Health Communication: Identifying how Patient Portals Impact Communication in Oncology

Abstract

Patient portals are becoming widespread throughout healthcare systems. Initial research has demonstrated that they positively impact patient-provider communication and patients’ health knowledge, but little is known about the impact of patient portals in the cancer setting, where highly complex and uncertain medical data is available for patients to view. To better understand communicative behaviors and perceptions of the patient portal and how it is utilized in oncology, in-depth, semi-structured interviews were conducted with 48 participants: 35 patients and 13 oncologists. Thematic analysis identified that portals help to enhance participation during in-person consultations, increase patients’ self-advocacy and builds rapport with providers. However, patients’ comfort level with reviewing information via the portal depended upon the severity of the test. Oncologists worried about patient anxiety and widening health disparities, but acknowledged that the portal can motivate them to expedite communication about laboratory and scan results. As patient portals become more widely used in all medical settings, oncologists should become more engaged with how patients are viewing their medical information and consider the portal within the framework of patient-centered care by valuing patients’ communication preferences.

Web-based applications that provide patients with 24-hour access to their electronic medical records (EMR), otherwise known as patient portals (HealthIT.gov, 2017), are increasingly being utilized by healthcare organizations (Ancker, Silver, & Kaushal, 2014; Hsiao et al., 2013). This increase is partly due to financial incentives provided to hospitals and clinicians to encourage portal usage under the Health Information Technology for Economic and Clinical Health Act (Blumenthal & Tavenner, 2010). Patient portals offer functional benefits to patients as they enable convenient patient access to EMR data from devices, such as personal computers or smartphones, they allow the ability to request medication refills, schedule appointments and they support secure messaging (Ammenwerth, Schnell-Inderst, & Hoerbst, 2012). These types of functionality have led to positive patient reports of using portals (Woods et al., 2013) and their availability has been shown to influence patients’ choice of provider (Walker, Meltsner, & Delbanco, 2015). Such portal access offers benefits to patient self-care as their use has been shown to, a) improve patients’ understanding of their condition (Alpert, Krist, Aycock, & Kreps, 2016), for example, providing a mechanism by which patients can confirm a “normal” test result (Keren, Muret-Wagstaff, Goldmann, & Mandl, 2003), b) promote self-management behaviors (Urowitz et al., 2012) and c) improve medication adherence (Sarkar et al., 2014).

Such evidence for patient benefits of portal use has been provided by researchers working in the primary care setting. For instance, portals aided disease management and increased the quality of consultations (Bartlett, Simpson, & Turner, 2012) and viewing records positively affected communication, enhanced knowledge, and allowed for greater participation (Woods et al., 2013). In this setting, patients typically access their EMR for a limited array of uses such as to review common laboratory tests from annual physicals or routine screening (Weingart, Rind, Tofias, & Sands, 2006).

Portal impact on Physician-Patient Communication

There is emerging evidence of portal usage in other health care settings suggesting that patients use portals in similar ways to those in primary care and experience similar benefits. Yet, primary care providers have different perceptions about how elements of medical records are displayed to patients compared to non-primary care providers, which impacts the timing of communication and presentation of information (Redd et al., 2015). For example, studies have revealed that pulmonologists (Tulu et al., 2016) and surgical specialists (Shenson, Cronin, Davis, Chen, & Jackson, 2016) found that patients were relying on the secure messaging options available in portals as a primary means to seek disease information and medical advice without in-person consultations. While asynchronous communication may be beneficial outside of normal office hours in certain instances (Shenson et al., 2016), researchers have questioned whether such portal use can accommodate more complex patient communication needs, since patients and providers desire a more personalized channel to ask and answer questions (Alpert et al., 2016).

Portals have also been touted as having the potential to enhance patient-centered communication (PCC) by increasing patient involvement through secure chats, and improving access to care by enabling diverse populations to view their records (Shaw & Ferranti, 2011). In fact, improving PCC by fostering greater patient involvement was one of the main motives for designing patient portals (Sands & Halamka, 2004). Moreover, portals facilitate the implementation of the Institute of Medicine’s appeal to improve patient-centeredness by providing patients with greater transparency through access to their own medical information. This in turn promotes patients’ ability to make more educated decisions in collaboration with their providers by being more informed about treatment plans and involving family members or caregivers if so desired (2001). Yet, in spite of these claims, there is evidence of barriers to improved PCC using portals, mainly due a) to their lack of user-friendliness and b) that some patients do not understand and are not prepared to interpret complex medical results (Giardina, Modi, Parrish, & Singh, 2015).

Portals and Oncology

As with primary care, the potential for positive outcomes of portal usage, noted above, also exists in the oncology setting. However, few studies have explored provider perceptions within this setting. One study examined the operational management of portals by nurses (Gerber, Shaalan, Duncan, Gill, & Craddock, 2017), while another found that oncologists were apprehensive about patient’s portal use due to the possibility for viewing sensitive tests (Alpert, Morris, Thomson, Matin, & Brown, 2018). Little is known about the impact portals have on patient-oncologist communication during their clinical interaction. Portals are increasingly being used as a method for cancer patients to quickly access potentially threatening disease and prognostic information (Gerber et al., 2014), without the benefit of a face-to-face consultation with their oncologist. However, several significant concerns have been expressed by oncologists about their patients’ use of portals. Oncologist question whether patients are able to fully understand the array of complex and potentially life threatening medical information contained in both pathology and scan reports (Rexhepi, Åhlfeldt, Cajander, & Huvila, 2016; Wiljer et al., 2010). These concerns are shared by providers in general, as they perceive that patients who are largely untrained to interpret such data may become confused (Callen et al., 2015) and highly anxious about their health condition without direct provider support (Boeldt et al., 2015). Particularly with cancer, clinicians’ lack of knowledge about how to use portals efficiently and concerns about the time and effort required to optimize the patient’s portal experience is significant due to the increased uptake and intensity of portal use by cancer patients (Laccetti et al., 2016).

Risk Information Seeking and Processing Model

To better understand the impact of patient portals used in oncology, the Risk Information Seeking and Processing Model (RISP) was applied as the conceptual framework for this study. RISP seeks to bring together components such as patient preferences for: a) information and decision-making, b) seeking and processing risk information and c) appropriate channels to communicate risk information (Griffin, Dunwoody, & Neuwirth, 1999).

The model has significant relevance to the current project as patient portals represent a new communication channel with the capacity to transmit risk information, which is defined generally as “things, forces or circumstances that pose danger to people or to what they value” (Stem & Fineberg, 1996, p. 215). Often, information-seeking and processing strategies are omitted from risk communication studies, making the RISP model highly relevant to utilize (Kahlor, Dunwoody, Griffin, & Neuwirth, 2006). In the context of this study, risk information includes prognostic or diagnostic information, such as a new cancer or a cancer recurrence. RISP is adopted from components of Eagly and Chaiken’s Heuristic Systematic Model of Information Processing (1993) and the Theory of Planned Behavior (Ajzen, 2005; Fishbein & Ajzen, 1975). Constructs from RISP, such as channel beliefs, information seeking behavior and information sufficiency were instrumental during the design phase of this study in order to guide the development of research instruments.

While there is some available evidence about the frequency and range of patient portal usage, the impact of changing oncologist-patient communication has been overlooked. Thus, the goal of this study was to understand the perceptions of oncologists and cancer patients about the potential impact of portals on such communication. Using in-depth semi-structured interviews with both oncologists and patients, we aimed to answer the following research questions:

RQ1: How do portals alter oncologist–patient perceptions of communication in oncology?

RQ2: What are patients’ and oncologists’ preferences for using the patient portal to communicate risk information?

RQ3: What is the impact of changing oncologist-patient communication, due to patient portal usage, on the patient-oncologist relationship?

Methods

Setting

This study took place at a National Cancer Institute designated cancer center in central Virginia. Over 70,000 patients are connected to the patient portal throughout the health system. Recruitment for this study occurred between May and September 2016 and interviews with both oncologists and patients were performed concurrently. This study was approved by the local Institutional Review Board of Virginia Commonwealth University.

Participants

Oncologists

Purposeful convenience sampling was used to recruit oncologists, in which individuals are identified that are especially knowledgeable about the phenomenon of interest (Creswell, Plano Clark, Gutmann, & Hanson, 2003). Members of the research team presented an overview of the study at hematology, radiation and surgical oncology service meetings. All 46 oncologists present at these meetings agreed to be contacted further about participation. Providers were mandated to use digital records, thereby making all oncologists eligible for the study. The research team requested a list of oncologists present at each meeting, and then a random number generator was used to select three oncologists at a time to be contacted for participation. Sampling concluded once thematic saturation was achieved (Glaser & Strauss, 2009). In total, 22 oncologists were randomly selected and sent an email inviting them to participate in an interview. Informed consent was reviewed individually with each oncologist and written consent was obtained from willing oncologists prior to participation.

Patients

Once oncologists joined the study, members of the research team accessed the oncologists’ clinic schedules to identify eligible patients. Patient eligibility criteria consisted of: a) had attended an appointment with a participating oncologist within the previous four weeks from the date of identification to increase the likelihood of recent portal interaction, b) enrolled in the portal c) fluent in English and d) at least 21 years of age. Purposeful convenience sampling (Creswell et al., 2003) was one again utilized as we were only interested in patients currently using the portal because our objective was to learn about their experiences as portal users. A random number generator was used to select five patients at a time to approach for participation. In addition, six patients who had reported a negative experience using the portal to their oncologist were referred to the study team. Eligible patients were mailed a letter summarizing the study along with an opt-out form and a self-addressed stamped envelope. Patients were asked to return the opt-out form if they were not interested in participating. If no opt-out form was received within two weeks of the date the letter was mailed, a member of the research team contacted the patient by phone to discuss study requirements, obtain verbal assent to participate and to set up a time for the phone interview. Sampling concluded once saturation of the themes was achieved (Glaser & Strauss, 2009), in which a total of seventy-two eligible patients were randomly selected and approached to participate. Patients received a gift card to thank them for their participation.

Procedure

This study employed qualitative in-depth respondent interviews (Lindlof & Taylor, 2010; Tracy, 2012). A semi-structured interview guide was developed by the research team using a set of predetermined open-ended questions. The unstructured nature of the interview allowed for flexibility and fostered the participant to lead the conversation, rather than the interviewer dictating discussion points. Integrating the seven main factors from the RISP model as an organizing framework, interview guides were constructed to address the overarching components of the model, such as a) information and decision-making, b) seeking and processing risk information and c) appropriateness of channels to communicate risk information. The semi-structured interview guides were modified to enable use in both patient and oncology groups, and were designed to prompt their personal and professional experiences about using the portal. For example, both groups were asked variations of the question, “how has viewing / inputting health information on the portal changed the way you interact with oncologists / patients?” Additional questions written for each group can be found in table 1. Interviews were administered by members of the research team, audio recorded and subsequently transcribed. The verbatim transcripts were uploaded to Atlas.ti v.7.5.17.

Table 1.

Selected Interview Questions

Beliefs about the usefulness of information in various channels
Patient: How has viewing your health information on the portal changed the way you interact with your oncologist?
Oncologist: How has patient access to the portal affected the way in which you communicate with and contact patients?
Capacity to learn
Patient: How are you able to understand the medical information presented on the portal?
Oncologist: What do you think are the implications of portal transparency for patients who may not have access to the internet and/or a computer?
Affective response to risk information
Patient: Did you want to take action after viewing information on the portal, such as contacting your oncologist or seeking additional information? If so, what actions did you take?
Oncologist: In what ways, if at all, have you modified your communication style and/or your process of informing patients of diagnoses and treatment regimens?

Data Analysis

The research team analyzed the transcripts by performing a thematic analysis to develop appropriate codes and identify patterns (Boyatzis, 1998). First-level codes focused on the basic activities observed in the data, otherwise known as primary-cycle coding (Tracy, 2012), was completed independently by two coders on three transcripts at a time. Coders then met to compare codes, discuss memos written down during analysis, and mention exemplar quotes. A final coding framework was developed based on agreed upon codes and code definitions developed during these meetings. Two members from the research team met once a week to compare codes of one transcript, to ensure the coding process was being conducted in a similar manner without major discrepancies. Meticulous analysis of coding themes, combined with weekly meetings to discuss initial themes using a process of constant comparison within and between interviews contributed to building larger-order thematic classifications (Charmaz, 2014). Although an inductive approach was used for analysis, final themes were reflected upon the constructs from the RISP model to ground our findings within its framework. To ensure qualitative rigor and to clarify differences of opinion (Morse, Barrett, Mayan, Olson, & Spiers, 2002), the full research team discussed discrepancies until there was consensus. In addition, member validations (Lindlof & Taylor, 2010) of emerging themes occurred with at least three participants within each group towards the end of the interview schedule. Analysis reached its conclusion upon

Results

Participants

Thirteen oncologists and 35 of their patients enrolled, which represents a 59% and 45% recruitment rate respectively. Of the 13 oncologists, eight were recruited from hematology/oncology, four from radiation oncology and one from surgical oncology. The average age of oncologists was 47 and most (54%) were female and white (77%). Interviews with oncologists averaged 28 minutes in length and were conducted face-to-face in private offices, except for one phone interview.

Patients were mostly female (60%) with an average age of 54 years. The most common conditions were breast cancer (24%), hematologic (16%) and gastro-intestinal (13%). The majority of patients (71%) had known about their cancer diagnosis for at least six months before joining the study. Less than half (43%) of patients had metastatic cancer and about a third (34%) had stage two through four cancer. Those who refused participation were on average slightly younger and male. Patient interviews averaged 17 minutes in length and all were conducted over the phone. Descriptive statistics of patients’ demographics can be found in Table 2.

Table 2.

Sociodemographic Characteristics of Patients (n=35)

Variable	Frequency (%)
Sex
Female	21 (60%)
Male	14 (40%)
Race
White	21 (60%)
Black	14 (40%)
Asian	0 (0%)
Household Income
Under $19K	7 (20%)
$20K - $39K	4 (11%)
$40K - $49K	1 (3%)
$50K - $74K	8 (23%)
$75K - $99K	4 (11%)
$100K+	4 (11%)
Prefer not to say	7 (20%)
Education
Some High School	1 (3%)
High School Graduate	2 (6%)
Some College	7 (20%)
Associate Degree	2 (6%)
Bachelor’s Degree	13 (37%)
Master’s Degree	7 (20%)
Professional Degree	2 (6%)
Doctorate	1 (3%)
Area of Test / Diagnosis*
Breast	9 (24%)
Hematologic	6 (16%)
Gastro-Intestinal	5 (13%)
Genitourinary	4 (10%)
Lung	3 (8%)
Sarcoma	3 (8%)
Skin	3 (8%)
Other	3 (8%)
Gynecologic	2 (5%)
Current Cancer Status
Metastasized	15 (43%)
Stage 2	4 (33%)
Stage 3	4 (33%)
Stage 4	4 (33%)
	Mean (Standard Deviation)
Age in Years	53.7 (10.8)

^*Each diagnosis/condition counted separately for patients with multiple diagnoses/conditions

Themes

Thematic analysis of the interviews revealed three overarching themes and related sub-themes representing both patient and oncologist interviews: 1) patient participation, 2) appropriateness to communicate risk information, and 3) greater emphasis on the patient-provider relationship.

Patient Participation

Patients generally found the portal useful to access their medical information. For example, a breast cancer patient said, “It’s good to have everything in one place.” Patients reported that such access enabled them to better prepare for their visit with the oncologist and thus took a more active role in the consultation, characterized by asking specific questions, fostering interdisciplinary conversations, and strongly advocating for their own care. All 13 oncologists agreed that patient access to their records was beneficial because it could better equip them with knowledge, although there was also concern about excessive workload implications. The sub-themes of a) portal access promotes patient control, b) self-advocacy, and c) engagement in care, demonstrate the specific manner in which the portal triggered additional communication.

Portal access promotes patient control.

Prior to the health system’s uptake of the portal, patients rarely requested to see their records, likely due to a lack of awareness. Patients were pleasantly surprised by the newfound transparency gained through using portals. A woman with cervical cancer said, “The fact that they’re even sharing any of this with me is great.” She continued, “It’s helped me feel like I’m in control.” Even the simple act of confirming an appointment was valued by patients who had been previously frustrated by cancelations.

Patients were also able to better manage their care by directly contacting providers. A colon cancer patient said, “I will email to refresh their memories”, such as scheduling a CT scan, resulting in “[The office] e-mailing me back and saying ‘yes you’re right, we’re working on it now.’” Relatedly, when patients had questions or concerns, the portal enabled them to feel as if they could easily reach their provider in a timely fashion. Before the portal, patients would leave messages for their oncologist, but were unsure if the message ever reached them. Now, a perceived sense of direct communication was established.

Self-advocacy.

Aside from viewing records and helping with administrative tasks, the portal made patients “more hands on” about dealing with their health. Patients used self-assured words like, “involved”, and “informed”, when describing how they felt after using the portal. A skin cancer patient proclaimed, “I’m more assertive and ask more questions.” Patients typically described themselves as passive when learning information from their oncologist for the first time. However, information from the portal made patients more actively involved and prepared them for discussions with the oncologist. A man with prostate cancer explained:

[The portal] helped me because now when I walk into his office, he’s not giving me information that I don’t already have access to. I can be ready to ask him specific questions rather than generalizing everything.

In addition to being more participatory during consultations, the ability to access and print their medical information empowered patients to facilitate encounters with various providers. For example, a Leukemia patient said, “Since doctors don’t talk to one another, often I’m the one who has to say, ‘how is this medication going to affect this?’, so it enables me to talk in a more informed manner with different doctors.” In one extraordinary example, a patient with lung cancer discovered a high sugar uptake level on a recent lab report, so she brought the report to another provider who eventually identified a mass which led to a complete hysterectomy. About the incident, she said, “It’s absolutely vital to have access to medical records so [patients] can seek outside help if they need it.” Aside from interactions with providers, patients who have undergone treatment for an extended period of time used information on the portal to maintain their health, like monitoring their diet.

Engagement in care.

Oncologists believed that patients using the portal were more participatory during consultations than patients who did not use the portal. A medical oncologist said, “I think it helps patients be more engaged in their health care” because they are seeking information about aspects of their own care. Overall, oncologists spoke favorably about the ability for patients to retrieve pertinent medical information because it allowed for the review of their most recent consultation. However, there was a consensus among oncologists that greater patient participation would result in increased workloads. A medical oncologist said, “I have twenty patients getting scans, but I can’t check twenty patient’s [portals] ten times a day.” This sentiment was echoed by all 13 oncologists, in which there was concern that patients would anticipate rapid communication, no matter the time of day.

Appropriateness to Communicate Risk Information

The views of both patients and oncologists about the appropriateness of displaying risk information revealed that both groups recognized that the presence of risk information in the portal affected their desire to use it. The views of each group were expressed through the following sub-themes: a) test type influences communication preferences, b) oncologists’ fear of losing control over disclosing information, and c) potentially widening existing health disparities.

Test type influences communication preferences.

Patients acknowledged that the type of information being accessed influenced their level of comfort in using the portal. A leukemia patient indicated that he would be upset by learning a diagnosis and said, “If I’m dying, I wouldn’t want that put in…I would want a call.” A breast cancer patient made the distinction about the type of test being conducted and said:

When I had my biopsy done, I did not see a report [on the portal]. I saw the doctor first. And that was very helpful. I would not have wanted to read the report first. But if it’s just blood work or something like that, that’s fine.

Oncologists fear losing control of the communication process.

It was very noteworthy that 54% of patients read the results of laboratory tests and scan reports ahead of the consultation. The possibility of patients learning a new diagnosis or discovering a significant change in their well-being was nearly universally feared amongst oncologists. As a medical oncologist stated, “The biggest concern I have is that patients will see information prior to meeting with the physician. Particularly about things like bad test results or things that might suggest progression of their cancer.” Oncologists believed that in-person communication was paramount, and purposefully scheduled appointments in advance to discuss the implications of lab tests and scans. A medical oncologist recalled the importance of face-to-face interactions:

A patient had a recurrence in metastatic disease and was having personal issues at home. We were able to sit down and spend a lot of time to figure out ways that we could help. If I was talking to her over the phone, she would’ve just broken down.

Concordant with patient views, oncologists believed that certain information was acceptable to be communicated over the portal, while other types of information were not. For instance, a medical oncologist described the differences between treatment discussions and conversations about prognoses:

I’ve got somebody five years out and they’re having hot flashes, we can manage that by [portal] communication…If I’m having to tell somebody you’ve got recurring cancer, you’re going to die in five weeks, I can’t do that over the computer.

Oncologists were concerned about patients’ ability to check their labs and scan reports, yet this fear was somewhat diminished because most oncologists believed that their patients had minimal engagement with the portal. A radiation oncologist said, “I must admit, I think it’s a minority of my patients who truly use it.” Similarly, a surgical oncologist recalled:

I haven’t gotten much feedback from patients…I have not had a lot of patients looking at their pathology reports. Those are the things that I think could be of most concern, is patients finding out the pathology, a week before they have an appointment.

Although oncologists were responsible for the content in patients’ medical records, few made any effort to probe patients and understand the role of the portal in their care. A medical oncologist wondered, “[The portal] doesn’t seem to have been overly positive for patients, but I’ll be honest, I haven’t taken a lot of time to actually ask them. Maybe they really like it? I don’t know.” Since there was no training program, many oncologists were unaware of the portal’s features and what was available for patients to access.

Potential to widen health disparities.

One of the reasons oncologists believed patients were not engaged with the patient portal was due to low education levels and lack of Internet access in the local area. A surgical oncologist believed that patients without the ability to log-in to the portal would be “left out.” A radiology oncologist said, “I think it sort of discriminates against certain people. I mean, I think if we really wanted it to be across the board, we’d need computers in the waiting rooms.” Taking this point further, a medical oncologist called the situation a “Two tiered system.” He continued, “You don’t just have to have access to it; you have to have sophistication to log-on.”

Oncologists were also concerned about their patients’ ability to understand the medical information presented within the patient portal. A radiation oncologist said about her patient population, “We don’t cater to the most literate and well educated patients…and the abbreviations that we use, I don’t know that they would have a clue.” However, oncologists were concerned generally about the complexity of information being presented regardless of education level or economic status.

Greater Emphasis on the Patient-Provider Relationship

Oncologists’ fears of patients unknowingly discovering a diagnosis or prognosis, as well as patients enhanced involvement from viewing their medical information, has impacted the patient-provider relationship. The first sub-theme, oncologist surveillance, indicates that oncologists were taking measures to expedite their communication with patients, while the second sub-theme, rapport with oncologists, describes how the portal facilitates closer relationships.

Oncologist surveillance.

The combination of oncologists’ trepidation with the patient portal and uncertainties about patient usage has caused them to develop strategies to control the flow of information and mitigate potential patient anxiety. When the patient portal was first introduced, oncologists envisioned increased workloads and unending emails. In response, oncologists have identified when to use the portal compared to other means of communication. These decisions are based on their perceptions of how specific patients will respond to reading risk information. For instance, a medical oncologist described the following experience:

I’ll [type], ‘your hemoglobin is stable’. I sometimes get five messages back. I’ll get a message, and I’ll respond. Then I get another one…If I had just called them, it would’ve been just done and over with. So I have learned who to send a message to and who to talk to.

To lessen oncologists’ fears of patients discovering a diagnosis or prognosis in the portal, oncologists informally observed which patients mentioned that they used the portal, and were more vigilant about contacting the patient as soon as a test result arrived. In fact, participation in this study prompted some oncologists to reconsider the role of the portal. A medical oncologist specializing in palliative care said, “I think that after this discussion, I may be more preemptive with discussing the portal with patients…then I could be better aware of who is using it.”

In addition to monitoring which patients were likely to use the patient portal, some oncologists’ allocated time during the consultation to explain how test results may be available before a clinician could get in contact. A medical oncologist recalled that initial results appearing in the portal may not be final:

I’ve warned patients that they may get scan results ahead of time and I warn them that it’s not just one thing to have to the scan, but you need to have the interpretation and explanation and that’s what I’m here for.

Oncologists thought it was important for patients to understand that diagnoses and treatment plans require careful analysis and therefore, information immediately available may be amended upon further scrutiny.

Rapport with oncologists.

Instant access to medical information made patients recognize the importance of their oncologists’ interpretation and presentation of results. Even though 30 of the 35 (86%) patients used the internet in combination with their portal, 21 (60%) acknowledged the limitations of only using external websites to conduct research after viewing the portal. For example, a lung cancer patient said, “I’d much rather hear from a doctor than look it up on Google…Doctor Google does not have the best bedside manner.” Another patient diagnosed with breast cancer said, “The Internet sometimes is overwhelming and gives me too much information.”

However, when patients had information from the patient portal in advance of appointments, they felt more comfortable interacting with their oncologist. A prostate cancer patient summarized how the portal transformed the rapport with his oncologist:

Most of my life, I’ve just been sitting there shaking my head. [With the portal], I’m looking at his note, looking at tests, looking up anything I don’t understand. I will always use that portal and I’m thankful for that portal. It’s changed our relationship.

In addition, patients believed that since they had more knowledge about their condition from viewing the portal, they were able to collaborate better with the provider. A patient with oral cancer declared, “We came up with the treatment that I have now” due to being able to view his record in advance and form a list of questions. A radiology oncologist agreed with this sentiment and acknowledged that the portal gives patients more ownership, making them, “A little bit more of a partner.”

Discussion

The emerging practice of providing patients access to their electronic medical records through online portals has been shown to benefit many patients in the primary care setting, particularly in areas of functionality, such as appointment setting (Irizarry, Dabbs, & Curran, 2015). The positive experiences of primary care patients using portals have underpinned efforts to extend portal use to other areas, such as oncology. However, unlike some other settings, cancer patients are potentially faced with complex risk information that may be difficult to understand and remember, and may provoke distress and anxiety and in turn, impact patient-oncologist communication.

Findings from the current study indicate that cancer patients report advantages to using portals, particularly at a functional-level, patients felt better able to control their appointments and appreciated having a direct link to their provider and care team. In addition, when armed with their test results and other relevant medical information from their records, cancer patients expressed a sense of confidence and empowerment enabling more active communication and further, expressed a sense of being a collaborator with their oncologist. Our findings are comparable to previous studies, which have demonstrated the potential for portals to improve patient education and autonomy (Ammenwerth, Schnell-Inderst, & Hoerbst, 2011), while also facilitating assertive communication (Rexhepi et al., 2016). Our results also correspond to findings from OpenNotes studies, which is an initiative to provide patients access to their written clinical notes via their portal. OpenNotes is currently utilized in 37 states and available to 17 million people (OpenNotes.org, 2017). The seminal OpenNotes study found that participating patients felt more in control of their care, and reported high adherence levels, while only a very small number reported confusion (Delbanco et al., 2012).

Our results varied when risk information was involved, such as a new diagnosis or a worsening prognosis. Some patients preferred to learn about non-threatening information through the portal, while others preferred more severe risk information to be conveyed personally by the oncologist. It is likely that these preferences were modified by patients’ a) perceptions of the level of risk involved, b) ability to process undesirable information, such as a poor prognosis, and c) sense of alliance and trust in their oncologist. This finding has implications for the role of portals to positively impact patient centered care (PCC) by helping oncologist to value, elicit and meet patient preferences (Epstein & Street, 2011), which are integral to providing high quality cancer care (Levit et al., 2013).

Our findings also align with the RISP model (Griffin et al., 1999), as the model was designed to explain variance in information seeking and processing specifically within the context of risk. RISP guided our questionnaire development, but results indicate that patients turned to the portal to seek out more information and perceived it as a verified and relevant channel. As all participants were users of the portal, perhaps these individuals crave any information about their health, even if it may elevate concern and induce stress. Therefore, the model should be utilized for further testing, to determine the specific factors that motivate information seeking behavior when risk information is involved. Moreover, the RISP model can be applied to other aspects of patient care, including engagement and advocacy.

In regards to stress, recent published studies have indicated that patients were not unduly distressed by viewing their test results using patient portals (Hall et al., 2014; McNamara et al., 2015; Rexhepi et al., 2016; Wiljer et al., 2010). This may be due, in part to a) disease stage and b) level of patients’ experience in dealing with cancer. In this sample, most patients were at least six months removed from learning a new diagnosis and thus, had undergone treatments and were educated about their disease. In addition, most (66%) were dealing with early stage disease. Further study should examine the degree of patient distress as a consequence of viewing risk information in their portal among patients learning of a new diagnosis, or being informed that their disease had entered a phase where treatment was no longer being given with curative intent. This finding also underscores previous literature on patient preferences that emphasize the need for oncologists to understand the way individual patients want to be involved in their care and make decisions about treatments (Brown et al., 2012).

Yet, as more than half of all patients viewed scans and lab reports prior to a face-to-face consultation with their oncologist, it is possible that portals are serving as substitutes for in-person discussions, which is counter to patient-centered care practices (Dias, Chabner, Lynch, & Penson, 2003). It is postulated that optimal patient-oncologist communication involves empathic discussions (Brown, Bylund, Eddington, Gueguen, & Kissane, 2010) that promote patient understanding, satisfaction, and well-being (Hack, Degner, & Parker, 2005). Future research could usefully explore whether the increasing frequency of portal use is altering preferences for information and lowering expectations of interactions with the oncologist. In light of our results, we suggest that clinical practice gain a better understand of how patients and oncologists can best use the portal to facilitate meeting patients’ communication preferences without negatively impacting the patient–oncologist therapeutic alliance.

Our data suggests oncologists were largely unaware of whether their patients were using the portal and did not know what content patients were viewing. This may be explained by; a) the lack of provider training; b) varying frequencies in how often study patients viewed their portal or c) that discussions about patient portal usage were not yet included in routine clinical care. To counter this, we suggest that oncologists receive support and education about how to optimize portal usage, guided by our extensive knowledge of effective communication skills training practices.

Finally, oncologists noted anecdotally that portals have the potential to exacerbate already existing health disparities through the potential for inequitable access. These concerns are supported by studies showing that older, economically disadvantaged and minority patients are less likely to use or register for portals (Gordon & Hornbrook, 2016; Siegel, Spies, Rodin, & Tu, 2016). Thus, future research is warranted to determine whether inequitable access to portals is limiting communication and widening disparities.

Limitations

Limitations included the possibility that results may not extend beyond the health system in which the study was conducted. The majority of patients in the health system were not enrolled in the patient portal, but all of the patients recruited were active portal users. Future research should probe individuals who purposefully chose not to enroll in the portal to understand their perspective. Similarly, the sample of oncologists may have included participants eager to express their opinion due to their own personal experience, which could have affected their viewpoint.

Conclusion

Patient portals are still a nascent technology, but have the ability to impact the way information is disseminated and understood. As patient portals become more widely used in all medical settings, providers should consider what patients might view and how it may be interpreted. The present study established general attitudes and perceptions of patient portals in oncology, indicating that usage may be different than in other medical settings, like primary care.