Production Systems and Unintended Outcomes

Introduction to the annual report

Achieving high quality from production systems is a global challenge, faced by every industry. According to the systems theory, the inputs will be transformed by the systems to generate final outputs (Mele et al., 2010; Von Bertallanfy, 1956). Every production system has an operational function that produce mix of services and products(Slack et al., 2013). The role of an optimal production system is to transform these inputs into desired outputs and minimise unintended outputs by working with the subsystems and elements. However, there is strong evidence that, complex organisations such as hospitals continue to produce unintended outcomes such as high level of prescription errors, planned care operation cancellations, infections, bed ulcers and longer hospital delays have been reported all over the world (Miller et al., 2007; Ryan et al., 2014; Stelfox et al., 2006; Wong et al., 2018). By utilising an extensive body of operation management literature, theories, concepts and principals, this study aims to explore the factors that influence the production of the unintended outcomes in a complex organisational setting. The thesis has chosen tertiary NHS hospitals in England, as the broader context for research with the focus interest in elective care and identified. On The Day Surgery (OTDS) cancellations as the undesired outcome, based on the theoretical foundation of Variability Methodology (VM) suggests that minimising Artificial Variation (AV) can reduce undesired outcomes, produced by organisations. This thesis aims to test the association and causal links between AV and undesired outcomes. When a planned care surgery is cancelled on the day of patient admission or the day of the surgery, it is defined as an OTDS cancellation. OTDS cancellations are the major and long-standing problem, faced by healthcare providers, policymakers, patients and healthcare professionals all over the world. OTDS creates substantial problems for the utilisation of theatres and clinical staff efficiency causing significant issues for healthcare providers (Schofield et al., 2005; Smith et al., 2018). The significant emotional impact, that the patients and families experience due to OTDS cancellations, has been emphasised by several researchers (Hovlid et al., 2013; Ivarsson et al., 2002; Lankoandé et al., 2017). Delays to their operations due to cancellation can be harmful to some patient cohorts (Argo et al., 2009; Gibney, 2018; Ivarsson et al., 2002). OTDS Cancellation Rates (CR) vary significantly, ranging from 1% to 23% across health care systems in Norway, the United States, Canada, New Zealand, the United Kingdom and Hong Kong (Yu et al.,2017).

Hospitals consist of an extensive number processes that are linked together (figure 1) by sub-systems and various working elements such as operating rooms, Critical care units and emergency department and many others (Barasa et al., 2017a). In a hospital, multiple healthcare professionals collaboratively deliver patient care for a patient with a wide range of clinical problems that need to be identified and treated accordingly. The complexity can be further augmented due to the hierarchies that exist between health professionals, the various epistemological perspectives of the different healthcare professional groups, shared resources and regulatory complications. The Complex Adaptive System (CAS) approach highlights the non-linear and interrelated nature of the processes and various components of healthcare delivery organisation, rather than limiting it to a cause and effect understanding. Some researchers Begun et al., (2003) and Kuziemsky (2016) describe healthcare organisation like hospitals as a CAS. The thesis aims to understand the overall research question: what factors related to variation influence ODTS cancellation in NHS hospitals. This report focuses on providing details of the following: overall research background, theoretical underpinning and application in all three studies that each consisting of clears objectives, research questions and methodology.

The report consisted of three sections:

Firstly, this report discusses the wider theories related to variation management, quality management approaches, and how various studies planning to study the complexity in organisations.

Secondly, explain theories that are used to explore in this thesis and theoretical application.

Finally, explain the scope of each study explaining the research questions, methodologies and theoretical contributions in each study.

In addition to the theoretical and methodological contributions, this study will be of great interest to NHS organisations, as OTDS cancellations `waste` millions of pounds, which are vital for all NHS organisations in the current economic climate.

The diagram is from the “Institute for Healthcare Improvement: Achieving Hospital-wide patient flow” report (Rutherford et al., 2017)

Operation management and managing variation

Operation management provide theories, concepts, techniques to design, analyse, and improve performance of the organisations and systems to deliver product and systems (Slack et al., 2013). Variations have been presented an opportunity for innovation, but managing variation has been a challenge to production systems. An extensive body of operation management literature has tried to address this challenge. Deming (1993) discussed the importance of managing variation as a principal component of managing production systems. Deming argues that the variation needs to manage holistically within the production system and introduce the concept of “system profound knowledge” encompassing four interrelated parts understanding of variation, theory of knowledge, psychology and appreciation for a system (Deming, 2018, 1993). Shewhart highlighted “assignable causes of variation” that originates from “special causes” may be found and eliminated (Shewhart, 1939, 1931). Using the concept of “statistical control” a process can achieve “stable and predictable” (page 6) variation (Shewhart, 1939, 1931). The concept of “statistical control” can be beneficial theoretical and practical contribution, but there is a reasonable possibility that the “statistical control” can still provide a high level of unwanted variation that leads to product and services out of the specification and with undesired outcomes. Deming argues that a process “that is not in statistical control has no definable capability: its performance is not predictable” (Deming, 1993, p. 99). Although “statistical control” provides predictable variation, there is no logical connection between control limits and specification limits (Deming. 1993). Taguchi provided a different approach to manage variation by optimising designs for performance, quality, and cost (Khan et al., 2014; Unal and Dean, 1990). Taguchi (1986) proposes that all deviations from the “target” result in a loss. The “target” is introduced as the aimed value of the quality characteristic and lies within the specification limits. The Taguchi argues that losses follow a quadratic pattern and small variations are probably of inconsequential importance and associated with slight losses. Deming (1993) promoted the idea of continuous improvement and argues that “the most important use of a loss function is to help us change from a world of specifications (meeting specifications) to a continual reduction of variation about the target, through the improvement of processes” (p. 217).

Managing variation

Several popular quality management approaches have to try to manage variation in production systems such as lean and six sigma. Six sigma (Harry and Schroeder, 1999; Magnusson et al., 2003) origins were reported in the 1980s within the improvement programs run by Motorola. In 1988 Motorola received a national quality award that leads to the popularity of six sigma among the business community (Pyzdek, 2001). Six Sigma’s core philosophy focuses mainly on reducing variability. The underlining assumption is that output variability is reduced by implementing a tightly controlled process and continuous improvement. Magnusson et al. (2003) reported that, six sigma uses a series of design, statistical, project, and customer and quality improvement tools. Some large companies like Volvo and Ericsson have reported significant multi-million savings based on the contribution from Six Sigma. Ability to gain senior management support and create cultural changes using highly trained Six Sigma practitioners has been reported as an essential success factors (Banuelas et al., 2002; Schroeder et al., 2007). Six sigma approach is likely to be successful when processes are well defined and relatively stable performance (Ertürk et al., 2016; Proudlove et al., 2008). The other criticisms for Six Sigma are the “stakeholder” aspects are poorly reported, fragmented to defined processes and functions, which can have implications for implementation (Proudlove et al., 2008). Lean is a widespread quality management concept that is extensively discussed by quality practitioners and scholars (Antony et al., 2019; Liker and Meier, 2006; Stone, 2012). The concept of lean focused on reducing variation and maintaining a fixed capacity, which leads to high utilisation of resources in the processes. Lean approaches rely on creating standardised and stable processes based on value-added activity in order to provide the best quality services or products as efficiently as possible. Lean philosophy is a customer driven and embraces continuous improvement approaches that support a simple and direct set of processes (Liker and Meier, 2006). Based on lean application, waste reduction and quality improvement in commercial production systems were widely reported (Dahlgaard‐Park et al., 2006). The main criticism against lean is that the approach lacks flexibility and less ability to deal with changing circumstance (Chay et al., 2015; Dahlgaard‐Park et al., 2006). Lean principles remain unchanged from manufacturing to services, but there is a clear need for adaption for the service industry, but what needs to be adapted is not clear (Gupta et al., 2016).

Several researchers argue that hospitals can use supply chain approaches to manage variation, and there is already lots of work completed in this area under supply chain management (Aronsson et al., 2011). Supply chain management applications are similar to the care pathway theory (Schrijvers et al., 2012). From the supply chain perspectives, the lean is used in high volume and low complexity setting. Maintaining a streamlined production process is a basis for a lean strategy, but in reality, this is difficult to achieve in big university hospitals (Borges et al., 2019). A study tries to implement in multiple hospitals concluded that “for less complex care processes (ENT and gynaecology), a large and sustained improvement was mainly the result of a better match between capacity and demand. For medicine, surgery, and paediatrics, which exhibit greater care process complexity, sustainable, or continual improvement were constrained because the changes implemented were insufficient in addressing the higher degree of complexity” (Mazzocato et al., 2014). A comprehensive review of lean applications in healthcare concluded, it is difficult to understand the positive impact due to lean (D’Andreamatteo et al., 2015). A systematic review explores the surgical processes found the potential in both six sigma and lean approaches but reported the need for high-quality studies with low risks of systematic bias to understand the positive effects both approaches (Mason et al., 2015). Radnor and colleagues argued that, the inherent dynamics presented in the healthcare sector provide significant barriers to implementing lean approaches(Radnor et al., 2012). Radnor and Osborne discussed that, lean can only achieve its potential in public services based on a dominant business model(Radnor and Osborne, 2013). There are significant differences in the healthcare delivery and production systems, which provides several limitations in the successful implementation of approaches like lean and six sigma. Although Lean and six sigma approaches have some success in managing variation in several commercial production systems, the effectiveness of these approaches to complex systems like in healthcare is inconclusive.

Studying variation in complex organisations

The general systems theory by Von Bertalaffy (1956) defined a system as a complex interaction of interactive elements. The main concepts of general system theory include the following: Subsystems or holisms, Open system view, input- transformed –output –model, system boundaries, negative entropy, feedback, hierarchy (Mele et al., 2010). Based on system theory, many authors have to define complexity as the interrelatedness of components in the system (Kannampallil et al., 2011; Simon, 1991, 1973). From broader social science perspectives, the key concepts of a CAS generally include embeddedness, co-evaluation, distributed control, self-organisation, emergence, unpredictability, non-linearity, phase changes, historicism, sensitivity to initial conditions, non-equilibrium, adaptation, nested systems, fuzzy boundaries (Kernick, 2006; Kuhn, 2009; Long et al., 2018; Manson, 2001; Plsek and Greenhalgh, 2001). Complexities in the system create difficulties in “computability” of various components in the system such as cognitive functions and behaviours, physical resources required or expanded (Kannampallil et al., 2011). Due to the nested systems, emergence and fuzzy boundaries can provide the number of methodological challenges to the research organisations that demonstrate CAS behaviours. CAS provides a fascinating descriptions of the systems, but theoretical applications can be challenging (McDaniel et al., 2009). This thesis aims to manage the challenges presented in studying a complex system like a hospital in several ways. Instead of attempting to control variables in a study, complexity theory directs the researchers to study the patterns of interactions and mechanisms within agents, and between agents that lead to developing outcomes (Anderson et al., 2005; Long et al., 2018). This thesis used the realist synthesis method to understand mechanisms and contexts that reduce OTDS cancellations in a hospital complex setting. Realist synthesis is designed to embrace the real social and organisational phenomena which have a very complex nature (Pawson et al., 2005). The realist methodology was chosen, because it is designed to manage the wide-ranging evidence from various sources and that can be used to understand various complexities surrounding surgery cancellations. The principals for realist synthesis and evaluation are well-documented (Greenhalgh et al., 2015; Wong et al., 2013) and are now widely accepted by reputable peer-review journals and research funding bodies (MacDonald et al., 2016). Another approach to study complex organisation is to use the concept of decomposition, which means wherever possible into smaller functional components and relationships between them (Kannampallil et al., 2011). The concept of decomposition is used by researchers (Ariyo et al., 2008; Groen and Patel, 1988) in many fields, and this thesis will use this concept in the research design of the empirical studies.

Some methodologist suggest that the complexity in the systems can be studied by using mixed methods research design, case studies (Litaker et al., 2006; Long et al., 2018; Walton, 2014;) Anderson et al., 2005; Barasa et al., 2017b)

This thesis aims to study complexity using realist methodologies, applying the concepts of decomposition and case study research design.

Variability Methodology and theoretical foundation

The variation in a healthcare organisation can be generated in many different levels: clinical, professional and flow, which can increase the complexity and adds cost to the healthcare system (Litvak and Long, 2000). In order to address the challenging tasks of managing variation, the IHO (Institute of Healthcare Optimisation) developed the Variability Methodology (VM) (Litvak and Joint Commission Resources, 2010). The VM identified two types of variation: natural and artificial variation (AV). The natural variation defined as the variation that has no control and AV, which can be controlled by healthcare design and by organising care (Litvak and Long, 2000). The AV can be presented due to the complication of scheduling practices, inappropriate of patient management and variation in care management of patients (Litvak and Joint Commission Resources, 2010). For example, two patients within similar diseases and recovery may spend different Length of Stay (LOS) due to AV. The underlining theoretical assumption for VM is by minimising AV that will lead to provide an optimal level of care and reduce undesired outcomes such as improve patient safety, staff satisfaction and income (Litvak and Joint Commission Resources, 2010). High elective scheduling variation can increase AV (Litvak and Long, 2000; McManus et al., 2003). The VM tries to identify, quantify and minimise the AV to optimise the hospital performance. In the peer-reviewed literature, application of the VM is completed in single centre studies in US hospitals and reported positive improvements such as nursing stress and patient safety (Litvak et al., 2005), net operating income and margin by 38% and 28% (Litvak et al., 2005; Litvak and Joint Commission Resources, 2010; McManus et al., 2003; Smith et al., 2013). This thesis aims to answer the overall research question: In a complex hospital setting, What factors influence ODTS cancellations? In order to achieve the research aim, the research enquiry has been broken down into three studies that each study is consisting of a clear objective, research question and methodology. The three research studies have clear links and provide valuable contributions to the main research question. Each study forms a chapter in this thesis. In thesis, it aims to explore the theoretical assumption of VM and find out the association and causal links between the AV and OTDS cancellations in the three studies.

The main three chapters as follows:

Thesis chapter 2- A realist synthesis: what contexts and mechanisms influence OTDS cancellations? (Described in section 3 and document B).

In this chapter, the mechanisms and contexts aim to reduce OTDS cancellations will be explored using literature, relevant documents and interviews with relevant stakeholders. The completed work in realist synthesis has identified nine mechanisms and several contexts, which is considered as “theories that explain” OTDS cancellations. These “theories” further developed into a middle range theory that explains the OTDS cancellations and applied into other complex systems.

Thesis chapter 3 - A multi-centred retrospective study to understand the variations in OTDS cancellation rates among NHS acute care providers (Described in section 4 and document C).

This chapter aims to covers number of exploratory and deductive research using the empirical data collected in five large NHS acute care trusts. In addition to the exploratory objectives, this chapter aims to apply 4Vs concept(Slack et al., 2013) to gain overview of the how organisational process work. As the final objective, the association between the AV due to elective care scheduling variation that exists in the hospital practices and OTDS cancellation rates due to bed unavailability will be explored.

Thesis chapter 4 - A multi-centred clinical speciality case study to identify the factors influencing OTDS cancellation rates (described in section 5).

By using the case study research design aims to understand what ways AV in elective care scheduling influence cancellation rates due to post-operative unavailability.

In the next section, the scope of the area of the research interest is explored by explaining consisted of national trends, the scope of the study and explaining the three identified three studies in detail.

Cancellation statistics and trends in the English NHS

There are a large number of retrospective descriptive studies with the limited scope of a single hospital or speciality, and several multi-centred (Schuster et al., 2011; Seim et al., 2009) studies also have been published in the peer-reviewed literature. Among the identified literature, no systematic, qualitative synthesis or realist synthesis was found. The literature searches completed suggest that there is a very little understanding of theories of surgery cancellations, and the driving factors for surgery cancellations. In 2017/18, NHS England recorded that 84,827 operations were cancelled on the day of the surgery across all NHS acute care providers in England, due to various reasons, which is the highest recorded number ever. Based on figures published by NHS England (National Health Services, 2015) in 2017/18 (n=84,827) there was a 24% increase in OTDS cancellations compared to 2013/14 (n=64,195). Between 2012/13 and 2017/18 in England, OTDS cancellation rates in England increased from 0.80% (2012/13) to 1.08% (2017/18). The continuing high number of hospital OTDS surgery cancellations shows that, there is a significant waste of resources created in the health system. The increase in OTDS cancellations could be a symptom of broader inefficiencies in the elective care system. There is a significant variation in the OTDS cancellation rates among English NHS trusts (McIntosh at al., 2012). The study analysed the OTDS hospital cancellation rates in English NHS trusts (n=107) from 2011/12 to 2016/17 and acute care trusts with less than 3000 operations per month were excluded from this stage due to low activity. Furthermore, truststhat only performed operations on children or women were also excluded due to their specialised case-mix (Proudlove et al., 2017). The analysis, which involved integrating data from 2011/12 to 2016/17, found that, the trusts in the top and bottom deciles had relatively consistent patterns of outstanding and poor cancellation rates (figure 1-1). During this six-year period, the NHS trusts currently in the best decile had reasonably consistently low hospital OTDS cancellation rates [average 0.50 % with a range of 0.22% to 0.91%] while those currently in the worst decile had reasonably consistently high rates [average 1.49% with a range of .59% to 3.58%). This consistency in performance provides an opportunity to investigate what the “high” performing trusts are consistently doing differently from the “poorly” performing trusts in the same healthcare context. The “high” and “poorly” identified trusts will be used in the work package 2 (section 4) to recruit NHS trusts.

Scope of the study

In England, elective care accounts for 18% of providers’ total annual expenditure, rising to 30% if outpatient spending is included (Monitor, 2015). Elective care operations cover all age groups and includehigh-risk patient groups like cancer and cardiology patients. This research study focuses on surgery cancellation that happened on the day of admission or the day of surgery (Figure 2-1). Apart from OTDS cancellations, surgery cancellations may take place at any point between the times the patient is added to the waiting list, up to the day of the operation. A minority of patients require admission the day before an operation; these cases are also captured in the scope of this study - means that sometimes patients are given the bad news about cancellations in an admission ward/unit while at other times they are told before arriving at the hospital. Some studies have mentioned cancellation decisions being made in the operating theatre (Sung et al., 2010).

A realist synthesis: what contexts and mechanisms influence OTDS cancellations?

Background

Despite the significant challenges that surgery cancellations bring to health providers, professionals and patients, the literature searches failed to identify any existing published systematic reviews, qualitative synthesis or realist synthesis that explain what strategies or factors influence the reduction of OTDS cancellations.

The rationale for Realist synthesis

Realist synthesis is designed to embrace the real social and organisational phenomena which have a very complex nature (Pawson et al., 2005). The realist methodology has the ability to understand various complexities surrounding surgery cancellations. A realist synthesis was preferred ahead of a more traditional systematic review because it is particularly well suited to the assessment of complex problem using a mixed body of evidence. Realist synthesis is explicitly theory-driven, recognising that it may be more fruitful to consider underlying programme theories about how and why particular strategies or mechanisms work in comparison what works or not. The lack of theoretically-informed work related to OTDS cancellation in hospital setting provided a further rationale for choosing a realist approach. The principals for realist synthesis and evaluation are well-documented (Greenhalgh et al., 2015; Wong et al., 2013) and are now widely accepted by reputable peer-review journals and research funding bodies (MacDonald et al., 2016).

Research questions

Given the limitations to the current knowledge and by applying a realist synthesis approach, this review attempts to address the following objectives and research questions. The objective of this review is to describe how the organisational strategies can reduce OTDS cancellations in different organisational contexts and develop practical recommendations for knowledge users. The study aims to answer the following research question: What organisational strategies, mechanisms and contexts reduce OTDS cancellations?

Theoretical and methodological contribution

The completed work in realist synthesis has identified nine mechanisms and several contexts, which is considered as “theories that explain” OTDS cancellations. There are clear patterns and links among these mechanisms. Using the realist concept of “retroduction” and the theoretical foundation from VM, this study aims to develop a middle-range theory based on identified mechanisms. The middle-range theory aims to explain how to reduce cancellations but can be applied to other complex systems. Famous sociology scholar, Merton (Merton, 1967)define the middle range theory as: “sufficiently abstract to deal with different spheres of social behaviour and social structure so that they transcend sheer description or empirical generalisation” (page, 68). Middle range theory development is used widely in the healthcare field (Gillespie and Marshall, 2015; Jagosh et al., 2012; Shearn et al., 2017) and is considered a significant theoretical contribution. In the healthcare setting, realist methodologies are widely used to evaluate complex interventions and to understand how various mechanisms influence outcomes. Researching interactions between various agents and identifying mechanisms in a system that impact on operation management capabilities is vital to understand throughput and other undesired outcomes but rarely research in an operational management setting. By applying realist methodologies into a establish operation management problem in a complex organisational setting and by developing middle-range theory, this study also aims to provide a useful methodological contribution to the operational management field.

A multi-centred retrospective study to understand the variations in OTDS cancellation rates.

Background

Currently there is limited literature on day of surgery cancellations in a wide variety of clinical specialities (Smith et al., 2018) and, in the absence of any multiple sites, hospital or trust studies in the literature in the NHS context, this exploratory section aims to provide a more in-depth understanding using empirical evidence from NHS acute care trusts. There is strong published evidence that, organisational contexts, such as a volume of elective surgery completed, influence CRs (McIntosh et al., 2012; Schuster et al., 2011). Therefore, in order to limit variation due to the number of patients treated the study focus on NHS acute care trusts who delivers more than 70,000 elective procedures. Using the exploratory analysis, this sectionaims to identify the similarities and differences among the five NHS trusts’ surgery cancellation rates across clinical specialities.

Theoretical applications

The topology of operations and processes - 4Vs application

In line with the system theory (Von Bertallanfy, 1956) all the systems have few things in common; they all take their ‘inputs’ like, raw materials, knowledge, capital, equipment and time and trying to transform them into desired outputs. Based on four V’s (Volume, Variety, Variation and Visibility) can be used to gain an overview of how operational processes work in an organisation or system (Slack et al., 2013). The concept of 4Vs is used in operation management to compare different commercial organisations in the same industry. Using the empirical data collected from five different NHS trusts, this study aims to apply 4Vs concept into hospital sites and main surgical clinical specialities to understand how overall OTDS cancellations rates and cancellation due to post-operative beds have any relationships. Each dimension in 4Vs can be calculated using the patient-level data collected from five different NHS trusts. The comparative analysis between the hospital sites and clinical specialities can provide details to compare various hospital sites and OTDS cancellation rates. This application (4Vs) provides the opportunity to look at hospital sites and clinical speciality abstractly, allowing various theoretical applications. We could apply the 4Vs to each hospital site and clinical speciality in each site while nesting them in an organisation and compare with OTDS cancellation rates. The comparative analysis between 4Vs and OTDS cancellation rates could help extend the concept of 4Vs concept further. For example, high scores of all four or some dimensions of 4Vs could lead to high complexity, which may have an association with high OTDS cancellation rates.

Variability Methodology

VM argues that minimising AV can reduce undesired outcomes of hospitals such as OTDS cancellations (Litvak and Long, 2000; Smith et al., 2013). Several single-centre studies implemented VM managed to demonstrate reduction of AV can reduce undesired outcomes in hospitals while increasing the desired outcomes (Litvak et al., 2005; Smith et al., 2013) – this suggests an association between AV and reduction of undesired outcomes. By calculating AV based on variation in scheduling and using OTDS cancellations rates, this study aims to understand the association between AV and undesired outcomes. The completed data collection by five NHS trusts and 15 hospital sites provide an excellent opportunity to test the association between the elective scheduling variations represents by AV and overall OTDS cancellation rates and different types of OTDS cancellations. Based on the results from the previous studies suggests to find positive correlation between AV and OTDS cancellation rates due to post-operative beds (McManus et al., 2003; Smith et al., 2013) but there are no empirical studies completed to understand the existing in AV due to scheduling variation and OTDS cancellations rates due to post-operative beds. The research study focuses on three different types of adult post-operative beds used in NHS hospitals: day care, inpatient and critical care beds. A measurement of variation will be calculated for each post-operative bed type, and the OTDS cancellation due to beach post-operative bed type is correlated in all 15-hospital sites (figure 3-2).

Methodology – multiple case study

Case study research design can provide a useful pragmatic epistemology that can support, phenomena acknowledging the “holistic view” of a complex system (Anderson et al., 2005) and help to facilitate the explorations of casual mechanisms between areas of interests (Gerring, 2017). OTDS cancellations are a little-studied phenomenon (Smith et al., 2018) embedded in complex organisational contexts. They are multi-factorial (Hovlid and Bukve, 2014) and the relevant variables and their inter-relations are as yet unknown. Case study research is the appropriate methodology in this context (Voss et al., 2002). After completing a comprehensive hospital productivity study, Castelli et al. (2015) concluded: “It would be worth focusing attention on those hospitals at the top and bottom of the rankings to identify specific drivers of differential productivity in those organisations” (p. 253). The case study development will be completed with the help of three data collection methods to improve validity and reliability. In this research, each OTDS cancellation will be investigated as an incident. A proforma (Appendix 1) will be used to collect the information in a consistent way to understand how and why the OTDS cancellations happened and how they might have been avoided. With a case study methodology, the context and experiences of actors like managers and clinicians are critical (Fisher, 2007). Multiple data collection methods, using questionnaire, interview, observation, and document analysis approaches helps to understand the complexity (Long et al., 2018). The anticipated data collection include two types of interviews: structured interviews to understand the care pathway practices in each of the clinical specialities with a broad cross-section of staff to understand the practices, production design and patient flow; and semi-structured interviews to understand the influence of various organisational and human factors, like clinical engagement and team working impact, on the ODTS cancellations.

Sample - Multiple case study

The case selection aims to identify two-hospital sites as units of analysis in a comparative case study design. From the analysis of the multi-centre study, two suitable cases will be identified. Case selection is based on deviant cases that are suitable for understanding casual links (Gerring, 2017). The criteria for one of the cases are a hospital site that demonstrates high level of scheduling variation high cancellation rates due to bed unavailability. The second unit of analysis will be the opposite of the fist criteria, which is a low level of scheduling variation and low cancellation rates (figure 4-1).

All of the above primary data will be collected as a part of the fieldwork, with the permission of the trusts, after relevant ethical approval. The reliability and validity of the case study research findings will be enhanced by a well-designed research protocol (Yin, 2009). In this section, the research protocol will be developed to maintain a consistent approach in the data collection and development of the case studies. A short pilot study will be conducted in each NHS trust to understand the local constraints before the main field study begins.

Case study development, analysis and completion

The case studies completed in this section will be compared using comparative case study analysis methods. The multiple case studies will be analysed to understand the association between the causalities regarding the cancellation rates for a clinical speciality in each trust, to reduce OTDS cancellations and to improve the productivity of elective care. The fieldwork is planned to start around December 2019, and the first draft of the full chapter is expected to be produced by July 2020.

Conclusion

In this thesis, it aims to explore the factors that influence the production of the unintended outcomes in a complex organisational setting. In the first instance, using realist methodologies mechanisms that generate unintended outcomes were identify and aims to develop a Middle Range Theory. In the second part, using empirical data, the thesis aims to test the theoretical foundation of VM by exploring the association of AV and undesired outcomes, produced by a complex organisational setting. Going further, using the comparative case study design, causal mechanisms between the AV and undesired outcomes produced by the organisation will be explored. The testing the theoretical foundation of VM is a vital part in understanding the theoretical and practical application of the production and patient flow. Each study is expected to publish a minimum of one peer-reviewed article, which aims to publish in Health management journals such as British Medical Journal of Quality and Safety, International Journal of Quality and safety and International Journal of Surgery. In the final analysis, by linking the findings and insights from all of the three studies, this thesis aims to make a series of empirical, theoretical and methodological contributions.