Introduction
Antimicrobial resistance is a growing public health crisis that is caused, in part, by unnecessary antimicrobial-prescribing. The overuse of antimicrobial therapy has been well identified as a concern across the spectrum of health care, including in palliative care and hospice care settings. Reference Shekhar1–Reference Juthani-Mehta, Malani and Mitchell3 In a study by Albrecht et al., during the last seven days of life, 27% of patients received at least one antimicrobial and 1.3% received three or more antimicrobials. Reference Albrecht, McGregor, Fromme, Bearden and Furuno4 More recently, Clark et al., identified that only 42% of hospice patients receiving antimicrobials were prescribed antimicrobials appropriately. Reference Clark, Halford, Herndon and Middendorf5
There are several barriers to improving antimicrobial-prescribing in palliative care and hospice care settings. Antimicrobials are frequently used empirically for end-of-life patients, as confirmatory laboratory tests can be viewed as burdensome for the patient. In addition, patients’ impaired cognition may make it difficult to assess for infection symptoms. Further, prescribers often mistakenly believe the possible symptom relief provided by prescribing antimicrobials for a suspected infection outweigh the potential antimicrobial-related harms for end-of-life patients. Reference Gaw, Hamilton, Gerber and Szymczak6 However, adverse effects are linked with antimicrobial use, including, Clostridioides difficile infection and the emergence of antimicrobial resistance. Reference Juthani-Mehta, Malani and Mitchell3,Reference Tagashira, Kawahara, Takamatsu and Honda7–9
Shared decision-making, which denotes the exchange of information between the patient, their caregiver, and their healthcare provider, is beneficial in end-of-life care conversations to help align medical care with patient preferences. The decision to initiate antimicrobials during end-of-life care is complex. Reference Albrecht, McGregor, Fromme, Bearden and Furuno4,Reference Gaw, Hamilton, Gerber and Szymczak6,Reference Lopez, Vyas and Malhotra8 It has been well documented in the literature that practice recommendations for antimicrobial-prescribing at end-of-life should involve shared decision-making. Reference Juthani-Mehta, Malani and Mitchell3,Reference Seaton, Cooper and Fairweather10 Specifically, joint discussions between healthcare providers and the patient/support caregiver regarding antimicrobial-prescribing, including their potential risks and benefits, as well as understanding the patient’s goals are key. 11 However, there is limited data on barriers and facilitators to incorporating antimicrobial-prescribing discussions into shared decision-making conversations.
We explored provider, patient, and support caregiver (eg, family member or friend) perceptions of barriers and facilitators to discussing antimicrobial-prescribing during the end-of-life period. Identification of potential barriers and facilitators to shared decision-making during the end-of-life period may guide antimicrobial-prescribing discussions.
Methods
Study design
We used an ethnographic approach for epistemological reasons to better understand barriers and facilitators to antimicrobial-prescribing decisions during end-of-life. Reference Schensul, Schensul and LeCompte12,Reference Waring13
Setting and participants
Purposive and convenience sampling strategies were used to identify semi-structured interview participants. The Veterans Health Administration is the largest integrated healthcare system in the United States, providing care to over 9 million patients. 14 Semi-structured interviews with healthcare providers and patients/support caregivers were conducted at two Midwest Veterans Affairs facilities, one acute care and one long-term care (ie, community living center) from November 2020 to May 2022.
Definitions
We defined hospice care as caring for a patient during the last 6 months of life and palliative care as caring for a patient after the diagnosis of a serious illness.
Data collection
Healthcare providers were invited to participate via e-mail and interviewed via telephone or video conference (CCG and EB). Patients/support caregivers were identified by hospital admission data. Patients who recently received a hospice care or palliative care consultation were approached and interviewed in-person (CCG and EB). Interview guides were developed by the research team with questions focusing on attitudes and beliefs about antimicrobial-prescribing during end-of-life care.
Interviews were audio-recorded and transcribed verbatim, lasting an average of 39:01 (healthcare provider) and 15:21 (patient/caregiver) minutes. All participation was voluntary. One patient/caregiver interview was not recorded, due to participant refusal; detailed notes were taken and used for analysis. Data collection ended when thematic saturation was reached. Reference Guest, Bunce and Johnson15
Ethical considerations
The study was approved by the Institutional Review Board at the University of Iowa (#: 201909834) and the Research and Development Committee at the Iowa City VA Health Care System. Verbal consent was provided prior to interviews.
Data analysis
Transcripts were analyzed using MAXQDA, a qualitative data program (VERBI Software, Berlin, Germany). Thematic content analysis Reference Pope and Mays16,Reference Agar17 was performed by an interdisciplinary team which included trained social scientists (CCG, EB, KG, and MM) and infectious disease clinicians (EP and DL). First, our analysis process involved reviewing three transcripts by the large group (CCG, EB, KG, MM, EP, and DL) and a codebook was developed based on a priori research questions and emergent content. Reference Ryan and Benard18 Eleven percent of the transcripts were coded via group consensus (CCG, EB, KG, MM, EP, and DL), a process that involved all team members coding transcripts prior to meetings where the final coding consensus was entered into MAXQDA during group discussion. The remaining 89% were coded by paired members of the analysis team (EB, MM, CCG). Paired consensus coding meetings started after the codebook was developed. Discrepancies were resolved by the larger team.
Next, we subcoded one of the most frequently applied codes “shared decision-making.” We consensus subcoded all text segments either in the large group or paired group setting into “goals of care” and “quality of life.”
Results
Fifteen prescribing healthcare providers and 13 patients/support caregivers who were currently receiving hospice care or palliative care treatment were interviewed regarding attitudes/beliefs about antimicrobial-prescribing and experiences during end-of-life care. Of the thirteen interviews, 62% were patients and 38% were support caregivers. Prescribing healthcare provider (47% female; 53% male) roles included physicians (N = 13; 40% serving in a leadership role) and pharmacists (N = 2), with an average of 4.53 years (SD:2.89) in their current role. Most patients (70%) were receiving palliative care, all were male, and average age was 73.9 years (SD:8.02). Three overarching themes emerged. Representative quotations can be found in Table 1.
Table 1. Quotations of key themes from semi-structured interviews with healthcare providers, patients, and/or support caregivers
Theme 1: Healthcare providers highlighted that shared decision-making with the patient and support caregivers was key for guiding optimal antimicrobial treatment.
“Goals of care” was defined as any discussion with the care team, patient, or support caregiver around the patient’s treatment plan, including impact on quality of life:
Your goals of care conversation is basically a blueprint or a map that you make with the patient that helps steer you during this course all the way to end-of-life. [Provider ID#13, Site 2]
To advise appropriately, providers need to understand the goals of the patient:
(…) what we discussed (…) all stems from that conversation about what is important to the Veteran. (…) what makes life meaningful for them. And then, that helps you advise them. Like, “Here’s what I think as a physician, you know, would be the next best step in terms of your care.” Which is still shared decision making, it just starts with a conversation about what’s important to them. [Provider ID#09, Site 1]
Yeah, they [providers] were trying to involve the plan of attack here. Finally decided that there is cancer on my liver. And they’re not sure how they’re gonna do, -treat it, the cancer. Coming up with options all the time, we discuss those options. [Patient ID#08, Site 1]
In some situations, the goal of a patient may be to prolong life to a desired timepoint, and this can influence the decision to prescribe antimicrobials:
(…) But if you have [end-stage] COPD and your son’s graduating from school in six weeks and their goal is to make it to there. In that patient we might want to prescribe an antibiotic [to treat pneumonia] in hopes of extending their life to reach their goal and after their son graduates, then usually they can just say, “Okay, I don’t need antibiotics because I know I’m gonna die of COPD and probably will have pneumonia, there’s a very high chance for it, and as it gets worse, it’s probably not gonna extend my life or increase my quality of life,” so we would choose not to do antibiotics. [Provider ID#13, Site 2]
In general, providers highlighted the need to recognize and shift their thinking away from treating infections and potential role for shared decision-making to help in doing so in the end-of-life setting:
There’s so many different factors that come into play with antibiotics, especially when people are trying to shift gears from “we always treat, we want everything” as far as treating every infection to recognizing that many infections are end-of-life events in terminal patients. [Provider ID#10, Site 2]
In 2017, the VA initiated a nationwide quality improvement project titled, Life-Sustaining Treatment Decisions Initiative (LSTDI) to promote personalized, patient-driven care for Veterans with serious illness by identifying, documenting, and honoring their goals of care preferences. Reference Foglia, Lowery, Sharpe, Tompkins and Fox19 The LSTDI is one mechanism to help guide the end-of-life shared decision-making process, as it was developed for primary care providers, or clinicians with the best relationship with the patient, to help ensure they have detailed goals of care conversations with patients. However, the life-sustaining treatment form can be challenging to navigate, when the different treatment options are presented like a menu of options rather than the life-sustaining treatment facilitating a larger discussion about what matters most to the Veteran:
(…) I’ve participated in discussions where it (…) sounds like you’re providing the family with a menu. (…) “Do you guys want antibiotics?”(…). [ProviderID#08, Site1]
Among the therapy options offered on the life-sustaining treatment form, antimicrobials are somewhat unique, as they are a function of standard medical care as illustrated in the following quotation:
(…) you absolutely have to talk to the patient themself and their family, if they have family, and they’re willing to include the family. (…) because again antibiotics are something that in standard medical care and the culture in general, are kind of an expectation. You know if there’s a bacterial infection then you treat it with, an antibiotic. So, to not do that and to not discuss, that possibility, that option to not use antibiotics in certain cases, in the hospice, I think many people would probably feel, like something really important wasn’t talked about. And the flipside of that is, you know, some people may not understand there is an option not to use antibiotics in certain cases. And they might choose that option if they knew that because antibiotics do come with risks. [Provider ID#14, Site 2]
The decision on whether to shift from prescribing to not prescribing antimicrobials relies heavily on end-of-life prognosis and the goals of care conversations providers have with their patients. In some situations, antimicrobials may merely prolong a patient’s suffering, which could motivate a provider not to prescribe: “if the patient is dying right now, (…) I’m not going to give antibiotics. It’s going to prolong suffering.” [Provider ID#4, Site 1]
Theme 2. Barriers to performing shared decision-making regarding antimicrobials include time constraints and uncertainty of end-of-life prognosis.
End-of-life goals of care conversations take time, as illustrated in the following quotations:
(…) we just don’t do this [end-of-life conversations] nearly enough. And there’s a lot of barriers, I think the biggest barrier the physicians always speak of is time. It does take time to have those discussions. [Provider ID#03, Site1]
An added barrier to having end-of-life conversations is the difficulty in providing the patient with an accurate prognosis of how much longer they have to live:
there’s a cultural barrier, which we won’t be able to change very quickly. (…) whoever is working with the patient long-term, as they’re aging, as they’re getting sicker, if it’s just something that we can remember to offer them as an alternative. (…) it’s hard, because you don’t want to offer something as an alternative until you’re really sure they’re, you know, at the end-of-life. [Provider ID#14, Site 2]
Highlighting this difficulty in prognostication, one provider recalled several patients who lived far longer than the medical team had predicted:
And the CLC [community living center] hospice unit was designed really to be for those that are very end-of-life, that need aggressive symptom management. Now, having said that, we had a number of Veterans, actually several, that lived over a year. Some that you think have only a few weeks that end up living a number of months. It sometimes becomes really challenging to look at what are goals of treatment. [Provider ID#10, Site 2]
Theme 3. Patient comprehension and trust help facilitate effective antimicrobial shared decision-making.
Patient/support caregiver responses predominately aligned with theme 3. From a patient perspective, important characteristics fostering effective shared decision-making include trust and closed-loop communication. Communication in which providers ensure patients are involved in the discussions and feel they’ve been heard, as well as to ensure they understand and comprehend the agreed upon care plan as illustrated by Patient ID#08, Site 1:
Well, they made sure I understood and asked me if I had any questions, and if I did, they did their best to answer them.
This patient went on to describe comprehension as:
Yeah, they [healthcare team] were trying to involve the plan of attack here. Finally decided that there is cancer on my liver. And they’re not sure how they’re gonna do, treat it, the cancer. Coming up with options all the time, we discuss those options. What can and can’t go wrong with how it affects you and all that good stuff.
Trust was often cited as an important component of shared decision-making. Involving the patient/support caregiver in their care discussions is a crucial part of the trust-building:
You have to trust somebody, you know? And if you’ve got a doctor that’s open with you that you can trust, you’re in good hands. If you don’t trust your doctor, you’re gonna worry yourself to death. So if you trust your doctor, and you know, you’ll get much better care out of it. [Patient ID#10, SITE 1]
Discussion
Main findings
Interview responses indicated healthcare providers recognize the potential benefit of leveraging shared decision-making and having a clear “blueprint” of the patients’ goals and wishes to guide antimicrobial-prescribing decisions; however, barriers included limited face-to-face time with the patient and uncertainty about the patient’s prognosis. Facilitators identified by patients and support caregivers largely centered on trust, comprehension, and feeling heard as important characteristics for effective shared decision-making.
Shared decision-making is considered a recommended practice for making antimicrobial-prescribing decisions during the end-of-life. Reference Juthani-Mehta, Malani and Mitchell3,Reference Seaton, Cooper and Fairweather10 Discussions surrounding patients’ goals of care can create opportunities for patients to make informed decisions about and contribute to their care, which has potential to alleviate provider challenges, as well as improve build patient trust.
Our findings identify healthcare provider and patient/support caregiver barriers and facilitators around shared decision-making regarding antimicrobial-prescribing at end-of-life. Previous studies have identified similar barriers to shared decision-making at end-of-life (eg, time constraints and prognosis uncertainty). Reference Gaw, Hamilton, Gerber and Szymczak6,Reference Bernacki and Block20 Other barriers identified in published studies include providers’ discomfort with having these difficult conversations with patients, Reference Gaw, Hamilton, Gerber and Szymczak6,Reference White and Andersen21 as well as residents’ lack of training, experience, and communication skills to conduct these nuanced end-of-life conversations. Reference Szmuilowicz, el-Jawahri, Chiappetta, Kamdar and Block22
Future research should include a broader range of healthcare settings, incorporate quantitative data to complement qualitative findings and explore the perspectives of various healthcare providers and delve deeper into the specific factors influencing antimicrobial-prescribing decisions in end-of-life care. Further, we can enhance our understanding of the complexities surrounding shared decision-making and antimicrobial use and develop targeted interventions to improve uptake of shared decision-making into antimicrobial-prescribing decisions during end-of-life care.
Strengths and limitations
Strengths of our study include in-depth qualitative interviews with healthcare providers as well as in-person patient/support caregiver interviews. A limitation is that our study was restricted to the Veteran Affairs Healthcare system. This could limit the generalizability of the results to other healthcare settings. The Veteran Affairs system may have unique characteristics, such as a specific patient population or organizational culture, that could impact the findings. Further, all patient participants were male. Antimicrobial use was not a qualitative interview recruiting strategy; thus, use was unknown for interview participants. Patient participants were hospitalized and very ill; thus, the study team sometimes had difficulty eliciting patient reflections.
Conclusions
Palliative and hospice care settings have unique challenges that can make it difficult to ensure guideline-directed antimicrobial-prescribing. Improving antimicrobial-prescribing at end-of-life is a priority to minimize antimicrobial-related adverse effects. Our findings highlight an opportunity for greater use of shared decision-making when discussing antimicrobial use during end-of-life care. Interventions should be designed to address the identified barriers to shared decision-making.
Acknowledgments
The views presented in this manuscript are those of the authors and do not necessarily reflect the position or policy of the Department of Veterans Affairs or the United States Government. We thank the VA healthcare providers, patients, and support caregivers who participated in our interviews for their time and willingness to share their experiences. Financial support for this work was supported by Research to Impact for Veterans (RIVR) grant #19-477.
Competing interests
All authors declare that there is no competing of interests with respect to the research, authorship, and/or publication of this article.
Open access
