Quality and patient safety surveillance systems in healthcare settings: a rapid realist review

Abstract

Objectives

Despite extensive efforts in data collection, quality and safety measurement remains a significant global challenge, with limited understanding of how and under what conditions quality and patient safety surveillance systems function effectively. With the aim of informing the development and effective functioning of quality and patient safety surveillance systems, a rapid realist review was conducted to develop a set of theories that address how, why, for whom and in what context quality and patient safety surveillance systems work.

Design

Rapid realist review to inform recommendations and intervention design for the monitoring and evaluation phase of the QS Signals Project, reported according to Realist and Meta-narrative Evidence Syntheses: Evolving Standards (RAMESES) guidelines.

Data sources

Initial programme theories were constructed based on data collected from key articles on quality and patient safety surveillance systems, consultation with an expert panel, informal meetings with a project team charged with developing a quality and patient safety surveillance system for maternal and infant health and a review of the project’s planning documents. A three-phase iterative search of PubMed, PsycInfo, CENTRAL, CINAHL and grey literature was conducted, including studies in healthcare settings across all patient groups.

Eligibility criteria

Documents were assessed for relevance (alignment with the theory under test), richness (depth of insight) and rigour (trustworthiness and coherence of data).

Data extraction and synthesis

Context–mechanism–outcome configurations were generated, iteratively refined and grouped under relevant programme theories to contribute to theory refinement.

Results

The review process resulted in the development of 11 final programme theories, identifying mechanisms operating at organisational and national levels. Effective systems were enabled by leadership commitment, organisational readiness for change and a supportive safety culture. Clear governance structures, including defined local and national roles, strengthened accountability and coordination. The establishment of multidisciplinary clinical advisory groups facilitated the selection of meaningful safety indicators. Sustainable financial investment and adequate human and technical resources were critical for implementation. Robust data governance frameworks enhanced trust, transparency and appropriate data use. User-centred system design improved data accessibility and usability, while feedback loops supported learning and continuous improvement.

Conclusions

Quality and patient safety surveillance systems function most effectively when supported by strong leadership, clear governance structures, adequate resources and a learning-oriented culture that enables the meaningful use of safety data. The findings emerging from this review provide comprehensive, practical and testable systems-level programme theories to inform future research on the development of quality and patient safety surveillance systems across diverse healthcare settings and international contexts.

STRENGTHS AND LIMITATIONS OF THIS STUDY.

• The Rapid Realist Review methodology enabled the inclusion of heterogeneous and context-rich evidence that would not typically be incorporated in a traditional systematic review.

• The engagement of an expert panel including IT, quality and safety, clinical, data experts and national leaders strengthened the search strategy, study selection and prioritisation of programme theories, thereby enhancing methodological rigour.

• Integration of context–mechanism–outcome configurations under shared mechanism themes was complex due to substantial variation in contexts, operational levels and system characteristics across included studies.

• The interlinked nature of the quality and patient safety surveillance system components made it methodologically challenging to attribute specific outcomes to individual components or to determine for whom they were effective.

• Limited outcomes data from recently developed digital quality and patient safety surveillance system improvements restricted the ability to establish robust causal mechanism–outcome links.

Introduction

Patient harm in healthcare settings due to unsafe care remains a significant global health issue and is a leading cause of death and disability worldwide.¹ Despite significant improvements in quality and patient safety, including the implementation of comprehensive infection prevention and control practices, digital quality measures for medication safety, reporting quality metrics including infections, sepsis, patient falls, antimicrobial resistance and developing evidence-based multifactorial care intervention standards, challenges in healthcare safety are ongoing.²,⁴ Global estimates indicate hospital-acquired infections range from 5–15% and up to 3% of patients experience preventable medication-related harm, at least one-fourth of which is severe or potentially life-threatening.⁵,⁷ This rate is a significant shortcoming considering the time and effort the healthcare system devotes to data collection through incident reporting systems, patient-reported outcomes, clinical care audits and organisational culture surveys.⁸ The majority of quality and safety data are unanalysed and many risky situations are buried in the reports, making it difficult for patient safety and risk analysts to detect them.⁹ WHO’s (WHO) Global Patient Safety Action Plan 2021–2030 recommends developing information flows to reduce avoidable harm and mitigate risks, reinforcing the need for effective systems to safeguard patient safety.¹ WHO emphasised that whatever data are used to assess a health system or organisation’s level of patient safety, the process must be strongly linked to learning and improvement. Thus, the primary role of quality and patient safety surveillance systems has been described as improving the quality and safety of care by learning from healthcare system errors.^{10 11} A well-functioning quality and patient safety surveillance system is expected to support the delivery of health services by ensuring the production, analysis, dissemination and use of reliable and timely information on the quality of care.^{1 8}

Quality and patient safety surveillance systems aim to learn from past mistakes, supporting a positive safety reporting culture, risk reduction and disseminating findings at the local, national and international levels.¹¹ Using health information technology for quality and safety measurements improves data quality, accessibility, medication safety, continuous monitoring, timely identification of quality and safety issues and interoperability across healthcare systems and enables the combination of data from multiple provider organisations to identify patterns not visible within a single organisation.^{29 12},¹⁴ Examples from the literature demonstrate that quality and patient safety surveillance systems aid in prioritising resource allocation to the areas of system weakness, prompting health professionals to take proactive measures while also serving as a foundation for scientific developments to enhance the healthcare system.^{8 12 15 16} However, the integration of heterogeneous surveillance systems and the lack of uniform standards make measuring quality and safety challenging.^{1 17} The main challenges include (a) the difficulty of setting error thresholds; (b) integrating data sets; (c) poor quality of data; (d) data access and sharing policies; (e) lack of multidisciplinary full-time, experienced team including data manager, project manager, co-investigators, health economists, physician, nurse, data analytics and clinical information services and administrators; and (f) organisations’ concern about potential breaches of health information.¹⁴¹⁸,²¹ These challenges limit the ability to fully understand the most effective methods for translating quality and patient safety surveillance systems into improved outcomes at both the provider and system levels.

In this regard, the WHO Global Patient Safety Action Plan 2021–2030 highlights the necessity for advancing methodologies to enhance safety measurement. This includes (a) increased application of theory and logic models, (b) more precise descriptions of interventions, their mechanisms and pathways to implementation, (c) a better understanding of both intended and unintended outcomes and (d) a more comprehensive approach to describing and measuring contexts and how they affect the success of quality and patient safety surveillance systems implementation.¹ To address these priorities, this rapid realist review synthesises the insights from the experiences, successes and failures of similar programmes and extracts generalisable findings to understand the key contexts, mechanisms and outcomes necessary for the development and effective functioning of quality and patient safety surveillance systems.²²,²⁵ By synthesising these insights, this rapid realist review aims to develop a set of theories that address how, why, for whom, in what context they work and what are the dominant outcome patterns in the identified contexts. Additionally, it aims to provide practical and transferable knowledge for various quality and patient safety surveillance systems operations, supporting meaningful system improvements.

Method

Design

This research was undertaken to inform of recommendations and intervention design for the monitoring and evaluation phase of the QS Signals Project. In Ireland, a substantial volume of quality and patient safety data is systematically collected each year.²⁶ The Irish Health Service Executive’s Patient Safety Strategy (2019–2024) commits to analysing patient safety information and data to assess the reliability of care processes, identify potential risks early and learn from failures and successes.²⁷ In line with this strategy, the Health Service Executive has initiated the QS Signals Project, which aims to develop a platform that integrates, analyses and displays quality, safety and operational data from various datasets in one centralised location to maximise insights into quality and safety in Irish maternity and neonatal healthcare settings.²⁸ Given the complexity of the intervention and the need for timely and utility focused insight, a rapid realist review was chosen as a pragmatic, theory-driven approach to generate evidence.²⁴ The Systems Theory lens was applied to understand the dynamic interactions of technical (reporting systems, data collection, networking infrastructure) and non-technical (clinical workflow, organisational policies, culture and management) factors influencing the development and functioning of quality and patient safety surveillance systems.²⁹

Expert panel

As this rapid realist review represents the first stage of a proposed realist evaluation of QS Signals Project intervention in Ireland, it was important to involve the Health Service Executive Quality and Safety Division. Members of this Division, the QS Signals project developers, coordinators and managers of the QS Signals project, were actively engaged in the process to ensure the rapid realist review addressed the programme-relevant gaps.²⁴ These knowledge users and experts will be referred to as the ‘expert panel’ throughout this manuscript.³⁰ The expert panel consisted of two quality and safety specialists from the project’s change management workstream, one academic expert in realist evaluation and two academic experts who were members of the clinical advisory group workstream responsible for the metric development of signals (see online supplemental appendix 1 for detailed information on expert panel involvement in this review).

Step 1: Identifying the initial programme theories

To inform the scope of the rapid realist review, key articles on quality and patient safety surveillance systems in any healthcare settings were identified using the keywords surveillance system, quality and safety in the Google Scholar and Pubmed search engine. One of the researchers (BE) conducted informal meetings with the project team members to get their opinions and practical experiences on the QS Signals operational process.²² Quality and patient safety surveillance systems vary in their data sources, validation, evaluation, operation (eg, active, automatic or cloud-based learning systems) and level (hospital, board, trust, national), making it impossible to apply a single universal method to reliably measure patient care quality and safety.^{1 11 17 31} Thus, the research team focused on papers in which quality and safety surveillance systems improve the flow of information and knowledge to promote risk mitigation, reduce unnecessary harm and enhance care safety, aligning with WHO’s Global Patient Safety Action Plan 2021–2030 and the strategic objectives of Ireland’s National Patient Safety Strategy (2019–2024).^{27 32}

The two-stage search was conducted across PubMed, CINAHL, PsycINFO, CENTRAL and SafetyLit, including grey literature and organisational websites, for example, WOS, Grey Literature Report, Google Scholar, OpenGrey.eu or Greylit.Org, Annual Reports of WHO, AHRQ, NHS and PSNet. Keywords focused on patient safety and quality surveillance systems, excluding public health, vaccination, COVID-19 and drug safety studies. Only English-language studies published from 2000 to the present were included, reflecting developments following To Err is Human.³³ Inclusion criteria targeted studies reporting context, mechanisms, outcomes or challenges of surveillance systems across any healthcare setting, with iterative refinement of search terms and inclusion/exclusion criteria throughout the review. The rapid realist review protocol provides more detailed explanation of the review process.³² (Online supplemental appendix 2 details the search terms.) Figure 1 provides an overview of the programme theory (PT) development phases used across this review.

Figure 1. The programme theory development phases. CIPTs, candidate programme theories; CMOs, context–mechanism–outcome chains; IPTs, initial programme theories; RRR, rapid realist review.

The six initial programme theories (IPTs) from the 12 candidate initial programme theories were constructed with the expert panel (Box 1).

Box 1. Prioritised initial programme theories.

IPT.1. Readiness for change

If there is (C) a clear, organised and evidence-based expectation from users; a culture open to change; project aims to deliver on an unmet ‘need’ at national and service delivery levels. This enacts (M), stakeholders’ understanding of and belief in the programme, and motivation to get involved; and results in (O) the change will be accepted more readily by users; users apply the change requirements more easily.

IPT 2. Leadership

If there is (C) a reliable and well-known national programme team leader; a leader who is a strong communicator; an experienced leader who knows the system. This enacts (M), trust, motivation, engagement and a sense of ownership by staff. And results in (O) greater buy-in of stakeholders.

IPT 3. Local governance model

If there is (C) a quality surveillance system in collaboration with local governance arrangements – which includes lead clinicians and managers and a safety champion with a complete understanding of the surveillance system, ongoing training for the staff on using the system. This enacts (M) ensuring early action and support for areas of concern; strong governance processes of the full implementation of the quality surveillance model and results in (O) optimising the operation of a robust and reliable governing body for patient safety; good oversight of clinical quality by taking quick, appropriate, objective, regular and national-based actions.

IPT 4. Establishment of a clinical advisory group to identify safety indicators and measures

If there is (C) a multidisciplinary Clinical Advisory Group (CAG) established with appropriate representation from disciplines and across the maternity networks under the leadership of the Clinical Director; extraction of appropriate measures/metrics for consideration from relevant sources (comprehensive review of international literature, consultation with subject experts, international and local guidelines/procedures/strategy documents, review of measures included in similar dashboards internationally); a co-design process facilitated by the project team that allows experts to drive the selection of measures through the use of appropriate methods (workshops, surveys, interviews) for inclusion in the suite of measures of Quality and Safety; an accurate assessment of the feasibility of identified measures by the project team with feedback to CAG members. This enacts (M) transparent, inclusive, and collaborative decision-making around the selection and prioritisation of indicators of quality and safety. And result in (O) a credible, acceptable, and useful set of measures.

IPT5. Financial investment and resource availability

If there are (C) sustainable financial and budget commitments to run the system. This enacts (M) ongoing development of a comprehensive infrastructure for data sharing to support the surveillance system. And result in (O) continuous improvement of the system.

IPT 6. Data governance

If there is (C) compliance with national and international data protection regulations, outlined through a Data Protection Impact Assessment and Data Sharing Agreements where required; the establishment of Data Governance structures (Data Governance Group and Data Steward Groups) to oversee the production of data management policies and standard operating procedures for the ongoing management of the programme. This enacts (M) trust that data shared will be appropriately managed ensuring data sharing; and result in (O) used to optimise insights generated from existing data-sets and action to improve the quality and safety of services.

In Step 2, the researchers conducted an extensive literature search. The search strategy and search string were detailed in the protocol for this rapid realist review, which was published elsewhere.³² Articles were included for full-text review based on the relevance criteria. Relevance is whether the study addresses/contributes to the theory under test, determined by the presence of traceable context–mechanism–outcomes and whether the quality and patient safety surveillance systems involved any phases or levels of these systems involving data collection, analysis, evaluation and system operation. Full-text articles were considered relevant if they included contextual details, mechanisms, strategies, processes or discussions of outcomes related to quality and patient safety surveillance systems, including their use as a model for signalling potential areas of concern or excellence. Papers focusing solely on data recording methods or on technical recommendations for recording systems were excluded. Papers in which the quality and patient safety surveillance system improvements corresponded to the types described in this review were also included. Studies focusing solely on vaccination-based or medical device surveillance systems were excluded.

The richness and rigour assessment of the 69 included studies: Richness (scored 0–4) reflected the extent to which studies detailed how the intervention was expected to work and described relevant contextual influences, while rigour (scored 0–2) assessed the trustworthiness of the study’s inferences and their coherence with the underlying theoretical framework.^{34 35} (A summary of these assessments is provided in online supplemental appendix 3.) A third reviewer (EM) checked the relevance, richness or rigour where uncertainties emerged.

All relevant studies and documents were reviewed to determine which PT they addressed and what additional evidence they presented to support or refute these theories. Using the completed extraction form, BE created context–mechanism–outcome configurations (CMOCs). These CMOCs were carefully documented, revised and amended as the testing strategy was further clarified. Neighbouring or rival CMOCs were grouped under the relevant IPTs until a final model was constructed, illustrating the potential pathways of the theories. The reviewer recorded which particular study addressed which theory or theories in Covidence to guide the iterative data extraction throughout the process.³⁶ The information used to draft the CMOCs for each study, along with an example of the CMOCs from the first IPT, can be found in online supplemental appendix 4.

Step 3: Analysing and synthesising the evidence

The IPTs guided the analytical process of identification of the CMOCs through iterative comparison. Integrating CMOCs was difficult because of the differences in the contexts, the operation level and the characteristics of quality and patient safety surveillance systems. Thus, initially, CMOCs were integrated under their related IPTs by focusing on their mechanism-resource (M1), as comparability of resource mechanisms proved easier. Other reasons for focusing on M1 as the first step in integrating CMOCs include the clarity of M1 in the included papers and its importance in addressing the main question for programme developers and the core team. In the following phase, the CMOCs were carefully assessed by giving equal importance to context (C), mechanism-resource (M1), mechanism-response (M2) and outcome (O). Abductive and retroductive analytical processes were used to determine how they support, challenge or refine the IPTs. The CMOCs were synthesised by using the following methods: juxtaposition—combining evidence fragments to propose a connection between them, reconciling—making two or more conflicting evidence fragments consistent, adjudication—making a judgement about the methodological quality and accounting for it with evidence, consolidation—bringing together into a more coherent whole, and situating a piece of evidence in a context (s) with other evidence fragments.³⁶ If the IPT1 had four CMOCs contributing to its refinement, support or rejection, the first row in the table represented IPT1, and the CMOCs are labelled 1.1, 1.2, 1.3 and 1.4. After the synthesis phase, if IPT1 had one final PT, the last row was PT 1. All rows between IPT1 and the final PT represented the CMOCs, illustrating the progress of the analysis.³⁷ (See online supplemental appendix 4.) In the process, insights from systems theory were employed to categorise the explanatory patterns that emerged across the studies at various levels. BE and EM reviewed all CMOCs and any discrepancies were resolved through consultation and discussion of the paper from which they were extracted.

Patient and public involvement

This study forms part of a larger research project on the development and evaluation of a quality and patient safety surveillance system for the Irish health service. Within this broader project, a public and patient involvement (PPI) group was established as part of a parallel workstream led by a separate research team, contributing to the selection and prioritisation of measures and indicators to be included in the system.^{38 39} However, PPI was not incorporated in this realist review (a decision taken owing to the technical nature of the realist methodology and the knowledge of the system that is required to interpret and refine programme theories). PPI will play a central role in the next phase of this rapid realist review, which is the development of the quality and patient safety surveillance system evaluation tool. In line with WHO Quality and Safety guidance,¹ the final selection of items will be made in consultation with the PPI group.

Results

A total of 639 potential records were retrieved, 607 remaining after the removal of the duplicates. 176 records were selected after the title and abstract screening stage, and 69 records were screened on full text. Finally, 19 papers that met richness and rigour criteria were included for data extraction (see figure 2).

Figure 2. PRISMA flow diagram. PRISMA, Preferred Reporting Items for Systematic Reviews and Meta-Analyses.

Articles included for context–mechanism–outcome extraction

A total of 47 context–mechanism–outcomes were extracted from 19 papers89 11 12 14 17 21 40,⁵⁰ that we rated as high for richness and rigour as these studies provided sufficient detail to inform theory development on quality and patient safety surveillance systems. From these 47 context–mechanism–outcomes, 22 demi-regularities were identified, which supported the refinement of the 6 initial programme theories into 11 final programme theories after consultative meetings with the expert panel. These 11 programme theories are grouped into 8 domains: readiness for change; effective leadership; local governance model; the establishment of a clinical advisory group to identify safety indicators and measures; financial investment and resource availability; data governance; system design; and national governance model (online supplemental table 1).

PT 1: Readiness for change

This theory explains that a positive learning culture in the organisation and national accountability mechanisms (C: Context) were central to fostering clinicians’ engagement with the quality and patient safety surveillance systems (M1: Mechanism-Resource). The commitment to improve safety and sense of responsibility was an important response to national accountability mechanisms (M2: Mechanism-Response), which resulted in increased information sharing, knowledge of trends related to patient incidents, organisational response to signals, and a positive reputation (O: Outcome).⁴⁰,⁴³

PT 2: Effective leadership

According to the PT2, active and focused organisational leadership is the mechanism central to fostering the clinical staff’s ownership of the quality and patient safety surveillance systems. This PT suggests that when a strong organisational follow-up and feedback mechanism is in place, supported by a national quality surveillance and improvement model that systematically monitors healthcare services across organisational, regional and national levels (C), meaningful system-level change becomes more achievable. Under these circumstances, if the senior administration assigns clear accountability to hospital leaders who actively and visibly prioritise safety, allocate resources, distribute responsibility across staff, communicate expectations transparently and monitor progress against defined targets in alignment with regional and national governance structures (M1), staff are more likely to experience a stronger sense of ownership, support and responsibility for follow-up actions (M2). As a result, a reliable and responsive system is maintained, characterised by complete organisational information flow, systematic closure of safety issues and continuous multilevel oversight that integrates quality improvement efforts into the national strategy (O).^{40 44}

PT 3: Local governance model

The PT3 suggests that the integration of a national patient safety surveillance system into routine job roles, supported by strong institutional capability and collaborative and robust organisational evaluation processes, creates the context for meaningful use of safety data (C). When safety information is rapidly shared, benchmarked, triangulated with other datasets and published at a local level (M1), staff are more likely to develop a sense of ownership, feel practically supported in monitoring safety issues and gain confidence in the organisation’s commitment to patient safety (M2). In such contexts, organisations move towards proactive use of information, anticipating risks, identifying priority areas early and initiating timely improvement actions, strengthening continuous learning (O). The context where patients and families can engage with the organisation’s governance body and clinical teams (C) enables effective engagement with the patients and their families in safety improvement (M1). When patients who experience harm and their families are actively involved in understanding incidents, discussing contributory circumstances, receiving transparent information about preventive actions and engaging with safety data (M1), this allows shared learning and professional motivation (M2). In such contexts, collaboration with patients and families drives system-wide safety and quality improvement, guides the development of effective error-prevention methods and contributes to transformative change in organisational leadership, culture and priorities (O).^{11 40 44 45}

PT 4: Establishment of a clinical advisory group

According to PT4, in the context where quality and patient safety surveillance system aims to examine the heterogeneous data sources (C), if a multidisciplinary quality and safety advisory group establishes a shared terminology, unified classification framework and common coding rules (M1), this can enhance interoperability and trust in safety data, promote collaborative problem-solving and facilitate early identification of inequities and system-level risks (M2). As a result, accurate comparisons can be made across measurement approaches and institutions over time, enabling the identification of risk signals, identifying appropriate precautions, the monitoring of intervention impact and strengthened accountability in protecting vulnerable populations and reducing disparities in care (O).^{17 21 46 47}

PT 5: Financial investment and resource availability

The PT5 explains that, in contexts where there are improvements in healthcare quality and safety surveillance through forensic big data reviews, and real/near-real-time surveillance systems are being implemented (C), if organisational leaders strategically invest in staff, IT resources, coding expertise and training for both clinicians and dedicated data teams (M1), staff can gain the support and capability to interpret data accurately, validate alerts and address data-related issues (M2). As a result, organisations complete the full cycle of problem identification, analysis, solution identification and implementation more effectively. Electronic health record queries function more reliably, false signals are minimised, high-risk patients are prioritised without increasing the burden on clinical staff and comprehensive evaluations are conducted (O).812 14 17 40 45 48,⁵⁰

In the context of low- and middle-income countries, where resources are limited and reliable population-based data are scarce (C), effective allocation of resources through system design, engagement with front line services and policymakers, and improvement of staff surveillance skills and securing long-term financial support (M1) enables programmes to overcome implementation barriers (M2). As a result, reporting systems provide reliable data and enable information sharing to assess strategies for reducing mortality and morbidity (O).812 14 17 40 45 48,⁵⁰

PT 6: Data governance

According to PT6, in contexts where hospitals, health systems and electronic health record vendors adopt information technologies and implement automated patient harm identification systems, and strict information governance policies are in place to ensure confidentiality, data protection and legality, (C) if patient safety data are anonymised, shared under clear agreements and analysed by an independent agency, this creates a legal and safe learning environment for collecting, analysing and sharing patient safety data with hospitals, clinical teams and patients, free from retribution, conflicts of interest or privacy and security concerns (M2). It also promotes public accountability, supports monitoring of patient protection measures and encourages participating authorities to access the patient safety database and establish collaborative relationships with the Central Office (M2). As a result, patient safety operations are planned and conducted in the best interests of patients and the public, key learnings from adverse events are disseminated, safety culture is promoted, electronic health record queries are reliable and monitoring of patient protection measures is enhanced (O).^{11 21 45}

PT 7: System design

PT7 suggests that in contexts where there is a national patient safety surveillance system providing clinical data abstracted from medical records with specific hospital inpatient safety measure rates (C), if the system uses information technology and statistical analytic techniques to analyse passive data (M1), then it enables reproducible and scalable error flagging by incorporating structured electronic health record data and maximising the use of routinely collected information to identify patterns and trends that might not be visible within a single organisation (M2). As a result, national-level patient safety data are generated, emerging trends are captured, and clinically relevant safety information is delivered to clinicians, ensuring a comprehensive approach to patient safety (O).^{9 12 16 17 47 48}

In contexts where there is a safe learning space for improving safety practices through an automated active hospital surveillance system of electronic health records (C), if the system applies algorithms to normalised data to generate signals of potential adverse events and develops predictive safety models validated through standardised chart reviews by trained nurse and physician reviewers (M1), then it enables a dynamic, real-time view of all-cause harm, predicts future adverse events by identifying patterns and risk factors in patient data and informs decisions to prevent harm to at-risk patients (M2). As a result, adverse events are detected more efficiently and systematically than manual chart reviews, allowing immediate intervention while patients are still admitted, supporting targeted quality and safety efforts, enabling effective planning to reduce future harm, healthcare costs and burdens and promoting a learning culture through trend analyses that inform hospital policy revisions (O).^{9 12 16 17 48}

PT 8: National governance model

According to this theory (PT8), in contexts where a quality surveillance and improvement model monitors healthcare services consistently at organisational, regional and national levels (C), if a clinical safety surveillance group is established to strengthen safety surveillance through government-level collaboration, engage policymakers to create actionable policies and foster a systems approach to safety culture with clear organisational roles (M1), then effective coordination, implementation, follow-up and information sharing on safety needs can occur at both national and local levels, with political commitment to continuous safety improvement, and organisations are supported in fulfilling their duties (M2). As a result, intelligence is shared and gathered from member organisations, thresholds and methods for identifying risks are agreed on, organisational and local concerns are identified and escalated in a timely manner, appropriate actions are determined, themes requiring national policy responses are highlighted, support is allocated to suitable organisations and information is shared among international stakeholders (O).^{42 44}

Discussion

This rapid realist review describes the robust and iterative process of developing programme theories which are more developed than the initial programme theories identified in the first phase of the review, as substantial evidence provided a more nuanced understanding of the interplay between the contexts, mechanisms and outcomes and between programme theories. The Systems Theory served as a guide to understanding what key aspects of quality and patient safety surveillance systems need to be established, how different factors can be effectively nurtured and coordinated to achieve desired outcomes and at which levels of the healthcare system these processes contribute to achieving desired outcomes.²⁹ The refined programme theories show that the components of surveillance systems such as leadership, governance and system design operate as interdependent context–mechanism–outcome configurations across system levels, where outcomes cannot be attributed to any single component in isolation. The findings suggest that programme theories are more likely to produce successful systems if they are conceptualised as an integrated package of interacting components and need to be tested across diverse healthcare contexts.

Readiness for change

This configuration aligns with the concept of safety culture, which refers to the shared values, beliefs and norms within healthcare organisations that influence safety-related behaviours and practices aimed at minimising harm.⁵¹ Safety culture dimensions, such as organisational learning and continuous improvement, have been identified as strong predictors of patient safety outcomes in the literature.^{52 53} Drawing on the framework proposed by the WHO (2020), understanding the social system’s capacity or resistance to change is critical in assessing whether safety improvements will be embraced and sustained.⁴⁵ Findings from our review configurations, supported by the existing theoretical literature, suggest that an organisational learning culture where health professionals can express their safety concerns and suggestions and a national context where accountability mechanisms are established emerge as core infrastructural factors that future quality and patient safety surveillance systems designers should consider. Similar insights have been reported in other patient safety initiatives, such as hospital fall prevention programmes, which demonstrate that positive workplace culture, clearly defined roles, supervisor support and ongoing training are essential for successful implementation and improved safety outcomes.⁵⁴

Effective leadership

In line with our results, evidence from the literature indicates that visible and strong leadership is critical in improving organisational safety culture since it encourages staff commitment.^{40 52 55 56} These leadership styles reflect effective leadership attributes, which are essential in fostering a positive safety climate. Key features of this leadership style include strong safety management practices, clearly defined roles and consistent follow-up on health staff involvement, directly influencing employees’ participation in safety efforts.^{54 57} Accordingly, in quality and patient safety surveillance systems, organisational leaders’ sponsorship should be ensured to make people accountable by providing clear and strategic direction and closing the safety loop in the follow-up action. The leadership component of the effective systems described by PT 2 is one of the mechanisms of the PT 8, and context for the PT 3, demonstrating the layered and complex nature of quality and safety surveillance systems through the interdependencies among the programme theories. By explaining how context and mechanisms interact to generate outcomes, this review provides transferable insights, emphasising that national quality and patient safety institutes/boards, as system enablers, require policies to enact their leadership capacities in governing networks across system levels and organisational hierarchies.²⁹

Local governance model

The PT 3 on Local Governance Model explains the success of the quality and patient safety surveillance systems from a sociotechnical perspective. This approach emphasises fulfilling end-user local priorities as they are central actors in shaping the direction of quality improvements.^{29 42 45} Moreover, developing an organisational structure by engaging patients and families is an essential mechanism in this review, which allows collaborative learning and motivates healthcare professionals to reduce the likelihood of a recurrence. WHO’s Global Patient Safety Action Plan 2021–2030 considers patient and family engagement as one of the seven actions and recommends engaging patients and families at all levels of healthcare to make healthcare services safer.¹ Thus, close coordination among all policymakers, leaders, healthcare providers and patients and their families to determine the effective and appropriate priorities in quality and safety improvement practice is essential.⁵⁸

Establishment of a clinical advisory group

The involvement of a clinical advisory group (comprising the relevant healthcare professionals) in the development of metrics has been found to be critically important in the literature and is also supported by our PT 4, as it enhances data interoperability across various patient safety measurement approaches and promotes clinical alignment.⁵⁹ Parallel evidence has been highlighted that defining which healthcare-associated infections data to collect requires input from clinical experts to ensure clinical usefulness, feasibility and alignment with infection prevention priorities.⁴ Allowing patient safety specialists to compare information from different measurement methods using a common language facilitates a consistent measurement approach derived from innovative electronic documentation systems.⁴⁷ This mechanism is supported by the WHO’s (2010) Framework for Action on Interprofessional Education and Collaborative Practice and Confetti’s (2022) detailed review that emphasises the importance of interdisciplinary collaboration in driving innovative and integrated safety strategies within healthcare systems.^{60 61} These findings highlight the need for active participation of care professionals in the metric development of quality and patient safety surveillance systems.⁶¹

Financial investment and resource availability

Investing in capabilities such as IT resources and staff is central to improving data quality and signal review at the organisational and national levels. Classen et al report that the successful application of such systems requires dedicated resource allocation and early evidence of cost reduction to integrate their existing workflow without adding burden to staff.¹² Consistent with Sociotechnical Systems Theory, the PT 5 suggests that funding for innovative technology in quality and patient safety surveillance systems is necessary but is not sufficient in its own. System improvements need to be accompanied by an adaptation strategy that invests in creating an innovation structure, accountability mechanisms, performance assessment and skills for the substantial uptake of innovations.²⁹

Although the included evidence reflects several different healthcare systems, low-income countries remain under-represented, which limits the explanatory power for these settings. The successful operation of such systems in these contexts is likely to be more challenging due to variable investment in information systems, technology and workforce expertise. The literature emphasises the importance of staff training, data quality improvement and highlights significant regional disparities in the implementation of surveillance systems.^{49 50} Further research is needed to explore these system-level and contextual determinants to inform more responsive and context-sensitive governance structures. In line with these results, future realist evaluations are needed to test and refine programme theories within these resource-constrained settings to better understand how context shapes implementation and outcomes.¹

Data governance

Engaging experts and stakeholders is a core element of realist reviews.^{24 30} In this study, a diverse group of experts including data scientists and IT experts was actively consulted at multiple stages of the review process. This active engagement of experts is a major strength of the realist review, enabling the provision of policy-relevant insights on data governance and system design, at a level of abstraction that can be transferred across settings.²²

Effective quality and patient safety surveillance systems require secure data management and governance policies to ensure patient safety and legal compliance.⁴⁵ In the WHO’s Action Plan, strengthening data-sharing channels between patient safety information sources is suggested for governments in establishing patient safety information systems.¹ The European Health Data Space Regulation establishes a harmonised and secure framework for accessing and reusing health data across the EU, laying the foundation for robust data governance that supports quality improvement and enhances patient safety systems through interoperability, transparency and trust.⁶² US patient safety organisations established by The Patient Safety Act offer a legally secure environment for analysing broader trends in care quality and safety data and providing feedback to organisations.¹² An interactive, safe environment where all users share potentially sensitive safety data, ensuring freedom from discovery, is an essential prerequisite to use, learn and bolster a well-informed safety performance.⁶³

System design

The interoperability of multiple levels of electronic health records is also essential in capturing emerging national trends to guide ongoing efforts to improve patient safety. The National Healthcare Safety Network, a US national-level patient safety surveillance system, provides actionable data on healthcare-associated infections.² These data demonstrated a substantial impact on hospital-acquired condition prevention, leading to 20 700 fewer deaths between 2014 and 2017.⁶⁴ Centralising health IT signals into a unified system is a foundational step for quality and patient safety surveillance systems, enabling stakeholders across the healthcare system to access data and drive improvement efforts. The Institute of Medicine emphasises the importance of reliable all-cause harm measurement systems to supplement or potentially replace manual record review, thereby facilitating timely clinician interventions to prevent or mitigate potential harm.⁶³ Similarly, recent surveillance system developments, including fast healthcare interoperability resources-enabled digital quality measures demonstrate how automated, patient-level and interoperable data capture can enhance real-time surveillance, improve data accuracy and support measurable patient safety improvements.^{2 12} On the other hand, the limited availability of outcome data from these recent digital surveillance system improvements in the literature may have constrained the robust causal inference between mechanisms and outcomes.

National governance model

Drawing on PT 8, outcomes such as effective coordination, shared intelligence and timely escalation of risks emerge from the interaction between contextual components and underlying mechanisms and their collective activation across system levels. While strong commitment from leadership and trusted champions is essential for effective operation of the surveillance systems, these represent only two elements of a broader national governance framework.^{29 65} Evidence from the literature indicates that system-level barriers, such as fragmented commissioning structures, misaligned incentives and policy–practice gaps, can limit alignment between policy intent and clinical practice, highlighting the critical role of national governance in coordinating and adapting incentives and facilitating knowledge flow across system levels.^{66 67} Further research is needed to explore these system-level governmental determinants, to inform more effective national governance structures in which leadership and trusted champions are understood as part of a broader portfolio, to improve quality and patient safety outcomes.

Conclusion

Quality and patient safety surveillance systems are most effective when supported by strong leadership, positive safety culture, clear governance structures, adequate resources and active engagement of stakeholders at local, organisational and national levels. Key mechanisms including clinical advisory groups, patient and family involvement, structured local and national governance, data interoperability and robust data governance enable coordinated action, continuous learning and evidence-based improvements in patient safety. Future research and implementation efforts should focus on developing predictive, real-time and digitally integrated quality and patient safety surveillance systems that are adaptable across diverse healthcare contexts, while ensuring engagement, ethical and legal accountability and sustainability at all system levels.

Data availability statement

Data sharing not applicable as no datasets generated and/or analysed for this study.