If an older person with equivalent needs can obtain timely home support in one municipality but waits substantially longer in another, the difference should be visible. If one area relies heavily on institutional care while another has developed stronger community capacity, national totals may hide an important service-model distinction.

Comparable municipal information can therefore support learning, but comparison needs care. Raw rankings can be misleading where population need differs substantially.

A municipality with higher expenditure may be inefficient, but it may also support an older population with greater disability or operate across a large rural geography. A municipality with fewer residential placements may have excellent home-based services, or families may be carrying more unsupported care.

Good comparative analysis adjusts the question. Instead of asking only which municipality spends most or uses most services, it asks why the pattern differs and whether outcomes justify the variation.

Scenario: two municipalities report the same activity but achieve different results

Imagine two Lithuanian municipalities of broadly comparable size. Each reports that 300 older residents received home-based social services during the year. At national level, the activity figures appear similar.

The underlying experience is very different.

In the first municipality, most people begin support relatively soon after assessment. Services are reviewed when needs change, and links with home nursing allow emerging health problems to be escalated. A significant proportion of people remain at home without an unplanned move to institutional care.

In the second municipality, the same number of people technically received support, but many waited substantially longer. Hours are concentrated among people whose needs have already become severe. Family carers fill gaps before formal support begins, and several people experience hospital admission during the waiting period.

An activity measure records 300 recipients in each municipality. An outcome-oriented system sees two different pathways.

The useful evidence would include waiting time, assessed need, intensity of provision, changes after review, hospital interfaces, family-carer experience and whether the person remained safely at home where that was their preference.

This does not mean creating an enormous national dataset for every individual. It means selecting measures capable of distinguishing service presence from service effectiveness.

From service outputs to person-centred outcomes

Long-term care has a particular measurement difficulty: success does not always mean improvement.

A person with progressive frailty may not become more physically independent. High-quality support might instead prevent avoidable deterioration, preserve decision-making, enable continued community participation or allow the person to remain at home for longer. For someone approaching the end of life, comfort, dignity and family support may matter more than functional improvement.

Outcome measurement therefore needs to respect different trajectories.

This is why outcomes-focused support begins with what matters to the individual rather than imposing a universal definition of progress.

Relevant long-term-care outcomes can include maintaining mobility, completing everyday activities, sustaining social relationships, reducing avoidable falls, improving nutrition, enabling a family carer to continue working, reducing distress or maintaining a preferred living arrangement.

Some outcomes are objective. Others require the person's own account.

A national evidence framework that measures only service activity will miss much of this. Equally, a system relying only on satisfaction questionnaires can miss safety, inequity or unmet need. Strong evidence combines administrative, professional and lived-experience information.

Individual assessment is the first data point in an outcomes system

Outcome measurement starts before a service begins.

If an assessment records only eligibility and service category, later review can determine whether provision occurred but not whether the person's life changed. A stronger assessment establishes a baseline: what the person can currently do, what difficulties exist, what risks matter, what family or community support is available and what outcome the intervention is intended to achieve.

This creates a logical chain between assessment, support and review.

For example, if home help is introduced because an older person can no longer prepare meals safely, the review should not merely confirm that visits took place. It should examine whether nutrition improved, whether the person can participate in meal preparation, whether risks changed and whether the intensity of support remains appropriate.

The same principle applies to integrated assistance, day social care, home nursing and residential support.

Good data quality and performance measurement therefore depend on the quality of frontline records. National dashboards cannot compensate for poorly defined needs or inconsistent recording at service level.

Integration requires information to follow the person

Lithuania's development of more integrated long-term care creates a practical test for information systems: can the relevant professionals understand the person's pathway across organisational boundaries?

An older person receiving home nursing and municipal social care does not experience their needs as separate health and social datasets. A nurse may notice that the person can no longer prepare food. A social-care worker may observe increasing breathlessness or confusion. Each observation may require action from a different part of the system.

Integration therefore requires more than a common referral route. Information has to move safely enough for professionals to coordinate care, while privacy and data-protection requirements remain respected.

This does not mean that every worker should have unrestricted access to every record. Access should reflect role, purpose and lawful information-sharing arrangements. The objective is appropriate visibility rather than indiscriminate data sharing.

The operational questions include whether professionals can identify who else is involved, whether important changes are communicated, whether assessments are repeatedly duplicated and whether an escalation reaches the service capable of responding.

These questions connect directly with digital records and information governance. Digitalisation can make coordination easier, but only if data standards, responsibilities and workflows are designed around the pathway rather than individual organisations.

Organisations considering comparable transformation can use the Digital Transformation Readiness Assessment to examine issues such as interoperability, digital capability, information governance and workforce readiness. It is not a Lithuanian regulatory instrument; its relevance lies in helping leaders test whether technology is supported by the operational conditions required to make it useful.

Scenario: the same deterioration appears in three separate records

An 82-year-old woman in Kaunas receives municipal help at home and outpatient home nursing. Her daughter also provides substantial informal support.

Over several weeks, the home-support worker records that the woman is eating less and appears more fatigued. A nurse separately notes reduced mobility. Her daughter contacts another service because she is struggling to manage increasing night-time needs.

All three parts of the system contain evidence of deterioration, but no single indicator is dramatic enough to trigger an emergency response.

If the information remains separated, the woman may continue until a fall or acute illness results in hospital admission. Each organisation could subsequently demonstrate that it maintained its own records correctly.

An integrated pathway treats the pattern as the important signal. Changes in nutrition, mobility and family-carer capacity are brought together, prompting reassessment. The response may involve clinical review, increased home support, rehabilitation or a different combination of services.

The governance lesson is significant. Data quality cannot be judged only by whether each record is complete. It also depends on whether information reaches the point at which a decision can be made.

For Lithuania's integrated-care development, that distinction will become increasingly important. The objective is not simply digital completeness but actionable continuity.

Measuring unmet need is as important as measuring provision

One of the hardest populations for any long-term-care system to measure is the group it is not serving.

Historical evidence has identified substantial unmet long-term-care need among older people in Lithuania. Current service utilisation has expanded, and comparative definitions have also changed, so older figures should not be treated as a description of today's exact level of unmet need. The policy issue nevertheless remains important.

Administrative systems naturally generate more information about people who enter them than people who do not.

A person receiving home care becomes visible through assessment, service and payment records. An older person relying entirely on an exhausted spouse may remain largely invisible until a hospital admission or crisis occurs.

Evidence about unmet need therefore needs to look beyond recipient counts. Waiting lists, rejected or withdrawn applications, repeated emergency use, carer surveys, population research and geographic access patterns can all contribute.

Low utilisation requires interpretation. It may represent low need, successful prevention or a service that people find difficult to access.

This is particularly important in rural areas, where transport, workforce availability and distance can affect practical access even where formal entitlement is the same.

Outcome accountability should consequently include the question: who is missing from the data?

Family caregiving is a major evidence blind spot

Family care is particularly difficult to capture because much of it occurs outside formal service systems.

Relatives may prepare meals, supervise medication, provide personal care, organise appointments, manage finances and remain available at night without those hours appearing as formal long-term-care activity.

If a service records that an older person receives two hours of formal support each week, the dataset can make the care package appear modest. It may not reveal that a daughter provides another 30 hours of assistance.

This creates two risks.

First, formal services can appear more effective than they really are because unpaid labour is carrying much of the pathway. Second, changes in family capacity may not become visible until the arrangement breaks down.

Evidence should therefore consider the sustainability of family support without treating relatives as another service resource. Questions about carer burden, employment impact, distance, health and willingness to continue can be important indicators of future risk.

The same principle applies to service-user feedback and co-production. People and families provide information that administrative systems cannot generate themselves. Their experience can reveal whether services arrive at useful times, whether care is coordinated and whether formal support genuinely reduces pressure.

Quality assurance needs a line of sight from service to system

Lithuania has established oversight arrangements for social services, including responsibilities associated with the Social Services Supervisory Department under the Ministry of Social Security and Labour. Municipalities also have responsibilities within the organisation, accreditation and oversight of services, while healthcare operates through its own quality and professional structures.

For long-term care, the challenge is to make those mechanisms collectively informative.

A service-level quality issue can indicate a local management problem. Repeated across providers or municipalities, it may reveal a workforce, funding or policy problem.

For example, missed home visits could initially be treated as isolated incidents. If the pattern increases across a municipality, it may indicate scheduling or recruitment pressure. If multiple municipalities experience the same problem, national workforce policy may need attention.

Governance therefore requires information to travel upwards without losing context.

Useful assurance should enable decision-makers to understand:

• whether people receive the services for which they have been assessed;
• whether provision is timely and continuous;
• whether significant incidents or complaints reveal recurring themes;
• whether staffing and capacity risks are affecting outcomes;
• whether particular population groups experience poorer access; and
• whether corrective actions actually improve the underlying problem.

Organisations examining the strength of comparable accountability arrangements can use the Governance Maturity Assessment to structure questions about responsibility, escalation and assurance. Its value in an international context is not to impose a UK governance structure, but to test whether evidence reaches the level capable of acting upon it.

Dashboards are useful only when measures trigger decisions

Dashboards have become a common response to data complexity. They can bring together service volumes, waiting times, workforce indicators, incidents and financial information in one view.

But a dashboard does not create accountability by itself.

A well-designed indicator should have a purpose, an owner and an expected response when performance changes. Otherwise the system can accumulate attractive visual reporting without altering decisions.

Consider a municipal indicator showing increasing waiting times for home support. The relevant questions are not limited to whether the figure is red or green. Decision-makers need to know which people are waiting, whether need is escalating, whether workforce shortages are responsible, what interim support exists and what action is being taken.

The Quality Dashboard Builder can help organisations structure the relationship between indicators, thresholds and governance response. For Lithuanian services, the measures themselves would need to reflect national requirements, municipal responsibilities and the local service model.

The strongest dashboards combine leading and lagging indicators. A hospital admission is a lagging event. Rising missed visits, workforce vacancies or repeated carer concerns may provide earlier warning.

This allows data to support prevention rather than simply retrospective reporting.

Workforce evidence should connect staffing to continuity and outcomes

Lithuania's formal long-term-care workforce remains comparatively small. OECD data for 2023 place the number of long-term-care workers at around 0.7 per 100 people aged 65 and over under the comparative methodology, far below the OECD average. Such comparisons require caution because workforce definitions and service models differ, but the underlying capacity challenge is significant.

Workforce data therefore need to go beyond headcount.

Vacancy levels, turnover, sickness, skill mix, geographic distribution and time spent travelling between people can all affect capacity. The same number of workers may support very different volumes of care depending on service design and geography.

Continuity also matters. Frequent changes of worker can affect trust, observation and the ability to recognise deterioration, particularly for people with dementia or complex needs.

Connecting workforce information with outcomes creates stronger intelligence. If a municipality experiences rising turnover at the same time as missed visits and complaints increase, the relationship deserves investigation. If a provider maintains continuity despite recruitment pressure, its operating practices may contain useful learning.

Workforce measurement should therefore help answer not simply “How many staff do we have?” but “What capacity and continuity can this workforce reliably deliver?”

Scenario: a dashboard shows stability while frontline pressure is increasing

A municipal service reports stable numbers of people receiving day social care at home. Expenditure remains close to budget, complaints are low and the headline dashboard therefore suggests a stable service.

Frontline information tells a different story.

Vacancies mean workers are travelling further and schedules have become tighter. Families increasingly provide support when visits cannot be extended. Staff record more cases where people's needs appear to be increasing, but service hours remain unchanged because additional capacity is difficult to source.

No single indicator has crossed a formal threshold. The risk sits in the relationship between them.

The municipality changes its oversight approach. Workforce vacancies, missed or shortened provision, reassessment requests, waiting times and family-carer concerns are reviewed together rather than separately. A pattern becomes visible before a major service failure occurs.

The response includes repriorising capacity, reviewing recruitment and examining whether some lower-intensity needs can be addressed differently without withdrawing necessary support.

The example illustrates why quality monitoring systems should be designed around operational risk rather than reporting convenience. Stable activity can coexist with deteriorating resilience.

Data can support prevention as well as accountability

The strongest evidence systems do not merely explain what happened last year. They help identify what may happen next.

Lithuania's demographic trajectory makes this increasingly important. The share of the population aged 65 and over is projected to increase substantially over coming decades, while the working-age population is expected to contract. Long-term-care planning therefore cannot depend solely on reacting to current applications.

Population projections can be combined with information about disability, service utilisation, workforce capacity and geography to anticipate future pressure.

At municipal level, the questions become practical. How many people are likely to require home support? Which settlements will have the greatest concentration of older residents? What happens if the workforce does not grow? How might expanding home-based services alter demand for residential care?

The Digital Twin Scenario Modeller offers organisations a way to explore interactions between demand, workforce and service capacity. It does not predict Lithuania's future or replace official planning. Its useful principle is scenario testing: decisions become stronger when leaders examine several plausible demand and capacity conditions rather than relying on a single forecast.

Predictive analysis also requires humility. Data can identify patterns, but it should not automatically determine an individual's entitlement or care pathway. Population risk modelling and person-centred assessment perform different functions.

Outcome evidence needs to influence funding decisions

Evidence becomes strategically important when it affects resource allocation.

Lithuania's long-term-care financing crosses organisational boundaries. Health-related services, municipal social services, state transfers, personal contributions and other support mechanisms do not all sit within one budget. This can make it difficult to capture the wider return from an intervention.

A municipality might invest in additional home support that helps prevent deterioration, while some financial benefit appears later through reduced hospital use. A healthcare intervention might improve mobility and reduce social-care dependency. Better support for a family carer may delay demand for formal services.

Activity-based financial reporting will not necessarily reveal these relationships.

Outcome evidence can help decision-makers understand value across the pathway. That does not require pretending that every social outcome can be converted into a precise monetary saving. It requires identifying where expenditure changes need, risk or future service use.

This is particularly relevant as Lithuania continues developing integrated long-term-care arrangements. Integration will be difficult to sustain if each funding stream is judged only by the activity purchased within its own boundary.

Person-level data need strong privacy and proportionality

Better integration should not become an argument for collecting unlimited information.

Long-term-care records can contain highly sensitive information about health, disability, cognition, finances, family relationships, living arrangements and personal care. Connecting datasets therefore increases both analytical value and governance responsibility.

Information should be collected for a clear purpose, access should reflect professional need, and people should understand how their data are used where applicable. Digital systems also need appropriate security, continuity and access controls.

There is a wider ethical question as predictive technology develops. Data may eventually help identify people at increased risk of falls, hospital admission or loss of independence. Such intelligence can support earlier intervention, but poorly designed models can also reproduce existing inequalities.

If historical data contain fewer services in rural communities because access was difficult, an algorithm trained uncritically on previous utilisation could interpret low use as low demand.

Human oversight therefore remains essential. Data can inform professional and policy judgement; it should not make complex social decisions invisible behind an automated score.

Learning should close the loop between evidence and improvement

Accountability is incomplete if data identify a problem but nothing changes.

This is where learning, incidents and continuous improvement become part of the evidence architecture.

Suppose several older people experience delayed access to support after hospital discharge. Recording each delay creates an incident history. Analysing the pattern may show that assessments are repeatedly completed after the person returns home rather than before discharge. Changing the pathway addresses the process. Measuring subsequent delays establishes whether the intervention worked.

The same cycle can apply to complaints, falls, medication concerns, service interruptions or repeated institutional admissions.

Strong improvement systems therefore connect four stages: identify the signal, understand the cause, implement a change and test whether the outcome improved.

Without the final stage, an action plan demonstrates activity rather than effectiveness.

Scenario: lived experience changes what the performance data mean

A Lithuanian residential social-care service reports high satisfaction from its annual questionnaire. Most residents select positive responses and formal complaints are rare.

On the surface, the evidence suggests strong experience.

A series of smaller group discussions produces a more nuanced picture. Residents generally like the staff but say they have limited influence over when they get up, when activities occur and how easily they can leave the service for community activities. Some people did not understand parts of the written questionnaire and had relied on staff assistance to complete it.

The service does not discard the satisfaction data. Instead, it recognises its limitation.

Future feedback uses more accessible methods, separates experience of staff from control over daily life and tracks whether changes increase individual choice. Managers also compare feedback with care-plan goals and community-participation information.

The result is a better definition of quality. High satisfaction remains valuable, but it no longer substitutes for evidence about autonomy.

This illustrates why accountability should include people's voices without reducing co-production to a survey score. Lived experience often explains what quantitative indicators cannot.

A stronger Lithuanian evidence model would connect different levels of accountability

The central opportunity is not to create one enormous national performance framework. Different decisions require different levels of evidence.

A frontline worker needs information that supports the next care decision. A provider needs evidence about safety, staffing, continuity and outcomes. A municipality needs to understand demand, access, capacity, quality and expenditure across its population. National institutions need sufficient comparability to identify systemic variation, plan reform and judge whether public investment is achieving its intended purpose.

These layers should reinforce one another.

National requirements that generate substantial reporting but little local value create administrative burden. Local data that cannot be compared or aggregated limit national oversight. Outcome measures that are too abstract for frontline practice become reporting exercises rather than care tools.

The strongest architecture therefore uses a relatively small set of common concepts while allowing detail to reflect local services and individual needs.

Governance should also establish what happens when evidence shows persistent variation. Transparency is useful, but accountability requires an escalation route, investigation and, where appropriate, support or corrective action. Wider quality assurance and governance become meaningful when decision-makers can trace an issue from data to explanation to action.

What international systems can learn from Lithuania

Lithuania's evidence challenge is not unique. Many countries have built health and social-care systems through different legislation, funding routes and institutions, leaving information fragmented along the same organisational boundaries.

The Lithuanian experience highlights why integration cannot be assessed simply by creating a joint service or shared policy. Integrated care also requires integrated intelligence.

The transferable lesson lies less in any particular Lithuanian database or administrative mechanism and more in the design principles.

Activity data remain essential, but they should be connected to need and outcome. Municipal comparison is valuable, but variation requires contextual interpretation. Digital interoperability can improve continuity, but information sharing needs proportionality and clear responsibility. Family and service-user experience can expose risks that administrative datasets miss. Workforce data become more useful when connected with continuity and quality.

Most importantly, evidence should lead somewhere.

A mature long-term-care system does not collect information because reporting is expected. It uses evidence to decide where capacity is required, which inequalities need attention, whether reforms are working and what should change when outcomes remain poor.

The future direction: from retrospective statistics to active system intelligence

Lithuania has an opportunity to develop long-term-care information alongside the care model itself rather than attempting to retrofit evidence after reform is complete.

As integrated long-term care develops, data standards can be designed around pathways. As home and community services expand, outcomes can be defined before activity becomes the dominant measure. As digital infrastructure develops, interoperability can be treated as an operational requirement rather than simply an information-technology project.

Future systems may make greater use of linked datasets, predictive analytics and artificial intelligence to identify patterns of demand or emerging risk. These are plausible directions rather than reasons to automate complex decisions prematurely.

The immediate opportunity is more fundamental: consistent information about need, access, continuity, workforce, quality and outcomes, connected sufficiently well to support decisions at person, provider, municipal and national level.

If that architecture develops successfully, Lithuania will be better able to distinguish whether additional expenditure is increasing activity or improving lives, whether community care is genuinely reducing dependency or merely shifting pressure to families, and whether national reform is reaching people consistently across different municipalities.

Conclusion

Lithuania's long-term-care reform is ultimately an accountability challenge as much as a service-design challenge. Expanding home support, strengthening integrated care and responding to demographic change will require substantial resources, but expenditure and activity alone cannot demonstrate whether the system is becoming more effective.

The stronger evidence model begins with the person. Assessment should establish what needs to change or be maintained; services should record what support was delivered; review should examine whether the intended outcome occurred. Provider and municipal information can then show whether access, continuity, workforce and quality are supporting those individual outcomes, while national analysis can identify variation and determine whether reforms are producing equitable results.

This requires health and social-care information to become more connected without erasing legitimate organisational responsibilities or weakening privacy. It also requires Lithuania to make invisible pressures more visible: unmet need, unsupported family care, rural access barriers and workforce fragility can all disappear behind apparently stable service volumes.

The decisive shift is from reporting to learning. Data become valuable when they trigger questions, evidence becomes useful when it changes decisions, and accountability becomes meaningful when persistent variation produces action. For Lithuania, that connection between national intelligence and everyday experience will be central to building long-term care that is not only larger, but more equitable, sustainable and demonstrably effective.