EHR Workarounds: Rise and Unfolding

(1)

EHR Workarounds: Rise and Unfolding

A user perspective on working around Electronic Health Record

systems: from reasons why to expected consequences

Master’s Thesis, MSc BA Change Management, University of Groningen, Faculty of

Economics and Business

Date: 20-1-2019

Supervisor: Dr. J. F. J. Vos Co-assessor: Prof. Dr. A. Boonstra

Tess Jonker (S2684934) T.l.jonker.1@student.rug.nl

(2)

1

ABSTRACT

Medical organisations increase their use of IT in the form of Electronic Health Record (EHR) systems. To overcome perceived hick-ups in workflow resulting from these systems, caregivers engage in work practices other than prescribed by the EHR, known as workarounds. This study adopted a user perspective to analyse workarounds as a response to EHR systems. Thereby, the objective of this study was to unravel the pathways of EHR workarounds from a user perspective. These pathways contain the underlying reasons of users to develop workarounds, and the consequences – what and for whom – users expect to follow from these workarounds. By means of qualitative analysis at the Internal Medicine- and EHR support departments of a large Dutch teaching hospital, a set of workarounds was identified. The results reveal 11 workarounds, which were predominantly performed by physicians. The workarounds identified in this paper were categorised as either performed while working with the system (in-system workflow sequence workarounds or in-system data entry workarounds), or by bypassing the system (out-system workarounds). Firstly, in-system workflow sequence workarounds were performed because of user- and system-related reasons, and were expected to result in benefits for the user and expected risks for the patient and the hospital as a whole. Secondly, in-system data entry workarounds were created because of system-, organisation-, and user-related reasons, were expected to result in benefits for the user and expected risks for the user, the system, the patient, and the hospital as a whole. Lastly, out-system workarounds were created because of system-, organisation-, and user-related reasons, were expected to have benefits for the user, the system, and the patient, and expected risks for the user and the patient. Based on the findings a model was designed to visualise the pathways of workarounds.

(3)

2

1. INTRODUCTION

Although enterprise systems are increasingly adopted in healthcare organisations, reaping the benefits that are supposed to follow from these systems continues to be challenging (Blijleven, Koelemeijer, Wetzels & Jaspers, 2017b). Enterprise systems are designed to effectively combine a number of department-specific programs into one, company-wide software application (Markus & Tanis, 2000). Yet, these systems are often characterised by a high complexity of operability, which limits its supporting principle as perceived by its users (Davenport, 1998). In healthcare, organisations increasingly adopt enterprise systems in the form of Electronic Health Record (EHR) systems (Van Den Hooff & Hafkamp, 2018). Interestingly, once employees feel an EHR system constrains and not supports their activities, they create so-called workarounds: they find ways to bypass these perceived constraints by adapting the EHR system or their use of it (Van Den Hooff & Hafkamp, 2018). Workarounds are adjustments of practices imposed by the system, temporarily or over a more lengthy timeframe (Alter, 2014). Due to the complexity of a hospital’s operative environments, workarounds are especially common in these settings (Zhou, Ackerman & Zheng, 2011).

However, despite their commonality, there is a need for expanding the understanding of workarounds (Blijleven, Koelemeijer & Jaspers, 2017a). This increase in understanding can be achieved by additional research on the reasons for workarounds to develop, as well as the consequences that go with them (Blijleven et al., 2017b; Friedman, Crosson, Howard, Clark, Pellerano, Karsh, Crabtree, Jaén & Cohen, 2014). First, earlier studies explained how workarounds are generally formed when users integrate with technology, as a response to conflicting goals and workflow bottlenecks (Alter, 2014; Blijleven et al., 2017b; Zhou et al., 2011). With that, to be able to execute a workaround, users should have the opportunity to adjust their prescribed work practices (Patterson, 2018). However, when considering what specific reasons lead to specific workarounds, research seems to be lacking (Van Den Hooff & Hafkamp, 2018). Understanding why workarounds are created is of critical importance for facilitating the often difficult decision of how to respond to workarounds (van de Weerd, Vollers, Beerepoot, & Fantinato, 2019). Second, the overall view on workarounds raises questions on the potential consequences they bring about. Whereas some researchers distinguish between beneficial and precarious workarounds (Ferneley & Sobreperez, 2006; Friedman et al., 2014), others describe how a single workaround can have both positive and negative implications (Beerepoot, Ouali, van de Weerd, & Reijers, 2019).This means that the relations between specific workarounds and specific consequences seem to be an area in need of additional research as well (Patterson, 2018).

(6)

5 system for failing, as it imposes work processes that are not in line with the way of working these users are accustomed to (Miller & Sim, 2004). To deepen the understanding of what makes caregivers engage in workarounds and to analyse whether they are aware of the related consequences, a user perspective is required. Miller and Sim (2004) argue that it is the caregivers’ attitude that plays a major role in the improvement of the alignment between a system and the organisation. This means that understanding the operability of the EHR through the eyes of its users enhances clarity of workarounds. Therefore, this study applies the user perspective to improve in terms of reasons for developing these workarounds and possible consequences.

Through unravelling the reasons and consequences of specific workarounds, this paper develops the pathways of workarounds containing their reason to rise and how they unfold. Thereby, we aim at finding an answer to the following research questions: What workarounds are created in EHR contexts? and Looking from a user perspective, how do these workarounds rise and unfold? To answer these questions, qualitative research was performed in a large Dutch academic hospital that implemented an EHR system two years prior to this research. Through interviews and observations within the Internal Medicine Department and the system’s support department, several workarounds were detected of which their underlying espoused reasons and possible consequences were unravelled.

The study’s contributions are threefold. First, this research adds to the current body of literature on EHR systems by improving an understanding of why EHR workarounds are adopted (Van Den Hooff & Hafkamp, 2018). Second, this research unravels the pathways of workarounds containing reasons for rise and expected consequences. Thereby, this study helps to increase the understanding of workarounds by systems designers, which in turn assists them in creating better fitting EHR systems, reducing the detrimental and costly inefficiencies they often bring about (Anthony, Campbell, Candon, Gettinger, Kotz, Marsch, Molina-Markham, Page, Smith, Gunter & Johnson, 2013; Seaman & Erlen, 2015). Third, uncovering the effects related to specific workarounds provides a basis for the decision-making on whether to discourage, temporarily encourage, or permanently encourage workarounds (Patterson, 2018).

(7)

6

2. THEORECTICAL BACKGROUND

The aim of this study is to clarify the mechanisms inherent in the performance of workarounds. Its line of argumentation follows Walker and Avant’s (2004) approach of concept analysis as reviewed by Butcher, Kagan, and Whall (2006). Concept analysis clarifies a phenomenon by identifying the uses of the concept, determining the defining attributes, and identifying reasons and consequences (Butcher et al., 2006). Morgan and Yoder (2012) explain the steps inherent in concept analysis as follows. First, to thoroughly define the concept, an overview of a variety of known definitions should be outlined. Second, the attributes that are most often connected to the concept should be identified. Third, an analysis of the concept’s underlying reasons show the circumstances preceding it. Last, the concept’s consequences show the resulting circumstances. Therefore, to create an overall understanding of workarounds, this part starts by uncovering definitions of workarounds in a variety of enterprise systems. Thereafter, this paper further confines to workarounds in Electronic Health Record (EHR) systems, to explicate workaround attributes, the known reasons for workaround creation and acknowledged consequences workarounds bring about.

2.1 The general concept of workarounds

(8)

7 As presented by Alter’s (2014) literature review, numerous definitions have been assigned to the concept of workarounds. Workarounds are closely linked to the work system of the organisation and the specific technology it implements (Alter, 2014). Therefore, one can expect that the number of definitions presented by literature is driven by the desire to fit each definition to the technological context in which the workarounds were studied. Nevertheless, as illustrated by Table 1, taking a closer look at these definitions shows only small differences in contents. In essence, they all cover a system use that deviates from the way it was prescribed, as a response to a perceived confusion or hick-up in workflow. This implies a shared agreement on the meaning of workarounds.

Table 1: Examples of workarounds for various types of enterprise systems

SYSTEM DEFINITION EXAMPLE

Enterprise Resource Planning (ERP)

System

“[A response to] ERP systems in order to counteract the perceived loss of power and identity arising from the use of the system, or trying to re-enact previous working practices prior the introduction of the system” (Ignatiadis & Nandhakumar, 2009, p. 5).

Make non-mandatory fields in SAP appear mandatory to avoid system errors (Ignatiadis & Nandhakumar, 2009)

Not logging off in the system to spare time (Ignatiadis & Nandhakumar, 2009) Supply Chain

Management (SCM) System

“[W]orkarounds are manifestations of employees disengagement from the monitoring technology” (Sobreperez, Ferneley & Wilson, 2005, p. 5)

Using personal data to create forecasting statistics (Röder, Wiesche, Schermann & Krcmar, 2014)

Create an overview by downloading data to a separate Excel file (Li, Haake & Mueller, 2017) Customer Relations

Management (CRM) Systems

“[W]orkarounds were expedient measures that agents

developed for dealing with the contradictory

requirements of using the new technology while fulfilling customer expectations” (Russell, 2007, p. 142)

Cheating or lying to the system when exposed to inappropriate screens (Russell, 2007)

Electronic Health Record (EHR)

Systems

“Workarounds are observed or described behaviours that may differ from organisationally prescribed or intended procedures. They circumvent or temporarily ‘fix’ an evident or perceived workflow hindrance in order to meet a goal or to achieve it more readily” (Debono & Braithwaite, 2017, p. 2)

Manually entering instead of importing patient data into a letter (Blijleven et al., 2017b) Not validating a medication order with the physician but rather with other nurses to avoid

fallout by the physician (Debono &

Braithwaite, 2017)

(9)

8 individual well-being: they could even make a difference between life and death (Halbesleben, Wakefield & Wakefield, 2008). This is further illustrated by the following examples of EHR workarounds: not checking the system for medication verification (Azad & King, 2008; Koppel et al., 2008; Van Den Hooff & Hafkamp, 2018; Zhou et al., 2011) dispensing medication before an order is confirmed in the system (Van Den Hooff & Hafkamp, 2018), and placing copies of patient IDs in several places to bypass the scanning of patient ID wristbands (Koppel et al., 2008). Studying workarounds in EHR contexts is especially interesting when acknowledging this difference in implications with workarounds performed in other systems.

2.2 EHR workarounds: definition and attributes

The second step of a concept analysis is to evaluate attributes involved with EHR workarounds. The goal of identifying attributes of EHR workarounds is to help clarify this phenomenon and to differentiate EHR workarounds from related concepts by elaborating on their key characteristics as well as characteristics that are not involved with EHR workarounds. This collective understanding is fundamental for explaining the mechanisms underlying EHR workarounds (Gibson, 1991). Like other enterprise systems, EHRs seem to be affected by the workarounds performed by its users. EHR systems are designed to comply with the pressures for increased efficiency and decreased expenditures that currently characterise the medical industry (Chandra, He, Liu & Ruohonen, 2013). To realise this, EHR systems aim to increase information sharing (Chandra et al., 2013), improve patient care (Van Den Hooff & Hafkamp, 2018), and enhance the quality of patient safety (Azad & King, 2008). This study adopts the following definition for EHR workarounds:

“Workarounds are observed or described behaviours that may differ from organisationally prescribed or intended procedures. They circumvent or temporarily ‘fix’ an evident or perceived workflow hindrance in order to meet a goal or to achieve it more readily” (Debono & Braithwaite, 2017, p. 2).

(10)

9 adaptations developed prior to the implementation of the EHR, or one-time adaptations performed only by mistake. Comparing the attributes of workarounds with how workarounds are previously defined, the definition by Debono and Braithwaite (2017) covers all aspects.

2.3 EHR workarounds: underlying reasons

The third step of a concept analysis is to explicate why workarounds are created. Where some reasons for workarounds are related to the system’s user, other reasons are related to the system itself or the organisation. Besides, workarounds can also be driven by a combination of these (Ferneley & Sobreperez, 2006). First, among user-related reasons are a lack of know-how with the specific EHR system or insufficient skills related to computerisation in general (Flanagan et al., 2013), resistance to the system (Ferneley & Sobreperez, 2006), a lack of time or a lack of motivation to execute the task at hand as desired by the EHR (Friedman et al., 2014). To illustrate, Flanagan et al. (2013) describe a situation in which a caregiver was unable to find a patient’s past test results in the system. The workaround this caregiver created as a response was to have the patient fax the test results instead. On the one hand, this illustration shows how insufficient knowledge or skills to efficiently work with the EHR system can lead to finding alternative solutions to continue workflow and support actual or perceived efficiency (McGann & Lyytinen, 2008). As this specific caregiver was unable to find information in the EHR, a workaround was created (Flanagan et al., 2013). On the other hand, the caregiver might also resist searching the system, for example to reduce the amount of time spent on this activity (Ferneley & Sobreperez, 2006). By having the test results faxed, the caregiver could spend the time he would otherwise spend on searching the system on a different task (Flanagan et al., 2013). From this illustration it becomes clear that workarounds do not necessarily have one single reason underlying its performance (Ferneley & Sobreperez, 2006).

(11)

10 that not only can workarounds be developed as a response to multiple reasons, these reasons could also be related to different domains (i.e. user- and system-related reasons).

The third type of reasons underlying workarounds are those related to the organisation’s facilities. The conditions of the physical environment of a caregiver’s workspace influence the way these caregivers work with an EHR system (Blijleven et al., 2017b; Rack, Dudjak, & Wolf, 2012). Examples of such conditions are light, sound, design, or temperature. To illustrate, Koppel et al. (2008) observed nurses taking a scanner from its cart to scan barcodes, as a response to the heavy noise present on the hallway the cart was positioned. Another example they show is how a lack of wireless connection of a scanner led to preparing, scanning, and transporting medications at a time that deviates from the time of administration (Koppel et al., 2008).

Based on this literature review, user-, system-, and organisation-related reasons (or a combination thereof) create conflicting goals, workflow bottlenecks, opportunity for development, and the integration of humans and technology, leading to the performance of workarounds (Patterson, 2018). The next section discusses possible consequences of these workarounds.

2.4 EHR workarounds: related consequences

The final step of a concept analysis is to find the consequences of workarounds. Several authors classified workarounds based on their consequences. For example, Ferneley and Sobreperez (2006) categorised workarounds as either essential, hindrance, or harmless. Another example is the typology by Friedman et al. (2014) who distinguished between temporary vs routinized, unavoidable vs avoidable, unplanned vs deliberate workarounds. By creating typologies, the authors make a distinction between hazardous and beneficial workarounds.

(12)

11 patient data that is unstable, unavailable, and unreliable (Kobayashi, Fussell, Xiao, & Seagull, 2005). As other departments rely on this data, workarounds in one part of the hospital potentially lead to negative implications for patient safety in other parts (Boudreau & Robey, 2005).

On the other hand, looking from a more positive perspective shows that workarounds identify system misfits and improve workflow. First, workarounds potentially show areas where the EHR system’s capabilities are not aligned with the organisation’s needs. It is through this that workarounds can be used to locate misfits between the EHR and the organisation and can thus be considered as a source for potential system improvements (Alter, 2014). To illustrate, writing allergy-related patient data in a separate note field instead of selecting one of the data options provided by the EHR system can indicate that the required data option is missing (Blijleven et al., 2017b). When caregivers voice such a workaround, system developers become increasingly aware of missing elements and can build-in the necessary data entry accordingly. In this way, workarounds contribute to a better fit between the EHR and the hospital (Van Den Hooff & Hafkamp, 2018). Second, by accepting a workaround as a method for a rare quick fix (Truex, Baskerville & Travis, 2000) or creative use (Ash, Berg & Coiera, 2004), workarounds actively contribute to a continuous workflow (Koppel et al., 2008). Caregivers can decide to perform workarounds in emergent situations if the EHR system does not recognise similar urgencies, to be able to continue the tasks they consider crucial (Kobayashi et al., 2005). For example, by providing a blood-needing patient in the Operation Room of a universally accepted blood and bypassing the EHR system’s requirements to provide only the correct blood type, because the patient’s blood type is unknown (Kobayashi et al., 2005). In this way workarounds are essential for saving lives, property or circumstances (Ferneley & Sobreperez, 2006), and contribute to time savings, decreases in medication failures, and compliance (Yang, Ng, Kankanhalli & Luen Yip, 2012).

(13)

12

2.5 The pathways from a caregiver’s perspective

(14)

13

3. METHODS

3.1 Research approach

Because we seek to increase the understanding of workarounds in terms of why users engage in specific workarounds and what consequences they expect their actions to bring about, we need to study workarounds from the caregiver’s point of view. Therefore, a qualitative research method is preferred over a quantitative one (Glaser & Strauss, 1965). To uncover the pathways of workarounds, a qualitative case study was performed. Case studies are usually performed to explore a phenomenon that is not (frequently) studied before. However, in some situations case studies “provide freshness in perspective to an already researched topic” (Eisenhardt, 1989, p. 548). In this light, case studies allow for a more comprehensive understanding of complex phenomena, by studying real-life phenomena through contextual analysis (Zainal, 2007). As case studies are a common method to study conditions and relationships, this method is highly suitable to explore the pathways of workarounds. More specifically, we chose to do an exploratory case study (Yin, 1981). We do not expect a single set of reasons for creating workarounds, nor do we predict a unified bundle of consequences of these workarounds. This means that we are actively exploring the concept of workarounds following the exploratory case study method as described by Yin (1981). Through triangulation, validity of the research was ensured. By reviewing literature, analysing documents, performing observations, and conducting interviews, potential biases in each of these methods were accounted for (Golafshani, 2003).

3.2 Empirical setting

The setting of this field study is a large Dutch teaching hospital, that counts approximately 14,000 employees. In December 2017, the hospital had purchased a new EHR system. Before this system was in place, each department operated with its own applications. The system was introduced through a Big Bang approach, meaning the hospital introduced the EHR and terminated the legacy system simultaneously (Vitari & Ologeanu-Taddei, 2018). At the moment this research was undertaken, two years after the EHR was introduced, the project was in its post-implementation phase. As workarounds are a post-implementation phenomenon (Bozan & Berger, 2018), this hospital makes up an appropriate research setting.

(15)

14 seem a good fit for explicating reasons for workarounds. Second, with the introduction of the EHR, the hospital initiated an EHR support department, comprising 33 support teams and an optimisation team. This department is created for the hospital’s employees to send all their EHR-related issues to and together, these teams are responsible for solving these requests. As workarounds are a response to issues with the EHR system, we expect the support department to be highly valuable in this research.

3.3 Data collection

The participants were selected following the purposeful sampling method (Palinkas, Horwitz, Green, Wisdom, Duan & Hoagwood, 2015). Thereby, the group of participants of this research complied to the following criteria: (1) all participants work either in the EHR support team or in the diagnostic department; (2) all types professions that work with the EHR on a daily basis are represented in the study sample; (3) the total body of participants is made up of a diversity in expertise, within the departments described in criterion 1. To maintain their privacy, all participants of this study were anonymised by codes.

Preliminary to the data collection, we conducted two exploratory interviews, one exploratory observation, and a document analysis to obtain full understanding of the hospital’s structure and progress in terms of EHR implementation. First, interviews were held with two members of the EHR support department. Second, a staff meeting by the diagnostic departments was observed. During this meeting, the present staff members were introduced to the EHR optimisation team and their upcoming support sessions. These sessions are one-on-one scheduled meetings designed to help the employees in this hospital improve their operations with the EHR. Third, document analysis was performed on the hospital’s governance for data registration. This explicit set of agreements on how to correctly work with the EHR served as a benchmark for identifying workarounds during the data collection.

(16)

15 Because these observations gave insight into the potential sources for workarounds by all employees, we took considerably more time during this observation. Table 3 shows an overview of the observed employees. With each observation, field notes were taken according to an observation scheme (Appendix A) and extensively summarised afterwards.

Table 3: Details participants observations

CODE POSITION EXPERTISE DURATION

(total 490 min)

O_PH1 Physician Acute medicine 60 min

O_PH2 Physician Nephrologist 60 min

O_PH3 Physician Head of department 70 min

O_PH4 Physician Nephrologist 60 min

O_SS1 Support staff Solution Centre 120 min

O_SS2 Support staff Training facilitator 120 min

Alongside the observations, a total of 17 employees was interviewed (Table 4). The interviews were conducted with members of the EHR support team and the hospital’s help desk, physicians, nurses, and administrative personnel. Interviewees were contacted by means of snowball sampling (Noy, 2008) to avoid possible biases. With that, a variety in interviewees in terms of opinion about the EHR and occupation was deliberately requested to control for both the potential of respondent’s bias and the possibility that some workarounds are occupation-specific. Of these 17 interviews, four were follow-ups with the employees observed earlier. Consequently, the first part of these interviews were based on the observations to verify the conducted information and ask for clarifications and reasons for the identified workarounds. All interviews were structured to leave room for probing. The semi-structured approach provided the opportunity to not only cover the central themes of this study, but also enabled the opportunity to discover aspects that were not determined primarily. After 17 interviews, no new information was retrieved, which means that saturation of the data collection was achieved (Glaser & Strauss, 1965).

(17)

16 questions (Burns et al., 2015). The third and final part of the interview is allocated to get in touch with the interviewee’s colleagues for further interviews. By this, we comply to the snowball approach as explained above (Noy, 2008). All interviews were recorded and transcribed.

To ensure reliability, a number of aspects were taken into account. First, all interviews and observations were done at different times of the day and on different days of the week. Second, field notes were taken during the data collection. These notes were used as a remembrance for the researchers to refer back to in further courses of the interview. This increased the depth of the interviews and clarified possible links embedded in the interviewees’ responses. Also, the field notes were combined and used as a supporting mechanism in the coding process, making sure any first thoughts during the interviews were taken into consideration when examining the data at a later stage in this research. Third, the interview transcripts and observation summaries were verified by the participants. All data was gathered in the native language of each observant and interviewee, either English or Dutch. The related transcripts and summaries were documented in the same language as collected. Last, the identified results were discussed and confirmed during reflective interviews, meetings, and presentations.

Table 4: Details participants interviews

CODE POSITION EXPERTISE DURATION

(total 715 min)

I_SS2 Support Staff Training facilitator 45 min

I_SS3 Support Staff Optimisation team 60 min

I_SS4 Support Staff Optimisation team 60 min

I_SS5 Support Staff Order complications 30 min

I_SS6 Support Staff Patient registration complications 20 min

I_HD1 Help desk Manager IT Help desk 40 min

I_PH1 Physician Acute medicine 35 min

I_PH2 Physician Nephrologist 40 min

I_PH3 Physician Head of Department, Researcher 60 min

I_PH4 Physician Nephrologist 45 min

I_PH5 Physician Acute Medicine 30 min

I_NU1 Nurse Flex worker 20 min

I_NU2 Nurse Nurse 25 min

I_NU3 Nurse Head Nurse 20 min

I_MMA1 Medical Administrator EHR core team 45 min

I_MMA2 Medical Administrator Kidney transplants 25 min

(18)

17

3.4 Data analysis

(19)

18 Figure 1: Data structure

Note: Letters with benefits and risks indicate to whom the users think these consequences apply. U = user; P = patient;

(20)

19

4. RESULTS

Based on the data analysis, 11 workarounds were identified. To unravel how the users view the reasons for, and related possible consequences of these workarounds, this section is divided into three parts. Each part represents a type of workaround as identified in the data. As explained in the method section, the existing typologies of workarounds seemed unsuitable for this research. Therefore, based on the data, a new typology was created. With this typology, a distinction is made between workarounds that are performed while working within the system (in-system workarounds) and workarounds that are performed through means other than the EHR, to accompany or bypass the system’s practices (out-system workarounds). In-(out-system workarounds are further segregated into workarounds that are related to the sequence of steps in a caregiver’s workflow (workflow sequence workarounds) and workarounds that are specifically related to the registration of data (data entry workarounds). The first part of this section covers two in-system workflow sequence workarounds; the second part covers five in-system data entry workarounds; the final part elaborates on four out-system workarounds.

Each part in this section is build up in the following structure. First, each workaround corresponding to the specific type is introduced. Second, we looked at the scope of each workaround. A workaround was marked as ‘performed’ when noticed by the researchers during an observation or when mentioned as performed by a participant during an interview. In case a workaround was mentioned by an interviewee with whom the workaround was detected during an observation, the density equals one. This is done to avoid a bias of duplication. Third the reasons for the performance of each workaround as espoused by the users is elaborated upon. Last, for each workaround, the consequences (positive and negative) are discussed. The data reveals that users expect the consequences of their workarounds to apply to themselves or other users, patients, the system, and the hospital. It is important to indicate that the benefits and risks as described are known or expected by the users, meaning that these effects could occur and not necessarily do occur.

4.1 In-system workflow sequence workarounds

The outcomes of the interviews and observations show two workarounds related to the workflow sequence imposed by the system: ignoring pop-ups and pre-starting a patient’s visit.

4.1.1 Ignoring pop-ups

(21)

20 “There are a few warnings that keep popping up, such as the one for allergy verification. These pop up so often,

I hardly notice them anymore. Imagine being in a conversation with a patient, and you want to look something up and that thing occurs on your screen. This means I have to interrupt my conversation to ask this question,

which I don’t want, so I close it. At a certain point you don’t even read what the pop-up says” (I_NU1). The pop-ups are part of the EHR’s warning system to impose the execution of tasks. However, the EHR seems to show pop-ups “in an unworkable and unnecessary frequency” (I_PH1), which is argued to slow down and be unsupportive of the work process. Physicians “simply don’t have the time during a consult to look at these things” (I_PH3). As a nurse disclosed: “Look, the problem is that it occurs all the time. If it would pop up twice a year there would be no built-up frustration and you’d think: Oh right, I do need to ask this again” (I_NU1).

An expected benefit of ignoring pop-ups affects the user, as a continuous workflow is preserved during a consultation: conversations do not have to be interrupted, and time is saved when not spending it on the announcement. What follows this workaround is an expected risk for the patient, as a “false sense of safety” (I_PH1) is created. The safety protocol built in the EHR is undermined as the nurses and physicians fail to read the warning signs. Besides, since closing pop-ups becomes a habit, there is a rising risk that the questions imposed by the pop-up are “never asked” (I_NU1), leading to jeopardising patient safety.

4.1.2 Pre-starting a patient’s visit

The second in-system workflow sequence workaround is starting a patient visit in the EHR, before the patient physically arrives to the appointment. This workaround was observed with two physicians and acknowledged by two other physicians during interviews. Despite the relatively low density in the data, one physician thinks “this happens a lot” (I_PH4). In preparation for a patient visit, many physicians want to pre-order treatments like blood tests to save time during the visit. The reason why physicians need to click the ‘start patient visit’ button, is because this act opens a new interface through which orders can be placed. Placing orders is only possible after the visit has started, “this functionality was not there before” (I_PH4). This means that the EHR is designed to ensure that tests and medications cannot be ordered by a physician before patients physically visit their consultation.

(22)

21

4.2 In-system data entry workarounds

Besides in-system workflow sequence workarounds, the data revealed a second type of in-system workarounds, namely those workarounds created with respect to entering data into the EHR. Five workarounds are related to this second type: copy-pasting, using separate text fields, leaving data fields empty, sharing login details, and registering false data.

4.2.1 Copy-pasting

The first in-system data entry workaround is copy-pasting data from one part in the system to the next, instead of registering data discretely in the system. For example, by selecting “a piece of text and copy-paste it into a letter” (I_PH4). This workaround was observed with two physicians and mentioned during interviews by three other physicians, three members of the support team, and one medical administrator. One reason underlying this workaround is the perception that “it will take too much time to register data discretely” (I_PH4). Another reason is that physicians are “scared because they are used to working in multiple systems and get nervous to register in one system and fear for losing data” (I_SS5).

This workaround is expected to benefit the user. Firstly, improved workflow and perceived time savings are achieved when copy-pasting, as opposed to discretely registering the data. Secondly, a better overview is created. As one medical administrator explained:

“Sometimes you don’t know all the medicines requested by patients, so I just take a look at their previous medications to make sure I have the right one […] then I can copy-paste that to the in-basket and let the physician prescribe it. […] Otherwise, I have to figure it all out myself, Google it and so on.” (I_MMA3).

Expected risks of copy-pasting are explained to influence the user, the system, and the hospital in terms of extra work, a loss of system support as the performers of this workaround miss out on the “convenience and support [the EHR] offers when registering discretely” (I_PH5), and the hinder it causes for conducting research from the data present in the EHR. As one physician explains: “you can create all kinds of reports if the data is registered discretely, you cannot do that if it is only plain text” (I_PH4).

4.2.2 Using separate text fields

(23)

22 “They are manually adding these patients, and they are using something called a speciality comment to just

write when they should discuss the patient next.” (I_SS2)

In some cases, underlying reason for the use of separate note fields is the lack functionality in the EHR, as it is “the only thing that is available now” (I_SS2). One nurse explained that if she wants to report a patient’s status, “you can only fill in a fixed number of words or characters” (I_NU2). The lack of space available to make notes makes her need “to add a separate note” (I_NU2). Another reason for this workaround is the lack of overview of the data for a particular physician or specialism: “[…] a problem list is a problem list, but I also see about fifteen other problems which I might not find interesting at all” (I_PH5). In this case, the separate note field is used to restructure issues in the problem list. A final reason is the lack of knowledge about the functionality of the system to click the right button. Therefore, “they just put it in a note” (I_SS2).

A possible benefit of this workaround is that through these separate notes, caregivers create a better overview for themselves in the form of a “good and concise summary of all relevant information” (I_PH5). The expected risks involved with this workaround affect the hospital and the user. Registering data in a separate text field hinders the possibility to generate data from the EHR for research purposes, as this information contains free text. A second risk is that the use of separate note fields leads to a potential loss of overview when applied extensively. As one physician explained: “It is a like a toilet paper roll from which all sheets are torn off, you have to read through 100 notes? 200 notes? Any patient experiencing the slightest inconvenience will have a note allocated to that” (I_PH3).

4.2.3 Leaving data fields empty

The third in-system data entry workaround is deliberately leaving data fields in the EHR empty, by not filling in specific patient information. This workaround was observed with one physician and mentioned during interviews with five other physicians. A reason for performing this workaround is the perceived restricting power the EHR has over the users as it demands certain information to be registered:

“I use [the EHR] in the way I want to use it, within certain boundaries. I appropriate a certain degree of freedom. […] There is always more than one way to skin a cat, and I know we have to work with [the EHR]. But

inside [the EHR], I believe everyone should be able to take their own paths, within the frameworks we set along the way.” (I_PH1)

(24)

23 “Yesterday I registered a patient that was already known by another specialism. This took me two hours. Two hours to order all this data in the EHR. Imagine that. Normally I would do that too, but I just wrote it down and

it cost me 10 minutes.” (I_PH3)

On the one hand, as becomes clear from the quote above, not filling in certain data fields in the EHR is expected to benefit the user as it saves the physicians time to spend on other activities. On the other hand, risks are expected for the system and the patient. One physician highlighted that the EHR “only works well if we all use it correctly” (I_PH5). Meaning that this workaround decreases the degree of support the EHR can provide. Moreover, not registering data leads to potential cascading errors in medicine prescriptions:

“Certain disciplines order an antibiotic treatment and do not include an end date, so it just stays in the system. The patient probably stopped taking the antibiotics a while ago, but the system says otherwise. This, in turn, has

consequences for my prescriptions” (I_PH1).

4.2.4 Sharing login details

The fourth in-system data entry workaround is sharing login details with other employees. This workaround was not observed and most interviewees denied sharing login details with colleagues. However, five members of the support team recognise this. One of them mentioned: “I find this very regretful, but I know these things happen” (I_SS4). Besides, a physician explained a situation in which this would occur:

“This is often through phone; the physician is busy elsewhere. I can imagine that when he picks up the phone and is asked ‘May this patient receive that medicine’, he might say: Just fix it because I’m busy” (I_PH5). Most interviewees explain that sharing login details is done either because of a lack of time or a lack of physical facilities, when “there is no computer at hand” (I_NU2).

All interviewees were hesitant about whether this workaround has possible benefits. One physician imagines “that it saves you a certain amount of work, but that does not weigh up against the risks” (I_PH5). While positive effects are hardly acknowledged, the interviewees did elaborately mention the risks of sharing login details with co-workers. One member of the support team highlighted the implications for patient safety:

“Assume that a person with insufficient knowledge orders or gives the wrong medication. Imagine a patient being allergic for that type of medication but receives it anyway. In the worst case, he dies. This can happen if

(25)

24 An expected risk for the user is that once login details are shared, there is no control over the potential abuse of these details “you never know if these details will be used again without your knowing” (I_PH1).

4.2.5 Entering false data

The final in-system data entry workaround is entering data that does not represent reality. This workaround was not observed, but explained by a physician and a nurse. As illustrated by the physician:

“I received the exact same pop-up as four weeks ago, to pay extra attention to the patient’s medication. Now I can’t continue, I need to enter new data into the system otherwise I cannot prescribe my medicines. So what

happens is, you are just going to make it up, you know, you just want to go on.” (I_PH1) The nurse mentioning this workaround explained she recognises it in the following situation: “even though I emptied the catheter bag at 11 pm, I enter this as emptied at 21.59” (I_NU1).

This workaround was performed deliberately to bypass the perceived restricting power the system has over the work process to avoid a delay in ordering medicines. As stated: “[The EHR] should not rule over us, it should help us. I mean, we need to use [the EHR], but it should not be the case that it imposes how I should do my job” (I_PH1). The nurse explained she has no other option than to perform this workaround as in the EHR “a day needs to be finalised at 10 pm […] That’s just how it is designed” (I_NU1).

An expected benefit of registering faulty information in the EHR in the nurse’s case affects the users, as there is no distorted image of the day before which improves workflow the next day. All activities of that day need to be registered before 10 pm to avoid “other alarm bells to start ringing” (I_NU1). The nurse did not expect this workaround to bring about dangerous situations, since “30 minutes or an hour difference is negligible” (I_NU1). The physician said that, as with ignoring pop-ups, entering faulty information may lead to a false sense of safety, negatively affecting the patient.

4.3 Out-system workarounds

(26)

25 4.3.1 Using paper

First, many interviewees recognise the use of paper for making notes. This workaround was observed once and mentioned twelve times during interviews with nurses, physicians, member of the support team and medical administrators. This means that each occupational group uses separate pieces of paper along the EHR:

“When I do my round of patient visits I always have a piece of paper with me. […] I rather write some keywords on paper, sit behind my desk, think about it, and register the information in peace.” (I_PH1)

As described above, the physician feels that first writing information down on paper and registering it in the EHR later helps him to process the information, so there is a lack of trust in his own abilities to directly register the data correctly. Another motive for using paper, is to keep eye contact with patients. As one nurse explained:

“The patient might be very nervous, in that case you do want to make eye contact. Another option would be to bring your computer, but then you would be talking to the screen instead of the patient. […] You look less at the

patient’s face, so you don’t see the impact of what you are saying. […] And the patient could feel less heard.”

(I_NU1)

The expected benefits following the use of paper affect the user and the patient. First, time is saved for the user to process and reconsider the information provided by the patient. Second, contact with the patient is preserved. A frequently expected risk affecting the user is that paper is vulnerable to “get lost or lie around” (I_NU3) leading to a loss of data. A loss of overview and extra work are also mentioned as expected risks of paper use. A risk affecting the patient is the possibility “to overlook items that are on the work lists in your computer, especially if you don’t know that patient well” (I_NU2), thereby jeopardise patient safety.

4.3.2 Using shadow systems

The second out-system workaround is the use of a system other than the EHR. Seven interviewees (two physicians, four members of the support team, and one medical administrator) admitted to using Microsoft Word and Microsoft Excel either as a substitute or as a complement to the EHR.

(27)

26 “Every day there are three, four, five notes added for each patient. These are all separate, so I cannot just scroll

through them. So, open, close, open, close. Then I also have to remember each note’s content. Therefore, I open a Word document next to it, so I can create my own file.” (I_PH3)

Another reason for the use of shadow systems is a lack of functionality in the EHR. This was elaborated upon by an employee of the EHR support team:

“[The EHR] does not support planning intakes. [The planners] keep an Excel file with the entire planning while you really just want to be able to do this in the EHR.” (I_SS5)

A final reason underlying this workaround is that the caregiver prefers a different lay-out than proposed by the EHR: “People make the whole letter layout in Word because they find the letters from the EHR ugly” (I_SS5).

An expected benefit for the users when using shadow systems is that better overviews are created. Also, by keeping an Excel file, the intake planners can schedule patient intakes a few months ahead which they were not able to do if operating solely with the EHR, thereby improving their workflow. Regarding the system, by acknowledging the deficiencies of the EHR, “improvements in the system can be made” (I_SS5). On the downside, an expected risk of shadow systems for users is the extra work that physicians have to undertake by entering information in both the EHR and their shadow system. Also, there is a risk of forgetting to register data in the EHR besides the shadow system, which leaves the system as not up-to-date.

4.3.3 Giving verbal consent for dispensing medication

The third out-system workaround is a physician giving verbal consent to a nurse for dispensing medication, only to order this medication in the EHR sometime later. This workaround was not observed, but mentioned during interviews with two nurses and one physician. To illustrate:

“What I sometimes try, and that depends a bit on the nurse I say that very honestly, I say: hereby you have my consent to carry it on. I will register the order later, or e-mail me in a moment that I have to order it. That's how

I do it” (I_PH2).

This physician continues by outlining the situation in which this would happen, which is often inconvenient:

(28)

27 Giving consent verbally happens because of a lack of time or physical facilities. One of the nurses explained a third reason for this workaround, namely: “no time, no motivation. […] If they [physicians] are at home and don’t feel like starting up the system they tend to give consent verbally” (I_NU3). Thereby, she states that this workaround is related to a physician’s willingness to order through the EHR directly.

Expected benefits of this workaround for the user are an improved workflow and time savings. For the patients, better care is expected to follow a verbal consent. As the nurse explained: “I think patient care is paramount, before the administrative part” (I_NU3). A possible risk of this workaround applies to the system. Forgetting to order the medication in the EHR later, meaning that “according to the system, this patient did not receive the medication” (I_NU3), leaves the system as not up-to-date.

4.3.4 Separating a scanner from its COW

The final out-system workaround entails separating a scanner from its COW to scan the wristbands of patients and the labels of infusion bags. However, nurses are supposed to bring the COW into the patient room and scan each infusion bag for every patient separately, while monitoring the COW’s screen. This workaround was not observed but explained by a nurse during an interview, who performs this workaround to not “wake up patients by COW’s noise” (I_NU3).

The expected benefit of this workaround is that patients are not disturbed by the COW’s noise during the night. However, taking the scanner into the patient room and away from the COW might jeopardise patient safety as the nurse is unable to notice any error that might appear on the screen when accidentally “scanning the wrong infusion bag, or the wrong patient” (I_NU3).

4.4 Results in summary

(29)

(30)

(31)

30

5. DISCUSSION & CONCLUSIONS

The aim of this study was to uncover the pathways of EHR workarounds by adopting a user perspective. This section discusses the research outcomes and is structured as follows. First, the pathway of each type of workarounds is explained. Then, the results for the separate elements of each pathway (espoused reasons and expected consequences) and the performers of identified workarounds are discussed. It concludes with theoretical contributions, practical implications, limitations, and suggestions for future research.

5.1 Three pathways of EHR workarounds

Figure 2 (next page) shows a visualisation of the pathways of each type of workarounds as found in this study. Firstly, in-system workflow sequence workarounds are developed in response to a user-related reason (Lack of time), and a system-related reason (Perceived restricting power of EHR). Looking at the consequences, the users expect in-system workflow sequence workarounds to create benefits for the user (Improved workflow; Time savings) and risks for the patient (Jeopardise safety; False sense of safety) and the hospital as a whole (Wrong billing).

Secondly, in-system data entry workarounds follow user-related reasons (Lack of time; Fear of losing data; Lack of knowledge), system-related reasons (Lack of functionality; Lack of overview; Perceived restricting power of EHR), and organisation-related reasons (Lack of physical facilities). What follows in-system data entry workarounds on the one hand are potential benefits for the user (Improved workflow; Time savings; Better overview). On the other hand, these workarounds carry expected risks for the user (Extra work; Loss of overview; Abuse of details), the system (Loss of potential system support), the patient (Jeopardise safety; False sense of safety; Cascading errors), and the hospital as a whole (Hindering research).

(32)

31

5.2 Espoused reasons for the rise of workarounds

First, this study shows that in some cases different types of workarounds are created for a similar reason. Likewise, similar workarounds are created for different reasons. This is consistent with findings by Koppel et al. (2008), who discussed the multiplicity of relationships between workarounds and their underlying reasons. Secondly, this study shows that workarounds are performed for reasons related to the system’s users, the EHR system itself, and the organisation, making this study’s results consistent with reasons presented in prior researches (Alter, 2014; Blijleven et al., 2017b). Interestingly, in this study, organisation-related reasons seem underrepresented in scope (Table 6). Where the users in this study acknowledge two organisation-related reasons to cause workarounds (Noise; Lack of organisational facilities), this did not happen frequently. Only five times, a caregiver mentioned working around the EHR for a reason related to the organisation. This is much lower compared to the frequency Black: Pathway in-system workflow sequence workarounds;

(33)

32 of a user mentioning user-related (71) and system-related (77) reasons. This underrepresentation can be partly explained by the fact that earlier studies that connect workarounds to organisation-related reasons like noise, focused primarily on workarounds performed by nurses (Koppel et al., 2008; Rack et al., 2012). This is because a nurse operates with equipment that makes sound, like barcode scanners, and not a physician (Rack et al., 2012). In this study, nurses constitute only a small part of the total body of participants, which could be an explanation for the minor presence of organisation-related reasons for workarounds in this study. Moreover, the infrequent referral to a lack of physical facilities is likely due to this study’s research setting, as large teaching hospital are often well-equipped with physical facilities.

Table 6: Number and Scope for Underlying reasons

UNDERLYING REASONS TOTAL NUMBER SCOPE

User-related 7 71

Organisation-related 2 5

System-related 3 78

Note: Total number = total number of reasons related to each category of reasons; Scope =

total number of times a user mentioned a reason related to each category

5.3 Consequences of workarounds

First, this study’s results reveal that users expect a series of consequences that constitutes a greater diversity in risks compared to benefits. For expected benefits, users mention six different consequences. Thereby, improved workflow is mentioned most often in this study. This benefit is expected to follow eight of the 11 workarounds, belonging to In-system workflow sequence and In-system data entry workarounds. This means that the majority of the workarounds that are performed by means of the EHR contribute to an improvement workflow. With that, no Out-system workaround was expected to result in an improved workflow. This seems legitimate, as this paper interprets Out-system workarounds to imply an additional item or task to the system (e.g. paper or shadow system), providing the caregiver with an extra administrative task to keep the system up-to-date. For expected risks, users mentioned 11 different consequences, over which the related workarounds are dispersed more evenly. Jeopardising patient safety is mentioned to be a negative effect for workarounds most often.

(34)

33

5.4 Performers of workarounds

Looking at the performers of workarounds, only two of the 11 identified workarounds are not mentioned by physicians, but by several other users: Bringing a scanner into the patient room and Sharing login details. Because scanning medication, blood bags, or infusion bags is part of the nurses’ set of tasks (Debono & Braithwaite, 2017), we can assume it is indeed unlikely that this workaround will be performed by physicians. Interestingly, no physician acknowledged to share login details with other employees. However, the members of the support team mentioned physicians do engage in this workaround. A possible explanation for this could be that through giving out login details, protected health information is shared and confidentiality is violated (Hassidim, Korach, Shrekberk-Hassidim, Thomaidou, Uzefovsky, Ayal & Ariely, 2017). Since this “unethical and dangerous” (Hassidim et al., 2017, p. 177) act is performed deliberately, physicians are likely to be hesitant to acknowledge it openly.

The results of this study show that nurses perform a less amount of workarounds than physicians. This is surprising, because earlier research suggests nurses to be “masters of workarounds” (Debono & Braithwaite, 2017, p. 2). A plausible reason for this is that the nurses in this study were significantly younger than the participating physicians. Age is proven to play a significant role in the intention of EHR use (Gagnon, Ghandour, Talla, Simonyan, Godin, Labrecque, Ouimet & Rousseau, 2014). This means that despite the fact that nurses are argued to be predominant performers of workarounds, the difference in age between the physicians and nurses in this study could have been the reason that the outcomes of this study prove otherwise.

5.5 Theoretical contributions

(35)

34 implementations (Soh & Sia, 2005). However, as illustrated by the expected risks that follow workarounds, this study shows that packaged software implementations come with inefficiencies too. Third, this paper has adopted an augmented approach compared to earlier studies. Other works studied workarounds in a non-empirical manner (Alter, 2014), or applied a managerial perspective (Beerepoot et al., 2019), or system’s perspective (Van Den Hooff & Hafkamp, 2017) to EHR workarounds. However, by looking from a perspective other than the user, these studies disregard the opportunity to unravel specifically why users engage in workaround behaviour, as well as their perceptions and awareness of the expected consequences of their acts. For that reason, this paper explains workarounds from a user’s perspective and thereby enriches prior research. Last, the exploratory method applied in this research has an advantage in the following way. Prior researches focus on the impact of workarounds on a predetermined subject, like patient care (Koppel et al., 2008) or efficiency and effectiveness (Blijleven et al., 2017b). The exploratory approach of this study enabled looking at the impact of workarounds to a greater extent, which exposed that EHR users expect workarounds to impact multiple subjects. Namely, this study revealed that a single workaround can have both beneficial and risky implications for themselves and other users, the system, and the hospital as a whole. Thereby, it questions other studies that label workarounds as either right or wrong (Ferneley & Sobreperez, 2006; Friedman et al., 2014).

5.6 Practical implications

Firstly, the results of this study show that all hospital employees should be aware of the consequences their workarounds expectedly bring about. As mentioned, the caregivers in this study expect all three types of workarounds to negatively affect the patient, whilst patient care is a hospital’s main concern. To mitigate against several risks as illustrated in this paper, hospitals could focus on controlling for the reasons for creating workarounds that lead to these risks. For example, this study revealed that the degree in use of giving verbal consent for dispensing medication could be diminished by improving the ease of use of the EHR system. If physicians are able to order medication through the mobile apps corresponding EHR systems, instead of only viewing patient statuses, caregivers carry the EHR system and all its functionalities in their pocket all the time. Through this, it will take a physician less time to order medication and creates the opportunity to order medication on the spot. Likely, this could lead to a decrease in performance of workarounds caused by a perceived lack of time or lack of motivation to find and start up a computer, or those caused by a lack of physical facilities like Computers on Wheels. Thereby, it should be noted that with the decrease of these workarounds, not only will this diminish the negative consequences that follow them, but also will the expected benefits of these workarounds automatically reduce.

(36)

35 power of the system over the medical work process. Before enterprise systems made their way into hospitals, physicians considered themselves as experts in the fields of diseases, medications, and treatments (Bate, 2000), as well as the ones responsible for setting the standards for medical performance (Mishra, Anderson, Angst & Agarwal, 2012). However, with the introduction of EHR systems, the knowledge previously owned by physicians is now also collected by a system which is available to all hospital employees. Since the system, and no longer the physician, is consulted for medical advice and procedural standards, power shifts from the physicians to the system (Beerepoot, Koorn, Van De Weerd, Van Den Hooff, Leopold & Reijers, 2019). With that, professional identity by physicians deteriorates (Mishra et al., 2012). This could be an explanation for the physicians in this study resisting the perceived restricting power by the EHR over their work process. To preserve their professional identity, physicians resist the power of the EHR and develop workarounds instead. Therefore, hospitals should find a way to improve the cooperation between physicians and EHR systems, while being aware of professional identity (Mishra et al., 2012).

Thirdly, hospitals are advised to make sure help desk employees have access to the EHR system. As was the case with this hospital, help desk employees have no access the EHR due to the Dutch privacy laws. Thereby, only physicians, nurses, and medical administrators can enter the system. This issue can be solved by a confidentiality contract. The help desk is an important body of a hospital as it is often the first channel for employees to reach out to when dealing with EHR-related issues. Moreover, the help desk is permanently established, as opposed to the temporary presence of support teams (Somers & Nelson, 2001). Unlike other hospitals (Beerepoot et al., 2019), the hospital that constituted the setting of this research had no clear guidelines for the decision of making adjustments to the system or providing additional training as a response to workflow hick-ups. Therefore, hospitals are advised to make the help desk more knowledgeable about the system and simultaneously provide them with a clear set of guidelines on how to respond to system’s issues and workarounds. In this way, they could offer more help and issues can be solved in smaller time spans. On the one hand, as solutions are likely to be provided quicker, the number of workarounds related to the system reasons will reduce and their negative implications will decrease accordingly. On the other hand, by including the help desk the body of employees working on solutions enhances the opportunity to continuously improve the EHR system, thereby creating a better alignment between the system and the hospital.

5.7 Limitations and future research

(37)

36 have affected the density and the type of workarounds as presented in this paper. Secondly, this study was performed at the Internal Medicine- and supporting departments of a large teaching hospital. Medical operators in surgical departments may create different workarounds and could have different reasons to do so. Therefore, it is debatable whether the workarounds identified in this study are representative for the entire hospital. Moreover, teaching hospitals tend to have better refined EHR systems because of the size of the supporting department (Blijleven et al., 2017b). As this could have influenced the type of workarounds, and thereby their pathways as presented in this paper, future researchers should be aware of this fact when studying these results in smaller hospitals. A final limitation of this study is that, despite the fact that expected benefits and risks are determined for each workaround, these consequences are limited to who or what they affect. Thereby, this research does not include the extent of the impact on the user, the system, the patient, or the hospital. We acknowledge the possibility that one can consider a single benefit to be as critical as multiple risks combined. Similarly, a single risk can be perceived as more significant than a combination of benefits. Unfortunately, the weight of each consequence was not discussed with the users of this study and therefore, this information could not be disclosed from our data.

(38)

37

REFERENCES

Ajzen, I. (2011). The theory of planned behaviour: Reactions and reflections. Psychology and Health, 26(9), 1113-1127.

Alter, S. (2014). Theory of workarounds. Communications of the Association for Information Systems, 34(55), 1041–1066.

Anthony, D., Campbell, A. T., Candon, T., Gettinger, A., Kotz, D., Marsch, L. A., Molina-Markham, A., Page, K., Smith, S. W., Gunter, C. A. & Johnson, M. E. (2013). Securing information technology in healthcare.

IEEE Security and Privacy, 11(6), 25–33.

Ash, J. S., Berg, M., & Coiera, E. (2004). Some Unintended Consequences of Information Technology in Health Care: The Nature of Patient Care Information System-related Errors. Journal of the American Medical

Informatics Association, 11(2), 104–112.

Azad, B., & King, N. (2008). Enacting computer workaround practices within a medication dispensing system.

European Journal of Information Systems, 17(3), 264–278.

Bagayogo, F., Beaudry, A., & Lapointe, L. (2013). Impacts of IT acceptance and resistance behaviors: A novel framework. International Conference on Information Systems (ICIS 2013): Reshaping Society Through

Information Systems Design, 1–19.

Bate, P. (2000). Changing the Culture of a Hospital: From Hierarchy to Networked Community. Public

Administration, 78(3), 485–512.

Beerepoot, I., Koorn, J. J., van de Weerd, I., van den Hooff, B., Leopold, H., & Reijers, H. (2019). Working Around Health Information Systems: The Role of Power. Fortieth International Conference on Information

Systems, 1–17.

Beerepoot, I., Ouali, A., van de Weerd, I., & Reijers, H. A. (2019). Working around health information systems: to accept or not to accept? Twenty-Seventh European Conference on Information Systems, 4, 1–15. Beerepoot, I., & Van De Weerd, I. (2018). Prevent, redesign, adopt or ignore: Improving healthcare using

knowledge of workarounds. 26th European Conference on Information Systems: Beyond Digitization -

Facets of Socio-Technical Change, ECIS 2018, 1–15.

Blijleven, V., Koelemeijer, K., & Jaspers, M. (2017a). Exploring Workarounds Related to Electronic Health Record System Usage: A Study Protocol. JMIR Research Protocols, 6(4).

Blijleven, V., Koelemeijer, K., Wetzels, M., & Jaspers, M. (2017b). Workarounds Emerging From Electronic Health Record System Usage: Consequences for Patient Safety, Effectiveness of Care, and Efficiency of Care. JMIR Human Factors, 4(4), 1–16.

Boudreau, M. C., & Robey, D. (2005). Enacting integrated information technology: A human agency perspective.

Organisation Science, 16(1), 3–18.

Bozan, K., & Berger, A. (2018). The Effect of Unmet Expectations of Information Quality on Post-Acceptance Workarounds among Healthcare Providers. 51th Hawaii International Conference on System Sciences, 9, 1–10.

EHR Workarounds: Rise and Unfolding