Learning from Incidents in Icelandic Care Services: Building Stronger Safety and Improvement Systems

A medication error in a nursing home, an unexpected injury during home support or a serious event affecting a disabled person may begin as a single operational incident. What happens afterwards determines whether it remains an isolated event or becomes useful intelligence about the wider care system. Iceland has formal routes for reporting serious events across healthcare and welfare services, professional and organisational responsibilities for safety, and national bodies able to investigate or supervise concerns. The more difficult task is ensuring that the learning travels back into everyday practice.

This matters across the Iceland Ageing, Long-Term Care & Community Support Knowledge Hub because Iceland’s care system is increasingly supporting people with complex needs across organisational boundaries. An older person may simultaneously depend on municipal home support, home nursing, primary healthcare, hospital specialists, relatives and digital systems. A disabled person may receive municipal services while also using healthcare and community support. Incidents occurring within these pathways may therefore expose weaknesses that no single organisation can understand in isolation.

The central safety challenge is not eliminating every adverse event. Care involves uncertainty, human judgement and legitimate risk. Strong systems instead recognise harm quickly, respond proportionately, understand contributing conditions, support the people affected and change practice where the evidence justifies it. For Iceland, the opportunity is to connect formal reporting with a deeper culture of learning from incidents so that recurring patterns influence workforce planning, service design, digital systems and national policy rather than repeatedly being managed as separate cases.

Incident learning sits across two different oversight systems

Iceland’s division between healthcare and municipal welfare services is reflected in its safety architecture. Healthcare providers and licensed health professionals operate within national healthcare legislation and the oversight responsibilities of the Directorate of Health. Welfare services, including important areas of municipal social services, disability support and services for older people, fall within the supervisory remit of the Quality and Supervisory Authority of Welfare, GEV, according to the legislation governing its functions.

The distinction matters operationally because an incident does not always respect institutional boundaries.

An older person receiving municipal home support may also receive home nursing. A medication-related event may primarily involve healthcare responsibilities, while concerns about missed assistance, communication or living conditions may involve the municipal service. A hospital discharge may create risks that only become visible after the person returns home. In supported living, an incident may involve welfare-service practice while also requiring clinical assessment from healthcare professionals.

A mature safety system therefore needs clarity about responsibility without allowing administrative boundaries to fragment the learning. The first question is who needs to respond to the immediate event. The second is who else needs to know what the event reveals.

Healthcare incidents have a formal national reporting structure

Icelandic healthcare legislation places explicit emphasis on patient safety and the reporting of adverse events. Healthcare institutions are expected to maintain systems through which adverse events can be reported, analysed and used to improve safety. Serious unexpected events can also require notification to the Directorate of Health.

This creates two connected levels of learning.

At provider level, incident reporting should enable healthcare organisations to recognise unsafe processes, investigate contributing factors and improve practice. At national level, serious events and wider safety information can give the Directorate visibility over risks that may extend beyond one institution.

The distinction between an adverse event and an unavoidable complication is important. Healthcare is not risk-free, particularly for frail older people with multiple conditions. A resident may fall despite proportionate prevention. A person may deteriorate rapidly even when care has been appropriate. Incident systems become distorted if every negative outcome is automatically treated as evidence of error.

Conversely, an outcome should not be dismissed as inevitable merely because the person was frail or clinically complex. The purpose of review is to determine what happened, whether the response was reasonable and whether anything in the care process, environment, communication or system increased the likelihood or severity of harm.

Welfare services also have duties when serious unexpected events occur

GEV’s supervisory framework provides a corresponding route within welfare services. Service providers within its remit are required to notify the authority of serious unexpected incidents, and GEV can receive complaints and information concerning the quality of services. Its supervisory work can include examination of documents, interviews, observation and direct engagement with people using services.

This is particularly important as Iceland develops more community-based disability support and seeks to enable more older people to remain at home. Safety cannot be monitored only in hospitals and nursing homes. Harm can occur through missed support, inadequate supervision, inappropriate restrictions, communication failures, unsafe transitions or a failure to respond when somebody’s needs change.

For organisations examining similar responsibilities, the wider principles of incident response, protection and escalation provide a useful analytical connection. The exact Icelandic legal process remains decisive, but the operational requirement is broader: immediate safety action and longer-term learning need to be treated as related but distinct tasks.

Not every incident needs the same response

A useful incident system is proportionate. If every minor event triggers a large formal investigation, staff can become overwhelmed and reporting may deteriorate. If only catastrophic events receive attention, repeated low-level signals can accumulate until serious harm occurs.

Services therefore need mechanisms that distinguish between events while retaining the ability to identify patterns. Relevant considerations can include:

• the actual and potential severity of harm;
• whether the person remains at risk;
• whether the event appears isolated or recurrent;
• whether several people or services could be affected by the same weakness;
• whether professional, legal or national reporting requirements apply; and
• whether the incident indicates a wider failure in staffing, systems, communication or service design.

The categorisation should guide the response, not determine the conclusion in advance. An apparently minor missed home-support visit may warrant limited individual review. Repeated missed visits affecting people with high dependency could indicate a significant capacity or scheduling problem.

Operational scenario: a medication error reveals a transition problem

An older woman returns to a nursing home after several days in hospital. Her medicines have changed during the admission. The discharge information reaches the nursing home, but the updated regimen is not fully reconciled with the medication information already in use. A dose of a medicine that should have been stopped is administered before the discrepancy is recognised.

The resident is assessed promptly and does not experience serious lasting harm. The immediate clinical response is appropriate, the event is recorded and the relevant professional and organisational processes are followed.

A superficial review could conclude that the nurse administering the medicine should have checked more carefully. A stronger investigation examines the whole transition. The hospital documentation, transfer timing, medication record, pharmacy arrangements, staffing level, handover and responsibilities for reconciliation are reviewed. Managers discover that the process relies on individual staff recognising differences between two information sources during a busy admission back to the home.

The improvement response therefore addresses the workflow rather than relying solely on additional staff reminders. Medication reconciliation responsibilities are clarified, discrepancies require resolution before routine administration proceeds where clinically appropriate, and recurrent transition incidents are monitored collectively.

This is where medicines, frailty and safety intersect with system design. The person who made the final error remains professionally accountable for their actions, but accountability does not remove the obligation to understand why the system made that error possible.

Investigation needs to move beyond individual blame

Care incidents often involve human actions. Somebody did not communicate information, a task was omitted, a judgement proved incorrect or an escalation occurred too late. It is therefore tempting to identify the individual closest to the event and treat their action as the cause.

That approach can produce an administratively neat conclusion while leaving the underlying risk intact.

A meaningful investigation considers the conditions surrounding the action. Was the worker competent for the task? Was the information available? Were responsibilities clear? Did workload make the intended process unrealistic? Was equipment functioning? Had the same problem occurred previously? Were digital systems interoperable? Did organisational culture make it difficult to ask for help?

This does not mean removing individual accountability. Deliberate misconduct, reckless practice or working outside professional competence require appropriate responses. But most safety improvement depends on distinguishing those situations from ordinary human error occurring within poorly designed systems.

The discipline of root-cause and thematic analysis is valuable when it is used to uncover interacting causes rather than to manufacture a single root cause for a complex event.

Psychological safety affects what the organisation can see

An incident-reporting system is only as useful as the willingness of workers to use it.

Care workers and healthcare professionals need confidence that reporting a genuine mistake, near miss or concern will lead to a proportionate response. A culture that treats every error as personal failure can unintentionally suppress the very information required to make services safer.

The opposite extreme is equally problematic. A learning culture does not mean that standards no longer matter or that harmful conduct has no consequences. It means that the response differentiates between deliberate wrongdoing, unsafe capability, system weakness and reasonable decisions that produced an unfortunate outcome.

Supervision and leadership therefore become safety mechanisms. Workers need opportunities to discuss uncertainty before it becomes an incident and to reflect after events without the process becoming either punitive or complacent.

For Iceland’s increasingly diverse workforce, psychological safety also has a language and inclusion dimension. Migrant workers or newly qualified staff may be particularly reluctant to challenge more senior colleagues or report uncertainty if they fear being judged as incompetent. A strong safety culture actively reduces that barrier.

Near misses can be more valuable than serious harm

Serious incidents naturally receive attention because the consequences are visible. Near misses are easier to overlook. Yet an event that almost caused harm can provide equally important information without the human cost of an adverse outcome.

Imagine a home nurse discovering that an older person has accidentally received two different sets of medication instructions after moving between services. The error is identified before the medicines are taken incorrectly. No injury occurs.

If the organisation records only events involving actual harm, the pathway weakness disappears from the safety record. If the near miss is captured, reviewed and compared with similar events, the system receives an early warning.

This principle is especially relevant in a small country. Iceland may not generate the same absolute volume of incidents as much larger health and care systems. Waiting for statistically large numbers before acting can therefore be inappropriate. A small number of similar events across different locations may already represent meaningful intelligence.

The strongest opportunity lies in combining formal serious-incident reporting with local mechanisms that make lower-level events and near misses visible. The objective is not to create an enormous reporting burden. It is to identify patterns early enough to prevent escalation.

Operational scenario: repeated missed visits become a system signal

A municipal home-support service experiences several missed or significantly delayed visits during a winter period. Individually, the incidents appear understandable. One worker is absent unexpectedly, another route is disrupted and a scheduling change is not communicated correctly. Alternative arrangements are made and none of the people affected experiences immediate serious harm.

Each case could therefore be closed as a minor operational disruption.

When the events are reviewed collectively, a different pattern appears. The service has very little spare capacity during evening periods. Replacement workers are being deployed across large areas, while the scheduling system does not reliably distinguish visits where delay is inconvenient from visits where it could create significant risk.

One person requires assistance to eat. Another needs help preparing safely for bed. A third has family nearby who can provide contingency support. Treating all three visits as equivalent hides important differences in consequence.

The municipality responds by strengthening contingency prioritisation, clarifying escalation when essential visits are threatened and reviewing whether staffing capacity is adequate for predictable absence and travel disruption. Workers are also given a clearer route for raising capacity concerns before visits are missed.

For organisations exploring comparable pressures, the Predictive Workforce Risk Module offers a way to examine how absence, vacancies, continuity and deployment pressure may combine into service risk. It is not an Icelandic reporting framework; its value lies in connecting workforce indicators with operational consequences before repeated disruptions become normalised.

Transitions are particularly important learning points

Many serious risks arise not within one service but between services.

Iceland’s health and welfare architecture makes this especially significant. Hospitals, primary healthcare, home nursing, municipal home support, rehabilitation, nursing homes and disability services can all contribute to a person’s pathway. Each may perform its own role appropriately while the transition between them remains unsafe.

Information may arrive late. Responsibilities may be interpreted differently. The receiving service may not understand how much somebody’s needs changed during hospital treatment. Equipment may not be available when the person arrives home. Family members may believe formal services are providing support that has not yet started.

Incident review should therefore ask whether the event exposes a pathway problem rather than only an organisational problem.

The wider principles of safe transitions between hospital and home-based services are relevant here. Repeated transition incidents may require action between organisations, not another internal policy within one of them.

This is also where governance becomes more difficult. One provider can change its own process. Cross-system improvement requires agreement about responsibilities, information and escalation across institutional boundaries.

People and families hold evidence that incident systems can miss

Formal records show only part of what happened.

A person using a service may know that the worker who arrived that day appeared unfamiliar with their support needs. A family member may have raised concerns several times before an incident. A nursing-home resident may explain that they stopped using a walking aid because it was repeatedly left out of reach. A disabled person may describe a restrictive practice that staff records present as ordinary support.

Listening to those accounts is not simply an act of courtesy after something has gone wrong. It is part of understanding causation.

Incident processes should therefore create proportionate opportunities for the person affected, and family or representatives where appropriate, to explain their experience. Accessible communication may be required for people with intellectual disabilities, dementia, sensory impairments or communication differences.

The process also needs to avoid treating family accounts as substitutes for the person’s own voice. Rights-based care requires attention to the individual’s wishes, communication and interpretation of the event.

Meaningful involvement can reveal risks that formal systems cannot easily quantify: loss of trust, fear of particular workers, repeated small disruptions, feeling ignored or a gradual reduction in independence after an incident.

Operational scenario: a fall raises a question about positive risk

An older man living in a nursing home values walking independently to the dining area. He has experienced declining balance and has previously fallen. Staff have discussed the risk with him and support his wish to remain mobile, using an appropriate walking aid and assistance when required.

One afternoon he falls while walking and sustains an injury requiring assessment.

The incident review could lead to a defensive response: require him to wait for staff before walking, increase restrictions and interpret any future independent movement as unsafe behaviour. That might reduce exposure to falls while significantly reducing autonomy and mobility.

A stronger review examines whether the existing plan was reasonable and whether anything changed. Staff discover that his walking aid had been moved during cleaning and was not immediately accessible. He attempted to walk without it rather than waiting.

The improvement is therefore not necessarily to remove his independence. The service addresses equipment availability, reviews his current mobility with relevant professionals and discusses the event with him. His preferences remain part of the decision.

The principles of positive risk-taking for older people are important because a safety system that responds to every adverse event by eliminating autonomy can itself produce harm. Incident learning should improve the quality of risk decisions, not automatically make them more restrictive.

Safeguarding and incident learning overlap but are not identical

Some incidents indicate possible abuse, neglect, exploitation or serious failures to protect a person. Those circumstances require the appropriate safeguarding and legal response rather than being treated solely as quality-improvement events.

Other incidents arise without abuse. A medication error, accidental fall or equipment failure may still cause serious harm but require a different response.

The distinction matters because organisations need to recognise when an incident should trigger protection, investigation or referral beyond the normal internal learning process. At the same time, safeguarding cases can generate important organisational learning once immediate protection and any formal investigation are appropriately managed.

Recurring neglect concerns may reveal staffing or leadership weaknesses. Financial exploitation may expose poor controls. Repeated unexplained injuries may indicate inadequate recognition or escalation. Restrictions imposed for convenience may reveal cultural rather than individual practice problems.

Strong safeguarding culture and leadership therefore connect protection with organisational visibility. The purpose is not to collapse all incidents into safeguarding, but to ensure that serious protection concerns are recognised and that wider learning is not lost.

Workforce competence needs to be examined without making training the default answer

Incident investigations frequently end with a recommendation for additional training. Sometimes that is exactly what is required. If workers do not understand a procedure or lack competence for a task, education, supervised practice and assessment may be necessary.

But “retrain staff” can also become an easy response to a problem whose real cause lies elsewhere.

A worker may know the correct procedure but be unable to follow it because staffing is insufficient. An electronic record may make essential information difficult to find. Responsibilities between professionals may be ambiguous. New workers may have received classroom training without adequate supervised practice. A policy may describe an ideal workflow that does not match the reality of the service.

Incident analysis should therefore distinguish knowledge from capability and operating conditions.

Where competence genuinely contributes, the response needs to go beyond attendance at training. Supervision, observation, practice assessment and continuous professional development can provide stronger evidence that learning has translated into practice.

This is particularly important in long-term care as residents and people supported at home develop more complex needs. The safety question is not simply whether sufficient workers are present, but whether the available skill mix matches the decisions and interventions required.

Digital systems can prevent incidents and create new ones

Digital records, scheduling systems, medication technologies, remote monitoring and electronic communication can improve safety when they make relevant information more accessible and reduce avoidable manual processes.

They can also create new failure modes.

An alert can be generated but not assigned to anybody. Information can exist in one system but remain invisible to another organisation. Excessive alerts can lead to important warnings being overlooked. Poor interface design can encourage selection errors. Automated scheduling can optimise travel while failing to recognise which visits require continuity or specific competence.

Incident investigation increasingly needs to ask not only what the worker did but how technology shaped the decision environment.

The stronger approach is to examine digital assurance as part of ordinary safety governance. Services should understand which critical processes depend on technology, what happens when systems fail and whether digital changes create unintended workload or risk.

Technology can support learning by making patterns easier to identify, but only if records are accurate and the organisation has the analytical capacity to interpret them.

Operational scenario: a digital alert exists but nobody owns the response

An older person living at home uses monitoring technology that can indicate changes relevant to their safety. During a weekend, the system generates an alert suggesting a potential deterioration in the person’s usual pattern.

The alert is transmitted successfully. Technically, the system has worked.

The problem lies in the workflow. The service responsible for monitoring assumes that clinically significant concerns will be reviewed by healthcare staff. The home-support team believes alerts are already clinically triaged before they become visible. The person’s family assumes that somebody is actively monitoring the technology continuously.

By the time a worker visits, the older person is clearly unwell and requires urgent clinical assessment.

The incident review does not conclude that remote monitoring is ineffective. Instead, it maps the pathway from signal to action. Ownership, operating hours, clinical escalation thresholds, acknowledgement and contingency arrangements are clarified. Information given to the person and family is also revised so that the service promise matches the actual response model.

The event demonstrates a wider principle: digital safety depends on operational accountability around the technology. A sensor can detect a change, but it cannot resolve ambiguity about who is responsible for acting.

Organisations considering similar technology-enabled pathways can use the Digital Transformation Readiness Assessment to structure questions about workflow, workforce adoption, information governance and digital resilience. Its purpose in this context is to test organisational readiness, not to replace Icelandic clinical or regulatory requirements.

Small numbers make thematic learning especially important

Iceland’s population scale creates both advantages and limitations for incident intelligence.

National bodies can potentially maintain relatively close visibility over serious events, while relationships between organisations may be less remote than in very large systems. However, individual services and municipalities may experience small numbers of particular incidents. Statistical analysis alone may therefore struggle to distinguish a meaningful emerging pattern from ordinary variation.

Thematic learning becomes especially valuable.

Several events do not need to be identical to reveal a common weakness. A missed home-support visit, a delayed hospital discharge and a medication discrepancy may all involve poor transfer of responsibility. Different falls may reveal recurring problems with night staffing or environmental design. Complaints, near misses and serious events may collectively expose communication barriers affecting migrant workers or service users.

The skill lies in coding and reviewing information in ways that preserve those connections.

National oversight bodies can contribute by identifying themes that individual providers cannot see. Municipalities and organisations can contribute detailed operational context. Neither level is sufficient alone.

Incident data should connect with other quality evidence

A service with a rising incident rate may be becoming less safe. It may also have introduced a stronger reporting culture. Conversely, a very low incident rate can indicate excellent performance or under-reporting.

Incident counts therefore require context.

Useful analysis connects them with other information such as:

• severity and potential severity rather than event numbers alone;
• complaints and service-user experience;
• workforce turnover, absence, overtime and continuity;
• interRAI or other relevant quality indicators;
• hospital transfers, falls, medication events or other clinical outcomes;
• inspection and supervisory findings; and
• whether previously agreed improvement actions remain effective.

The Quality Dashboard Builder can help organisations exploring comparable governance questions avoid interpreting incidents in isolation. The objective is not to create a universal score, but to combine enough evidence to distinguish changing risk from changes in reporting behaviour.

Closing the loop is the real test of incident governance

Many organisations can demonstrate that an incident was reported and investigated. Fewer can demonstrate that the resulting action produced sustained improvement.

An action plan may state that a policy was updated, staff were reminded or training was delivered. Those are implementation activities, not evidence that the underlying risk has reduced.

Closing the loop requires a later question: did the change work?

If missed visits led to revised scheduling controls, managers should examine whether essential visits are now more reliable. If medication reconciliation changed, subsequent transition incidents should be reviewed. If workers received new training, supervision should establish whether practice improved. If a digital workflow was redesigned, testing should confirm that alerts reach an accountable person within the intended response time.

This is where embedding learning into everyday practice becomes more important than producing the investigation report itself.

Repeated incidents after an action has supposedly been completed are valuable governance evidence. They suggest either that the intervention was ineffective, that it was not implemented consistently or that the original analysis did not identify the real cause.

Local learning needs a route into national improvement

Iceland’s safety architecture has the potential to support learning at several levels. Individual organisations understand the operational detail. Municipalities can identify recurring pressures within local welfare services. The Directorate of Health can see healthcare safety issues with national relevance. GEV can use complaints, serious-event notifications and supervisory evidence to identify patterns within welfare services.

The stronger opportunity lies in making those levels interact.

A risk that repeatedly appears in one service may require local improvement. A pattern occurring across several organisations may indicate a national workforce, guidance, digital or service-design issue. Persistent variation between areas may require closer examination of resources or implementation. A serious event may expose a gap that requires changes beyond the organisation where it occurred.

This does not mean that every incident should travel through the entire system. Information should remain proportionate and respect confidentiality and legal requirements. But there needs to be a route through which sufficiently important learning can influence decisions above the level of the individual provider.

Organisations examining whether their own governance arrangements create that visibility can use the Governance Maturity Assessment to structure questions about accountability, escalation, assurance and learning. It is not a substitute for Icelandic oversight; it provides a practical framework for examining whether information actually changes organisational decisions.

Transparency needs to support trust rather than defensive reporting

People using services and families reasonably expect organisations to respond openly when significant harm occurs. Staff also need clarity about how incidents are handled. National bodies need sufficiently reliable information to identify risk.

Transparency therefore has several audiences.

It does not require publishing every operational detail or compromising privacy. It means being able to explain how safety concerns are recognised, how serious events are reviewed, what improvement follows and whether recurring risks are being addressed.

At system level, aggregated learning can help demonstrate that reporting serves a public purpose. If workers and organisations perceive incident reporting only as exposure to sanction, defensive behaviour can follow. If reporting is visibly connected to practical improvement while serious accountability remains credible, the system has a stronger basis for trust.

The balance is important. Learning cannot become a reason to avoid responsibility, and accountability cannot become so punitive that risks disappear from view.

From incident management to a learning system

The most mature safety model does not begin when harm occurs. It uses incident evidence to reshape the conditions in which future care is delivered.

That can mean changing staffing models after repeated continuity failures, redesigning transitions after recurring medication discrepancies, strengthening specialist support after a pattern of complex clinical incidents or revising digital workflows when technology repeatedly creates ambiguity.

It also means connecting incident learning with service planning. If several municipalities experience similar difficulties sustaining evening home support, the issue may be broader than individual rota management. If nursing homes repeatedly encounter the same transition risk from hospitals, a shared pathway response may be more effective than multiple local procedures.

This is the deeper value of continuous improvement: operational events become evidence about system design.

For Iceland, population scale could make that connection particularly powerful. A relatively small system can potentially circulate learning between national institutions, municipalities and providers more quickly than much larger and more fragmented systems. But scale alone does not guarantee learning. Information still requires ownership, analysis, communication and implementation.

International learning: the event is local but the cause may be systemic

Iceland’s formal reporting arrangements reflect its own healthcare legislation, welfare-service responsibilities and national institutions. Countries with regional governments, insurance systems or much larger provider markets will organise oversight differently. The Icelandic mechanism therefore cannot be transferred as a ready-made model.

The underlying principles are more widely relevant.

First, serious incidents need formal accountability, but formal reporting is only one layer of safety. Local organisations also need to learn from lower-level events and near misses.

Second, investigation should preserve individual professional responsibility without assuming that the person closest to the harm is the complete explanation. Staffing, information, technology, environment and organisational design frequently influence the outcome.

Third, incidents at organisational boundaries require shared learning. A transition failure cannot always be solved by improving only the sending or receiving service.

Finally, action is not the same as improvement. Updating a procedure or delivering training closes an administrative task; it does not demonstrate that risk has reduced.

Other systems can adapt those principles without replicating Iceland’s supervisory institutions. The transferable lesson lies in building a credible route from event, to understanding, to action, to evidence that practice became safer.

Conclusion

Iceland already has important foundations for learning from harm: statutory responsibilities within healthcare, serious-event reporting, national patient-safety oversight, GEV supervision of welfare services, complaints processes and increasingly structured quality information. The strategic opportunity is to connect those mechanisms more consistently with everyday operational improvement.

That requires incident systems that people trust enough to use, investigations capable of looking beyond immediate human error and governance that recognises recurring patterns across workforce, technology, transitions and service design. It also requires the experience of people and families to remain visible. An investigation may establish what happened administratively while still missing what changed in the person’s confidence, autonomy or sense of safety.

As care becomes more complex and increasingly delivered across homes, municipalities, healthcare institutions and community services, the most important incidents will not always fit neatly within one organisational boundary. Iceland’s small scale can be an advantage if local learning reaches national decision-makers and national intelligence returns to services in a form that changes practice.

The strongest safety system is therefore not the one reporting the fewest incidents. It is the one able to recognise risk honestly, respond proportionately, learn across boundaries and demonstrate that the same insight has altered what happens next. That is how incident reporting becomes more than accountability after harm: it becomes infrastructure for safer care.