Understanding drivers of emergency department visits among patients from populations disproportionately affected by acute care use during cancer treatment: insights from qualitative interviews Authors: Jennifer Elston Lafata, Ethan Gibbons, Solomon Ayehu, Soroush Fariman, William A Wood, Jacob N Stein, Benyam Muluneh, Emily M Ray Published: 2026 ABSTRACT AbstractPurposeEmergency department (ED) visits are frequent during cancer treatment, burdensome for patients and caregivers, and disproportionately affect those from marginalized groups. While many such events may be preventable, rates in practice remain stubbornly high, and patient perspectives on the drivers of ED use remain underexplored, particularly among marginalized patients.MethodsWithin a quality improvement initiative aimed at reducing sociodemographic disparities in ED use, we conducted semi-structured interviews with adults receiving systemic anticancer therapy at a large academic medical center. Eligible participants were from marginalized populations with an ED visit within 30 days of treatment. Interviews explored symptom experiences, care-seeking decisions, and perceived barriers to outpatient support. Transcripts were analyzed using iterative thematic analysis to identify pathways to and underlying causes of ED presentation.ResultsAmong 11 eligible patients approached, 10 participated, all of whom were Black and/or Medicaid insured. Two pathways to ED presentation were identified: self-referral and referral by a health care professional. Underlying ED visit drivers, namely symptom onset or escalation accompanied by worry and uncertainty, barriers to accessing trusted clinical advice between visits, and limited effective self-management strategies, were consistent across pathways. Patients normalized treatment-related symptoms, relied on improvised self-management strategies, and sought guidance from informal advice networks, including family members and peers, which shaped their care-seeking decisions.ConclusionsPatient interviews revealed multiple, interrelated drivers of ED use during cancer treatment that extend beyond symptom severity alone. Findings identified multiple targets for improved supportive care processes and strategies to reduce preventable ED visits during cancer treatment.Keywords: Quality improvement, In-depth patient interviews, Chemotherapy, Unplanned hospitalizations, Symptoms, Self-management, Access, Family caregivers Introduction Emergency department (ED) visits and unplanned hospitalizations during cancer treatment are common [1, 2] and increasing in frequency and acuity [3]. For patients and caregivers, these events not only drive medical costs but are also a source of distress, burden, and reduced quality of life [4–13]. Despite estimates that 20–51% are preventable [3, 14–18], efforts to reduce these events during cancer treatment have demonstrated variable effectiveness in practice [19–22]. Acute care events during cancer treatment disproportionately impact marginalized patients [4, 8, 23–25] with increased use among racial and ethnic minorities as well as those who are uninsured or receive Medicaid [6–8, 26, 27]. Additional nonclinical risk factors include low socioeconomic status [27, 28], rurality [9, 10], and low educational attainment [9]. Few studies have explored how people with cancer make decisions about seeking ED care [29]. Published qualitative studies [30–34] offer limited guidance on how to address modifiable drivers of ED use during cancer treatment, especially for sociodemographic groups who experience a disproportionate burden. A notable exception is a recent study by Hong and colleagues that explicitly explores ED use among Spanish language speakers [35]. As part of a quality improvement (QI) project at a large, academic medical center, we observed higher ED use among patients with Medicaid and those who are Black or American Indian compared to privately insured or non-Hispanic White patients. To better understand potential drivers of those disparities, we conducted interviews among patients from these groups with the goal of identifying and developing patient-informed supportive care solutions to reduce preventable emergency care during cancer treatment. Methods Study design and setting Semi-structured interviews were conducted between October 2024 and June 2025 at a large, academic medical center in the Southeast United States (US). Like most academic medical centers, chemotherapy education from clinical pharmacy practitioners, clinical navigation support, and a 24/7 telephone triage program were available. Because interviews were designed for local quality improvement, the project was determined by The University of North Carolina at Chapel Hill Institutional Review Board to be nonhuman subject research (#23–2818). Institutional operational standards for patient privacy were maintained. Partial funding was received from the National Center for Advancing Translational Sciences, UM1TR004406 UNC-Chapel Hill Clinical and Translational Sciences Award (CTSA). Clinical trial registration was not applicable. We followed the Standards for Reporting Qualitative Research (SRQR) during manuscript preparation [36]. Patient eligibility, identification, and recruitment Eligible patients were aged 21 years or older, received systemic treatment at the medical center in the 30 days prior to their ED presentation, and were Black, American Indian, and/or Medicaid insured. Those with electronic health record (EHR)-documented evidence of cognitive impairment, incarceration, or needing a translator were excluded. As time permitted, a QI nurse with a background in emergency medicine reviewed nightly operational reports to identify patients receiving systemic anticancer therapy who had presented to an ED. Patients were eligible regardless of tumor type or stage if they had a scheduled return to the infusion clinic at a time when a student research assistant (an undergraduate student in public health who also worked as a certified nursing assistant in the medical center) was available to conduct interviews. Consequentially, not all potentially eligible patients identified through the nightly reports were approached for participation. Eligible patients were approached during their next scheduled infusion visit. Patients were invited to take part in one semi-structured, in-person interview. All participants provided verbal consent and agreed to audio recording. Our goal was to recruit 8–10 patients or until we reached data saturation, defined as the point at which no new insights or information emerged from two successive interviews [37]. Interview guide development The interview guide was informed by Turnbull and colleagues’ conceptual model for emergency care-seeking in which the decision to access emergency/urgent care is shaped by both individual- and social network-levels as well as the broader socio-temporal context [38]. Interviews began by asking about the circumstances surrounding the decision to visit the ED. Participants were asked to describe experiences leading up to their ED visit, including any communication with providers, symptom management, and unmet needs. Participants were then asked about input received from family and friends to shape their help-seeking and decision-making behaviors, whether they perceived their ED visit as avoidable, and any additional education or care management they wish they had received. Data collection Semi-structured interviews, approximately 30 min in length, were audio-recorded using a digital recorder. All interviews were transcribed verbatim and de-identified prior to analysis. Patient diagnosis, age, sex, race, and insurance status were abstracted from the EHR by the QI nurse; a medical oncologist (ER) reviewed the medical record to obtain cancer diagnosis, chief complaint, and final ED visit diagnosis. Analysis Transcripts were analyzed by three team members (EG, JEL, and SF) using iterative thematic analysis techniques and constant comparative methods [39]. Collectively, the team has background and expertise in public health, cancer care delivery research, quality improvement, oncology pharmacy, and supportive care. Team members independently reviewed four transcripts, meeting after each one to discuss and refine initial thematic patterns. Following this review, they developed a preliminary coding manual that incorporated a template of provisional codes [40]. Two additional transcripts were then consensus-coded to verify the completeness of the codebook and to establish coding consistency. During this process, two codes were further subdivided for clarity, and one new code was added. The remaining transcripts were coded independently by two reviewers each, with discrepancies resolved through discussion and, when needed, a third reviewer. Coded data was managed in Microsoft Excel. After the first eight interviews were coded, team members met with the oncology clinicians supporting the project (BM, ER, JS, WW) to ensure clinical validity and to begin labelling themes and their connections. Two additional interviews were conducted and similarly dual-coded, which confirmed that data saturation had been reached. The connecting of codes to identify themes and subthemes was achieved via an iterative and reflective process in which coded quotes from each interview were read and reread by all members of the research team to ensure grounding in all interviews. We did not share interim findings with participants but the resulting themes that were selected by the QI team for intervention development were validated during discussions with the oncology services’ Patient and Family Advisory Board. Results Participant characteristics Eleven patients were approached for an interview; ten agreed to participate. Participants ranged in age from 37 to 61 years and were Medicaid-insured, Black, or both (Table 1). Table 2 summarizes patient chief complaints upon ED presentation and ED diagnosis. Table 1Participant characteristics (N = 10)CharacteristicN Mean age53 (11) yearsSex Female7 Male3Race Black9 White1Medicaid Yes2 No8Cancer type Breast4 Colorectal4 Other2 Participant characteristics (N = 10) Table 2Cancer and ED diagnoses among the patient sampleStudy IDDiseaseChief complaint or reasons for presentationEmergency department findings and/or discharge diagnosis001Breast cancer, stage IIBNeck and arm swelling; shortness of breathIndeterminate; ruled out blood clot; treated for headache; no neuroimaging needed002Colon cancer, stage IVPain with provider concern for strangulated umbilical herniaCancer-related pain, no acute abnormalities on CT or labs003Rectal cancer, stage IVSyncopeIndeterminate; ruled out myocardial infarction and tachyarrhythmia004Breast cancer, stage IIVision problems, numbness in fingertipsUveitis and scleritis likely in the setting of chemotherapy/immunotherapy005Esophageal cancer, stage IIICough, congestion, shortness of breathCOPD exacerbation and pneumonia006Hepatocellular carcinoma, stage IVUrine retention, constipation, chest painConstipation; ruled out myocardial infarction007Rectal cancer, stage IIIShortness of breath, sensation of throat closingChemotherapy-induced cold sensitivity; incidentally found to have hyponatremia, suspected water intoxication008Breast cancer, stage IHigh blood pressure, headacheHypertension, chronic benign headache009Breast cancer, recurrent stage INausea, vomiting, abdominal crampingChemotherapy-induced nausea and vomiting010Colon cancer, stage IVNausea, vomitingChemotherapy-induced nausea and vomiting Cancer and ED diagnoses among the patient sample Thematic analysis Patients described two pathways to their ED presentation: (1) self-referral without engagement of a health care professional and (2) following a health care professional’s referral. The latter included referrals occurring at scheduled infusion appointments or between scheduled appointments. Regardless of the pathway, the factors contributing were similar (Fig. 1). Patients consistently described the sudden onset or escalation of a worrisome symptom that raised concerns about possible consequences (Theme 1: Symptom onset, escalation, and worry). Steps taken after symptom onset varied not only by perceived symptom severity and worry but also by patients’ beliefs regarding care access (Theme 2: Challenges accessing trusted medical advice), their own self-management strategies (Theme 3: Self-management challenges and context), and the influence of informal advice networks (Theme 4: Informal advice networks shape self-management strategies and care-seeking decisions). Each theme and subtheme is described below with additional illustrative quotations (Table 3). Pathways to the emergency department (ED) during cancer treatment Table 3Themes, subthemes and illustrative quotesThemes and subthemesIllustrative quotationsTheme 1: Symptom onset, escalation and worry“Granted the temperatures were like 24° and it said…it would feel 14° … When I felt everything going numb, I panicked … So, my windpipe … really started to close and I was gasping … I don’t know how much was a side effect to any of the treatment itself … versus it was just damn cold … but I ended up having a panic attack. So that’s why I came to the emergency room cause in the moment I couldn’t tell … I literally felt like in that moment I could have died … mainly because of the panic. … You know, telling myself to breathe was like huge in that moment … I was able to really talk myself into my breathing, trying to control my breathing because that was even scarier than my hands being cold. So, between me leaving home…and midway through somewhere in there, it was enough of the distraction that whatever was going on with the numbness with my hands was gone… by the time I got here, that had faded away. But it was just that initial shock. … So ultimately I determined that, yes, it was more of a panic attack… than a reaction to the medicine I was getting.” Patient 007“I did because I really couldn’t see anything. … I couldn’t. I could as far as my mail even me. Wearing readers I could not read my mail or anything like that. In order to read, I will sometimes have to have a magnifying glass to be able to read my mail. So I was very concerned about that and plus I wasn’t driving” Patient 004“I went two weeks ago because I was having like some migraines on my right side in the when I woke up that morning and I had been having high blood pressure issues in the last two months, two to three months. And I had never had issues with my blood pressure before. And so I was a little bit worried about the pressure on my head ‘cause I couldn’t move it and I was worried that I might I might be having a stroke or aneurysm or I just, I just didn’t know and wanted to get checked.” Patient 008Theme 2: Challenges accessing trusted medical advice Subtheme 2.1: Trust in their cancer care team, distrust in others“…I was told I had a UTI. I was told I had hernia and it was neither one. So, I went back to [other] hospital, I said something ain’t right… they misdiagnosed me and I could have died because the guy told me when I got here [medical center], he said … You need emergency surgery.” Patient 002“We had appointments coming up anyway up here. And I said I’ll just wait. … As long as I’m not hurting like I was, I’ll wait and, see them up [at the cancer center] because I don’t believe anything these people down here say anyway… [The cancer center is] just probably 2 1/2 h [away]… If possible, I just go straight [to the cancer center]. … I’m here every other week anyway.” Patient 005 Sub-theme 2.2: Barriers to clinical advice in between visits“I’ve called up here and haven’t gotten an answer a few times…I was in a kind of downward turn at that point where I think I was calling, and I just did not get an answer… And so, I just decided not to call. I really didn’t want to have to call 911. But, I think, in that moment I, I also, because I did not get an answer. I was trying to self-medicate…” Patient 007“… the team contacted me and they [Cardiology] wouldn’t have had an appointment for about like 3 weeks. So if I was able to get an appointment sooner then I wouldn’t have had needed to go.” Patient 001Theme 3: Self-management challenges and contextSubtheme 3.1: Difficulties discerning expected versus concerning treatment-related“if anything, it was maybe just feeling nauseous…but I kind of figured that [part] would come with it during the chemo anyway. So, I really didn’t think too much of it.” Patient 003“I get dehydrated sometimes because sometimes I lose my appetite and I can’t eat. Rather it’s from this medicine or the cancer … I go through cycles where I can’t eat… I just adjust and do what I have to do and keep moving.” Patient 006“… half the time they don’t even know I’m going through pain. But I keep that smile on my face… I’d have generally experienced a lot of pains like that. If it’s a different type of pain now, I’m going to say something… if it’s a pain that comes up, I’ll deal with it… up to a certain point. But if they don’t disappear, I gotta go.” Patient 002“… I kept saying this is going to pass. Yeah, I’m going to get over this. But it didn’t. It began to get worse, so I knew I needed to just seek medical attention.” Patient 004Subtheme 3.2: Ineffective and poorly tolerated supportive medications“And it was the same thing, fatigue, but I didn’t have a fever or anything like that. It got worse throughout the night … and I was just very, very weak. Couldn’t hold anything down. Vomiting, you know, like the yellow bile. The medications did not work … not … alleviating any of the symptoms.” Patient 009“So normally I take my [over the counter and prescription pain] medication. I combine them together and then for the nausea, I have three different types of nausea medication that I’m on that I normally take… but for some reason this time nothing was working.” Patient 010“I weaned myself off of the [pain medication]… it made me feel like I was in a tunnel. And it was hard for me to walk.” Patient 007Subtheme 3.3: Reliance on improvised self-management strategies“I go through sickness to fatigue, you know, all that stuff. But you know, I always put my TV on. I find the funniest stuff to look at.” Patient 002“the enema wasn’t working. I had been taking the [laxative] that I got from my doctor. Nothing was working. And so, I know people who unclogged themselves. So, I was like, I had never done it before, but I was like, I was going to try because it was, I was just getting really worried and nervous and so I did it myself. It was painful. It was horrible and I felt sick afterwards. I felt weak, I felt dizzy.” Patient 006Subtheme 3.4: Desire for autonomy and self-management skill-building“And they [ED staff] told me … what to take, … which I got right there to kind of help … I try to do a little more as far as with the knowledge, you know, having that, you know, so I can try to avoid having the same incident once again.” Patient 003“This makes us feel better to do our own research and understand ourselves.” – Patient 008’s caregiver “I want to know what I’m talking about.” Patient 005“… depending on the pain, I really try to process what I’m feeling, how I’m feeling it, what I’ve done and try to consider those things in my choice…I have these choices of medications for this reason… I’m trying to figure out if whatever I’m feeling is this going on because I’m anxious? … Am I overthinking something? And is that the cause? Or do I just need to chill and relax? Should I take this muscle relaxer? Would that help the tumor, #1, from being angry, you know, what’s triggering [the pain]? … I also don’t want to overthink something. And if, if, if there’s something that I’m feeling or I feel like I need to take something, I kind of want the advice that I can do without having to come in.” Patient 007Theme 4: Informal advice networks shape self-management strategies and care-seeking decisions“what I’ve learned by talking to somebody else with cancer is that… I’ll make my plate and then I’ll get like a third of the way through and I can’t really taste my food anymore… So I was talking to someone else and they said they learn just to have a three-ounce piece of meat… and a little bit of food and that’s it.” Patient 006“I reached out to my sister-in-law who’s a nurse and she had already been saying you need to get your blood pressure checked, you need you need to talk to a doctor. And I said, well, I can’t until I go in and see my primary. So, she had already kind of advised me to go to urgent care, just to see if they feel like I need to be on like a low dosage prescription for blood pressure ‘cause I never was on any medication for my blood pressure.” Patient 008“But both of my daughters are, I mean not doctors, nurses… They knew that all weekend I had been sick, but by Sunday I wasn’t feeling it better. They told me I need to go to the ER.” Patient 010 Themes, subthemes and illustrative quotes “Granted the temperatures were like 24° and it said…it would feel 14° … When I felt everything going numb, I panicked … So, my windpipe … really started to close and I was gasping … I don’t know how much was a side effect to any of the treatment itself … versus it was just damn cold … but I ended up having a panic attack. So that’s why I came to the emergency room cause in the moment I couldn’t tell … I literally felt like in that moment I could have died … mainly because of the panic. … You know, telling myself to breathe was like huge in that moment … I was able to really talk myself into my breathing, trying to control my breathing because that was even scarier than my hands being cold. So, between me leaving home…and midway through somewhere in there, it was enough of the distraction that whatever was going on with the numbness with my hands was gone… by the time I got here, that had faded away. But it was just that initial shock. … So ultimately I determined that, yes, it was more of a panic attack… than a reaction to the medicine I was getting.” Patient 007 “I did because I really couldn’t see anything. … I couldn’t. I could as far as my mail even me. Wearing readers I could not read my mail or anything like that. In order to read, I will sometimes have to have a magnifying glass to be able to read my mail. So I was very concerned about that and plus I wasn’t driving” Patient 004 “I went two weeks ago because I was having like some migraines on my right side in the when I woke up that morning and I had been having high blood pressure issues in the last two months, two to three months. And I had never had issues with my blood pressure before. And so I was a little bit worried about the pressure on my head ‘cause I couldn’t move it and I was worried that I might I might be having a stroke or aneurysm or I just, I just didn’t know and wanted to get checked.” Patient 008 “…I was told I had a UTI. I was told I had hernia and it was neither one. So, I went back to [other] hospital, I said something ain’t right… they misdiagnosed me and I could have died because the guy told me when I got here [medical center], he said … You need emergency surgery.” Patient 002 “We had appointments coming up anyway up here. And I said I’ll just wait. … As long as I’m not hurting like I was, I’ll wait and, see them up [at the cancer center] because I don’t believe anything these people down here say anyway… [The cancer center is] just probably 2 1/2 h [away]… If possible, I just go straight [to the cancer center]. … I’m here every other week anyway.” Patient 005 “I’ve called up here and haven’t gotten an answer a few times…I was in a kind of downward turn at that point where I think I was calling, and I just did not get an answer… And so, I just decided not to call. I really didn’t want to have to call 911. But, I think, in that moment I, I also, because I did not get an answer. I was trying to self-medicate…” Patient 007 “… the team contacted me and they [Cardiology] wouldn’t have had an appointment for about like 3 weeks. So if I was able to get an appointment sooner then I wouldn’t have had needed to go.” Patient 001 “if anything, it was maybe just feeling nauseous…but I kind of figured that [part] would come with it during the chemo anyway. So, I really didn’t think too much of it.” Patient 003 “I get dehydrated sometimes because sometimes I lose my appetite and I can’t eat. Rather it’s from this medicine or the cancer … I go through cycles where I can’t eat… I just adjust and do what I have to do and keep moving.” Patient 006 “… half the time they don’t even know I’m going through pain. But I keep that smile on my face… I’d have generally experienced a lot of pains like that. If it’s a different type of pain now, I’m going to say something… if it’s a pain that comes up, I’ll deal with it… up to a certain point. But if they don’t disappear, I gotta go.” Patient 002 “… I kept saying this is going to pass. Yeah, I’m going to get over this. But it didn’t. It began to get worse, so I knew I needed to just seek medical attention.” Patient 004 “And it was the same thing, fatigue, but I didn’t have a fever or anything like that. It got worse throughout the night … and I was just very, very weak. Couldn’t hold anything down. Vomiting, you know, like the yellow bile. The medications did not work … not … alleviating any of the symptoms.” Patient 009 “So normally I take my [over the counter and prescription pain] medication. I combine them together and then for the nausea, I have three different types of nausea medication that I’m on that I normally take… but for some reason this time nothing was working.” Patient 010 “I weaned myself off of the [pain medication]… it made me feel like I was in a tunnel. And it was hard for me to walk.” Patient 007 “I go through sickness to fatigue, you know, all that stuff. But you know, I always put my TV on. I find the funniest stuff to look at.” Patient 002 “the enema wasn’t working. I had been taking the [laxative] that I got from my doctor. Nothing was working. And so, I know people who unclogged themselves. So, I was like, I had never done it before, but I was like, I was going to try because it was, I was just getting really worried and nervous and so I did it myself. It was painful. It was horrible and I felt sick afterwards. I felt weak, I felt dizzy.” Patient 006 “And they [ED staff] told me … what to take, … which I got right there to kind of help … I try to do a little more as far as with the knowledge, you know, having that, you know, so I can try to avoid having the same incident once again.” Patient 003 “This makes us feel better to do our own research and understand ourselves.” – Patient 008’s caregiver “I want to know what I’m talking about.” Patient 005 “… depending on the pain, I really try to process what I’m feeling, how I’m feeling it, what I’ve done and try to consider those things in my choice…I have these choices of medications for this reason… I’m trying to figure out if whatever I’m feeling is this going on because I’m anxious? … Am I overthinking something? And is that the cause? Or do I just need to chill and relax? Should I take this muscle relaxer? Would that help the tumor, #1, from being angry, you know, what’s triggering [the pain]? … I also don’t want to overthink something. And if, if, if there’s something that I’m feeling or I feel like I need to take something, I kind of want the advice that I can do without having to come in.” Patient 007 “what I’ve learned by talking to somebody else with cancer is that… I’ll make my plate and then I’ll get like a third of the way through and I can’t really taste my food anymore… So I was talking to someone else and they said they learn just to have a three-ounce piece of meat… and a little bit of food and that’s it.” Patient 006 “I reached out to my sister-in-law who’s a nurse and she had already been saying you need to get your blood pressure checked, you need you need to talk to a doctor. And I said, well, I can’t until I go in and see my primary. So, she had already kind of advised me to go to urgent care, just to see if they feel like I need to be on like a low dosage prescription for blood pressure ‘cause I never was on any medication for my blood pressure.” Patient 008 “But both of my daughters are, I mean not doctors, nurses… They knew that all weekend I had been sick, but by Sunday I wasn’t feeling it better. They told me I need to go to the ER.” Patient 010 Theme 1: Symptom onset, escalation, and worry Treatment-related symptom escalation, accompanied by worry and anxiety, at times amplified by patients’ connections to prior test results or past medical experiences, drove an urgent need for reassurance that led patients to the ED. Across participants, treatment-related symptoms, including becoming “really, really constipated” with dehydration (Patient 006), “vomiting” and being “very, very weak” with fatigue, while medications “did not work” (Patient 009), “cold sensitivity that resulted in a panic attack” (Patient 007), “vision problems” with eyes that were “very red” (Patient 004), and blood pressure that had “been elevated” in ways never experienced before (Patient 008), were accompanied by patient-described anxiety and an urgent need for reassurance. For example, people reported worrying that the symptom might mean a “stroke or aneurysm” (Patient 008), a cardiac event triggered by seeing “calcifications” on a recent scan (Patient 006), a complication so frightening that “[I] could have died” (Patient 007), or a fear, “am I having a reoccurrence?” (Patient 006). One patient experienced cold-sensitivity neuropathy that escalated into a panic attack and ED visit because she could not distinguish it from something life-threatening. By the time this patient arrived at the ED, her numbness had resolved. Regardless, the panic, which had been compounded by uncertainty about whether her symptoms were treatment-related, led to an ED visit. Theme 2: Challenges accessing trusted medical advice Trust in their cancer care team, distrust in others Patients described trust in their cancer care team and a distrust in other places, making the medical center the only place they believed could “go straight to the problem” (Patient 002). This distrust was often grounded in experience (e.g., stemmed from prior misdiagnoses) and resulted in hesitancy seeking care from facilities near their home. Some held a preference for seeking care exclusively at the medical center even if their condition demanded immediate care. For example, one patient (Patient 005) described how they would decline care from closer hospitals even during a cardiac event, stating “I’d rather die in the ditch out front of that hospital than to die in that hospital.” Another patient voiced a commitment to keeping all care within a single network because “everybody’s on the same team, so everybody can see what’s going on here and what’s going on there. That’s important” (Patient 006). This desire for integrated care from their trusted providers led patients to not seek care locally in between scheduled visits, but instead to “just wait” for upcoming appointments. Barriers to clinical advice in between visits Patients described challenges between visits communicating with their care teams at the medical center. These challenges stemmed from a variety of reasons including previous unsatisfactory experiences with after-hours support, uncertainty regarding whom to contact, appointment unavailability in non-oncology clinics, and even to “avoid bothering the care team” (Patient 005). One patient noted that, while a navigator’s number was available, “those are daytime hours that I can [reach them] and I just hate having some kind of question after hours because of likelihood that I can get it answered or even on the weekend” (Patient 007). Other patients acknowledged not even thinking about their care team because it was after hours—“it was a Saturday and I didn’t even think of my health care team” (Patient 006). Lack of timely outpatient appointment availability outside oncology presented a parallel structural barrier. For example, one patient who received what they perceived to be a concerning cardiology test result attempted to arrange follow-up but was told that no cardiology appointment was available for three weeks. When asked whether the ED visit could have been avoided, that patient’s response was unambiguously “yes:” “if there were available appointments” (Patient 001). Theme 3: Self-management challenges and context Difficulties discerning expected versus concerning treatment-related symptoms Patients described an expectation that feeling unwell was an inherent and unavoidable part of undergoing cancer treatment, leading them to normalize symptoms and delay seeking support. Patients framed their symptoms as “part of chemo” and therefore not worth attention, even when interfering with daily functioning. Patient 006 described cycles of appetite loss and dehydration and how these were a recurring feature of treatment rather than a reason to seek help. For some patients, symptoms were concealed from those around them: Patient 002 described absorbing pain in silence while maintaining an outward composure. In addition, patients described intentionally “pushing through” discomfort with the belief that symptoms would eventually pass on their own. By viewing their symptoms as something they had to “deal with,” patients attempted to self-manage symptoms that might have benefitted from interventions or guidance from their cancer care team. As such, difficulty distinguishing symptoms that can be safely managed at home from those that require medical intervention contributed to delayed care-seeking. Ineffective and poorly tolerated supportive medications Patients described challenges managing treatment-related symptoms because the supportive medications available to them were, at times, insufficiently effective, poorly tolerated, or both. Patients reported how prescribed anti-nausea, pain, or bowel medications did not provide adequate relief. Patients also reported limiting or avoiding supportive medications altogether because the side effects were perceived as equally or more distressing than the original symptoms. These experiences, none of which included reaching out to a care team for additional guidance, contributed to worsening symptom burden and escalation to acute care. Reliance on improvised self-management strategies Patients described using a variety of informal or improvised strategies to manage their symptoms. These approaches included distraction techniques, lifestyle adjustments, and self-directed remedies. One person described their own nutritional workaround when prescribed supportive approaches fell short: “with chemo, even if you can’t eat, you got to force… Watermelon gave me the base coat… I started eating watermelon and I’m telling [myself], ‘I can eat what I want’” (Patient 002). While such strategies sometimes helped and were seemingly essential components of self-management, they were often insufficient to address escalating symptoms, and in some cases introduced additional risk. Without additional support from their cancer team to assess which strategies may be appropriate in which contexts or to offer further guidance, patients had limited ways to adapt to what they were trying before things became worrying. Desire for autonomy and self-management skill-building Patients expressed a willingness and desire to self-manage their symptoms. They described actively seeking to improve their self-management skills, drawing on education from previous ED visits, or experimenting with learned strategies to prevent future exacerbations. They voiced a sense of personal responsibility that extended beyond symptom management to encompass self-advocacy, family protection, and the avoidance of unnecessary escalation. One patient articulated: “I preach self-care and self-advocacy, and so I’m going to advocate for myself more than anyone else is going to… ultimately I am responsible for me to ask questions, to do research.” The same patient viewed autonomous management as both an act of care toward family—“I don’t want my family to feel burnt out or feel guilty… what I can manage on my own, I manage on my own”—and a deliberate first step before seeking emergency care: “I felt like I was gonna try and take care of it before I made a decision to call 911” (Patient 006). This drive for self-sufficiency was balanced by awareness that they did not always possess the knowledge or skills to know about their disease and treatment: “…I do try to be mindful … if I’m overthinking … I don’t try to self-medicate or do something on my own without having someone verify ….” (Patient 007). This desire for autonomy, paired with recognition of self-care boundaries, shaped how patients navigated when to manage symptoms independently and when to seek professional input. Theme 4: Informal advice networks shape self-management strategies and care-seeking decisions Patients reported turning to informal advice networks to guide their symptom management strategies. These networks, including family members, friends, others with cancer, or informal information sources, provided practical suggestions that patients reported trying. Patients’ trusted advisors frequently helped interpret symptoms and decide whether medical attention was needed regardless of whether they had prior health training or relevant experiences. Thus, as trusted partners, regardless of their background, network members served as active decision makers who shifted patient perspectives by insisting on escalation, directing them or even driving them to the ED, or overriding the patient’s initial assessment of their symptoms. Advice from network members with healthcare training was seemingly viewed as authoritative, lending creditability to their recommendations and influencing care-seeking decisions, albeit not described as substantively different from advice provided by nonmedical informal advisors. Discussion We identified a consistent set of drivers underlying ED use during cancer treatment beyond the presence or severity of symptoms alone. Whether self-referred or clinician-directed, ED visits were precipitated by symptom onset or escalation accompanied by worry, uncertainty, and an urgent need for reassurance regarding safety or treatment-related complications. Barriers to timely, trusted clinical advice between scheduled visits, particularly during evenings and weekends, further constrained patients’ perceived options and positioned the ED as the most reliable source of immediate care. In parallel, patients described how their normalization of treatment-related symptoms, limited effectiveness or tolerability of supportive medications, and reliance on improvised self-management strategies—including waiting for upcoming oncology appointments to see their trusted team—contributed to delayed care-seeking until symptoms became unmanageable. Informal advice networks, including family members, peers, and caregivers, frequently served as surrogate triage systems, shaping symptom interpretation and escalation decisions. These findings suggest that potentially preventable ED use during cancer treatment reflects not only challenges in symptom control, but broader unmet needs related to communication, reassurance, accessibility, and support for safe and effective self-management outside the ED. Our findings align with and extend prior quantitative and qualitative work demonstrating that acute care use among oncology patients is frequently associated with poorly controlled symptoms such as pain, gastrointestinal distress, fatigue, and respiratory complaints, particularly during periods of active systemic therapy [3, 12, 30, 32, 33, 35, 41, 42]. Our qualitative findings add to this literature by illustrating how symptom burden interacted dynamically with patients’ interpretations, emotional responses, and perceptions of care access among a population of marginalized patients. These findings are consistent with recent qualitative findings [30, 35] as well as with conceptual models of urgent care–seeking behavior [38] that describe an initial phase of “illness work,” during which symptoms were appraised, normalized, and managed independently. During this phase, patients commonly tolerated discomfort as an expected consequence of treatment, engaged in trial-and-error self-management, and delayed seeking help until symptoms worsened to the point of becoming worrying. Importantly, findings like those of others [35] also illustrate how, despite supportive service availability, patients face challenges recognizing when symptoms warrant escalation, uncertainty regarding which resources to contact, and a hesitancy or difficulty accessing available support, highlighting how the presence of supportive care alone can be insufficient to ensure patients from marginalized groups can navigate and address symptom management needs. Symptom escalation alone seemed to rarely trigger ED use. Rather, patients described a transition toward seeking reassurance once uncertainty about symptom meaning, safety, or progression exceeded their perceived capacity to self-manage. At this point, perceptions of real-time access to trusted clinical advice, shaped by prior experiences with after-hours availability and responsiveness, became central to decision-making. This transition maps closely onto the “navigation work” described in Turnbull and colleagues’ framework [38], as well as prior qualitative studies, in which patients rely on socio-environmental context [42, 43] and lay networks to interpret symptoms and determine when medical care is warranted [33, 35]. Our findings underscore how informal advice networks functioned as triage systems. Family members, especially those with health care experience or peers living with cancer, often played a role in prompting escalation, reinforcing prior work demonstrating the influence of social context on ED utilization, [38] and perhaps pointing to opportunities to target educational programming or other interventions for this audience. Importantly, the need for reassurance emerged as a central driver of ED use. Patients sought care because they were unable to determine whether their symptoms signaled a medical emergency and lacked accessible pathways to obtain that determination elsewhere. These findings echo prior studies which highlight how uncertainty, fear, and limited perceived access to acceptable alternatives contribute to emergency care use even when symptoms may be clinically manageable in alternative outpatient settings [30, 33, 35, 42]. Taken together, our findings point to several actionable intervention opportunities. First, proactively providing patients with clear, cancer-specific guidance regarding symptom expectations and decision thresholds, delivered before crises occur, may help reduce anxiety-driven escalation. Simple, actionable decision rules and anticipatory guidance could help patients distinguish expected treatment effects from warning signs requiring urgent evaluation. Second, improving patient awareness of and access to all-hour trusted, coordinated clinical advice may preserve outpatient pathways as viable alternatives to ED care. Increasing clarity for patients and caregivers around whom to call and how to reach them, by strengthening patient communication with outpatient navigation and nurse triage teams, as well as increasing access to urgent outpatient appointments, may serve as important mechanisms to better meet patient needs in the outpatient setting. Such educational interventions must be paired with system-level supports that ensure reliable availability and consistent, timely clinical responses when patients seek guidance. Supportive care strategies that acknowledge patients’ desire for autonomy while strengthening self-management skills and system responses are likely to align with patient needs and preferences. The breadth of factors influencing ED use identified here and elsewhere [30, 35] suggests that no single intervention is likely to meaningfully reduce potentially preventable acute care utilization across all oncology populations and care settings. Rather, findings support a multifaceted approach that combines evidence-based strategies for upstream education, communication, and supportive care interventions with local adaptations not only to specific operational contexts but also patients’ lived experiences and access perceptions. Accordingly, our health system is implementing a cancer care learning collaborative modeled after the Institute for Healthcare Improvement Breakthrough Series framework [44], through which oncology programs will receive evidence-informed guidance and insights from these interviews while testing and evaluating small-scale changes tailored to local settings. This approach recognizes the complexity of cancer-related acute care utilization and the need for iterative learning across diverse patient experiences and practice environments. Limitations Several study limitations should be noted. First, we relied on a convenience sample of patients with a subsequent infusion visit at one academic medical center. While we sought to identify a diverse cohort of patients experiencing disproportionate acute care use at this medical center, we were unable to recruit any American Indian patients for this study. (Notably, the one patient who declined participation was American Indian.) The sample raises the possibility that not all relevant themes were identified, particularly those held by patients who did not participate, whose experiences or perspectives may differ. Additionally, researchers’ perspectives may have influenced data interpretation. The study was designed to understand experiences among populations disproportionately affected by acute care utilization, specifically Black and Medicaid-insured patients. Although several identified themes, such as uncertainty surrounding symptom interpretation, need for reassurance, and reliance on informal support networks, may plausibly extend beyond these groups, the present study was not designed to assess similarities or differences across different populations. As such, we cannot determine the extent to which these findings are transferable to other marginalized populations or to the broader population of patients receiving cancer treatment. Conclusions Interviews with patients from populations disproportionately affected by acute care utilization revealed multiple, interrelated drivers of emergency care use during cancer treatment that extend beyond symptom severity alone. These findings identify potential targets for supportive care-focused interventions in these populations and provide hypotheses for future research examining whether similar drivers influence emergency care use in broader oncology populations. Author contribution JEL and ER conceptualized the project. All authors assisted with analyses and thematic identification. JEL, EG and SA drafted the initial manuscript, including tables and figure. All authors critically reviewed, revised, and approved the final manuscript. Data availability De-identified interview transcripts are available from the corresponding author upon reasonable request, subject to ethical and privacy considerations. Declarations Competing interests Partial (approximately 15%) support for this work was provided by UM1TR004406 UNC-Chapel Hill Clinical and Translational Sciences Award (CTSA). The remainder of funding came from Institutional Support. WW reports consulting for Lantern Health, Teladoc Health and American Society of Hematology Research Collaborative, and equity with Koneksa Health. Footnotes Publisher's Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. References Associated Data Data Availability Statement De-identified interview transcripts are available from the corresponding author upon reasonable request, subject to ethical and privacy considerations.