Abstract
Objective: The aim of this study is to explore whether specific ethical questions arise with the use of a shared electronic health record (EHR) system, based on the daily experience of primary care physicians (PCPs).
Methods: In this qualitative research project, we conducted 14 in-depth semistructured interviews with PCPs in a tertiary hospital setting.
Results: We identified 4 themes: 1) PCPs describe the EHR as a medicine with side effects, for which they provide suggestions for improvements; 2) A shared record raises ethical questions related to autonomy and trust; 3) Although use of the EHR often disturbs rapport with the patient, it can also support the patient-doctor interaction when it becomes an active part of the conversation; 4) A shared EHR may cause health care providers (and their relatives) to avoid seeking help for sensitive issues.
Discussion: PCPs fear access to results could cause confusion and anxiety in patients, resulting in tensions between autonomy and beneficence. Improved efficiency and quality of care with a shared EHR relies on doctors trusting each other's input to avoid duplicate tests. However, this might compromise a fundamental skeptical attitude in practicing medicine, and we should be aware of a risk of increased confirmation and anchoring bias.
Conclusion: The EHR is considered to be a work in progress—EHR design could be improved by examining physicians' coping strategies and implementing their suggestions for improvement. Ethical questions related to autonomy, trust, and the status of records that belong to doctor–patients need to be considered in future research and EHR development.
- Electronic Health Records
- Health Information Exchange
- Medical Ethics
- Medical Informatics
- Physician-Patient Relations
- Primary Care Physicians
- Qualitative Research
- Quality of Health Care
- Tertiary Care Centers
Already in 1982, Mark Siegler described confidentiality as a “decrepit concept.”1 He estimated that somewhere between 25 and 100 health professionals and administrative personnel legitimately accessed a patient's record during a standard hospital stay, noting that a conflict arises between our traditional understanding of medical confidentiality and our desire to provide better and more comprehensive care to the patient.1–2 The electronic health record (EHR) allows for much more information to be stored and accessed by a wider circle of both medical personnel and patients. At the same time, it has become a complex and multifunctional tool also serving other functionalities such as financial reimbursement, risk management and quality and safety improvement.3 Although the EHR has been in use for quite some time now, its effect on the quality of care remains unclear.4⇓⇓⇓⇓–9 Correspondingly, its impact on the patient-doctor relationship also remains a topic of research.10,11 Two issues often arise: time spent on the computer for documentation takes away time directly spent with the patient, and the EHR can have a negative impact on the quality of the interaction (or rapport) with the patient.12,13 However, patient satisfaction does not seem to be negatively influenced by EHR use.14 With this study, we search to understand which ethical questions remain relevant today within a shared electronic record, specifically related to the triadic relationship between patient, doctor and EHR. Although several studies focus on challenges of EHR use, far fewer actually consider the lived experience of the physician.
Objective
The central aim of this qualitative research is to understand how primary care physicians (PCPs) integrate the EHR into their workflow and whether specific ethical questions arise while sharing medical information both with other practitioners and with patients.
Materials and Methods
Theoretical Framework
This study adheres to a framework of technological mediation and postphenomenology. A postphenomenological approach to technology combines the empirical orientation of Science and Technology Studies with a more conceptual philosophical analysis.15 A technology is then perceived as something that mediates the way we experience the world.16 This also applies to the moral role of technologies: “by mediating human experiences and practices… technologies mediate moral decisions and actions.”17 In the classic example of the gun debate, technological mediation does not assume that “guns kill people,” nor that “people kill people,” but the “gunman,” a hybrid of human and nonhuman elements, kills people.17 The gunman becomes a new entity with specific moral options. In this sense a technology is neither a moral agent, nor a neutral instrument: it is a moral mediator. The central perspective is that of the technologically mediated subject.17 This theoretical framework informed the development of interview questions and served as a background for data analysis, adhering to an ethical-constructive technology assessment approach.18
Design
Within a qualitative research protocol, semistructured interviews with PCPs were studied using thematic analysis. A thorough literature search informed the interview guide (see Appendix). Although the initial questionnaire contained questions referring to the moral concepts of autonomy and confidentiality, it also allowed for participants to address other themes. The interview guide was developed through intensive deliberation with all authors until consensus was reached. Responses from previous interviews were integrated in the follow-up questions presented to succeeding participants in an iterative process. The first author conducted all interviews. She is a Belgian primary care physician (PCP) with a Master's degree in Philosophy who has completed a course in qualitative research methods in Public Health, received additional training in interviews with experts, and gained experience interviewing doctors in a previous study. The research protocol was approved by the Institutional Review Board of the University of Pittsburgh (IRB Number PRO17010359).
Participants
Participants were recruited in an academic general internal medicine outpatient practice embedded in a tertiary hospital in Pittsburgh, PA, through a purposive, convenience-based sampling process.19 The practice includes 46 faculty physicians. Physicians use Epic software and patients can access certain portions of the EHR and message their doctors through a portal system. The second author facilitated recruiting physician participants. The first author presented the study at a regular meeting of PCPs, after which they were given the opportunity to sign up for an interview. The main characteristics of the sample are shown in Table 1. Fourteen physicians consented to an interview. All interviews took place between February and May 2017 and lasted between 30 and 74 minutes. Data saturation was reached after 11 interviews, but 14 interviews were completed to establish a dataset that is both rich and thick.20–21
Characteristics of the Sample
Data Analysis
All interviews were recorded digitally, transcribed verbatim and coded in NVivo 12 (QSR International, Burlington, MA) by the first author. Three interviews were also studied and coded by a second researcher (MS) and elaborately discussed until consensus was reached. The first author continued the coding process with the established code book, expanding with new codes when necessary. The procedure of including a second researcher (MS) extended throughout the data analysis process and combined peer debriefing with a systematic audit trail to assert trustworthiness and credibility of the data analysis.22 The final coding scheme consisted of both deductive (based on the literature search and interview questions) and inductive (emerging from the interviews) codes. The thematic analysis process adhered to the 6 phases described by Braun and Clarke.23 All coauthors provided input at different stages of this process. Moreover, both the Standards for Reporting Qualitative Research (SRQR) and Consolidated criteria for reporting qualitative research (COREQ) guidelines for reporting qualitative research were observed.24–25
Results
Three major themes and 1 minor theme were identified: 1) PCPs describe the EHR as a medicine with side effects, for which they provide suggestions for improvements; 2) a digitally shared record raises ethical questions related to autonomy and trust; 3) although use of the EHR often disturbs rapport with the patient, it can also support the patient–doctor interaction when it becomes an active part of the conversation; 4) minor theme: a shared record may cause care providers (and their relatives) to avoid seeking help for sensitive issues. A selection of quotes per theme is available in Table 2.
A Selection of Quotes Per Theme
A Medicine with Side Effects
All participants paint a carefully nuanced picture of the electronic record. They acknowledge it as an effective tool with a positive impact on the quality of care. They highly value having medical information available “at their fingertips”: in real time, searchable and easily accessible. They also deem the information to be more accurate. The record helps them in preventive care as well as chronic disease management with extra functionalities of checklists, reminders, and smart algorithms. Some doctors emphasize how the EHR facilitates patient population management and reaching quality targets. Not only does the EHR solve many practical issues occurring with paper records, it also informs the PCP about everything that happens to the patient outside of the consultation room:
“I've actually been here long enough to remember the paper chart days, and I remember the transition to the EHR. I can do a comparison and I know so much more about what happens not in my office now. I know what happened in the emergency department last week. I know what the consultant told them. I would not have known these things. The patient could not really tell me these things a lot of the time, so I was in the dark a lot, and I feel like that is not the case anymore.” (PCP 7)
However, many “side effects” were also mentioned during the interviews. First, 9 physicians describe how the EHR interrupts the conversation with the patient and disturbs rapport. Second, half of the participants experience alert fatigue due to the many pop-ups that are not always useful to them. Third, the EHR is described as “bulky” or “clunky,” because of the phenomena of overdocumentation and copy-paste notes. Navigating all the available information requires a lot of clicking, leading to the experience of the EHR not being user-friendly. Fourth, PCPs generally think the EHR is time-consuming because it requires a lot of extra work. Some interviewees specify that this extra work is a consequence of the modern requirements of practicing medicine: teamwork, prevention and chronic disease management all intensify the need for registration. And fifth, the record is experienced by half of the interviewees as rigid, both in how difficult it is to change certain information, and in one-size-fits-all alerts that are not adapted to individual patient characteristics.
Overall, the record is described as a work in progress. All but 1 physician also provide suggestions for improving the above mentioned side effects, such as the addition of cross-referencing, risk calculators and scheduling, and an upgrade of search methods, messaging, automation and overall user-friendliness. Moreover, many physicians have developed coping strategies to integrate the computer into the conversation with the patient. This is further explored in the third theme. Two doctors emphasize the importance of including physicians into the design process of EHR development.
Ethical Questions
With (certain parts of) EHRs being accessible to patients, long-existing ethical questions mainly involving autonomy and trust are being reformulated within the digital patient-doctor relationship.
Conflict between Respect for Autonomy and Beneficence
Twelve interviewees describe situations in their daily workflow related to the EHR where they experience a conflict between the principles of respect for autonomy and beneficence. When medical results are shared electronically, doctors worry that patients may find it difficult to interpret these results correctly, potentially causing confusion and anxiety. Some PCPs suggest providing an explanation with the results. Others argue it would be better to select which information becomes accessible, although they acknowledge this could result in a paternalistic attitude:
“I think—and I do not mean this in a patronizing way—I think patients would be better served if we could pick and choose information that we give them. Now I know that could be abusive, you could delete or not send them things. But… I still think that they do not need every single solitary thing in the chemistry panel that has no relevance to them.” (PCP 2)
The majority of interviewees are not comfortable with patients adding or changing information, unless under strict conditions (such as source identification, separate sections, or limited to certain types of information, for example from self-tracking activities).
There is no clear uniformity in the answers to the question what purpose patient access should serve. Some mention patient education and engagement, while others highlight a better compliance as a preferred outcome.
One PCP described a case that fleshes out the ethical conflict between respect for autonomy and beneficence at the level of a shared EHR. He talked about his strategy to include the background story of a patient to his notes, a patient who was oftentimes written off by medical practitioners as “difficult.” However, when he came to understand her background of rape, prostitution, and homelessness, he could see her in a different light. By documenting this in the EHR and making it accessible to colleagues, he hoped they too could see her in a more compassionate way. When asked whether he had discussed this approach with the patient, he responded:
“That is a good question, I've never asked her specifically if she knows that it is in there… That is a really interesting ethical question. I've never thought about it, because my goal in documenting it was to decrease the judgment that she would get. And I think that it probably will if someone reads it. But… I never thought to ask her if that is something that she would want redacted… I think there would be a lot of emotions involved that would have to be overcome for her to appreciate the benefits of the documentation.” (PCP 9)
This clearly shows how some doctors struggle between respect for autonomy—by consulting their patients' preferences and providing them with a comprehensible explanation—and beneficence, where they act in the best interest of the patient but risk a certain level of paternalism.
Trust Versus Skepticism
EHR use impacts trust both in the patient–doctor relationship and in doctor–doctor interactions. Regarding the patient-doctor relationship, some PCPs note that a minority of patients do not trust EHRs (or the way they are used and shared) because so many care providers and other personnel can access their information. Some physicians use the information that colleagues made available as a way to double-check what patients tell them:
“It is also helpful when patients try to hoodwink you by saying: “oh, that doctor said you should give the prescription”—they give you information based on what their needs are. And you recognize that is not the same as what was documented by the other physician.” (PCP 8)
Contrarily, another participant thinks patient access to the record may actually increase trust and compliance in patients.
When it comes to doctor-doctor interactions, 3 participants describe how they are skeptical of information in the record added by colleagues:
“I always say: take it with a grain of salt. That history is only as good as what the person before you heard—the history has nothing to do with what was said.” (PCP 14)
PCP 1 also mentioned how he could be put on the wrong track when reading a patient's record before seeing them: often the face-to-face contact would lead him to a very different conclusion than the one he had in mind solely based on the information in the record. In this sense, the diagnostic process requires a certain level of skepticism.
EHRs' Disturbing and Supporting Impact on the Patient–Doctor Interaction
“I enter the room, introduce myself to [the patient] and do not touch the computer. It has been a personal decision I made a long time ago, that on engaging with the patient not to focus on the computer. First the patient. Second, it has to do with establishing rapport with the patient.” (PCP 5)
As mentioned in the first theme, a majority of participants feel the EHR disturbs rapport within the patient-doctor conversation, mostly by interfering with eye contact and breaking the flow of the narrative. Twelve physicians comment on the design or configuration of the examination room and how this may hinder (or facilitate) computer integration into the conversation.
To avoid a negative impact, PCPs describe a diverse set of coping strategies: studying the record before the encounter (“chart prepping”), avoiding the computer during (part of) the conversation, giving patients an explanation for what they are doing while interacting with the computer, or jotting down notes on paper instead of using the EHR.
On the other hand, the EHR can also be a tool that supports the conversation. Despite the hindrance computers cause, most physicians provide examples where they share the screen with patients: to discuss results, to look at health maintenance overviews or to reconcile medication lists. This active engagement of the computer into the conversation with the patient seems to counterbalance the negative effects of computer use. About half of the participants provide examples of beneficial computer practices, such as using information in the EHR as a conversation starter, or as a tool for patient education.
“The EHR can summarize a lot of data and serve as the starting point for a very constructive conversation between the patient and the physician that I think can actually cultivate trust in the relationship.” (PCP 13)
Sensitive Information: Care Provider as Patient
When talking about handling sensitive information in a shared record, 6 physicians spontaneously mentioned examples of care providers or their close family members seeking medical help.
“A good example is: one of my patients was battered by her husband who is a faculty member here. That would be very sensitive. So, those would be… what I consider to be sensitive.” (PCP 6)
“I think some patients are very concerned about some of their information [being] available, especially if they are health care providers themselves.” (PCP 8)
These participants were especially careful about handling sensitive information of physicians working at the same center. The awareness of “unfettered access to charts” (PCP 5) by physicians withheld some of sharing sensitive information in the record. Two participants also mentioned examples related to family members of faculty requesting information being removed from their records, one related to sexual orientation and another to antidepressant medication. Another interviewee was worried about physicians getting individual insurance in their next job, if certain information—such as use of antidepressants—became accessible to third parties.
Discussion
The interviewed physicians describe the EHR as a “medicine with side effects.” We see a similar analysis of benefits and concerns in previous research.26 First and foremost, participants view EHR use as beneficial to quality of care. This is in line with other studies that focus on physicians with EHR experience,27,28 although another interview study found the opposite.29 Two doctors rightfully note that a more intensive collaboration between users (both physicians and patients) and EHR developers could prove beneficial in creating a more efficient digital record and avoiding problems.
Autonomy Versus Paternalism
Our participants are struggling to find a balance between beneficence and respect for autonomy.30 Previous research has shown that care providers are skeptical of giving patients granular control over their EHR information, fearing it could damage the patient–doctor relationship and negatively impact quality of care.31 Patients, on the other hand, value individual control and privacy.32 Moreover, some physicians consider a better patient compliance to be the most important outcome measure for patient portal systems. Patient empowerment and compliance do not always coincide and EHR design will probably vary depending on which goal is prioritized. Although patient portals have been active for some time now, it is not clear whether patients also understand the information in the EHR—although this is a vital prerequisite for autonomy and informed consent.33 Doctors in other studies express similar concerns about potentially harming patients by creating confusion or anxiety, especially when interpreting lab results.26,34
We argue that these findings urge us to rethink which goal a shared EHR should serve: patient empowerment? Compliance? Or something else? Second, we should consider how patient portals can benefit patients in a meaningful way, through user-friendliness, comprehensible information presentation, and support for vulnerable patients.
Trust within Doctor-Doctor Interactions
EHR integration aims to improve quality of care by decreasing redundancy and by avoiding unnecessarily repeated tests.6,26,35 To reach this goal, trust between care providers is essential. Many aspects of EHR design aim at exchanging information that needs to be trusted by the next care provider. This study shows that skepticism, understood as an attitude aimed at reducing anchoring and confirmation bias, could be an underestimated aspect of the diagnostic process. A certain level of skepticism remains an important diagnostic tool that may not get the attention it deserves in terms of digitalization of the record. In medicine, a vital part of the job is to identify omissions, which sometimes requires “starting all over again.” One may wonder whether the current EHR design risks increasing anchoring and confirmation bias. Interestingly, the solution may also be digital in the form of decision aids. A deliberate reflection protocol and diagnostic checklists have shown small but consistent benefits in reducing errors.36
Active Integration of the EHR into the Patient-Doctor Conversation
When a physician is fully absorbed by the computer screen and loses focus on the patient, the EHR may undermine rapport. On the other hand, an active use of the computer in the conversation which also engages the patient, can prove beneficial. This is in line with previous research results: active information sharing allows for more patient engagement with the EHR, reducing the negative impact of the computer on the patient-doctor interaction and improving collaboration.37 Several strategies such as using the computer to facilitate conversation, adjusting room design and maintaining eye contact while typing are evidence-based recommended practices, which are also acknowledged here.11,38 Despite concerns some PCP expressed, sharing notes with patients does not seem to negatively affect the patient–doctor relationship.39
Effect of the EHR on Care Providers' Help-Seeking Behavior
Several participants highlight difficulties in treating sensitive issues presented by doctor–patients, especially when it concerns faculty working at the same hospital or their family members. An increasing body of literature shows that physicians often resort to self-treatment and experience multiple barriers in help-seeking behavior.40⇓–42 This study reveals that digitally sharing information might actually be another, rather unexpected, barrier in help-seeking behavior of medical care providers, one that intersects with previously examined barriers such as a feeling of embarrassment, cultural issues, and confidentiality questions.42 Should physicians' records be given a more stringent confidentiality status? More research is needed to understand how information sharing through EHRs affects help-seeking behavior of medical personnel.
Conclusion
This study, based on the daily experience of PCPs working with their EHR system, identified ethical questions related to patient autonomy, trust, and help-seeking behavior of doctor-patients within the triadic relationship of patient, doctor and shared EHR. These findings could inform both EHR design and future research, and is one of the first steps into creating an ethical framework for a shared EHR.
Acknowledgments
The authors thank all participating primary care physicians for the time and effort they put into this interview study. We also thank 3 anonymous reviewers for their useful comments and suggestions.
Appendix
Interview Guide PCPs
Oral Informed Consent Script for PCPs:
We want to thank you for your willingness to participate in this study.
We are asking PCPs to help us by participating in interviews as part of a research study focusing on data sharing in and through the EHR. We would like to hear from you how you experience working with the EHR in your day-to-day practice, what strengths and weaknesses you see, and how you think about sharing information electronically with other care providers and patients.
Your participation will include being interviewed for approximately 45 minutes. The principal investigator, Tania Moerenhout (me), will conduct the interview. It will be audio recorded, but you will not be identified on this recording; interviews will be stored anonymously. If you prefer not to answer any of the questions, you may say so and the interviewer (I) will move on to the next question. Your insights from the interview will only be used for study purposes and to produce a report in a scientific journal which may be presented at scientific conferences. Your name or other identifiable information will not be mentioned. Your participation is completely voluntary: in case you want to withdraw from the study, you can do so at any time without any negative consequences. There are no particular risks involved in this study. Should you have any other questions or comments afterward, you can contact either the interviewer or the supervisor of this study. You will not be provided any incentive to take part in the research.
Do you agree?
Background questions:
Age; gender; years in practice; type of practice; hours of patient contact per week (consultation blocks); use of EHR program (which program, use of on-line appointment system); use of digital patient interactions (e-mail, Amwell, etc.).
Topic 1: General Attitude Toward EHR, Use of EHR
➣ If you think about the last contact you had with a patient: Could you walk me through the different steps of the visit, and how you used the computer?
o Prompt: If you think about the start of the conversation, as of which point do you use the computer?
➣ How do you integrate the use of the computer in the conversation with the patient and/or with other care providers?
➣ How does the computer affect your interaction with the patient?
o Prompt: what efforts do you put in to reach (positive) or overcome (negative) this effect?
o Do you engage patients in this computer work?
o Do patients ask questions about their EHR?
Topic 2: Impact on Quality of Care
➣What influence would you say does the EHR have on the quality of care you provide?
➣ What are the main benefits of the EHR use to you?
➣ What are the main problems or pitfalls?
Topic 3: Data Sharing; Shared Confidentiality (with Other Care Providers); Patient Access
➣Thank you for your answers. I would like to talk a bit more about sharing information with other care providers and patients on-line. Let us start with other care providers. How do you experience the online sharing of patient information with other care providers?
➣ When talking about digitally sharing patients' information with other care providers, 1 question that often comes up is how to handle sensitive information.
• First of all, what do you consider to be sensitive information in the EHR in this context?
• How do you handle this in the data sharing process?
• Prompt: specify sensitive information if necessary (reproductive/sexual health, substance abuse, mental health information, genetic information, domestic violence (NCVHS, see Caine et al, 2015))
➣Patients can also access their EHR. They can check certain parts of their record on-line, through the patient portal. Do you know what patients can or cannot see on the patient portal?
• How do you feel about patients accessing their records?
• Could you describe your experience with this function?
• According to you, what features or functions would an ideal patient access system have?
• Prompt: add, hide, remove information?
➣ Given these functions of data sharing with colleagues and with patients, does this influence the way you use the EHR?
Topic 4: Multifunctionality of EHR
➣ We have just discussed 2 functions of the EHR, being data sharing with colleagues and patients—of course it is used for more than that. I have created cards with 5 functions of the EHR, on which I would like to hear your input. According to you, how important is each function? Would you add any other functions, or leave out a certain one? (5 functions: storing of patient information/patient care, payment system (billing), quality-of-care control, generating research data, patient access/sharing of patient information).
Topic 5: Digital Double/Digital Patient: Future
I have one final question to round up this interview. Imagine you would treat a patient solely based on the EHR without knowing the patient beforehand/in real life: how would you respond to this? On what kind of conditions would you consider this?
Ok, thank you for your responses! Before we wrap up, would you like to add anything else?
Notes
This article was externally peer reviewed.
Funding: TM is supported by a Clinical PhD fellowship (Research Foundation Flanders–Fonds Wetenschappelijk Onderzoek: Grant Number 1700618N). MS is supported by a BOF grant (Ghent University: Grant Number BOF15/GOA/011).
Conflict of interest: none declared.
To see this article online, please go to: http://jabfm.org/content/33/1/106.full.
- Received for publication April 28, 2019.
- Revision received August 20, 2019.
- Accepted for publication August 20, 2019.
