The Health information system and its consequences for team interaction

(1)

The Health information system and its consequences for team

interaction

The evolvement of system use during its post-implementation phase and the effects on

team interaction in a hospital

6 July 2016

Marnix Marie

S1806238

m.p.w.marie@student.rug.nl

Master Thesis

MSc. Business Administration – Change Management

Faculty of Economics and Business

University of Groningen

Supervisor

Dr. J.F.J. Vos

Second supervisor

Dr. M.A.G. van Offenbeek

(2)

2

Acknowledgements

(3)

3

ABSTRACT

(6)

6

INTRODUCTION

In healthcare, communication between nurses and physicians has been marked as one of the main difficulties in the process of ensuring the safety of patients. A system that can open up such communication and coordination is a Healthcare Information System (HIS) (Hung, Chen, & Wang, 2014). A HIS can improve the collaboration and coordination among the different groups (e.g. nurses, physicians and supportive personnel), for instance, by providing the groups with information about what other team members have performed on the patient (Chiason, Reddy, Kaplan & Davidson., 2007). A HIS is a repository of information about the health of a patient, which is stored, transmitted and accessible by multiple (authorized) users (Samy, Ahmad, & Ismail, 2010). A HIS provides the backbone for collecting, compiling and utilizing information on patients and allows for the integration of teams (Raghupathi & Tan, 2008). This connectivity supports clinical decision making, cross-disciplinary communication and continuity of care (e.g. Brailer, 2005; Jensen, 2013). Especially, since a medical center is a professional bureaucracy with mostly autonomous professionals, who have substantial freedom to practice their task without extensive consultation (Mintzberg, 1983). Such a HIS can therefore improve the care that is given, since it enhances connectivity between cross-disciplinary functions, while it has been found that poor patient outcomes often result from insufficient collaboration between physicians and nurses (Chiason et al., 2007).

(7)

7

system are important (Brown, Massey, Montoya-Weiss, & Burkman, 2002). This fits a sensemaking perspective, since the users interpret the system, its purpose and the way it can be used for their work (Jensen & Aanestad, 2007).

This cognition or sensemaking is an individual phenomenon, but for team activities to occur, a shared understanding is required (Balogun & Johnson, 2005). It also demands a shift from the individual to the collective level, in order to understand complex organizational dynamics (Guiette & Vandenbempt, 2013). Despite this, most studies of sensemaking overlook the fact that sensemaking is a team-based effort (Balogun, Bartunek & Do, 2015). Yet, collective sensemaking lets employees interpret (technological) information via interaction with others, construct explanations to understand this information and respond to it as a team (Akgün, Keskin, Lynn & Dogan, 2012): this will create a common understanding (Christiansens & Varnes, 2009). So, this collective sensemaking appears relevant to understand and explain how teams perform under changing conditions (e.g. use of an implemented HIS) (Guiette & Vandenbempt, 2013). Within hospitals, sensemaking as a team occurs via verbal communication between the different hospital staff members (Kitzmiller, Anderson & McDaniel, 2010). So, the way hospital staff members as a team make sense, influences their choices and their actions, such as utilizing a system (Thomas, McDaniel, & Anderson, 1991). Furthermore, due to team sensemaking, a better control of action comes into being, since any new information is discussed as a team (Balogun & Johnson, 2005). Therefore, via team sensemaking, a team can understand what is happening, learn from each other and solve problems, which can have an influence on the way a health information technology is utilized (Kitzmiller, et al., 2010).

It appears that the discussed literature emphasizes on the necessity of interaction between the professional healthcare groups, which not only improves patient care but also influences the utilization of a HIS. It also explains how the use of such a system can occur via team sensemaking. However, the literature appears to be limited when it comes to the evolvement of system use during the post-implementation phase, as argued by Burton-Jones & Straub (2006). Maruping & Magni (2015) add to this that according to recent research on post-implementation use of technology, the true gains from collaboration technologies can be realized when the users explore the various system features and try to incorporate these features in their work practices. However, the literature does not explain how such exploratory behavior can be promoted when users are embedded in teams, which can create team sensemaking, and actions are interdependent (Maruping & Magni, 2015).

(8)

8

How does the use of a Health Information System evolve during the post-implementation phase and how do the interactions between and within the professional health care groups play a role in this?

A primary contribution of this paper will be to increase our understanding about the way system use evolves, during the post-implementation phase of a HIS, and how the collaboration and communication between and within the different healthcare groups affect this process. The following section will discuss the theoretical framework.

LITERATURE REVIEW

This section will cover the four topics, namely the Health Information System utilization, adaptive system use in the post-implementation phase, team interaction and team sensemaking. The former two topics will be explained from an individual perspective, whereas the latter two topics will explicitly focus on the team level. These building blocks, as introduced in the introduction, lead to the research framework shown below (Figure 1). The health information system utilization will be discussed now.

Health Information System utilization

Nowadays, it is hard to think of health care without information and communication technology (ICT) (Ammenwerth, et al., 2004) Besides, adopting new technologies is essential for many organizations to maintain their competitiveness nowadays (Edmondson, Bohmer, & Pisano, 2001). Already in 1992, Sandiford, Annett and Cibulskis (1992) argued that 25-40% of the health workers’ time was spent on filling out forms and collecting data and that half of that time could be saved by using an information system. Recently, there has been a great increase in demand for IT in healthcare, in order to achieve a

(9)

9

better and quicker information flow, and better services. The benefits of utilizing such technology can be substantial (Jensen & Aanestad, 2007). A HIS is an example of such an IT system in healthcare. The aim of HIS is to contribute to high-quality and efficient care of patients, improve cost-effectiveness and integrate teams (Haux, 2006; Chiasson & Davidson, 2004; Raghupathi & Tan, 2008). However, unless the system is assimilated into the medical practices in ways that it improves the healthcare, the benefits seem to appear limited (e.g. incremental, automation improvement) (Chiasson et al., 2007). Seizing these benefits can be challenging, since the healthcare systems are technically complex and are considered to be less developed than for other industries. Furthermore, the institutional structure (e.g. a dual administrative structure of medical personnel and administration) increases the complexity of developing information systems for medical professionals (Chiasson & Davidson, 2004).

Besides these challenges and its contribution to healthcare, a HIS can also introduce some problems. There is the problem of user acceptance, which concerns the difficulties that users have with the numerous procedures and interfaces (Haux, 2006). Secondly, according to Gattiker and Goodhue (2005), the impact of an information system, will be influenced by the interdependence and the differentiation that exists between sub-units in the organization. This impact can differ between the different healthcare groups (e.g. doctors, nurses, and supportive personnel). Whenever these groups interpret the technology in different ways and these differences are not discussed, they may lead to unintended and misaligned expectations (i.e. usage of the IS system), contradictory actions (e.g. doctor, nurses, and supportive personnel not using the system in the same way) and unanticipated consequences (e.g. resistance and skepticism) (Orlikowksi & Gash, 1994). Such breakdowns can negatively affect the working process (e.g. collaboration) and decision making, which can create adverse side effects in the process of care (e.g. harm for the patients) (Ammenwerth, et al., 2004). Finally, Health Information Systems that require doctors to enter their orders in the system, rather than handwriting them or dictating them to nurses or supportive personnel sometimes offends their sense of status and therefore fails to gain their acceptance (Markus, 2004). Thus a critical success factor can be to ensure the buy-in and collaboration (e.g. via acceptance and aligned expectations) from the clinicians. This buy-in is, according to Kaye, Kokia, Shalev, Idar & Chinitz (2010, p.173) the ‘make or

break factor’. All in all, the optimal utilization of a health information system is dependent on the

interrelation of the system’s functioning with the work of health care professionals (Berg, 1999).

Adaptive system use in the post-implementation phase

(10)

10

implementation, but that it may improve in time. Orlikowski (2000, p. 425) argues that this performance drop is due to a misfit between the user and the system: “Technology per se cannot

increase or decrease the productivity of workers’ performance, only use of it can.” Developments in system use over time

When a firm introduces an information system, employees do not simply start using it, right after its introduction. They can be overwhelmed by its complexity. Nonetheless, over time, the system can become part of someone’s everyday work, which is called a process of assimilation (Yamauchi & Swanson, 2010). These authors further mention that the assimilation of technology requires individuals to develop own routines and to learn by doing. Cawsey, Deszca and Ingols (2016) strengthen this by stating that when systems are used to promote the desired behavior, the users’ attitudes towards it may adjust over time in the desired direction, as they live with the new system or as they mention “have them live with new structural or systemic arrangements”(Cawsey et al., 2016, p.172). Berg (1997) explains this as a co-evolution between the users and the system. Users adopt their way of working to a system, while the system use is adjusted to their work routines, each can change the other. This means that the users’ norms, practices, beliefs and values evolve along with the use of the system. So, it needs time before a system is accepted, and to assess the usefulness of a system and the post-implementation behavior, it requires looking beyond the implementation phase (Ehie & Madsen, 2005), Nonetheless, it is important to prevent the wrong ‘habits’ from being learned, since any habits that are formed when you start using a new information technology could be difficult to change later on (Orlikowski & Gash, 1994). Hence, there is a need for continuous improvement during post-implementation phase, with maintenance and support (i.e. adjusting or adding of new functions), and ongoing reengineering of business processes (Ha & Ahn, 2014).

Triggering adaptive system use

(11)

11

control. However, automatic system use can be mentally efficient, since it does not require the user’s attention (Polites & Karahanna, 2013). Regarding the former type of use, Sun (2012) introduces adaptive system use (ASU), which is the revision of which and how a certain system can be used over time. According to Venkatesh and Bala (2008), this system use can depend on perceived usefulness, the extent a person thinks the system can improve their performance, and perceived ease-of-use, the extent a person thinks the system use is clear and straightforward. These perceptions can change due to experience (i.e. over time) with the system.

Furthermore, Sun (2012) found certain ‘triggers’, which are situations that people encounter when using a system. These triggers can be any situation of change where there is a contradiction (i.e. lack of fit) between the user, between different activities or between different phases of an activity. For instance, a new situation is a contradiction of a current and this new situation (e.g. when a system is implemented that requires new tasks) (Sun, 2012). The users try to interpret and react to these triggers, which changes their behavior. Finally, facilitating conditions (i.e. support one gets from his environment) could moderate these possible triggers (Sun, 2012) and facilitates the use of the system (Venkatesh & Bala, 2008).

The influence of support and training

Nicolau (2004) argues that in general, organizations need to engage in a number of support activities, such as post-implementation-review (PIR). This concerns the question what was successful and what needs to be improved about the system. Ha and Ahn (2014) also stress the use of training during this post-implementation stage. This training facilitates users to cope with the changes and improve their expertise with the system. This can motivate users to be more cooperative, collaborative and communicative. User manuals, user training, super-users (experts of the system) and peer support are sources of support that can help nurses and doctors during this phase (Jensen & Aanestad, 2007; Venkatesh & Bala, 2008). Finally, Venkatesh and Bala (2008) also argue that training and support during post implementation leads to greater system success and user acceptance (i.e. more favorable perceived ease of use and usefulness) of complex systems, such as a HIS. However, there can be differences in the way nurses and doctors view this support during the post-implementation phase. For instance, nurses highlight the importance of being informed about the project by management, whereas the doctors rely on their peers (e.g. colleagues) instead of receiving general information (Jensen & Aanestad, 2007). The different perceptions on the initiatives by the different groups, has to be taken into account.

(12)

12

specialties. This means that a successful application of a system, that influences this interaction, requires conforming to the preferences of the healthcare groups (Mittman, Tonesk, & Jacobson, 1992). Hence, there needs to be a fit between the system, user and organization (e.g. department). For instance, physicians appear positive towards a system, as long as the system is compatible with their professional norms and values, and provides benefits for their work (Kaplan, 2001) . However, the physicians seem reluctant when they are forced to be more precise in writing their orders via a system (Berg, 2001). Nurses appear positive as long as the system is not highly normative (e.g. cause standardization, rigidity, predictability and lack of autonomy) (Kaplan, 2001). This is emphasized by Berg (2001), who states that nurses like to retain the influence they have in helping the physicians, which can be limited by a HIS, since such system increases automation. Furthermore, this automation may cause some nurses to worry about patients’ privacy due to digitalization, while others worry about the increase in the administrative and digital workload (Van Offenbeek & Vos, 2016). Finally, these authors also mention that user friendliness is essential for the administrative personnel, in order to successfully integrate the system in their work.

Therefore, as users gain more experience with a system, they might discover the unique features it offers (Hiltz & Turoff, 1981). So, for this research, it is important to consider the use over time. In the beginning, the user can be unfamiliar with the system, but after they experience the system, receive support, and maybe encounter certain triggers, they might change the way they interact with it (Jasperson, Carter, & Zmud, 2005). As Sun (2012, p. 454) states: “An enriched understanding

of people’s revisions of system use and triggers is of great value to information systems research.”

These different interpretations of the various professional health care groups are an interesting topic for further investigation, in order to determine whether there is synergy between the primary work process (e.g. patient care activities), secondary work process (e.g. administrative support) and the HIS (Berg, 2001).

Team interaction

A team is a group of two or more people that interact and adapt to certain shared values (Ashmos & Nathan, 2002). Effective teamwork facilitates collaboration, coordinated effect and task accomplishments. It also allows for learning, due to shared expertise and social interaction (Kitzmiller,

et al., 2010). This teamwork, as mentioned earlier, between nurses and physicians seems a key factor

(13)

13

affects the individuals’ use of it (Chiason et al., 2007). So it appears that also inter-team interaction is associated with learning and positive organizational outcomes (e.g. system use and patient care) (Nembhard & Edmondson, 2006).

Whenever there is a reduction in interaction, for instance due to an information system, it increases the possibilities of errors because of miscommunication and fewer team-wide conversations, about the planning and coordination of the patients (Campbell, Sittig, Ash, Guappone, & Dykstra, 2006). So, a HIS could increase the speed of communication, but can also cause misunderstanding, for instance due to the complexity of communicating via the system, which can lead to an increase in medical errors (Koppel, et al., 2005). Therefore, for successfully using a HIS there is agreement on the notion that face-to-face teamwork and communication are required (Edmondson et al., 2001). However, according to Orlikowski and Gash (1994) there can also be a negative effect of group membership (e.g. a team), due to the different interpretations, meanings, and beliefs among teams (i.e. team sensemaking). This can create differences between teams and therefore possible conflict. Nonetheless, this possible conflict can be managed through inter-team collaboration, which enhances the effectiveness of teams (Dreu & Van Vianen, 2001). Jordan et al., (2009, p. 7) emphasize this as follows: “Practice staff and clinicians may fully understand the specifics of an improvement effort, but

it is through conversations that they produce a shared vision (i.e. team sensemaking) of how a given intervention will improve care of their patients and will enhance real adoption of a change”. So, via

team interaction, team sensemaking can occur that can enhance the adoption of a change, since this is a communicative practice where people get triggered by interacting with others (Akgün, et al., 2012).

Team Sensemaking

As mentioned earlier, this topic will be on the team level, since the way the different teams make sense of the implemented system, determines how an information system is used and/or accepted. Akgün et

al. (2012) state that team sensemaking is an ongoing process of meaning construction that reflects how

(14)

14

some as not. Thus, whenever a team uses a technology, they will evaluate which features offer value to them, allowing for a shared understanding (Saeed, Abdinnour, Lengnick-Hall, & Lengnick-Hall, 2010).This requires continuous sensemaking (Bansler & Havn, 2004).

In addition, whenever a technology is adopted, team sensemaking has an impact on the motivation among the different professionals (Jensen & Aanestad, 2007). Maruping & Magni (2015) show that this collective motivation as a team helps to form the post-implementation behavior of individuals and the way they use technology during this stage. Because when team members, as a collective, make sense of their work as being meaningful, they are more likely to use a system that supports this work (Maruping & Magni, 2015). This ensures that individuals engage in exploration (of a system), since they know using the system is meaningful to the team. So, by looking at team sensemaking, the interpretations about the system and its influence can be observed, or as Orlikowski and Gash (1994, p.33) argue: “Examining people’s shared cognitions around technology can be a

powerful means of articulating and tracing the influence of information systems in organizations.” It

is therefore important to understand how the different teams make sense of the implemented system, since sensemaking drives the acceptance and use of an information system (Saeed et al., 2010). Based on this theory analysis, it shows that the four topics as mentioned at Figure 1 are interrelated and influence each other. Moreover, several additional subjects have been added, namely face-to-face communication, the intra-team communication and the three factors that influence the adaptive system use: triggers, support and the use over time. This led to the more detailed research framework seen below (Figure 2)

(15)

15

METHODOLOGY

This section describes the methods that were used to collect and analyze the data. First, the research approach will be explained, followed by the research site and data collection. Finally, the data analysis will explain how the gathered data are analyzed.

Research approach

There is still much unknown about how system use evolves during the post-implementation phase and how team interaction affects this process. This lack of knowledge is a trigger for theory development (Aken, Berends and Van der Bij, 2012). The goal of theory development is to observe the phenomenon via case studies, which can be single or multiple cases and also on numerous levels of analysis (Yin, 1984). Dasgupta (2015) adds to this that a qualitative research focuses on ‘how something happens’. This research explores the effect of how system use evolves during post-implementation and how team interaction plays a role in this. Moreover, these case studies can be used to provide description and generate theory (Eisenhardt, 1989). This research focuses on the latter, since it relates to generating theory about an unknown phenomenon.

For this research, two cases were studied on two levels, namely the three teams within two hospital departments and the two departments itself. Furthermore, a cross-case analysis will be used, which compares insights from the two cases (both teams and departments) with each other (Aken, Berends, & Van der Bij, 2012). Afterwards, the results will be compared with existing literature and any new or different insights are turned into propositions. These findings and propositions can be found in the discussion and conclusion section.

Research site

The study was conducted within a large teaching hospital in the Netherlands, in which a HIS was implemented several years ago. So far, the system supported the organization with process optimization and an increase in effectiveness. However, recently, the three healthcare groups from department X raised their concerns about the reduced personal inter-team interaction (i.e. face-to-face contact). The three groups felt more and more alienated from each other and they worried that this could have serious consequences for the patient outcomes. This study investigates how this use has evolved after the implementation and in what way the team interaction has played a role in this. The case is relevant, since a system can increase collaboration between teams and therefore patient care, but only when there is buy-in from all teams (Kaye et al., 2010).

(16)

16

Data collection

In order to observe the problem in practice (Aken, Berends, & Van der Bij, 2012), documents about the HIS (e.g. manuals) were analyzed, followed by thirteen in-depth, semi-structured interviews. This combination and complementation of multiple sources is triangulation (Baxter & Jack, 2008). The respondents were from both departments and from each team, two or three respondents were randomly chosen, to ensure the broadest possible view on the topic. Furthermore, to achieve reliability, Aken et

al. (2012) was used. Reliability ensures that the results of a study are independent of characteristics of

study and can therefore be replicated in other studies (Swanborn, 1996). According to Aken et al. (2012), there appear to be four sources of bias: researcher, instrument, respondents and situation. The first was covered by having a third party review the interview questions and the codes that were deducted from the interviews. This led to better and less biased insights. The second bias was covered by using both documents about the HIS and the interviews, which allowed for triangulation and therefore reliability (Baxter & Jack, 2008). The third bias is prevented by randomly choosing respondents from each department and each team, ensuring that each team is represented by at least two persons. The final bias was avoided, since the interviews were held in an office where nobody else was able to hear what the interviewee is saying.

Validity is about actually capturing what was intended to study and report what has been seen or heard (Ali & Yusof, 2011). There are three types of validity: construct, internal and external validity. The first is about the extent to which the instrument measures, what it needs to (Healy & Perry, 2000). This is covered by letting a third party check both the interview and the codes. The internal validity is about having justified conclusions about relationships (Aken, Berends, & Van der Bij, 2012) and the extent to which changes in a dependent variable can be attributed to a controlled variation in the independent variable (Healy & Perry, 2000). This is covered with triangulation. The external validity will be harder to guarantee, since the research will merely be conducted in a medical center. Therefore it is unclear whether the same results can be expected in other organizations.

Interviews

The interview questions were developed on the basis of the literature analysis and the core subjects. The questions were checked by both professional and academic experts. Finally, the protocol was tested via a pilot interview with one of the experts from the medical center. This resulted in some preliminary codes that were processed in the protocol. However, this pilot interview was only used to make some adjustment to the (wording of) questions, the results itself were not used in this report. The final protocol can be found in Appendix I.

(17)

17

second part of the interview covered the interaction between and within the interviewee’s team and how this influenced their opinion (i.e. to what extent team sensemaking was apparent), so the interviewee was asked about their view on the opinion of their and other team[s]. An example of such a question: “To what extent does the opinion of your team influence you?”

The interviews gave insights about the (evolving) system use during the post-implementation phase, and the effects of team interaction. The interviews were conducted with representatives from both departments and all three groups, and lasted between 45 and 60 minutes. Each interviewee received a short introduction about the topic up front. The confidentiality of the respondents was secured, since their names are not mentioned in any of the transcripts or analyses. In addition, abbreviations are used to refer to the respondents: doctor (D), nurse (N), medical administration (A), and both departments (‘X’ and ‘Z’). Hence, D1X is doctor one, from department ‘X’. The interviews were recorded, after the interviewees gave their consent to do so. Afterwards, the interviews were sent to the respondents, to give them the opportunity to make any final changes. In addition, the interviews were conducted and transcribed in Dutch, since the respondents were all native Dutch speakers. Furthermore, by transcribing the interviews in Dutch, it is assured that no expressions or sayings are wrongly translated; since translating might require an interpreter of the language (English in this case) in order to convert certain expressions sufficiently (Moerman, 1996). Therefore, only the meaningful quotes were translated in English.

Timeline

As this research aimed to investigate how the system use evolved during the post-implementation phase, it was important to use a timeline. The timeline was used to refresh the memory of the respondents, since most events occurred several months or years ago. The recurring memories were then used to get a better grip on the development of the system use. So, on this timeline, the most important events and incidents were portrayed, that occurred from the start of the system until now and of which some could be ‘triggers’. This perspective provides the possibility to consider the consequences of system use and can highlight the “triggers that can modify either the set of initial

conditions or the object of resistance” (Lapointe & Rivard, 2005, p. 479). These triggers can be, as

mentioned earlier, contradictions which result from the consequences of the system use, the actions of others, and reactions of system advocates to resistance behavior (Lapointe & Rivard, 2005). The timelines were constructed with the help of the HIS documents, the manager of one of the departments, and four IT-experts that helped implement the system several years ago. For the timelines, see Appendix VIII.

Data analysis

As mentioned by Eisenhardt (1989, p. 539), analyzing the data is ‘the heart of building theory from

(18)

18

analysis are used for a single study (Yin, 1984). First there is the team level, consisting of the three healthcare groups and then there is the department level, consisting of department ‘X’ and ‘Z’. These two cases with their multiple levels are compared. In order to do so, a within-case analysis was used first. This allowed the researcher to discover unique patterns for each single case and to make sense of the amount of data that was gathered. This within-case analysis also strengthened the input for the cross-case analysis, since the researcher became familiar with each single case, before starting comparing (Eisenhardt, 1989). Afterwards, a cross-case looked for patterns and differences between the departments, whereas a cross-subcase analysis was used to compare the teams. As mentioned, two somewhat opposite departments were chosen. These polar cases, allowed the topics of interest to become “transparently observable” (Eisenhardt, 1989, p. 537), which helped to look for within-group similarities and intergroup differences (Eisenhardt, 1989). This was all done with a retrospective approach, which involves pursuing past events to give a clear and detailed picture of what took place (Dasgupta, 2015).

After transcribing the interviews, the transcripts were analyzed and coded. Deductive codes were constructed, based on detailed reading of theory, whereas inductive codes were derived from the interviews. First, ‘open coding’ was used (Mortelmans, 2013), which is about dividing the transcripts into smaller parts, based on the topics or subjects. This ‘open coding’ was applied after the first pilot interview and led to the isolation and separation of the themes that could help answer the research question: the HIS utilization, system use and interaction. The second step was axial coding, which combined the lose codes, that were generated at the previous step, into certain concepts or categories. This led to five different categories: HIS utilization, adaptive system use, interaction, team sensemaking and patient care. The final step was about connecting the categories to each other, in order to come up with theory (Mortelmans, 2013), which is presented in the result section.

(19)

19

RESULTS

The results are presented in this section, organized on the base of the health information system utilization, the adaptive system use, interaction, team sensemaking and the patient care. An overview of the codes, which gives a clear overview of the discovered variables, can be found below. Based on the results, an answer will be given on the research question. Since the research was conducted within two departments, both departments and its teams will be covered. First, a comparison will be made within the departments [within-case], followed by a comparison between the teams [cross-subcase] and departments [cross-case].

Health Information System [HIS] utilization

Based on the findings, it appears that there are two variables that, taken together, explain the HIS utilization at the hospital. These are the system functionalities and the changing work.

System functionalities

The system, PoliPlus, supports the healthcare groups by allowing them to store patients’ information and retrieving this information more efficient than they could before. As argued by one of the nurses:

Category label Subcategories Definition Codes

Health Information System utilization [Deductive] System functionalities, Changing work [Deductive]

The utilization of a health information system is dependent on the interrelation of the system’s functioning with the work

of health care professionals (Berg, 1999).

Efficiency, Information storage, System possibilities,

Automation [Deductive] Use of patient files, Work load, Dependency [Inductive]

Adaptive system use [Deductive]

System use over time System support Triggers Changing perceived usefulness [Deductive] ‘

The revision of how a system can be used over time (Sun, 2012).

Peer support, Training, ICT Support, Usefulness, Ease of

use [Deductive] Gradual learning, Loss of knowledge, Autodidactic,

Introduction of new functionalities, Workarounds,

Intensity of use, Uniform use [Inductive] Team Interaction [Deductive] Inter-team interaction Face-to-face communication [Deductive] Work climate [Inductive]

Conversation (i.e. communication) between and within the professional healthcare groups (Jordan et al.,2009)

System communication, Intra-team interaction, Distance

[Deductive] Contact moments, Feedback,

Necessity, Formality [Inductive]

Team sensemaking [Deductive]

Team mental model [Deductive] Team sensemaking

moments [Inductive]

Collective sensemaking can allow employees to interpret (technological) information via interaction with others, construct explanations to comprehend it and respond to it collectively( as a team) (Akgün, Keskin, Lynn & Dogan, 2012)

Shared cognition [Deductive] Cohesion, (Multidisciplinary)

Team meetings [Inductive]

Patient care [Deductive] Changing primary process Risks [Deductive] Urgency [Inductive]

The aim of HIS is to contribute to high-quality and efficient care of patients (Haux, 2006; Chiasson & Davidson,

2004).

Process support, Privacy, [Deductive] Time pressure,

Reliance [Inductive]

(20)

20

“It is saved in one place and you can always retrieve it [..]I can find it straight away on any computer,

which makes it a lot easier” [N2X]. This allows them to spend less time searching for files. Besides

saving time, the system also offers others possibilities for the teams. For the nurses and medical administration, the system is more used as a ‘Show Box’ [in Dutch: Kijkdoos]. They use the system to retrieve information. The doctors are the ones who also put information into the system, such as the patients’ status. However, since the doctors are using the system more intensely, they are also the ones complaining about missing functionalities. For instance a doctor complains about the following: What

can I do with a system when I cannot add my echo images to it? [..] this has already been an issue for so long. That is incomprehensible for ‘Z’” [D2Z]. Another doctor adds: “I can read it [the letter], but cannot change or add something. When we are in a big team, where we have to see a patient with many different disciplines, all four of us would like to add something” [D2X]. When looking at the

new functionalities that have been added, it appears that the nurses and administrative personnel are more positive about the added functionalities. However, the doctors are less positive about it and do not see the added value of expanding the system with new functionalities. As argued by one of the doctors: “Meanwhile they have been sticking plasters, by adding new functionalities.”[D2Z]. Finally, another doctor adds to this: “It has become a building with separate small buildings, it is a

monster.”[D1X]. So, it appears that there are major differences between the teams, but not between

the departments. The changing work

The possibility to store and retrieve information also changed the way of work; except for the nurses who are still working with their own paper file. Both medical administration and doctors argue that this information should also be digitalized in order to have patient files complete. Due to this automation of work, the work load for the doctors increased, which is confirmed by all three groups. Now, the doctors have to type their own letters, as mentioned by a doctor: “The doctor’s job consists

for 60% out of seeing patients and for 40% out of typing.”[D2Z].

As for the rest, there seem to be large differences between the departments. Especially for the medical administration of department ‘X’, the work has become automated: “Boring, I find it terribly

boring. At half past four we have squared eyes from watching at the screen. You only sit behind a desktop, everything is digitalized.”[A1X]. For department ‘Z’ the medical administration did not

complain about this automation. Furthermore, department ‘X’ already digitalized all files, whereas department ‘Z’ still appeared to be working with the paper files. For doctors at ‘X’, this digitalization appears to be a necessity and that their job depends on it; it is something you cannot stop, as argued by a doctor: “It is what it is. Times are a changing. The time when secretaries sat beside you at the

policlinic to write things down, is no longer.”[D2X]. The other two groups also see the inevitability,

(21)

21

are changing digitally, which also means that new systems will come. I am preparing for this and so are others.” [A2Z].

Adaptive System Use

The system use can be classified into four categories: system use over time, system support, triggers and changing perceived usefulness.

System use over time and system support

Most interviewees acknowledged that their understanding of the system evolved gradually and that they learned most from simply doing it. None of them thought the system was difficult to understand, merely because the functionalities were added one by one, as argued by one of the nurses: “It is easy

that it [PoliPlus] is expanded bit by bit with functionalities you work with [..] then you work with it regularly and you understand it.”[N2X]. In addition, both medical administration and nurses were

dependent on their own skills, in order to grasp the system, since they did not receive any training, it was merely learning by doing [autodidactic]. The doctors however, did receive training on their first day as a doctor: Both other groups feel like they should need to receive more training in the future, when the EPR comes. Whenever people did receive support, it was primarily from their co-workers. This peer support seemed important to all teams, as mentioned by one of the medical administration: “If I am not sure anymore how it works, I will ask a colleague.”[A1Z]. Finally, the support from ICT was not thought highly of, especially not by the doctors, as mentioned by one of them: “I will never

call the helpdesk again [..] It does not work, you are on hold way too long, they have unprofessional personnel, which is not their fault. Sometimes you are helped ok, but most of the times you are not.”[D1Z]. Whenever a problem would arise, all respondents would rather ask their colleagues than

the ICT department. The only difference between the departments was between the medical administrations. At department ‘X’ they argued that they were losing knowledge over time, due to the automation of their work. They no longer typed the letters, which meant that they were no longer confronted with medical terminology. This was not mentioned by the other department.

Triggers

The introduction of certain functionalities changed the way of working for the three groups. With the

digital patient registration [in Dutch: digitale decursus], the doctors no longer need to write down

their notes on a piece of paper; they can store them digitally. As argued by one of the doctors: “What

made the difference is the digital patient registration. It started out as a scratch-pad and turned into a system in which you can really maintain and record patient files.”[D1X]. The letter dispatch [in

Dutch: brief afhandeling] only affects the doctors and medical administration, since the doctors have replaced the medical administration in typing the letters, as argued by one of the medical administrators: “Now they [doctors] write the letters digitally and send them to us. The only thing we

(22)

22

needs to write on a piece of paper or tell the nurses/and the medical administration what should happen with the patient (which types of research, when the next appointment should be, etc.), but he/she can simply fill this in on a digital file, which not only saves time, but also prevents errors, as one of the nurses explains: “Especially this one [DAF] is convenient, because the doctor does not need

to drop by anymore in this office [..] This form is entered by the doctor, but we execute the tasks.”[N2X]. The introduction of these functionalities, led to the gradual abolishment of the paper

patient file at department ‘X’, which affected the doctors and medical administration most. It became clear that this department worked more intensively with all three functionalities, whereas the department ‘Z’ still relied on paper for some proceedings.

Changing perceived usefulness

The system is easy to use for all interviewees: none of them experiences difficulties with its functionalities. Also, all interviewees understand the system, in order to do their jobs. Most of them are satisfied with the amount of functionalities that are available to them and do not prefer to learn any more, since their job does not require them, as argued by a nurse: “I understand the parts I work with

[..] if you do not need it, it is a waste of your time to be occupied with it.”[N2X]. Some interviewees

would rather learn more about the EPR than about PoliPlus, since the former is expected to replace PoliPlus within the next one-and-a half year. Finally, age was also a factor that influenced the way employees use and perceived the system. As argued by two employees from the medical administration: “The younger the people, the better they can work with PoliPlus [..] but the older get

confused when they have to do something new or different in PoliPlus”[A3Z]. The older employees

that were interviewed (fifty-five or older) agreed to this, so it seems that age does play a role when it comes to learning and understanding a new system.

However, there were also differences in the way the interviewees perceived the system’s usefulness. The administrative personnel would like to have the possibility to adjust any mistakes that are made by the doctors in their letters, which could prevent possible mistakes, as stated by one of the medical administrators: “That is a danger; we leave it [typing of letters] up to them. When the doctor

makes mistakes, it is very sloppy [..] but we no longer check it.”[A2X]. The doctors did not perceive

the system as user friendly, while it is quite slow and consists of multiple ‘building blocks’. Furthermore, the system did not allow doctors from different disciplines to work together in a single file, whereupon doctors decided to work together in a Word document for certain multidisciplinary issues, which is acknowledged (although disapproved) by one of the doctors: “This causes people to

come up with the idea to work in a Word document, where multiple people can write in at once. If I would be head of ICT I would say: ‘this is not the way you should work in my hospital [..] everybody looks for their own solution for a system that is not ok”[D2Z]. As mentioned before, the doctors use

(23)

23

have a (slightly) different way of working with the system, as argued by a doctor: “Every doctor uses it in a different way. We have not made any agreements on how to use the system. Some write the things that have happened to the patient on top, while others write it on the bottom. There is no uniformity.” [D1Z]. The system also offers templates, although the doctors disagree on its usefulness, since it takes time to install them. For instance, one of the doctors argues: “We do have templates, but I find that like

baking pancakes on a campfire, maybe if you have a lot of spare time and nothing to do. If you need to know something quick it is too much of a hassle.”[D2X]. This lack of uniform use, offers challenges

for the patients’ care, since every doctor can report the information in a different way. Finally, the nurses and medical administration use the system more as a ‘show box’, so their system use is more or less the same.

There were also some differences between the departments. For department ‘Z’, the nurses and doctors were frustrated that PoliPlus could not cooperate with the other two systems that they need in order to operate, as argued by a doctor: “The system should be supporting the process and prevent

mistakes, but it is far from doing that.” [D2Z]. Working with these additional systems was also time

consuming. The doctors at ‘X’ agree that the system has a lot of annoyances, but that it still is an addition. So, it seems that the system is of less support for department ‘Z’ than for ‘X’. Furthermore, all three groups at ‘Z’ still write information down in a paper file (i.e. workarounds), as acknowledged by a nurse: “We not only work with PoliPlus, we also still work in the paper files. That is difficult

sometimes; one doctor puts it in PoliPlus, while the other puts it in the paper file.” [N2Z]. For

department ‘X’ it seems that only the nurses still work with their paper file.

Team interaction

The interaction between and within the teams consists of three categories: inter-team interaction, face-to-face communication and the work climate.

Inter-team interaction

There seem to be two within-case similarities. First, it seems that there is an absence of feedback between teams about the way the system can be used better/differently. Most agree that they would be open to feedback, but that it never happens, since everybody has their own part in PoliPlus and sticks to that. Secondly, the nurses and doctors appear to be able to communicate with the other two disciplines via PoliPlus, while the administrative personnel can only read the messages from the doctors, since they are not able to reply via PoliPlus.

Then, there are quite some differences between the two departments. For instance, the communication between the different teams happens increasingly via PoliPlus, especially for department ‘X’. The communication of the doctors from ‘X’, with the medical administration, happens mostly via the system, as argued by a doctor: “That [communication] has become different. We still

(24)

24

increase in the system communication, there are also less contact moments between the doctors and the medical administration at department ‘X’, as argued by one of the doctors: “Back then, I still had

to go to the medical administration to write my letters and to consult what to do” [D2X]. For

department ‘Z’, the doctors still see the medical administration on a regular basis, since they still use the paper files, as mentioned by a doctor: “Since we still work with a paper file, we still drop regularly

at the medical administration.”[D2Z]. The nurses appear to speak less with the medical

administration, but still as much with the doctors, since they have to consult about the patients. This is another reason for the doctors at department ‘Z’ to keep in touch with the medical administration, as argued by one from the medical administration: “We have a lot of consultation moments, also with

doctors and nurses [..] When there is an emergency, you beep the doctor and then consult with him.”[A2Z]. So their role seems somewhat different from the role of their colleagues at department

‘X’.

Face-to-face communication

For the intra-team interaction, the teams still depend on face-to-face communication; the system is used more as a way to secure information about a patient. As argued by one employee from one medical administrator: “We work with multiple colleagues, when I do something today, I should be

careful that my colleague does not do the same thing tomorrow, so that is why I put it in the digital patient list.”[A3Z]. Especially in a dynamic environment like a hospital, decisions have to be made

quickly and, according to both departments, face-to-face communication is then required, as argued by a doctor: “For every acute moment, you need to stick your heads together.”[D1Z]. This is acknowledged by the nurses, who seem to communicate face-to-face with the doctors and, somewhat less, with the medical administration.

This changing role of the medical administration, regarding the face-to-face communication, is also the only large difference between both departments. For department ‘X’ the medical administration is not consulted. This face-to-face communication is therefore being missed by the medical administration at ‘X’, as argued by one of its employees: “They [doctors] should walk by

more often. You treat each other differently when you speak to each other face-to-face, instead of sending a message via a system.” [A1X]. The medical administration at ‘Z’ still speaks face-to-face

with the doctor, due to the paper files, and this is appreciated by both ends, as reinforced by a doctor:

“It is necessary to speak each other face-to-face, since this is important for your process of working.”

[D2Z]. Work climate

Regarding the work climate, there only seem to be differences between the departments. The working climate appears to be more formal at ‘X’, especially regarding the relationship between the doctors and medical administration, this is acknowledged by one of the medical administrators: “It has become

(25)

25

why the atmosphere in our team is sometimes not too good anymore.”[A2X]. This is because the

doctors do not visit the medical administration as much as they used to. Due to this diminishing person contact, both parties find it increasingly difficult to speak about non-work related topics. For ‘Z’ the climate is still more informal, since there is regular (physical) contact between the teams, yet this is also slowly changing due to the digitalization.

Distance also appears to influence the way the teams interact. At department ‘Z’, the office of the medical administration is situated close to the policlinic, which is acknowledged by one of the employees: “You have the consultation hours in the offices here [..] doctors walk in and out to get

things done, so communication goes really easy. There is close contact and I have the feeling that when our department would be somewhere else, it would be less personal and more via the phone.”[A2Z]. For ‘X’ however, the medical administration is situated somewhat remote, as pointed

out by a doctor: “They are quite remotely located. They are at the end of the hallway, so naturally I drop by less often.”[D1X]. Both departments acknowledge that due to the digitalization the distance

between the doctors and medical administration is (‘X’) or will become (‘Z’) larger.

Team sensemaking

During the interviews, it became clear that the different teams experienced a different team mental model via their team sensemaking, which took place during the team sensemaking moments.

Team mental model

It seemed that all groups had their own shared cognition due to their sensemaking. The nurses appeared to have a shared cognition regarding the system use. According to one of the nurses, this influences her way of working and the groups cohesiveness: “If your team works in a certain way, it is

hard not do the same [..] to that extent, you are dependent upon what your team wants.”[N2X]. The

medical administration was quite positive about the system. It also appeared that this shared cognition, ensured that their group cohesiveness was large and that they would help each other out, as one of the mentioned: “I do not mind doing something for the doctor every once in a while, but I think that in

general all functionalities of a certain discipline should stay with that group [..] look who’s job it is to do it and that group should find their own solutions.”[A3Z]. Finally, most doctors shared a negative

opinion about the system, since they encounter many problems when using the system, as argued by one of them: “I think a lot of my colleagues [doctors] will say: ‘this is the system we work with, we no

longer bother about it, we just work it’. I do not think that there is anybody who is positive about the system.”[D2Z]. For them, the system sometimes prevents them from doing their job (i.e. not being

able to share information with other disciplines and errors). In this, they have a ‘common’ enemy, which in a certain way, strengthens their cohesiveness.

(26)

26

departments, as reinforced by one of them: “It is true that when everybody is negative in your team,

you are influenced by this. That is also because we have such a close group. Sometimes that is a pity, since you do not dare to say that you think different about some things.” [A2X]. In addition, the

doctors from ‘X’, were somewhat positive about the system, one of the doctors even seemed really positive, while the other felt that there were still quite some annoyances. For department ‘Z’ the doctors were more annoyed, mainly because they had to work with two other systems next to PoliPlus, which was more time consuming.

Team sensemaking moments

It seemed that all groups discussed their issues during their weekly meetings, especially the doctors and to a lesser extent the medical administration. For the nurses, their meetings were reserved for discussing patients, rather than the system, since they are not working with PoliPlus on a regular basis. Finally, it appears that the inter-team learning about the system only happened in the ‘hallways’, as argued by a doctor: “That [learning from each other] has to happen in the hallways, then you will pick

it up.”[D1Z].

The main difference between both departments regarding the team sensemaking moments, are the multi-disciplinary meetings. At department ‘X’ there is a weekly multidisciplinary meeting, where every team is represented and different topics are discussed (e.g. planning of patients). Department ‘Z’ does not have these multi-disciplinary meetings, only meetings between doctors and nurses about the patients, and this is being missed, as argued by one of the nurses: “I think it would be good to sit

around the table with multiple disciplines and discuss certain problems you encounter. In that way we can see how everybody thinks of it.”[N1Z].

Patient care

In theory, the health information system supports the patient care and it seemed in practice that three variables determine whether the system supports the care for patients: the changing primary process, the urgency and the risks.

Changing primary process and urgency

For the medical administration the system removed the control function that they had before. Due to this, there is no final check on the letters for any errors and lab results are sometimes not communicated towards the patient. This possibility of errors is acknowledged by the doctors who also argue that they are not the ones who are educated to write letters. However, they agree that the system supports the process by having all information in one place and that forms are filled in correctly and readable, which can be a problem on paper: “We used to write on paper, where you had to put a

patient number on top. This consists of seven numbers, so you can imagine how easy it was to make mistakes in this. With the new form this is not possible anymore [..] that is the benefit of digitalization, on paper you could write something unreadable or misplace an X.”[D2X]. The nurses agree that

(27)

27

that there are less misunderstanding and more overview, since face-to-face communication can sometimes be interpreted differently. All in all, face-to-face communication is essential in a hospital, especially for the doctors and nurses, since it is an acute job where time is of the essence. For example, three out of four doctors agree that the system can take up too much time due to errors, slow functionalities or typing during a consult. This time [often minutes] is subtracted from the limited time the doctors have to spend on the patient, as mentioned by a doctor: “In an acute job as ours, this

[system not functioning] bothers us a lot. Even patients sometimes say: ‘is it not working again?”[D1Z].

The main difference between the departments resides with the doctors. The doctors at ‘Z’ argue that there is a lack of overview, due to the separate systems, the paper files and the increased workload: There is no overview [..] Everybody just copies and pastes. If there is a message saying that

the patient is crazy, then it will be in her file, the rest of her life.[D1Z]. The doctors feel like this lack

of overview and control is a serious threat for the safety of patients, for instance when they are not reminded by the system of certain lab results. For department ‘X’, this is less an issue, since the doctors only work with PoliPlus and not with other, separate systems or paper files.

Risk

All three groups argue that they rely upon the system, which can lead to risks. For instance, one doctor skips her morning meetings in order to start early with PoliPlus, in case some functionalities would not work. This can be a risk, since the doctor misses certain, maybe viable, information about patients. The medical administration share this opinion and agree that doctors spend too much time on administrative tasks and that face-to-face communication works a lot faster when you are in a hurry, as mentioned by one of them: “Last night, someone sent me a ‘check list’ [in Dutch: ‘bolletjesformulier’]

that said ‘urgent’. I had to do something that morning, while I was on the phone all morning. When you would stand beside me at eight in the morning, to ask me to call a patient, it could be solved straight away.”[A2X]. In addition, the nurses argue that there is a lot of reliance on the system, which

sometimes can be a risk (e.g. when someone enters the wrong information, it will be in the system forever).

The system can also be a burden for some doctors regarding privacy, since the information is widely shared amongst different disciplines, some information can be shared that may not be shared, as argued by one doctor: “I am all for privacy and the thing here is, when you have someone who is

abused [..] psychiatrists sends me a letter with her information [..] I cannot scan this letter just for my specialism; I can only scan it publicly, which means that once this woman comes in with a broken leg, the emergency room immediately sees this information. That is why I keep these letters in a paper file.”[D2Z]. The nurses also acknowledge this risk. One of the nurses and a doctor stressed that one

(28)

28

increases the risks of people looking for information about patients they do not require. This can become a bigger issue when the hospital switches to an ERP that shares information nation-wide.

Finally, the main difference between the departments is regarding their ideas about the digitalization. The medical administration at department ‘Z’, realizes that there is more and more reliance on the system, which makes them fear for their job. Department ‘X’ is already in a somewhat advanced stage when it comes to the digitalization, which makes them less fearful about it (i.e. they already passed this phase). Furthermore, the nurses at ‘X’ agree that the system works faster than a paper file, which saves time, yet it also forces them to spend more time behind a computer and less time on a patient, whereas the nurses at department ‘Z’ seem to work less with the system.

DISCUSSION AND CONCLUSION

In this section, the most important findings will be highlighted, discussed and compared with existing literature. For a clear overview, see figure 3. Afterwards, the theoretical and practical implications will be presented, followed by the limitations, future research and finally the conclusion.

Discussion

First, the study shows that face-to-face communication is still a necessity for the care of patients, as argued by Edmondson et al. (2001). Doctors and nurses have to consult about patients, while the system is used to store what is consulted. However, for the medical administration, face-to-face communication only improves the overall process, but does not affect patient care, which was not accounted for by Edmonson et al. (2001). The research also shows that due to the system there is less physical communication, especially between doctors and medical administration, which bothers the medical administration and makes them more skeptic about the HIS. This reduced communication confirms Jordan et al. (2009) and Gattiker and Goodhue (2005), who argue that relationships between teams are the primary source of system functioning and when groups have different opinions about the system and when these are not discussed, it leads to resistance and skepticism. Furthermore, the study shows that distance plays a major role in the quantity of face-to-face communication: the medical administration of department ‘Z’ was located near the policlinic, while their colleagues of department ‘X’ were more isolated. This adds to Brown and Duguid (1991), who argued that it is difficult to learn from others or interact, when you are distanced from each other. Due to this distance, feedback from other teams about the use did not or seldom happen; everybody focused on their part of the system. So, a HIS does not always allow for the integration of teams and cross-disciplinary functions, as argued by Raghupati and Tan (2008) and Chiason et al. (2007), especially when teams are distanced from each other.

The Health information system and its consequences for team interaction