Unintended Consequences of Electronic Health Record Adoptions
The Role of Collective Affordance Actualization in Changing Medical Workplace Activities
Master Thesis
MSc BA Change Management
Faculty of Economics and Business
1
Abstract
Over the past decade electronic health records (EHRs) have been increasingly adopted in healthcare. Their potential to increase transparency and efficiency, while at the same time decreasing costs and errors, have made EHRs a much sought-after technology. Previous research has uncovered that EHR implementations can have a major impact on the work activities of its users. However, the mechanisms that enable the impact EHRs have on their users remain understudied. By using a collective affordance actualization perspective on EHR usage this study attempts to uncover some of the mechanisms that determine changes in work activities in response to EHR adoption. To account for the ecosystems in which the collective affordance actualizations take place this study employs the cultural-historical activity theory (CHAT) framework. The outcomes of this study on one hand confirm the effects of collective affordance actualization on more known outcomes such as increased administrative burden and on the other hand expose changed activities in the realms of personalized care, collaboration, and data utilization.
2
1. Introduction
In recent years, an increasing number of healthcare institutions around the world have shifted from paper based clinical record keeping to Electronic Health Records (EHRs) (King, Patel, Jamoom & Furukawa, 2014). The increase is not surprising given that advertisement of EHR software developers is laced with promises of improved patients’ health and care, maximized revenue, and protected payments (EPIC, 2020). All outcomes which are increasingly sought after by healthcare providers (Ash & Bates, 2005). EHR developers learned to respond cleverly to today’s emphasis on patient centric care as the following slogan of world’s largest EHR software provider, EPIC, illustrates: “with the patient at the heart” (EPIC, 2020). Multiple studies underline that EHRs potential is no empty promise. Supposedly, digitalization in healthcare would improve efficiency and effectiveness (Grimson, Grimson & Hasselbring, 2000) and standardize procedures and protocols (Pathak et al., 2015). Furthermore, EHR adoption could potentially result in cost reductions (Choi, Lee & Rhee, 2013), more patient oriented care (Finkelstein et al., 2012), increased patient safety (Lin, Jha & Adler-Milstein, 2018), and accelerated clinical research and predictive analyses (Hersh, 2007). However, the mechanisms determining whether initial EHR goals are obtained currently remain unknown (Atasoy, Greenwood & McCullough, 2019).
Despite the high potential attributed to EHRs, many EHR implementations often either fail to meet expectations or deliver results other than intended (Yu, Zhang, Gong & Zhang, 2013). Therefore, uncovering which mechanisms effect healthcare organizations in meeting EHR implementation goals can help prevent unnecessary costly investments, foresee unintended outcomes, and ensure a better fit (Agarwal et al., 2010).Studies that looked at outcomes of digitalization in healthcare revealed a variety of possible reasons behind failure to unlock EHR’s potential. Quantitative research shows varying results with regards to the objectives associated with implementations such as costs (Himmelstein, Wright & Woolhandler, 2010) and quality of patient-oriented care (Zhou et al., 2009). Whereas from qualitative research we know that implementation processes can be complex (Boonstra, Versluis & Vos, 2014), often suffer from a gap between design and reality (Heeks, 2006), and that people and procedure resources pose primary barriers during implementation and adoption (Gesulga, Berjame, Moquiala & Galido, 2017). Furthermore, Nguyen, Bellucci and Nguyen (2014, p.792) suggest that failures might be the result of the possible existence of: “a lack of sociotechnical connectives between the clinician, the patient and the technology in developing and implementing EHR”. Not only might failures be explained because of sociotechnical misalignment, other studies have uncovered that changes in clinical workplace activities due to EHR implementations might play a role as well (Carayon et al., 2015; McGeorge et al., 2015).
3
argued that increased interaction between IS and organizational research can help to explain the reshaping of organizations due to (digital) transformations. Ammenwerth, Iller and Mahler (2006) affirmed the pursuit for incorporating both angles when they found that fit between individuals, task and technology is critical for IT-adoption success. By studying the interaction between user and EHR and the changes in medical activities this entails, this study employs the theory of affordances. An affordance approach particularly draws attention to the influence technology and its environment have on behaviors and practices (Fayard & Weeks, 2014). More specifically Burton-Jones and Volkoff (2017) propose to use affordance actualization, which is effective use of the system, to study the actual effects of affordances. The goal of this study is to uncover the mechanisms that underlie changes in clinical workplace activities following EHR implementation by looking at EHR affordance actualization. Therefore, the question central to this research is:
‘How does electronic health record affordances actualization change clinical workplace activities?’
Qualitative research on a rich set of interview and observation transcripts has been carried out to answer this question. The research departs from a constructivist stance and employed several constructs of informed grounded theory (Charmaz, 2006). In addition to the theory of affordances (Volkoff & Strong, 2013), this research draws on the Cultural Historical Activity Theory (CHAT) (Engeström, 1987). CHAT is used to incorporate broader organizational aspects rather than mere technical. Simultaneous use of the theory of affordances and CHAT in this context is endorsed by findings from Pope et al. (2013, p.11) who concluded that digital innovation in healthcare must be understood both as “computer technology and as a set of practices related to that technology, kept in place by a number of actors in particular contexts”. Combining CHAT with theory of affordances allows for uncovering the relationship between what EHR allows its users and how this fit with the environment in which the activity is executed.
4
medical setting. In the following chapters the theoretical foundations, methodology, results, and discussion of this study will be set forth.
2. Literature Review
This chapter elaborates on the theoretical frameworks and relevant literature that are employed to study the mechanisms that determine whether EHR goals are met and to study changes in clinical workplace activities following EHRs implementation. First, the literature on digital transformations in healthcare will be discussed in more detail. Subsequently, the function and usability of an affordance perspective and what is known about EHR affordances will be explained. Since this study is on an organizational level, special attention is rewarded to the stream in literature that bridges between individual and collective affordance actualization. Lastly, the employability of CHAT for this study’s purpose will be set forth and different streams in literature are discussed to give direction as to which line of reasoning and thought are followed in this study.
2.1 Digital Transformation in Healthcare.
Healthcare has been the stage for multiple digital transformations over the past few years. Burton Jones et al. (2020, p. 1) define digital transformations in healthcare as: “large [and] complex interventions, such as whole-of-hospital or multi-site system implementations that involve major changes to how organizations function”. In recent years, digital transformations in healthcare included the introduction of electronic health records (EHRs), personal health records (PHRs), and clinical data exchange (Blumenthal and Glaser, 2007). Of these, EHR is the most widely implemented information system (Hansen & Baroody, 2019). EHRs come in various forms and the term relates to a variety of electronic information systems used in the healthcare sector (Boonstra, Versluis & Vos, 2014). Hansen and Baroody (2019) classified EHR as a form of enterprise information system (EIS). They argue that EHR is “marked by its use of multiple functional roles/units within an organization, the automation of workflow elements, and the incorporation of preconfigured usage and interaction patterns” (p. 2). EHRs often come with build-in functionalities designed to assist with decision making, such as clinical decision support systems (CDSS) and alerts (Pope et al., 2013; Kizzier-Carnahan et al., 2019). It is therefore often said that EHRs are more than mere digitized paper files (Adler-Milstein & Bates, 2010).
5
and Boonstra and Govers (2009) identified that additional drivers for EHR implementation are the complexity of medical data, data entry problems, security and confidentiality concerns, multiple objectives (learning and care), and the varied medical workforce. Exact numbers are not available but given the potential benefits of EHR implementation it might be surprising to find out that up to 70 percentof EHR implementations fail or experience serious backlog (Leviss, 2011).
Even though the potential and motives for implementation are plenty, academics have voiced their concern about EHRs not being able to meet the needs of today’s rapidly changing healthcare environments (Evans, 2016). Furthermore, EHR implementation often induce negative side effects such as potential weakening of social ties together with reduced social inclusion as a result of IT-enabled reduction of face-to-face contacts (Chen, 2013), and decreased efficiency due to poor information systems (IS) interfaces as well as information overload (Franssila, Okkonen & Savolainen, 2015).To this day IS research concerning EHR has mainly been occupied with explaining factors that drive adoption and innovation (Venkatesh et al., 2011), data entries, privacy concerns and patients’ usages of EHR (Angst and Agarwal, 2009). A time study by Carayon et al. (2015) found that the use of EHR technology has a major impact on the work activities of medical personnel. Their two main findings concern an increase in time spent on clinical review and documentation and more emphasis on clinical review and documentation in general. Even though they present evidence for changing workflows, Carayon et al. (2015) acknowledge that the motives and mechanisms behind the changes remain unclear.
Clearly, EHR implementation does not one on one result in the promising outcomes in which brings according to previously mentioned academics. The impact that changed activities have on the differing outcomes might shed a light on variety of outcome possibilities. Blumenthal and Tavenner (2010) sum the challenges of IS implementation up nicely: “it is not the technology that is important, but it is the effect. Meaningful use is not a technology project, but a change management project”. Considering the combination of information entry and retrieval features in EHR, features designed for decision making assistance and the lack of knowledge on how implementation results in change, examining the impact of EHR implementation on work activities through the affordances EHR pose becomes increasingly interesting in these rapidly changing times.
2.2 Electronic Health Records Affordances.
6
actualization of an affordance. The path between existence and actualization can be described as two stages. The first is perception through a recognition process followed by adjustment of behavior on the perceived opportunity for action (Bernard, Recker & Burton-Jones, 2013). The affordance perspective lens ensures provision of a better language for describing how practices are shaped and modeled through structure and setting (Fayard & Weeks, 2014).
Application of the theory of affordances has recently gained ground in IS literature. For instance, in combination with environmentally sustainable digital transformations (Seidel, Recker & Vom Brock, 2013) or in the context of enterprise social media (Leonardi, Huysman & Steinfield, 2013). The revival is driven by the notion that using an affordances lens provides new insights in explaining the consequences of IT artefact usages in organizations and accompanying organizational changes (Leonardi, 2013). Researchers increasingly start applying an affordance perspective in EHR research. The most extensive work thus far is by Strong et al. (2014), who in their research on understanding effective use identified eight affordances which they describe as: “central to analyzing how the use of an EHR results in organizational change” (p. 67). The data from this study is analyzed based on the eight affordances by Strong et al. (2014) which are:
“1) Capturing and archiving digital data about patients 2) Accessing and using patient information anytime from anywhere 3) Coordinating patient care across sites, facilities, and providers 4) Standardizing data, processes and roles 5) Monitoring organizational operations 6) Substituting healthcare professionals for each other 7) Incorporating rich information into clinical decision making, and 8) Shifting work across roles” (p. 67).
As described earlier, there is a distinction between the presence of an affordance and the actualization of an affordance. The definition by Strong et al. (2014, p. 70) of what affordance actualization is, is widely accepted in IS literature: “the actions taken by actors as they advantage of one or more affordances through their use of technology to achieve immediate concrete outcomes in support of organizational goals”. A focus on affordance actualization allows to emphasis on the how and why of EHR use outcomes rather than on what outcomes occur through EHR use (Strong et al., 2014).
7
analyzed are performed jointly with others or are otherwise influenced by activities of others (Fayard & Weeks, 2007)
Academics in IS affordances research increasingly agree on including organizational context as this might help to explain how and why IT-associated change unfolds (Strong et al., 2014; Evans, Pearce, Vitak & Treem, 2016). Kaptelinin and Nardi (2012) go even further and reason that affordances are the result of three-way interaction between person, mediational means, and environment. For this study that would mean between healthcare professional, EHR and the hospital setting in which interaction between the former two occurs. To further unravel the situational effects on affordance actualization and changing clinical activities this study employs the cultural-historical activity theory (CHAT). Whereas affordance approach allows to study the influence technology has on behaviors and practices (Fayard & Weeks, 2014), CHAT is used as a framework for data analysis to go beyond merely technical aspect of usage by incorporating broader organizational aspects (Allen, Brown, Karanasios & Norman, 2013).
2.3 Cultural-Historical Activity Theory.
A mere affordances perspective does not do justice to the sociocultural context in which medical work activities take place. Since the social context appears to be of great importance in the adoption and implementation of EHR (Kruse, Kristof, Jones, Mitchell & Martinez, 2016; Ben-Zion, Pliskin & Fink) the affordance perspective in this study is extended by the cultural historical activity theory (CHAT). Initially developed by Russian cultural-historical psychologist Vygotsky, brought to the West by Michael Cole, and popularized by Yrjo Engeström (Nussbaumer, 2012) CHAT concerns understanding the dynamics of practices in activities (Greig, Entwistle & Beech, 2012). According to its founding father. the central idea of the theory was that an activity cannot be examined without its context, in this study the hospital setting with its formal and informal rules and own dynamics (Vygotsky, 1978). CHAT’s focus is on an object or outcome of activity – patient care –as the aim towards which subjects
– medical specialist, resident and nurse (specialist) – work together to meet an identified need which it Figure 1 Activity system (adapted from Engeström, 1999)
8
captures in an activity system (Blackler, 1995). Figure 1 illustrates the activity system in which EHR is used as a tool to enable medical staff to exercise patient care in the context of a hospital.
CHAT is object oriented and tool and culturally mediated (Lin, Tan & Chang, 2008). The objective patient care is carried out by the subject (medical specialist, resident, or nurse specialist) and mediated by the tool (EHR) and by the social historical context made of rules, community, and division of labor. Rules in healthcare contains both the formal and informal rules such as protocols, norms, and value (Mills & Spencer, 2005). CHAT community with regards to hospitals refers to “wider community of practitioners” (Greig, Entwistle & Beech, 2012 p. 306).
The additional value of CHAT lies in its ability to uncover drivers of change, in CHAT change is stirred by contradictions in the activity system (Engeström, 2001). CHAT namely, facilitates exploration of differing values of those involved in an activity that reinforces their views (Blacker, 2009). The theory highlights the role contradictions play in the realization of new or reshaped activities (Greig, Entwistle & Beech, 2012). Allen, Brown, Karanasios and Norman (2013 p. 840) refer to contradictions as: “anything within the system that opposes the overall motive of the system, the aim or purpose that subjects within the system are individually or collectively striving towards. CHAT allows to deduce changes in activities by exposing contradictions.
Another often used definition of contradiction comes from Kuutti (1996, p. 34) who describes a contradiction as “a misfit within element, between them, between different activities, or between different developmental phases of a single activity (…) [and so] manifest themselves as problems, disrupters, breakdowns and clashes”. The different levels that Kuutti (1996) refers to in his definition of contradictions refers to the gradation which Engeström (1987) had brought to the fore earlier. Engeström and Sannino (2010, p. 7) redefine the levels of contradictions as follows:
9
Figure 2 Visualization of the different contradictions which can be present in activity systems (adapted from Engeström, 1987)
Figure 2 displays where and how the different levels of contradictions are situated in an activity system. In this figure the triangle in the bottom left depicts the main activity, the one under analysis. When contradiction arise between constituents a tension is generated that in term provides opportunity for change in activity (Blacker, 1993). Contradictions become driving forces from the moment they are dealt with in such a way that an emerging new object is identified and turned into a motive (Engeström & Sannino, 2010).
The contradictions that this study focuses on are those that derive as the result of EHR affordance actualization within the activity system of a hospital. Literature on attempts made to employ CHAT together with the theory of affordances to explain the social context in which affordances in the field of IS exist is scarce (Karanasios, 2018; Wolff-Piggot & Rivett, 2016). Furthermore, those studies were conducted at an individual level and with only focused on predetermined affordances. The present study aims to contribute to IS and organizational literature by bringing these to theories together. The following chapter describes the methods used to answer the research question.
3. Methodology
10 3.1 Theoretical Approach
The aim of this study is to discover the impact EHR affordance actualization has on activities in the clinical workplace. Due to unforeseen data collection restriction as a direct consequence of the global Covid-19 pandemic it has not been possible to collect firsthand data. Access was initially provided by an academic hospital but was reversed as the hospital had to forbid access to non-essential visitors. This study employs a constructivist orientation, meaning that the theories and concepts that I will construct in the following chapter are based on interviewees construction of stories that try to explain their perception of reality and their interaction with the environment (Corbin & Strauss, 2008). The constructivist orientation and the open-ended nature of the research question justify the use of qualitative research methods (Edmondson & McManus, 2007). The qualitative data used has been gathered by two researchers for the purpose of studies into EHR’s relation to interprofessional collaboration and organizational learning. Both studies utilized a multi-method approach that enabled stronger theory development and validation of concepts (Eisenhardt, 1989; Jonson & Jehn, 2009). The used data collection methods were observations, interviews, and secondary data sources. The latter consisted of data provided by EHR developers on the applicability of their software.
Several concepts of informed grounded theory proved useful for discovering new insights and linkages in the data (Charmaz, 2006). Contrary to traditional grounded theory (Glaser & Strauss, 1967), informed grounded theory allowed for inclusion of existing theories such as theory of affordances and CHAT. Both theory of affordances and CHAT suit the constructivist orientation of this study. Affordance can be studied from a constructivist stance since they are a product of interpretation and their effectuation depends on perception of possibilities the IT artefact, EHR in this case, holds to users in their environment (Hutchby, 2001). The same applies to CHAT. Allen, Karanasios and Slavova (2011, p. 776) argue that working with CHAT: “provides researchers a theoretical lens to account for context and activity mediation” and therefore enables a better understanding of the context in which EHR users construct their behavior and meaning. In the following sections elaborate on the setting and participants and the data collection and analyses of the present study.
3.2 Setting and Participants
11
Hospital 1 is a large regional hospital which employs around 6500 people of which almost 500 are medical specialists. The hospital holds more than 1100 beds and more than 600.000 patients visit the hospital per year. In 2018 hospital 1 transitioned to a new integrated EHR, HiX by Chipsoft, after working with several separate in house developed systems for several years. Hospital 2 is a smaller regional hospital with around 3000 employees of which 230 are medical specialists, 650 beds, and more than 400.000 outpatient visits per year. In early 2016 hospital 2 fully transitioned to EPIC after switching to electronic records in 2015. Finally, hospital 3 is a large teaching hospital with close to 11500 employees of which almost 700 medical specialists, it holds 1500 beds and serves more than 850.000 outpatient visits per year. In 2017, after working with a diverse set of information systems, the hospital transitioned to EPIC as their integrated EHR.
3.3 Data Collection
Three different healthcare professions and occupations are central to this study. These are medical specialist, resident, and nurse (specialist). The choice for including multiple groups comes from earlier findings that suggest that different groups have different perspectives on the implementation of EHRs (McGinn et al., 2011). Hospitals pose a good fit for studying the effect of IS implementation on changing work activities because due to continuous development of medical knowledge and know how, adapting and innovating is a central aspect of this sector. The data used for this study was compiled of 48 pages of observation transcripts and 188 pages of interview transcripts. The durations of the interviews are approximately similar with interviews ranging from anywhere between 21 and 60 minutes of which the majority last around 35 minutes. Table 1 presents an overview of the observations and interviews per hospital and per profession. Interestingly the majority medical specialist who participated in the previous studies are male. The data of hospitals 1 and 3 do not include a single female medical specialist except for one female surgeon in hospital 3. This study makes the same distinction between medical specialists and surgeons, since the study from which the surgery department data stems concluded that there are differences in how internal medicine and surgery department change their workflow in response to EHR implementation. Most observation data have been acquired during consultation, multidisciplinary consultation, handover, and ward rounds. Beside some smaller activities, these activities seem to demand the largest sum of time spend by the subjects of this study.
Table 1 Overview Observations and Interviews
12 Nurse (Specialist) 2 100 65 1 5 2 Medical Specialist 2 50 69 2 10 Resident 2 100 68 1 9.5 Nurse (Specialist) 2 100 60 2 7.75 3 Medical Specialist 5 0 210 4 9.75 Surgeon 3 33 76 - - Resident 1 0 22 2 10.25 Nurse Internal Medicine (Specialist) 4 100 110 1 8.5 Nurse Surgery (Specialist 3 66 91 1 3 Total 26 52 930 17 80.25 3.4 Data Analysis
All observation and interview transcripts were uploaded to Atlas.ti and labelled based on researcher, method, profession, and hospital. Since the observation and interview data was secondhand and I was unfamiliar with the data at first, I made use of the three-stage coding process as set forth by Strauss and Corbin (1990). The stages consisted of open coding, axial coding, and selective coding. The purpose of the first round of coding was to become familiar with the topic and the angles which the interviewees pursued. Throughout this stage the activities in which EHR is used and possible contradictions were identified and initial storylines to pursue appeared. During the subsequent stage of axial coding, I refined the initial codes and created subcategories for which the affordances as described by Burton-Jones and Volkoff (2017), the actors of CHAT by Engeström (1991), and the redefined levels of contradictions by Engeström and Sannino (2010) served as input for deductive codes. At last inductive relationship were identified between the core categories.
13
4. Findings
Data revealed a misalignment between the initial promises of EHR and firsthand experiences of healthcare workers operative in post EHR implementation hospital settings. Multiple collective affordance actualizations have been identified as well as contradictions these trigger as to how medical workplace activities are executed. Some of the contradictions produced by collective affordance actualization were expected based on analysis of previous literature, others were not. The analysis of the data took place on the basis of the eight EHR affordances identified by Volkoff and Strong (2017) and described in the literature review namely: access and use, capture and archive, coordination, standardization, monitoring, substituting, incorporation, and shifting roles. With emphasis on the identified affordance actualizations which pose the biggest impact on clinical workplace activities. In this chapter the collective actualization of affordances identified in the data are discussed, followed by the contradictions they cause and subsequent changes in medical workplace activities. These findings are structured based on how care is delivered (standardized versus personalized; reporting versus quality of patient contact) and interprofessional interaction and collaboration (coordination versus hierarchy; data utilization versus vast data amounts).
4.1 Care Provision.
The perception of potential of EHR adoption differs between the initiators of the implementation and EHR end users. In general, end users express that one of the main goals of their work is to deliver patient centric care. To that end, they view EHR usage as an instrument and not a goal: “I view EHR more as a tool to capture practices of patientcare (…) but to a certain extent, currently it’s too complex and demanding” (medical specialist, H1). From this quote and supporting data across most interviews it becomes evident that hospital personnel does not want to be controlled by EHRs. While initiators often seem to employ EHRs for the purpose of generating generality of care and transparency of costs as the following quote expresses: “this system [EPIC] is not made for doctors. According to me, this system is built for controlling everything. (…) The largest disadvantage of the system is that, in my view, no one has ever looked at whether it is developed with the user in mind” (medical specialist, H3). Analysis of the data on how care is delivered resulted in the discovery of two contradictions as the result of collective affordance actualizations. The first concerns the tension between imposed standardization by EHRs and desired personal delivery of care by medical personnel. The second concerns the friction between reporting and documentation affordance actualization and quality of patient interaction. The following sections elaborate on how collective affordance actualization changes workplace activities related to how care is delivered.
14
patient care can be delivered in a personalized fashion. On the one hand delivery of care is standardized as the result of collective incorporation affordance actualization in terms of adhering to alerts and pop-ups. And on the other hand, through collective actualization of standardization affordances by following protocols and pre-established questionnaires. Since alerts and pop-ups are often part of protocols and pre-established questionnaires, both collective incorporation and standardization affordance actualization are analyzed here simultaneously. Before claims about collective actualization of incorporation and standardization affordances can be made, we first look at the different ways and degrees to which the affordances are actualized in the data.
The data is clear about the fact that most interviewees share some reservation as to whether standardization serves patient care. The following quote is exemplary for how most interviewees look at standardization imposed by EHRs: “I think that standardization plays too big of a role. I understand that it has to do with keeping things manageable but individual patient cases are not always easy to standardize. You cannot always proceed in the same way” (resident, H1). Interviewees acknowledge the need for some standardization but emphasize that complete standardization does not result in more adequate patient care. Some interviewees go one step further and condemn all forms of standardization imposed on them. Especially with regards to pop-ups and alerts in the data, the comments are generally very negative: “pop-ups are annoying. Indeed, the system does increase standardization, but I think those pop-ups are a big problem. Pop-ups only do one thing and that is ensure safety, and safety per definition does not serve efficiency” (medical specialist, H3). The interviewee acknowledges that pop-ups could potentially increase safety, but he is nevertheless negative about their application as it enforces standardization on him. However, the negative stance of this interviewee towards standardization as being inherently bad is not shared by all interviewees, as the quote in the beginning of this section already illustrated.
An additional issue indicated with pop-ups and many pre-established questionnaires is that for each patient you need to click through the same alerts and pop-ups which is said to affect patient-oriented care as the following quote illustrates: “The way in which we have to order an infection control is outrageous. The system makes no difference between males and females. Whatever it is you order (…) you always need to answer the question whether someone is pregnant or not (…) that question should have never been asked.” (medical specialist, H3). The incomprehension this medical specialist talks about is found in other user groups. Nurses experience similar frustration because of standardization imposed on their work activities by EHRs:
15
After which she adds that: “you are not a good nurse (…) when you need all the steps laid out for you in an EHR" (nurse, H1). What the nurse says here is that incorporating and standardization affordance actualization in terms of obeying instructions derived from EHRs disqualifies the users as a decent medical professional. Though this specific nurse might not completely adhere and therefore not actualize the affordance, from the data it becomes clear that many other interviewees do and therefore that their work activities change. Whereas before users would highly relate on their judgement of the situation and arrange their work accordingly, they now generally execute simply what the EHR tells them to do. The following quote perfectly illustrates the consequence of collective incorporation affordance actualization on personalized care:
“for the less experienced doctor [dictated work instructions] might be very useful. However, I do think that if you are always told [by the EHR] to follow protocol A, you will never consider that protocol B might be better suited for a particular patient. And if you are always told to use protocol A, do you then even know that there are other options?” (medical specialist, H2).
Once more, this quote illustrates that whereas EHRs are designed to support users in executing their work, end users often seem to perceive this more as a threat then as actual assistance. Both the quote by the nurse as by the medical specialist indicate that they perceive that their professionality, knowledge, and skill is questioned by the EHR. They are also both aimed at new entrants to the medical workplace. It can therefore be said that collective actualization of incorporation and standardization affordances mainly takes place at the lower levels of seniority in the medical setting. There is fear that future healthcare professional will lack critical thinking skills once incorporation affordance are indeed collectively actualized as is already the case with junior medical professionals who are found to be more likely to adhere to alerts, pop-ups, protocols, and pre-established questionnaires.
16
“Clinical view is very important. It is something which I always had but has recently become something we can work on together so to speak. (…) You used to be dependent on ward rounds or on the moment you could spar with a medical specialist [to develop your clinical view] which some people did and other did not. But you can also read [EHR] reference [to deepen your clinical view]” (medical specialist, H1).
Even though there is a small group of EHR users that can see the benefits of collective incorporation and standardization actualization the majority foresees that it will result in near robotic execution of clinical work activities without consideration of the patient’s situation. Collective actualization of incorporation affordances, such as acting on alerts, pop-ups, protocols, and pre-established questionnaires, contradicts the reliance on professionals’ clinical view and consequently changes workflow and harms the personalized attention patients are used to receive.
Reporting versus quality of patient contact. One of the initial reasons for EHR implementation was the increased external pressure to comply with registration requirements: “in every patient contact that I have I use my EHR. Also, to record my registrations, I am obliged to do that, to register everything I do or discuss with the patient” (nurse, H2). This interviewee expresses that she must actualize some reporting affordances, she is obliged to do so. All interviewees acknowledge that they use EHR for this purpose and most agree that in essence EHR affords adequate possibilities to do so: “I think the EHR is perfectly suitable for reporting, documenting and all of that. (medical specialist, H3). To that end EHR users actualize different capture and archive affordances EHR pose but always with the intention to register what has been said, observed or discussed. The vast array of ways in which the capture and archive affordances are actualized while always with the same objective in mind indicate collective capture and archive affordance actualization. However, actualization of capture and archive affordances appears to be very time-consuming. One interviewee estimates the time he spends on reporting at almost four fifths of his total time spend at work:
“there is a lot of administration that you need to report. Recently, there has been a study conducted in Amsterdam in which they followed residents. They found that 80% of their time was spend in either meetings or administration. And I think that is about right. A workday approximately lasts for ten hours, if I spend two of those on patient contact that is a lot” (resident, H3).
17
data comes across: “honestly, I hate to spend much time looking at my screen when I have patient interaction” (medical specialist, H2). So, both patients and EHR users are aware of the disturbing effect computer usage for the purposes of actualizing capture and archive affordances has on patient interaction during consultation.
However, not only does collective capture and archive affordance actualization influence the contact EHR users have with patients during consultation, there are also known instances where actualization has impact on contact with patients who stay at the hospital: “It does happen on busy days that I simply do not have the time to get patients out of their beds because of all the red tape. On those days I leave with a very unsatisfied feeling, then it feels as if I fell short” (nurse, H1). Clearly, actualization of capture and archive affordances pose a challenge for everyone involved. Immediate and precise administration of patient information has always been a key task of medical personnel and is supposed to improve through EHR usage. However, medical personnel find themselves increasingly struggling with how to marry actualization with adequate patient interaction: “there are so many steps you need to go through, also steps that add no value. They expect us to fill out the whole patient history which simply cannot be done during consultation (…) resulting in less time to spend per patient” (medical specialist, H3). An often observed change in work activities to solve the issue of spending time behind a computer during patient interaction is to postpone the data entry to a later moment. However, such a decision does not come without consequences:
“You need to be aware that you do not sit behind your computer the whole time. I try to consciously move the computer away so I can talk with the patient. There is nothing that seems more annoying to me than that the caregiver sits behind his keyboard the whole time. (…) What I do is perhaps also not always the best solution, but I often report afterwards. There might then be crucial things that you forget, but I take those for granted, because I rather have enough time to spend with the patient” (nurse, H2).
18
archive affordances without harming the quality of patient contact have in common that they include freeing time on other activities.
Earlier in this chapter collective standardization affordance actualization has been discussed considering adherence to protocols and pre-established questionnaires. The same applies to data capture and archiving: “I think in theory you should be able to work faster through standardized data entries. You only register once and then you should be able to see it in any way, whether in clinic or in notes. Once data is stored in one place, preferable recorded discreetly, you can use it in other places as well”. (medical specialist, H3). One explanation for the time spent on reporting following collective actualization of capture and archive affordances lies in the proceedings required for actualization: “[EHR] usage requires a lot of clicking and arranging” (resident, H3). As for now, collective actualization of capture and archive affordances has potential to increase the patient information available but burdens EHR users with additional time management issues. EHR users are forced to change their work activities to free time for conclusive reporting. As a result, the quality of patient contact deteriorates slightly, and users experience more stress and pressure in their work activities. Figure 3 summarizes the findings on changes in clinical workplace activities related to delivery of care.
Figure 3 Overview of changes in activities regarding provision of care
4.2 Professional Interaction and Collaboration.
19
implementation the hospital was starting to become a more integrated organization: “we have had compartmentalization over the past 30 years, meaning that medical specialists worked parallel to each other. The last few years we have worked to improve this” (medical specialist, H3). Unfortunately, EHR implementations seems to have nullified some of the progress that has been made in integrating coworkers in all three hospitals. Analyzing the data on interaction and collaboration resulted in the discovery of two contradictions derived from collective affordance actualization. The first concerns the improved coordination possibilities parallel to redefined hierarchical structures. The second is about utilization of data and cast amount of data available. The remainder of this subsection elaborates on the actualization of which collective affordances resulted in these two contradictions and how interactive and collaborative activities in the clinical workplace changed in response to these contradictions.
Coordination versus hierarchy. Besides affording exchange of patient data within hospitals, another
important role for EHRs is providing means for collaboration, directing work, and facilitating communication. The following quote illustrates the role of both data exchange and collaboration affordances in interaction and collaboration:
“As a nurse I use EPIC [EHR] at the start of my shift to inform myself about current patient situations and to report my findings. But I also use EHR to communicate with my colleagues, both with other nurses as well as across disciplines, mainly with medical specialists. Think of additional reporting, clarifying problems or uncertainties, or providing input to be used during ward rounds” (nurse, H3).
The interviewed nurse lists several activities EHR facilitates which used to be executed manually or via the phone before the implementation of EHR. The interviewee describes actualization of EHR coordination affordances. In the data there are many examples of ways in which interviewees use EHR features to coordinate medical activities. Among other ways, interviewees describe the use of the order functions to arrange work between medical specialist and nurses and observational data shows the use of EHR features to coordinate work with general practitioners and pharmacists. Previous ways of coordinating have become obsolete and are only turned to in cases of system failure: “the resident finds out that he misplaced an order and calls the nurse to make it right” (resident observation, H2), when the receiver does not understand the order and follows up on it face-to-face or via phone: “you ordered something, but should I really administer this [medicine]?” (medical specialist, H3) , or to emphasize the importance of executing the order straight away: “when I enter an order I always need to call them if I want them to take care of it immediately” (medical specialist, H3). Due to widespread actualization of different coordination affordances for the purposes of coordinating work among medical personnel there is collective actualization of coordination affordances.
20
specialist redefines the hierarchical structure and limits the extent to which they could exercise control over their work. For instance:
“if we [nurses] would have liked blood sugars to stop we would consult with a medical specialist to have the blood sugars stopped. We used to have an activity list in the front of our binders, in which we would draw two stripes, so it was clear for us that we no longer had to do that [administer blood sugars]. Now a medical specialist needs to order to stop the order (…) and if they do not do so immediately it might happen that they forget to do it at all” (nurse, H3).
The quote illustrates how nurses have become more dependent on the medical specialist to execute their work. Both groups share a mutual goal namely, patient care, but collective actualization of coordination affordances sometimes harms achievement of this goal. Collective coordination affordance actualization widens the gap between nurses and medical specialists, sometimes even to the extent that the two no longer physically meet: “the implementation of EHR has resulted in deterioration of collaboration in the hospital. Previously medical specialist would frequently visit us, now we depend on them returning our calls or messages if want things clarified or when order must be requested” (nurse, H3). Most nurses indicated that they highly valued the physical interaction they had with medical specialists before the implementation of EHR. One nurse indicated that actual communication is now disappearing since: “because well, you are supposed to find out about everything yourself in EPIC [EHR] right?” (nurse, H3) after which she adds that she prefers: “face-to-face interaction with the ward doctor, to personally hear what needs to be done and how to proceed”. Even though the remarks of this nurse are indicative of how most nurses perceive this, there is an exception to preference of face-to-face interaction in the data. Preference of face-to-face interaction seems to be closely related to the social skills of the medical specialist: “one is more social then the other, which does have an effect. You can deliver better care if you are better tuned to each other” (nurse, H3). Just as the quality face-to-face interaction was dependent on the social skills of the sender and receiver, so is also the quality of interaction via EHR dependent on the technical skills of the people involved.
21
Whereas nurses experience collective coordination affordance actualization by medical specialists as belittling, medical specialists generally perceive the ordering process from a completely different angle as the following quote illustrates:
“for instance [nurses] can only administer paracetamol when there is an order to do so. That way for nurses it is clear they really need to do so instead of communicating via the phone ‘do this, give that medicine’. Before they could have felt a little unsafe about it, but because now it is in there and the medical specialist put it there it should be alright because he [medical specialist] did so consciously” (medical specialist, H2).
From this quote it becomes apparent that medical specialists actualize EHR's coordination affordance by using it to distribute work orders. However, it also becomes painfully evident that both groups perceive the benefits of actualization differently. Medical specialists actualize coordination affordances because they think this helps nurses with their activities and prevents errors in execution. Whilst nurses experience that actualized coordination affordances negatively influence the way they go about in their work. However, collective affordance actualizations do not only have negative effects on the interaction and collaboration in hospitals. In fact, data exchange, a key aspect of EHRs, is said to have even improved a little due to access and use; capture and archive collective affordance actualizations: “data exchange goes a little easier now” (medical specialist, H3) because: “information is more easily accessible and broader information is available” (medical specialist, H2). The final subsection goes into more detail on the effect of data availability and utilization on clinical workplace activities.
Data utilization versus vast data amounts. For medical specialist, availability of correct and up to date
patient information on which to base their decisions and proceedings is key. Earlier on we looked at EHR's affordances to collectively capture and archive data. Here we look at EHR'’s affordances to collectively access and use data. Having the most accurate data at hand can make all the difference at times in which decisions on how to act must be rushed. One interviewee described EHR’s role herein as follows:
“[EHR] helps in that sense that it provides quick information on a patient and his history. You need that to make an estimate on how to proceed. If a patient shows up with abdominal discomfort and you see that the same health issues appeared more often of the past few months, then that requires a different approach as to a patient with no previous complaints. Access to such information is now only two clicks away” (resident, H2)
22
“It’s such an enrichment that we can have insight into the notes of medical specialists from everywhere. We can look at medical photographs, have instant access to blood test results, everything is so much more complete which allows for better interpretation of a patients situation and whether we should get a medical specialist involved or not” (nurse, H3).
Access to information might initially not come across as something to do with collaboration and interaction. But as the previous quotes illustrates the collaborative aspect of access to and availability of correct data is in the hands of co-workers. The quote is also illustrative for the vast amount of diverse data that can be accessed and used. Collective access and use affordance actualization enable better access to and better use of patient data. However, collective access and use affordance actualization also brings about a new challenge namely, handling vast amounts of data and selective usage. One resident expressed the challenge as follows: “now we have to consider, do we want to read every last bit of information or do I trust the last letter and pick it up from there” (resident, H1). With the last letter the resident means the note of the last person who saw and assessed the patient. The other option the resident speaks about means reading through all patient data history to get a picture of the patient yourself. The following quote shows that reading through all patient data history can be near impossible: “Image someone staying at the hospital for five weeks. Every day the dietitian adds notes to a patient’s file but so does the nurse, physiotherapist, and everyone else., that means 35 days times 10 notes a day that is 350 notes. Good luck getting an overview of that” (medical specialist, H3).
In addition, not only has it become increasingly more difficult to filter the required information, the unnecessary information also has potential to distract users from their jobs: “EHR allows to get to the heart of a problem faster (…) but it also introduces a whole load of new annoyances that distract and divert attention from the root cause” (medical specialist, H2). Interviewees from different hospitals voiced the same concerns: “Everything that is relevant can be found in the EHR but also a lot of information that is not relevant” (resident. H1). The limited time often at the disposal of the EHR user makes it more difficult to learn and understand a patient’s trajectory and history. Whilst the ability to judge patients’ situations is a highly valued skill and thought to residents and nursing students in all three hospitals.
“I tell [my students] to look at the reason why a patient has been admitted and to find out about all the research that has been done. You can find everything the medical specialists have agreed on in HiX [EHR], but that requires some digging. Such information comes separate from the nurse output. But you need to find it out because it helps you to understand why you need to execute your nursing procedures” (nurse, H1).
23
data. As one interviewee put it: “we often blame [use of wrong data] on the EHR, but the EHR is nothing more than an electronic system. If you put garbage into it, garbage will come out. If you put nothing in, nothing comes out. We are to blame” (medical specialist, H3).
Fortunately, the data also contains multiple examples of cases in which access and use affordance actualization enriched the work activities of users. On interviewee described how a more tech savvy colleague had been able to really benefit from EHR use: “he can make graphs that show patients temperature, moisture lists, all those kinds of things. It allows for visualization of the patient’s situation’. (resident, H2). In term visualization allows for better assessment of the situation and ultimately in better suited care. Overall, we can conclude two changes in activities as the result of collective access and use affordance actualizations. First, in general EHRs provide a more inclusive and more complete picture, allowing users to offer better continuity of care. Second, collective access and use affordance actualization results in more byproducts which make it harder to filter the required information and request better filter skills of EHR users. Figure 4 summarizes the findings on changes in clinical workplace activities related to interaction and collaboration.
24
5. Discussion
The goal of this study was to uncover the mechanisms that underlie changes in clinical workplace activities. To that extend that research question this study attempted to answer was: “How does electronic health record affordances actualization change clinical workplace activities?” This chapter respectively discusses this study’s general findings and meaning, contribution to the theory, strengths and limitations, practical implications, and directions for future research.
5.1 Principal findings and meaning
The findings show that collective actualization of EHR affordances impacts the clinical workplace activities of EHR users beyond the expectations that derived from the study of previous literature. Based on previous literature it was anticipated that implementation of EHR would not simply result in actualization of the promised potential EHRs pose to medical professionals and healthcare institutions such as hospitals, but it remained unclear which mechanisms determined which goals of EHR implementation would be achieved. The present study embroidered on both findings by Nguyen, Bellucci and Nguyen (2014) who hinted that absence of sociotechnical connectiveness played a role in failure to unlock EHR’s full potential. As well as on recommendations by Carayon et al (2015) and McGeorge et al. (2015) to look at the role of changing medical workplace activities when attempting to uncover mechanisms that play a role in meeting promising EHR potential.
The main findings of this present study concern the discovery of the large role reserved for collective actualization of EHR affordances. Strong et al. (2014) were the first to extensively describe the main affordances and their actualization that EHR features pose for her users. The data available for this study has been analyzed based on the eight affordances they completed with the collective angle proposed by Strong et al. (2014) and put into context by employment of Engeström’s (1987) CHAT. Analysis of the data uncovered misalignment between motivation for implementation (results expected by both the initiators and end users) and actual impact measured by focusing on work activities. Previous studies create the expectation that EHR implementation and adoption will result in increase of quality of patient-oriented care (Zhou et al., 2009) and improved improve efficiency and effectiveness (Grimson, Grimson & Hasselbring, 2000) to name just a few benefits. Whereas the present study argues that those two expected outcomes do not go hand in hand.
25
behavior is found in all aspects of medical workplace activities ranging from reporting to administering care. These findings are in line with findings by Pathak et al. (2015) that healthcare would become increasingly standardized but contradict findings by Finkelstein et al. (2012) who argued that patient care would become more patient-oriented following EHR implementation. All considered, this study argues that patient care becomes significantly more standardized as opposed to patient oriented considering collective affordance actualizations such as the aforementioned. Second, the present study found that collective actualization of capture and archive affordances changes the work activities of medical professionals that involve reporting and administration of administered care.
Furthermore, this study looked at the impact on activities concerning interprofessional interaction and collaboration. The findings support Chen (2013) who described that EHR implementations would lead to reduction of face-to-face contacts. Collective affordance of monitoring and incorporation affordances are found to increase coordination in hospitals but simultaneously negatively impact the perceived level of hierarchy. Especially nurses were found to experience an impact due to collective actualization by medical specialist. Much effort that went into bridging the two professions and establishing unwritten rules concerning the execution of medical work has been nullified by implementation of EHRs. And finally, collective actualization of access and use affordances is found have potential to benefit decision making of EHR users and by making a diagnosis. However, as Carayon et al. (2015) already found, the time EHR users spend on clinical review has increased dramatically following the implementation of EHRs. This study confirms that user’s time management is complicated and indicates that collective actualization of access and use affordances is one of the reasons. Because of the indicated contradiction healthcare professionals are not better or less equipped in terms of data utilization following EHR implementations.
Taken together, this study looked at the impact of collective affordance actualization on clinical workplace activities whilst incorporation the influence of the social-historical context in which they are executed. By doing so, this study uncovered multiple contradiction that followed the implementation of EHR in hospitals. These contradictions spurred to the changes in clinical workplace activities and helped to understand the mechanisms that lead to these changes.
5.2 Theoretical contribution
26
there were always some who did and some who did not actualize or even identified a certain affordance. Furthermore, this study affirms the claim by Fayard and Weeks (2007) that organizational activities are often jointly performed and therefore ask for simultaneous analyzes. Examples are the impact of reporting on utilization of data at a later stage and coordination on delivery of care.
Finally, answered to the call to include organizational context in analyzing IS affordances (Strong et al., 2014; Evans, Pearce, Vitak & Treem, 2016). By employing CHAT besides theory of affordances this study has been able to place the actualization of affordances in their sociohistorical context. The limited available research on the combination of these two theories make this study a valuable contribution to this stream in literature. The dual use of both theory of affordances as well as CHAT proved to be very helpful in understanding how users acted on contradiction regarding the execution of their workplace activities and in the consequent changes following the contradictions.
5.3 Strengths and limitations
By combining two previously compiled data sets this study is based on an extensively rich observation and interview data. The strength of this research is that not only multiple hospitals but also different departments within some of these hospitals are included. By doing so the findings of this study are relatively generalizable for hospitals that use other EHRs as those represented in the data. Furthermore, the inclusion of nurses and residents besides the in the existing literature often preferred stance of the medical specialist (McGinn et al., 2011) allowed for a more complete picture. Especially, since the work activities of these three groups are often interrelated.
Several limitations within this study need to be considered. First, the set up in which this study has been organized did not allow for longitudinal data collection. Only cross-sectional interview and observation data was available and therefore no longitudinal conclusion can be drawn. This makes the conclusion a little debatable because it is difficult to speak of changes in activity systems when only the situation at one point in time is available. However, the way in which the interviews were conducted and the consciously asked question about the previous situation did allow for careful conclusions to be drawn. Additionally, when interviewees are asked to recall about a previous situation different biases can have occurred over time through which they can give a distorted recollection. Second, even though the data was compiled of interviews and observations conducted at different hospitals, causing two different EHR systems to be represented in the data, careful consideration is required in generalizing the findings. Different EHRs might allow for different affordances to be actualized. Furthermore, to extend the findings to other IT artefacts and contexts in which knowledge is essential to execute work asks for even more precaution. If professionals in other contexts recognize themselves in the findings of the present study that might indicate that the findings are generalizable to their context, but generalization should be proceeded with caution
27
The implications for practice are threefold. First and foremost, the findings of this study should incentivize healthcare managers to revise the initial goals they envisioned for EHR implementation. This study has found that EHR affordance actualization results in workflow changes that complement the hospital’s objectives but also in several that undermine those objectives. Interestingly the undermining changes in workflow are often unintended and unforeseen. Now that we know that collective EHR affordance actualizations can bring about changes in medical work activities, managers should evaluate whether the changes are worth the goals they pursue. Furthermore, this study highlights the need to consider the need to not misconceive the activity system in which you implement the EHR. Managers should analyze the fit between the EHR, subject, rules, community, division of labor, and objectives.
Second, the findings of this study might inspire EHR developers to think beyond features and focus on the different ways in which features might be used by end users to execute their workplace activities. In software development it is critical to evaluate current performance to know what and where to improve. This study shows that merely looking at adherence to intended use does not tell the whole story. Developers should go beyond data analyzation by actively engaging with the end users and invite them to share about how they use the software and how this impacts their work. Only then can software be developed that not just dictates workflow but facilitates and compliments it. And finally, this study gives insight to EHR end users on why they sometimes get frustrated and annoyed by using the software. During implementation more emphasize should be on the side effects of usage. EHR should not be introduced as a mere replacement of paper patient records but the truth should be shared about the consequences of EHR usage.
5.5 Future research
28
to identify and actualize the affordance. Further research would thrive by a broader defined definition of the concept. Finally, researchers are encouraged to study whether other knowledge professions alike that of the medical professional such as pilots or accountants experience the same contradiction following the implementation of EHR. It would be interesting to find out if and IS implementors can learn from other disciplines.
6.
Conclusion
29
