Clinical handovers: Operationalizing and testing the handover performance matrix

(1)

MASTER THESIS

Clinical handovers: Operationalizing and testing the handover

performance matrix

MSc Technology & Operations Management, MSc Supply Chain Management

University of Groningen

Rommert Rijpkema

S2047519

Supervisor: Prof. Dr. J. de Vries

Co-assessor: MSc. A.C. Noort

3

rd

_{assessor: Prof. Dr. Ir. C.T.B. Ahaus}

(2)

2

Abstract

(3)

3

Preface

I would like to pay my gratitude and great thanks to my supervisor Prof. Dr. Jan de Vries, for his effort and guidance throughout the process of writing this thesis. Furthermore, I would like to thank Piet Penninga for giving me insight is his current research trajectory and helping me in understanding the intentions and implications of the handover performance matrix. I would also like to thank the staff at the Martini Hospital, the staff at the Isala Hospital and the general practitioner for their time and contribution to this research project. And finally, I want to thank my co-assessors, MSc. A.C. Noort and Prof. Dr. Ir. C.T.B. Ahaus for their time.

(4)

4

1. Introduction

With the aging population as a result of the baby-boom generation, we see a vast increase in healthcare costs. Additionally, rare and expensive diseases come to our attention trough growing media presence, which intensifies the debate about the affordability of the sector. Costs of this sector have risen from 70 billion in 2006 to 96 billion in 2016 in the Netherlands (CBS, 2017), which is an increase of 37 % over a period of 10 years. Increased patient complexity and higher costs put pressure on healthcare organisations to manage their processes more efficiently. Among this wide variety of healthcare processes, handovers are critical to the efficient transmission of information and responsibility. This process is complicated by the increasing specialization of healthcare units which requires more interaction and cooperation when handing over patients (Cohen and Hilligoss, 2009). In reality, this can result in deficits in communication, a lack of coordination and harm cost efficiency as well as optimal care for the patient (Olsen et al., 2013).

Extensive research has been conducted on the handover process, but the focus is often on particular attributes of the process or directed towards a specific patient group and not on the general performance of the handover (De Vries and Huijsman, 2011). Cohen and Hilligoss (2009) support this and conclude that there exists no universal way of analysing the performance of the handover process. When the handover performance is difficult to analyse and therefore challenging to improve, different types of failures can occur. These failures could result in an increased likelihood of adverse events, which can be the cause of life-threatening situations, unnecessary treatments, unintended re-hospitalization and eventually extra costs (Hesselink et al., 2014).

Since handover performance has a significant impact on the efficiency of a healthcare organisation and patients’ safety, Penninga (2018) developed the handover performance matrix (HPM), which describes handover processes and analyses performance. Penninga (2018) assumes that a handover process can be described by the level of complexity and connectivity, resulting in a matrix consisting of four types of handovers. These first two types are protocol determined and patient determined handovers, where connectivity is aligned with complexity which is considered to be efficient. The second two types are overperformance and underperformance, which are considered to be inefficient and where connectivity and complexity are not aligned. Even though this matrix gives insight into the organisation and performance of handovers, it has not been operationalized, and is therefore difficult to apply to an actual handover process.

(6)

6 process can be measured and categorized according to the HPM. Subsequently, a multiple case study was done to test the operationalized HPM on actual handover processes, and to assess whether it is a useful and adequate tool to describe and analyse the handover processes. This research project will provide researchers with an initial operational handover performance matrix and guide them in future research on handover performance. Ultimately this research project tries to contribute to increased efficiency of the handover process which is critical in achieving cost efficiency and patient safety.

This paper starts with a literature review on the current state of the healthcare sector and describe several matters within the handover process that affect efficiency. Additionally, the handover performance matrix is extensively discussed. After elaboration on the theoretical background, the methodology of the research project is explained, followed by the literature review and multiple case study subsequently. The results of the research project are then discussed and concluded to answer the following research questions.

RQ1: How can the handover performance matrix be operationalized?

This question is focused on developing an operational method to apply the HPM to actual handover processes.

RQ2: Is the handover performance matrix an adequate tool for analysing and describing handover processes?

(7)

7

2. Theoretical background

It is difficult to overlook the fact that the healthcare sector is a dynamic and ever-changing system where societal, medical and technological factors play an essential role in its current performance. With the aging population, more elderly with chronic diseases require medical attention, which puts more pressure on healthcare organisations (Eggins, Slade and Geddes, 2016). At the same time, additional pressure is put on the length of stay of patients which demands organisations to be efficient in the use of resources (Minkman, 2011). Furthermore, the increasing specialization of departments and clinicians calls for better interaction between different units (Cohen and Hilligoss, 2009). The number of processes that need to be coordinated by these units make the system complex and difficult to manage. And even though technological innovations provide healthcare organisations with the capacity to process more patients, this does not solve all complexities. The last decade showed that hospitals are increasingly focussing on patient logistics, clinical pathways, vertical integration and data interchange (Aptel and Pourjalali, 2001). These measures are examples of the implementation of industrial supply chain management practices in the healthcare sector and how they are used to improve patient care (Young et al., 2004). More often, management practices and perspectives from operations and supply chain management are integrated into healthcare organisations. Moreover, since studies on these practices and perspectives move towards a more customer and individualized focus, it becomes more interesting to examine its application in the healthcare sector (Meijboom, Schmidt‐Bakx and Westert, 2011; Dobrzykowski et al., 2014).

2.1 Clinical handovers

The handover, or handoff process is one aspect of the healthcare sector which is affected by the increased complexity. Different definitions of the handover process exist in previous literature. First of all, it can be described as the process where information and responsibility are transferred from one unit to another and where patients are moved between or within units (Vidyarthi et al., 2006). Bates et al., (2014:142) describe the handover process as ‘the transfer of understanding from sender

to receiver’. Where responsibility, authority and information about a patient are transferred to the

receiver to assure the continuity and quality of care, after the patient is moved to another department or organisation. However, the definition adopted in this research project is formulated by Cohen & Hilligoss, (2009: 494) as ‘The exchange between health professionals of information

about a patient accompanying either a transfer of control over, or of responsibility for, the patient’

since this elaborate explanation seems to best capture the complexity of the process. During this exchange, important information can be lost, and misunderstandings about responsibility can occur, leading to adverse events concerning the condition of the patient (Cohen and Hilligoss, 2009; Clarke

(8)

8 patient. The criticality of this process combined with recent insights from operations and supply chain management studies triggered interest in the handover process and scientific publications significantly (Flemming and Hübner, 2013). According to Cohen & Hilligoss (2009) one major factor amplifying this development is the implementation of a new requirement in the National Patient Safety Goals of the Joint Commission which stated that there should be a ‘standardized approach to handovers including an opportunity to ask and respond to questions’ (Cohen and Hilligoss, 2009). This development is accompanied by several technological and societal trends such as the aging of the population, demanding consumers that ask for the use of technological innovations and pressure from governmental and group initiatives (Benham-Hutchins and Effken, 2010).

The handover process is prone to several barriers, which can prevent human or physical resources in achieving increased effectiveness. Solet, Norvell, Rutan, & Frankel (2005) identify four barriers which hinder the quality and content of handovers. The first barrier is the physical setting of the handover process. For an adequate face to face handover, a quiet and private space is needed to assure quietness and assure confidentiality of the patient’s information, but availability is often limited. The social setting is the second barrier, which refers to the ease clinicians feel that they can share information. The status difference, for instance, can create a barrier in the adequate sharing of information. The third is the language barrier, and it can occur when clinicians have different nationalities or when they speak different dialects. Not only the origin of clinicians is the cause of this barrier, also different departments with particular specializations may have diverse ways of notation, which can cause loss of information in the handover process. The last barrier is the channel of communication, which makes a distinction between direct and indirect communication. Direct communication involves the face to face meetings that clinicians have when a handover is done, and where the full range of communication is available, meaning all verbal and nonverbal means in a conversation. Indirect communication, on the other hand, is communication by phone or other mediums. This can cause the receiver to make his/her own interpretations and assumptions about what the sender actually meant. Next to these barriers, Solet et al. (2005) identify one more barrier, which is ‘time & convenience’, which is often not sufficiently available and therefore affecting the quality of the handover.

(9)

9 the handover process may have a less accurate indication of their performance, which makes improvement difficult. Additionally, the availability of information technology can be a barrier. This can either be caused by the availability of information technology as means to communicate or caused by the variety of systems and their complexity to use them. Finally, the role and responsibility of the patients themselves is often not clear. Like to which extent they feel part of the handover process and act on it. This factor influences the handover process, as more assertive patients improve the efficiency.

(10)

10

2.2 The handover performance matrix

To gain a better understanding of the process and to deal with the complex nature of the handover process and heterogeneity of the organisations, the HPM was designed, which is shown in figure 2.1. It assumes that handover processes can be described according to two dimensions, connectivity and complexity, which leads to a categorization of four types of handovers. The next section examines how the HPM was developed, discusses the attributes of both connectivity and complexity and explains characteristics of the different types of handovers.

Figure 2.1: The handover performance matrix

(11)

11 sharing of information (Penninga, 2018). Integrated technology in the handover process can have different applications. For example, Li et al. (2013) argue that a lack of standardized information sharing is often the cause of errors and the occurrence of adverse events in the healthcare environment. Likely, integrating technological systems could prove to be useful in standardizing the handover process. Another application of these systems is to support the information transfer between clinicians by mediating the communication process through e-mail, telephone conversation or the implementation of electronic health records (Patterson, 2012). This could overcome problems that occur due to a lack of time or availability of clinicians. However, there are some drawbacks to the implementation of technological systems in the handover process. Penninga (2018) mentions that standardization may be too rigid for specific handovers and therefore the degree to which technology can be used in the process differs. Additionally, Benham-Hutchins & Effken (2010) argue that the successful implementation of technology in an organisation is determined by existing workflow processes and communication. Also, the level of confidence clinicians have in the system determines the successful implementation of technology. When clinicians in an organisation believe that a system is not reliable, the implemented technology can become a barrier in the handover process (Coiera, 2000). Besides the integration of technology, the attribute of information sharing and communication is also supported by partnerships, which is the third attribute of connectivity, and can be reflected in several ways in the handover process. For example, different healthcare organisations with different specializations can carry various perspectives on processes, which can lead to issues in the handover process. When clinicians in the handover process enter with different opinions, consensus is needed on the advised course of action (Bates et al., 2014). The means to achieve these partnerships lie in the training of the clinicians and create mutual understanding (Penninga, 2018). These practices should create an environment of trust and generate a sense of ‘teamwork’ to ensure continuity of care (Clarke et al., 2012). However, partnerships do not only concern the creation of common ground or mutual understanding, they also connect the organisation of the different healthcare units. The alignment of policies and having agreements on various aspects of the handover process are also important factors in creating partnerships (Benham-Hutchins and Effken, 2010). Ultimately these partnerships can decrease barriers between units, aid the information sharing and communication and even ease the process of implementing technological solutions.

(12)

12 different specialists, often located at different locations, and for a variety of resources. These issues led to the identification of the two attributes of complexity, which are uncertainty and variety. Firstly, uncertainty can refer to the uncertainty about the diagnosis of the patient or the course of action of the patient. Furthermore, it can be related to the availability of resources such as the availability of clinicians or necessary equipment. Secondly, variety can be found in the number of comorbidities of a patient, which may require a variety of clinicians and resources and can complicate the handover.

(13)

13 high detail. Underperformance may even be worse, since in this situation, connectivity is too low for a highly complex patient. This may seriously harm the patient’s health and may be the cause of adverse events. An example can be a complex handover, where no electronic protocols are present, and written charts need to be used. These charts are time-consuming and prone to interpretation errors, which can have greater impact on complex handovers, than on less those with less complexity.

In short, the healthcare sector is becoming more complex with the aging of the population and increasing comorbidities. At the same time pressure is put on organisations to improve the efficiency of resource usage and reduction of costs. Departments are increasingly specialized which calls for more integration, especially at the handover process. Nevertheless, the increased complexity makes it challenging to implement standardized solutions in the transfer of patients from one unit to another. Barriers in the handover process ask for more extensive research and a better understanding of the handover process, which Penninga (2018) tried to achieve by the development of the HPM. While this matrix seems promising, it has not been operationalized yet, and is consequently difficult to use when analysing actual handovers. Therefore, this research project aims to operationalize the HPM and test whether it is an adequate tool in describing and analysing the handover process and thus contribute to a better understanding. The research questions are as follows.

RQ1: How can the handover performance matrix be operationalized?

This research question composes the first part of the results where literature is studied to find the factors that describe the attributes proposed in the HPM. It aims to develop a set of variables that clinicians can score on a five-point Likert scale, which together form an initial operationalization of the HPM

RQ2: Is the handover performance matrix an adequate tool for analysing and describing the handover process?

(14)

14

3. Methodology

To be able to bring the HPM into practice and perform a study on its usability, it needs to be operationalized. This operationalization can be done in many different ways, but in this case, a literature review was used. Ideally, a comprehensive study involving different methods should be used to gain insight into the dimensions of the HPM, but since the time span of this research project was limited, a literature review was best suited. Also, this first attempt in operationalizing the HPM has the purpose of providing directions for future research such as a case study. A literature study may therefore act as a starting point from which the dimensions of the HPM can be investigated. This literature review was of an integrative nature (Cooper, 1984), meaning that literature was studied to verify and refine existing theories and that knowledge from different academic sources was used to gain new insights in the HPM regarding operationalization. These insights were used to identify variables that describe the attributes of both connectivity and complexity. Cooper (1988) classified research according to six characteristics which define the literature review. These characteristics are focus, goal, perspective, coverage, organisation, and audience and determined the framework of this literature review. The focus was on practices or applications since this research project analysed how the dimensions of the HPM are applied in practice and which variables can be used to measure the attributes. The goal was to integrate existing research and synthesize them to develop variables that describe the attributes of the HPM. This was done with an espousal of position perspective, meaning that after collection of data, the researcher decided on the most relevant variables from his perspective. In terms of coverage, a representative approach was applied. Since time constraints limited the amount of scientific literature that could be gathered, a sample of the literature was used to identify variables that describe the different attributes. Furthermore, this review was of conceptual organisation, whereas it combines research on the same concepts and groups them to identify the different variables. Finally, the audience of this research consists of general scholars with no specialization in clinical handovers, since its goal was to provide an understandable initial operationalization that can assist future researchers to carry on the development of the HPM.

(15)

15 screened, selected by their abstract and assessed on quality. Subsequently, the fourth step was (iv) doing the actual review and analysis. And finally, (v) the results were interpreted and presented in this paper. The sketched framework and review guidelines are the base in developing the literature review shown in Table 3.1.

Step What Description

i Planning and formulation The theory and rationale behind the healthcare sector, clinical handovers and the HPM were described in the theory section which emerged into a research question. A planning was made to guide the researcher in the process. Since the time span of this research was limited to 2,5 months and this research consist of two related but separate studies, realistic planning was crucial.

ii Literature search A literature search was done by searching for academic papers in different databases. The categories that were used are:

1. Papers with content that identifies factors that can describe information & communication in clinical handovers.

2. Papers with content that identifies factors that can describe integrated technology in clinical handovers.

3. Papers with content that identifies factors that can describe partnerships in clinical handovers.

4. Papers with content that identifies factors that can describe uncertainty in clinical handovers.

5. Papers with content that identifies factors that can describe variety in clinical handovers.

Databases that were used are ScienceDirect, EBSCOhost and Google Scholar. Keywords used were: Handover, handoff, healthcare information systems, partnerships, information and communication sharing, connectivity, complexity, handover variety, handover uncertainty. Forward and backward searches were used to expand the article base.

iii Screening, selection & quality assessment.

Articles were screened and abstracts were read, from where articles were either included or excluded.

Inclusion criteria: • Peer-reviewed • Published after 2000

• Related to dimensions of the HPM

Articles are assessed on quality, based on a full-text review regarding three issues: Rigour, credibility and relevance (Dybå and Dingsøyr, 2008)

iv Review and analysis Reviewing and analysis were done by reading the full text, highlighting relevant passages and categorize them by use of open, axial and selective coding (Strauss and Corbin, 2008) which resulted in a set of variables for each attribute described in the HPM.

v Interpretation and presentation The results of the coding were interpreted and variables describing the attributes of the dimensions described in the HPM were identified. The results are comprehensively described in this paper and the identified variables are presented in a table.

Table 3.1: Literature review

(16)

16 To conduct this part of the research project, a multiple case study approach was selected as most appropriate. The use of case study methods in testing the HPM was chosen, since it allows to study the phenomenon in its natural setting (Voss, Tsikriktsis and Frohlich, 2002). This means that the HPM was applied to actual handovers between units. The case study approach is also appropriate for small sample studies (Voss, Tsikriktsis and Frohlich, 2002), which was, due to time constraints, the case in this research project. The time limitation in combination with the organisational structure of healthcare organisation and handovers also constrained the ability to collect multiple sources of evidence. From the six sources of evidence Yin (1994) describes, only interviews were used to collect data on the different cases. The reason for this, is that triangulation of different collection methods (Voss, Tsikriktsis and Frohlich, 2002) was difficult to achieve, due to the nature of handover process and lack of archival records or documentation relevant to this research. First of all, observation proved difficult to arrange due to the organisational nature of the handovers, where multiple disciplines do their part of the handover at different times. Additionally, privacy reasons made it complicated to attend face-to-face handovers. Also contacting all parties and conducting interviews with everyone involved in the handover proved not possible to arrange within the time limit, due to the occupation of the clinicians from different disciplines. This resulted in the fact that the cases were limited to a single interviewee.

(17)

17

Table 3.2: Case details

One remark needs to be made about the nature of case 4. Because shift to shift handovers deal with multiple patients at the same time, and the research and interview protocol were not designed and prepared accordingly, case 4 does not describe one specific handover moment. However, the attributes and identified variables of the HPM were discussed with the interviewee to gain insight into the usability and adequacy of the operationalization. These results were also taken into account in the final analysis.

Semi-structured interviews were conducted to bring the operational HPM into practice. Before the interview, a brief summary of the nature and purpose of this research was provided, and the HPM was introduced shortly to make the interviewees familiar with it. The interviews were divided into categories related to the attributes of the HPM. Questions related to the variables of the attributes of the HPM were asked where the interviewee had to score the variable on a five-point Likert scale. Next to each of these questions, interviewees were asked to clarify their chosen score and to give their opinion on the usability of the variable. The semi-structured interview then allowed the interviewer to relate to examples that were given and ask for more in-depth information. Finally, the interviewees were asked to reflect on the interview and indicate whether they missed some crucial aspects. New information that was gathered and seemed relevant was also implemented in the interviews that followed. These interviews were recorded, and notes were written down during the interview. The interview protocol can be found in Appendix A.

When analysing data, two steps can be identified: Within-case and cross-case (Eisenhardt, 1989). In this research project, first the within case data was analysed by describing the background of the cases and examining the scores and remarks on the variables of each interviewee. The total score on both dimensions determined the position on the HPM, where both axis range from zero to the sum of all variables times the maximum score of five. Subsequently, the cross-case analysis was done by looking for patterns between the cases. While patterns between cases can indicate that an overlooked factor in the HPM is essential, responses from single cases were given the same weight.

Case 1 2 3 4

Clinician interviewed Physiotherapist General practitioner Nurse Coordinating Nurse Type of patient CVA patient Patient with multiple

comorbidities at the end of his life

CVA patient N/A Oncology

department Type of handover Physiotherapist in a

hospital to physiotherapist in a rehabilitation home General practitioner to nurse in neighborhood care Nurse in hospital to Nurse in care home

Shift to shift: Coordinating nurse to nurse within

(18)

18 The reason for this, is that the number of cases is small, and handover processes are different and complex to such a degree that every comment can potentially be of same relevance. The multiple case study results were documented, interviews were analysed and relevant passages were selected and categorized by open, axial and selective coding (Strauss and Corbin, 2008).

(19)

19

4.1 Results literature review

This section covers the result of the literature review of the operationalization of the HPM. First, the connectivity dimension is discussed with its attributes: integrated technology, partnerships and information and communication. Second, the complexity dimension is discussed with its attributes: uncertainty and variety. Finally, table 4.1 is provided to give an overview of the identified variables that describe them.

4.1.1 Information & communication

To facilitate an efficient handover and have an efficient transfer of information and communication, Lee et al. (2016) identified ‘communication culture’ as an important factor. This determines the degree to which a communication culture is present that allows clinicians to actively participate in the handover process. This means that they are willing to speak up, ask questions and provide feedback (Lee et al., 2016). This makes the handover process more interactive, which is important when information asymmetries exist between clinicians. Clinicians in the handover process may have different knowledge, expertise and experience, which can initially form a barrier in communications but is also an opportunity to learn from each other (Cohen and Hilligoss, 2009). Therefore it is important that the involved clinicians create an interactive handover culture where questions are asked, participants learn from each other and feedback is given. Eggins, Slade and Geddes (2016) also argue that the idea that ‘a quick handover is a good handover’ should be replaced by ‘an interactive handover is a safe handover’ and therefore this seems to be a relevant variable of information and communication.

To be able to communicate in the first place, certain interactive communication channels should be present. These describe the way clinicians communicate with each other in a handover process. Communication channels can be classified into two categories: direct and mediated communication (Solet et al., 2005). Direct communication is a face to face meeting where involved clinicians make use of verbal and non-verbal communication. Mediated communication is the sharing of information through other channels like papers, computerized systems, e-mail and telephone conversations. These methods can also be combined into hybrid forms, where for example a telephone conversation is used as a follow-up on a face to face meeting (Cohen and Hilligoss, 2009). Each medium has its advantages and disadvantages and is different for each case, but clinicians seem to have a preference for interruptive mechanisms, so face to face meetings and telephone conversations above e-mails and voicemails (Coiera, 2000).

(20)

20 process (Cohen and Hilligoss, 2009). This is an important factor of information sharing in a handover process since it can determine the willingness or openness of clinicians to speak up or ask questions (Cohen and Hilligoss, 2009). Power structures do not only exist between different layers of experience of the clinicians but also between teams. These clinicians and teams, often feel the need to appear capable and talented, and therefore they are hesitant in asking for help, which may have adverse effects on the performance of the handover (Cohen and Hilligoss, 2009). This variable is different from communication culture, in the way that power structure can also affect the organisation of the handover and not only openness of communication.

Next to the ways that information is transferred and clinicians communicate, the content of the handover is also variable. The ‘relevant handover content’ describes the amount of relevant information that is shared. Each patient requires a different amount of information to be shared, and therefore communication standards need to flexible in conveying exactly what information the other party requires (Cohen and Hilligoss, 2010). Often the information is categorized into ‘reasons for admission’ and ‘active problems and suggested course of action’ (Solet et al., 2005) however the details of this information can differ significantly. It is important for the performance of a handover that for each patient the right amount of relevant information is shared in sufficient detail.

4.1.2 Integrated technology

Supporting the sharing of information and communication in the HPM, is the integration of technology. Several studies indicate that the presence of electronic handover protocols is an important variable supporting the transfer of information. These protocols contain a set of standardized ‘boxes’ that need to be ticked or filled electronically. These protocols make sure that a certain amount of information will be provided and not forgotten. This digital recording of information has proven to increase handover communication and decrease the chance of miscommunication and associated errors (Li et al., 2013). It is essential that these records are online available and can be accessed from anywhere by clinicians in preparation of the handover (Van Eaton

et al., 2004). Van Eaton et al. (2004) also stress the importance of regularly updating the records and

automated data processing and incorporation. When these system of digital data recording are presents and integrated into organisations they can have a positive effect on the performance of the handover which means improved continuity of care and reduction of adverse events (Palma, Sharek and Longhurst, 2011).

(21)

21 always be possible due to convenience issues like time constraints, physical distance and occupation of clinicians (Solet et al., 2005). However, interaction with the other clinicians might still be necessary, and therefore different technological options can be used. This can be verbal communication such as a telephone call, set of voice recordings (Cohen and Hilligoss, 2009) which are exchanged between clinicians or even a video conference (Eggins, Slade and Geddes, 2016). But also non-verbal electronic communication channels can be used such as e-mail or other communication applications (Eggins, Slade and Geddes, 2016).

Finally, the use and integration of an electronic health record (EHR) system can be a significant facilitator of information sharing and communication. Apker et al. (2014) stress that these systems are a valuable tool for the effectiveness of a handover. These systems are software applications used in one or more healthcare organisations. This variable does not explicitly focus on the presence of such systems but more on the usability and compatibility with the system which the other party in a handover uses. Within the department, clinicians usually have access to the same system, but between organisations, a different system can be used which can make the sharing of information more difficult. On the other hand, when they are integrated, they facilitate effective patient handover and reduce the risk of adverse events due to an error in communication sharing (Palma, Sharek and Longhurst, 2011).

4.1.3 Partnerships

The second attribute that supports the sharing of information and communication is partnerships. Research indicates that partnerships between parties in a handover process can be based on a shared mental model (Lee et al., 2016). Variables such as a shared vision (Lee et al., 2016) common goals (Hilligoss and Cohen, 2013) and shared information expectation (Apker, Mallak and Gibson, 2007) are incorporated within this variable. Such partnerships can help clinicians to find a balance between thoroughness and efficiency in a handover process (Lee et al., 2016) and strive towards a minimal number of errors in communications that can cause adverse events regarding patient safety (Apker, Mallak and Gibson, 2007). Also a shared culture and background is an element of a mental model that can increase mutual understanding by collaboration and interaction by clinicians (Flemming and Hübner, 2013; Eggins, Slade and Geddes, 2016).

(22)

22 differences in expression of disagreement, variation in addressing senior clinicians and misunderstanding of verbal and non-verbal communication to communicate significant information (Eggins, Slade and Geddes, 2016). When clinicians share the same language and communication patterns regarding verbal and non-verbal communication, the process is enhanced significantly, and the risk of miscommunication is reduced (Benham-Hutchins and Effken, 2010).

Furthermore, Effective communication can be supported by having a common policy , which is also a determinant of partnerships. According to Gupta, Bevan and Vasudev (2014), handovers should not only be the responsibility of clinicians but should also be the responsibility of managers who form and control policies concerning the handover process. These policies have an impact on how the communication is performed, so when there are differences in policies between units, this can affect the quality of the handover (Benham-Hutchins and Effken, 2010). As Cohen and Hilligoss (2009) describe, handover policies are often created with other goals than improving the efficiency of a handover process. Hence, well aligned policies between different units in the handover process can have a positive effect on communication and the sharing of information.

The sharing of information and communication, as mentioned, can become more difficult through increasing comorbidities which increases complexity. This can also mean that more clinical disciplines are involved in the treatment of the patient and therefore also involved in the handover process. To support information sharing an communication, multidisciplinary consultations can be a form of partnerships to increase the connectivity of the handover. When taking the quality of handovers into account, these consultations need to be considered (Jeffcott, Ibrahim and Cameron, 2009). To be able to accurately transfer all information, meetings, where all the relevant clinicians are present, are desired. The presence of these consultations can increase collaboration and help achieve optimal outcomes of the handover process (Freitag and Carroll, 2011).

Finally, trust was identified as an important variable in describing partnerships. In communication and transfer of patient information, trust is an essential factor in achieving optimal outcomes (Clarke

et al., 2012; Apker et al., 2014). Mistrust can be a result of differences in orientation concerning

(23)

23

4.1.4 Uncertainty

Uncertainty is one of the two attributes of complexity in the HPM. Literature indicates that complexity in the handover process can be caused by several uncertainties. First of all, the course of action can cause uncertainty. For instance, when there are more treatment options for the patient having a specific disease, it might not be certain what the next step in treatment is (Mayor, Bangerter and Aribot, 2012). Another factor that contributes to this type of uncertainty can be the background and socio-economic status of the patient. A patient may have to choose between relatively cheap and expensive treatments but is partly bound by their socio-economic background, which can occur when they do not have health insurance or cannot afford it (Safford, Allison and Kiefe, 2007). This also means that the course of action is uncertain.

Closely related to the uncertainty about the course of action is the uncertainty about the diagnosis of the patient. The uncertainty of the diagnosis has a significant influence on the content of the handover and complicates it (Mayor, Bangerter and Aribot, 2012). A diagnosis may not be apparent, or the patient has vague symptoms of which clinicians are not entirely confident. This can cause them to make assumptions on what the diagnosis might be (Cohen and Hilligoss, 2009). While this seems closely related to the first variable, the example mentioned in the previous section distinguishes diagnosis related uncertainty from the course of action related uncertainty. An uncertain diagnosis can be the cause of an uncertain course of action. However, due to other circumstances like patient demands or socio-economic status, a definite diagnosis does not mean that there is no uncertainty about the course of action.

(24)

24 (2012) classified acuteness into four categories with increasing uncertainty in the task to be performed. So the higher the acuteness of the patient, the more task uncertainty can be experienced. This variable is different from the course of action uncertainty, since this uncertainty is caused by the difficulty of the task and acuteness of the patient and not by a variety of treatment options.

4.1.5 Variety

The second attribute of complexity is variety, and research led to the identification of three possible causes. First of all, variety of comorbidities of the patient. As mentioned before, comorbidities are the parallel occurrence of diseases and disorders in a patient. A greater number of comorbidities can increase the severity of the disease and additionally make it more difficult for clinicians to correctly diagnose the patient and determine the course of action (Schaink et al., 2012). Therefore an increase in comorbidities can make the handover process more complex.

Besides that, variety can also be found in the involved clinicians. Due to complex diseases or comorbidities, a larger variety of clinicians may be involved in the treatment of the patient ,such as a medical specialist, a nurse and a pharmacist. All these clinicians need to do their part of the handover, which is often done separately. Besides that, these clinicians have different information needs and a varying amount of experience and technical knowledge, which complicates the handover process (Eggins, Slade and Geddes, 2016). Finally, due to a greater number of involved clinicians, it can become more challenging the hear the other parties, and the chance that the handover is interrupted can increase (Eggins, Slade and Geddes, 2016).

(25)

25

Attribute Variable Description

Information & communication Communication culture Measures to what degree involved clinicians are willing to participate and speak up.

Interactive communication channels Measures to what degree interactive communication channels are used ranging from face to face meeting to paper charts.

Power structure Describes to which degree power

structures affect the communication Relevant handover content Measures whether relevant content is

present in the sharing of information. Integrated technology Integrated electronic handover protocols Measures whether protocols are

present and integrated between healthcare units.

Integrated electronic communication channels

Describes to which degree electronic communication is part of the handover such as telephone calls or

videoconferences.

Integrated electronic health record Measurement for the degree to which EHR systems are present and

integrated between healthcare units.

Partnerships Shared mental model Measures the presence of common

goals, shared vision, shared culture and information expectation. Language and communication patterns Described the degree to which

clinicians use the same language and ways to communicate verbal and non-verbal.

Common policy Measures whether policies of different

units are aligned. Multidisciplinary consultations Describe to which degree,

multidisciplinary consultation are present when needed.

Trust Measures the amount of trust

between units.

Uncertainty Uncertainty about course of action Measures uncertainty that can arise

due to multiple treatment options and a patient’s socio-economic status. Uncertainty about diagnosis Measures uncertainty caused by the

inability of clinicians to make a correct diagnosis and therefore have to make assumptions.

Uncertainty about available resources Describes the uncertainty originated from the unavailability of clinicians and equipment such as beds.

Task uncertainty Measures the uncertainty caused by

task difficulty and acuteness of the patient.

Variety Variety of comorbidities Measures variety of comorbidities and

the implication on complexity due to increased severity and difficulty to diagnose the patient.

Varieties of involved clinicians Measures the variety of clinicians involved in the process and the effect of their information needs on complexity.

Varieties of needed resources Measures impact of the variety of resources involved on handover complexity.

(26)

26

4.2 Results multiple case study

This section will cover the results of the explorative multiple case study. Each case is described separately, which is done according to three steps. First, the background of the case is described and handover procedures are explained. Next, the scoring of the variables is presented and elaborated upon. The visual representation of this scoring can be found in Appendix B. Finally, the results and responses of the interviewee are analysed. At the end of this section, an overall analysis combines the results of the four cases and this is presented in a table to clarify the findings.

4.2.1 Case 1

The first case that was selected, was the handover between a physiotherapist in the neurology nursing department in a hospital and a physiotherapist in a nursing home. The patient that was transferred was a patient who had a cerebrovascular accident (CVA) and was treated initially in the hospital. For rehabilitation purposes, the patient needed to be transferred to a nursing home where another physiotherapist would continue the rehabilitation and treatment of the patient. The handover was done in multiple parts, meaning that each involved clinician transferred information from their discipline without consulting the other involved parties. In this case, only one discipline was studied, and only the physiotherapist in the hospital was interviewed. This physiotherapist gathered general data from the EHR system and added discipline-specific information she deemed necessary. Of this information, a printout was made and sent to the physiotherapists in the nursing home. In this case, the physiotherapist did not know the other physiotherapist but handovers of this kind, with this specific nursing home, had occurred before. After the transfer of information, a telephone call was made to clarify some parts of the handover and give additional information. The interview with the physiotherapists produced a score on both dimensions of the HPM, and the results can be found in table 4.2.

Attribute Variable Score

Information & communication

Communication culture 5

Interactive communication channels 3

Power structure 5

Relevant handover content 1

Integrated technology

Integrated electronic handover protocols 4

Integrated electronic communication channels 3

Integrated electronic health record 3

Partnerships

Shared mental model 1

Language and communication patterns 4

Common policy 2

Multidisciplinary consultations 1

Trust 4

Total score connectivity 35/60 Uncertainty

Uncertainty about course of action 2

Uncertainty about diagnosis 1

Uncertainty about availability resources 1

(27)

27

Variety

Variety of comorbidities 4

Varieties of involved clinicians 3

Varieties of needed resources 3

Total score complexity 16/35

Table 4.2: Scoring of handover case 1

Regarding information and communication, this case scored a relatively high score. The handover culture was described as very open, even though the handover was for the most part done by a printout of information and the clinicians did not know each other. Also, power structure seemed to have no influence on the means of communicating, and handover content was brief. Integrated technology did play a supporting role since the EHR system was used to produce general patient information, but this was not compatible with both organisations. However, a protocol was used as a checklist for the transfer of information, but this protocol left space for the physiotherapist to write down information as elaborate as she wanted. Furthermore, the transfer of information and communication was supported by a phone call. Partnerships received a relatively lower score, which had to do with the fact that the two parties did not know each other and the handover was between two organisations. This resulted in a low degree of a shared mental model and common policies. On the other hand, this kind of handover had occurred before and took place within the same geographical region, which resulted in a higher score on trust and language and communication patterns. As mentioned, the handover was done between two physiotherapists while other clinicians were also involved with the patient, therefore no multidisciplinary consultation was present. On the complexity dimension, uncertainty about the variables was high but did not complicate the handover and thus was scored low. On the other hand, a variety of comorbidities and required resources did complicate the handover, which resulted in an intermediate score on complexity as well. These scores described this handover as ‘overperformance’ according to the HPM.

(28)

28 The uncertainty attribute also resulted in issues, but only in terms of formulation.

Physiotherapist: ‘In this case, there was a certain amount of uncertainty of the diagnosis, but that did not affect the complexity of the handover.’

Also, the scoring of other uncertainty variables resulted in the same issue, that uncertainty might be present but not necessarily affects the complexity of the handover. This formulation issue however, was resolved in the interview and scores were given on the degree to which uncertainty affects the complexity of the handover process. Variety, similarly, faced the same formulation issue and additionally identified another problem.

Physiotherapist: ‘For me, it is difficult to say whether the variety of involved clinicians increases complexity because I only do my part of the handover and it might be different for the others involved.’

Due to the composition of the handover process, it was not clear whether the variety of comorbidities and resources made the handover process more complex. This resulted in an unreliable score since not all involved clinicians were asked to score the handover process. But if they were, it might still be the question of whether they are able to score the effect of variety on handover complexity.

Finally, the physiotherapist commented on the language gap between the theoretical operationalization of the HPM.

Physiotherapist: ‘All these questions are very theoretical, and you do not know how a handover works in reality.’

Physiotherapist: ‘We try to implement things that exist in theory in practice, but then the translation to language on the work floor is made, and that is not yet done here. People on the work floor do not really understand what is meant.’

This made it harder for the physiotherapist to score the variables since she had trouble to understand what was meant.

4.2.2 Case 2

(29)

29 instructions of the GP. This means that the handover, in this case, was more a command than just the transfer of information. The GP also inserted the information and instructions in his digital system, which could later be accessed by him or his assistant, to answer questions and provide clarification. But in this case, that need did not occur. This case also only interviewed only one party in the handover, which was the GP.

The interview with the GP produced a score on both dimensions of the HPM, and the results can be found in table 4.3.

Interactive communication channels 3 → 1

Power structure 4

Integrated electronic health record 1 → 3

Partnerships

Common policy 4

Trust 3

Task uncertainty 2

Variety

(30)

30 no multidisciplinary consultation as indicated by the GP. Concerning uncertainty and variety and their effect on complexity of the handover, these were all scored low except the uncertainty about the availability of resources. This patient needed different electronic devices and specific pumps and medicine, which caused the handover to be more complex. The final scores on both dimensions describe this handover as ‘overperformance’.

When reflecting on the usability of the operationalization of the HPM, a couple of issues were identified of which the first concerns power structure.

GP: ‘The fact that the handover is more a command, is caused by the differences in authority.’

GP: ‘I do not think that the power structure gets in the way of the handover, but the handover is more a command and not a consultation without obligation.’

This indicates that power structure determined the way the communication is structured, so a GP who is giving a nurse a command. On the other hand, the GP indicated that this does not necessarily come in the way of communication. This could mean that the variable was not specified well enough and that the effect of power structure on connectivity could be emphasized. Furthermore, the GP noted that it would be interesting to see what the nurse would score the effect of the power structure, which could result in a more reliable score.

When assessing integrated technology on usability, this case also showed an issue. This has to do with the fact that all scores were given a one, since no technology was used. Even after correcting one variable, this still gives a low score of connectivity at the end of the line, while this may not necessarily mean that the handover was not performed well. The GP noted:

GP: ‘The ones who fill out the interview should indicate the weight of the variables or that when you have a scale like this that, you ask the person to pick the two variables most important to them to assign double the weight.’

This implicated that assigning weights to all variables could give a more realistic score on both dimensions. This allows, for example, to compensate for the absence of technology use by establishing partnerships.

When analysing partnerships, the main issue was the absence of interviews with all involved parties. The GP indicated that there could be a significant difference in trust between the parties and also commented on the shared mental model.

(31)

31 This causes difficulties in assessing the usability of the variables. However, the GP indicated that these variables are essential factors in handover processes and determinants of connectivity.

Additionally, the GP commented to the following regarding multidisciplinary consultations.

GP: ‘You have a general practitioner and a nurse who is, in essence, multidisciplinary, but I would not call this a multidisciplinary consultation.’

This indicates that this variable might not be specified well. The literature review describes this variable as a measurement to which degree all involved disciplines are present. But in this case, it is interpreted as a meeting with more than one discipline and that caused confusion in the scoring of the variable.

Regarding uncertainty, a similar issue occurred as with case 1.

GP: ‘funny thing is that I do not really take uncertainty with me in the handover. I make my own decisions beforehand, and therefore the handover is quite certain.’

So there might be a significant amount of uncertainty, but this is mostly taken away beforehand and therefore does not complicates the handover process. Similarly, the variables of the variety attribute generated the same response.

GP: ‘The patient does have comorbidities but in my opinion that does not affect the handover and thus does not make it more complex’

GP: ‘Many resources are present and needed but this does not increase the complexity of the handover’

This formulation issue made it more difficult for GP to assign a score to these variables.

Another vital remark was made by the GP concerning the usability and adequacy of the operationalization of the HPM.

GP: ‘A complex variable is the opinion of the family, patient and bystanders. Of course, specialists think certain things, but that is more medical and technical. I can think of something, but if there is a partner that cannot say goodbye or children who need to come all the way from America or do not agree at all, that makes it much more complex.’

GP: ‘Also the degree to which the family has insight into the handovers. In the handover book I leave there at home, and which is read by the partner, I cannot say: I hope she dies tonight. But you want to communicate that it could be over quite fast and that the family accepted that and therefore you describe it a little woolly because you know the family can read it. While if I could communicate it directly, in a one on one meeting without the family, I would be much more direct which could clarify the handover.’

(32)

32 A final general remark on the usability was made by the GP.

GP: ‘This all seems a little artificial. The handover seems more flexible and when you need to place something in a model you feel like you need to make choices. But on the other hand, you force people to make choices which makes it a little less subjective.’

This can illustrate the gap between the academic and theoretical nature of the operationalization, and refer to the language issues between researchers and clinicians. Therefore this might need to be improved in order to increase understanding among clinicians and receive more reliable scores.

4.2.3 Case 3

The third case that was selected concerns a handover that took place between a nurse in the neurology department of a hospital and a nurse in a nursing home. The nurse in the hospital was interviewed to score the handover process and comment on the usability of the operationalization. The patient that was transferred, was a patient who experienced a CVA and needed to be treated further after being in the hospital for a couple of days. In this handover process, a third party was involved called the ‘transfer point’. This third party’s task was to search for nursing homes in the area and find a bed and resources to treat the patient further. This party also visits the hospital once in a while to discuss process-specific issues, and might make a phone-call for clarification about a particular handover. After all requirements were met, the nurse in the hospital wrote the handover in the EHR system in the hospital. This system already contained an elaborate amount of information and a protocol was used to add additional relevant information to the handover. After all the information was inserted into the system, a printout was made and send to the nursing home. The interview with the nurse produced a score on both dimensions of the HPM, and the results can be found in table 4.4.

Interactive communication channels 1

Power structure 5

Integrated electronic health record 3

Partnerships

Common policy 4

Trust 4

(33)

33

Variety

As can be seen in table 4.4 above, communication culture and interactive communication channels are scored the minimum number of points. Because the handover was done by a print out of the information in the EHR system, no communication culture was present. Furthermore, power structure did not affect communication, and relevant handover content was considered intermediate. Integrated technology was moderately present, since an electronic handover protocol and an EHR system were used, but not integrated between the parties involved in the handover process. Partnerships were scored relatively high, where there was an evident presence of a shared mental model, similar language and communication patterns, common policies and trust. Since the handover was done solely on paper, no multidisciplinary consultation was present. The handover, overall, scored nearly the minimum amount on the complexity dimension, which resulted in the fact that this handover can be described as ‘overperformance’.

A couple of issues occurred when analysing the usability of the HPM according to case 3. The first was the involvement of the ‘transfer point’ that manages the availability of beds and resources. This third party increases connectivity between the hospital and the nursing home by making agreements with both parties and reducing uncertainty at the hospital regarding the availability of resources. However, the current operationalization does not incorporate this when scoring the handover. Additionally, trust as a variable was more challenging to score In this case.

Nurse: ‘it is hard to say if I trust the other party, I assume that they perform everything in a decent manner, but I am not sure of it. I do not think about it a lot.’

From a professional perspective, the nurse had to trust the other party and assumed things to go right, but she did not seem to fully trust them personally. This complicates the variable in the way that trust might not be there, but since there is no time and place to worry about it, it does not necessarily impact connectivity. However, this can impact the reliability of the score.

(34)

34

4.2.4 Case 4

This case could not be studied as the previous three cases, in the way that one handover moment was chosen to test the HPM. This had to do with the nature of the handover, which is a shift to shift handover by nurses of the oncology department in a hospital, where they transfer multiple patients at once. The nurses work shifts of eight hours and then transfer all the patients at the department at once. The coordinating nurse who is handing over the patients writes a ‘day list’ in Microsoft Excel and makes a printout to support the handover, which is done face to face. These handovers take place in a separate and quiet room, where nurses can take time to do the handover. The nurses to whom the patients are transferred, prepare the handover by looking in the EHR system to read the relevant information about the patients. Within the department, nurses are organized by seniority. This means that there is a difference in task authority between the nurses and thus different tasks are handed over to different nurses. When the transferred information is not clear after the handover, clarification is given by a telephone call.

While no variables could be scored, this case did give insight in relevant comments on the usability of the HPM. Firstly, the coordinating nurse commented on the preparation of the handover as an essential factor in connectivity.

Coordinating nurse: ‘I only hand over what is necessary, the rest was written down earlier by me and should have been read by the other party.’

The EHR contains a significant amount of patient information about current and previous treatments and personal details. Preparing this beforehand is essential, and when the other party is not sufficiently prepared, this can affect communication in the handover process. This is currently not included in the operationalization, and thus an essential factor that could be missing.

Another factor that seemed to be missing is the effect of time and place of the handover.

Coordinating nurse: ‘The handover always takes place in a separate room. Where we sit down, we never do the handover standing that would be exceptional, so we take our time. The other party to whom the patient is transferred has already read the patient, and sometimes we do the handover with the patient, so he is part of the process.’

Coordinating nurse: ‘We try to do complex handovers at bedside, so that the colleague which you transfer the patient not only receives the information in theory, but can actually see it.’

(35)

35 nurse mentioned that it could happen that, during a handover between nightshifts, nurses are more fatigued, which makes them prone to errors in interpreting information. Also weekends can cause issues due to the absence of specific clinicians. Additionally, she mentioned that bedside handovers are ‘the best’ handovers because the other clinicians can see and talk to the patient at the same time. In the current operationalization, these factors do not enhance connectivity since they are not measured by a variable.

Concerning the complexity variables, some critical context seemed to be absent. Firstly, at the end of a shift, not all tasks may yet be performed, and a nurse might still be waiting for a doctor to visit the patient. This can cause uncertainty about the course of action. Currently, this is not specified in the variable of ‘uncertainty about the course of action’. Furthermore, the coordinating nurse noted that, when scoring the ‘variety of involved clinicians’ variable, a higher score can also be caused by the fact that multiple nurses are involved in a shift to shift handover, due to the distribution of task to nurses with a different level of experience. This is also not indicated in the explanation of the variable. Finally, the coordinating nurse remarked on the content of the handover concerning the details of the information that is transferred. The variable as is, suggests that more information means more connectivity. The nurse however, indicated that it often happens that the other party provides too many details so that she needs to filter out information

Coordinating nurse: ‘I like short handovers, and what complicates things for me is that when colleagues do the handover very elaborately.’

When working shifts at the nursing department, there can be thirty handovers at once, which means that if more information is provided than needed, vital parts may be overlooked. This suggests that more detailed information might not necessarily lead to improved connectivity, but can result in the opposite. This means that this effect is currently not taken into account in the operationalization and can affect connectivity.

Clinical handovers: Operationalizing and testing the handover performance matrix

MASTER THESIS