Shared decision making in home hospice nursing visits: A qualitative study

Introduction

Shared decision making is a process wherein a choice is jointly made by a healthcare provider or team of providers and a patient or family member. (1) This joint decision making has become the standard of practice in healthcare delivery. The underlying philosophy is to remove the tradition of paternalistic medical practice that assumes the doctor knows best, and create a model where patients are responsible partners in the decisions related to their care. There are numerous models and definitions for shared decision making, especially related to advanced cancer. (1–3) The lack of agreement regarding a theoretical framework and definition have made comparisons across studies very challenging. (1) As a result, measuring the concept and practice of shared decision making is also difficult. (4)

Shared Decision making was first defined in a research report by the Presidents Commission for the Study of Ethical Problems in Medicine and Biomedical and Biobehavioral Research in 1982. (5) That definition was based on mutual respect and partnership. The commission stated that the health care professionals invited patients to participate in a dialogue where the professional helps the patient understand their medical situation and the various choices and the patient states their wishes. (5) Years later in 2006, Makoul and Clayman did an extensive literature review of the shared decision making literature.(6, 7) Their review concluded that there was no standard definition of shared decision making within the literature. They identified 20 different concepts in the literature related to shared decision making and found 60% of the articles failed to include any conceptual definitions of shared decision making. One third of the papers which did have conceptual definitions did not ground them in any previous literature.(1)

While there is agreement that hospice and palliative care are important settings for shared decision making, there are few studies in this environment. A systematic review of shared decision making in palliative care concluded that decision making between providers and patients and families has rarely been observed. Additionally, they found that the process fails to reconcile the patient perceptions with a third party assessment of the shared decision making option. The review also discovered that palliative care research tended to have patients who were improving throughout the nursing home rather than ones that patients had declined in.(8)

Makoul and Clayman used the concepts from their review of shared decision making definitions to identify nine essential elements for a shared decision making process. (1) These elements are defined and explained in Table 1. They purposefully chose the term elements to avoid placing responsibility on either the patient or provider for the specific task. Makoul contends that for shared decision making to occur patients and providers first identify a problem and discuss options available. A discussion of the risks and benefits follow as well as a discussion about how the options relate to patient values and preferences. Following this, the patient’s or family member’s self-efficacy or ability to follow through on the options is examined before the provider makes a recommendation from among the identified options. The assessment of the extent to which all parties understand the presenting problem and the courses of action available should occur before a decision is made or it should be deferred. Finally, a follow-up plan to evaluate the decision or to make a deferred decision should be standard practice.(1)

Table 1:

Makoul and Claymans Nine Essential Elements for Shared Decision Making (1) and frequency of use in visits

Essential Element	Conceptual explanation	Frequency n(%)
1Define/explain the problem	Patients, family, or providers define and/or explain the problem which needs to be addressed.	64 (98%)
2Present options	A series of options or alternatives is presented to be considered. Providers should review the options which exist and patients and family should present options which they may know of.	31(48%)
3Discussion of pros/cons	A discussion of pros/cons, benefits, risks, costs should be discussed as providers and patients may have differing perspectives on the importance, especially convenience and opportunity costs.	21(32%)
4Patient values/preferences	Patient’s values and preferences need to be clarified through discussion of ideas, concerns, and outcome expectations.	39(60%)
5Patient ability/self- efficacy	Discussion of patients ability, self-efficacy, to follow through with a plan helps determine viability of options.	14(22%)
6Provider knowledge/recommendations	Physician or provider knowledge and affiliated recommendations in the context of the specific decision need to be explained.	45(69%)
7Check/clarify understanding	Both parties should check understanding of facts and perspectives throughout	5(8%)
8Make or explicitly defer decision	Decisions may not always be possible upon first discussion and additional work may be needed so they can be deferred to another time.	20(31%)
9Arrange follow-up	Physicians and patients should arrange a follow-up to track the outcome of decisions which were made or to reach a decision which was deferred.	32(49%)

1. Makoul G, Clayman M. An integrative model of shared decision making in medical encounters. Patient Educ Couns. 2006;60(3):301–12.

Hospice care was established on the philosophy that patients and families are a part of the interdisciplinary team and have the right to participate in decisions related to their care. (9) Despite this focus, little research has been done exploring the shared decision making process in hospice. A recent study found that hospice family members want to be part of the hospice team. (10) Washington and colleagues (6) conducted a multi-method study evaluating the involvement of patients’ family caregivers in team decision making as they conducted interviews with providers and observed team care plan meetings. The researchers identified barriers to team shared decision making that included time constraints, communication skills, unaddressed emotional needs of family members, staff attendance and involvement, and a lack of a structured shared decision making process. (6) However, no studies were found that observed hospice shared decision making during hospice care delivery with individual providers.

The goal of this study was to apply Makoul’s essential elements for shared decision making to encounters between hospice nurses and hospice patients and family members during home visits. We structured the operational definition of shared decision making after the Washington hospice team study. (6) In that study, shared decision making was defined as a process wherein a choice is jointly made by hospice staff members and a hospice patient or family member. (6) Using Makoul’s essential elements (1) and Washington (6) hospice specific definition, we addressed the question: What elements of a shared decision making process are found in home hospice nursing visits?

Methods

We conducted a secondary analysis of a sub-sample of audio tapes collected from a nationwide national hospice study. (11) Details of the larger study are available elsewhere (11) however, in brief, the national study recruited nurses from hospice agencies in four different regions of the US. The hospice nurses were asked to volunteer to audio record their home visits with cancer patients. They were asked to wear digital recorders during their home visits for approximately two weeks. At each visit the nurse identified herself, the patient, and the date before entering the home. The patient and those present were reminded they were being recorded and that the recorder could be turned off at any time. Recorders were collected by study staff and data was uploaded to a secure server by study staff. With the approval of the University of Missouri Institutional Review Board and a signed Data Use Agreement with the University of Utah, 65 nurse visits were selected as a stratified sub-sample for maximum variance based on years of nursing experience, nurse race, and nurse gender. This number was selected as it allowed data saturation. The selected visits were regularly scheduled and did not include initial admission visits, emergency on call visits, or visits where the patient died. The audiotaped visits were transcribed verbatim.

The transcript utterances were analyzed using the pre-established coding frame of Makoul’s elements of shared decision making. Nvivo software was used to organize coded themes. Methodological rigor was maintained with co-coding and ongoing peer debriefing. (12) Two coders were trained and separately reviewed the transcripts and coded pertinent utterances using the nine elements. All coding was reviewed by the first and second author concurrently and consensus was reached on each code. The findings were reviewed and discussed among the research team throughout the process.

Results

The demographic and practice characteristics of the 65 nurses in this study are found in Table 2. There were 18 hours and 21 minutes total minutes of audio, with a mean visit time of 38 minutes. Nurses on average had slightly more than three years of hospice experience with a range of 0–21 years. Despite efforts to obtain maximum variance, the majority of nurses were female (91%) and white (74%).

Table 2:

Summary of Demographics of Sample (n=65)

Demographic	Frequency (Percentage)
Gender Female Male	59 (91%) 6 (9%)
Race White Black/African American Asian Other Unknown	48 (74%) 4 (6%) 2 (3%) 2 (3%) 9 (14%)
Degree Associate’s degree Diploma Bachelor’s degree Graduate degree Unknown	41 (63%) 3 (5%) 17 (26%) 3 (5%) 1 (1%)
	Mean (Range)
Age	44.2 yrs (25–69 yrs)
Nursing Experience Mean Years as RN Mean Mean Hospice RN Mean	13.6 yrs (1–46 yrs) 3.3 yrs (0–21 yrs)
Visit time Mean length of visit	38 mins (11–95 mins)

Analysis found the use of all the elements of shared decision making in the nursing visits, however they were not all used in every visit. Table 1 summarizes the frequency of use of each element. Only two (3%) nurse visits contained all nine elements, and another 5 (8%) demonstrated all but one of the elements. Nearly one quarter (22%) used 6–7 elements, and 28% used 4–5 elements. Eighteen percent used less than half of the elements, and nearly one-third (31%) of visits had no elements of shared decision making. The most frequently used element was defining the problem and the least used was the assessment of patient and family understanding.

Element 1: Define the Problem

Defining a problem was communication that sought to clarify, assess, or understand a specific concern. Given the hospice focus on the patients and the family as the unit of care, we did not differentiate between patients and their caregivers when coding the elements or examining the problems identified in the visits. The 65 nurse visits identified 271 different types of problems (an average of 4 problems per visit). Table 3 details the nature of the problems defined in these encounters. An illustration of defining of the problem can be seen in the conversation below during which the nurse assessed the patient’s pain.

Table 3:

Problems Identified in Hospice Nurse Visits (N= 65)

Problem defined in visit	Frequency of the problem
Pain	19
Nutrition/Hydration/Weight	11
Shortness of Breath	16
Balance/Ambulation/Fall	12
Bowel/Bladder	9
Wound	7
Fatigue	6
Edema	5
Psychosocial	5
Sleep	4
Financial issues	3
Swallowing	3
Mental Status	2
Infection/Thrush	3
Vomiting	2
Eyesight	1
Hiccup	1
Itchy	1
Palsy	1
Speech	1

Nurse: How many of your breakthrough [pain medications] did you have in the last 24 hours?

Patient: Let’s see, I had one last night I had 3 yesterday.

Nurse: [Are you] doing them one at a time or two at a time?

Patient: Last night, it was two at a time.

Nurse: How about during the day?

Patient: During the day it depended.

Nurse: How intense was [the pain]?

Patient: I want to say between 3 and 4

Nurse: Okay

While discussion of this problem was initiated from the nurse, other concerns were initiated by family members or patients.

Element 2: Identify Options

The second element in Makoul’s model, identifying options, was defined as communication that presents various solutions for a defined problem. Options were specifically identified in nearly half (48%) of all the nurse visits. The intent of the model is that more than one option is identified to allow a choice. In nearly one third (33%) of visits only one option was presented. Two or more choices were presented in 15% (10) of all the nurse visits. An example of this can be seen in the following visit when the nurse gives options for a physician assessment of a skin problem—either describing the problem to the physician or taking a photo.

Caregiver: So this is something that ought to be looked at? I don’t know, do you think the doctor ought to come and check him out?

Nurse: I can describe it to him and see what he recommends. The other thing that we do from time to time, like with wounds and things we take pictures of the wounds and upload it onto their file. Oh so he could [see it]

Nurse: Then he can have access to it or if he wanted to kind of take a quick look at it I could send it to him over the phone.

This example is illustrative of the creativity sometimes necessary in identifying options in an isolated home setting.

Element 3: Discuss Risks and Benefits

Communication that examines or provides education regarding the specific advantage(s) and/or disadvantage(s) of a treatment or behavior is defined as the discussion of risks and benefits, the third element in the model. Discussion of risks and benefits was identified in 32% (21) of all the nursing visits. Both risks and benefits were discussed in 9% (6) of all visits. Discussion of only benefits was observed in 12% (8) of all nursing visits, and 11% (7) discussed only risks. An example where risks and benefits of an option were discussed can be found in this discussion between a nurse and patient:

Patient: How can someone gain weight like that while not eating? [Is it] because of my medicine?

Nurse: The steroids can do that. You have puffiness in your face, but it’s not like some of my patients with congestive heart failure. Their legs become so swollen … you know, a lot of water weight.

Patient: Yeah.

Nurse: You’ve got a tiny trace of edema in your feet.

Patient: Umm hmm

Nurse: I can see [swelling] on your face. It is a little puffy.

Patient: It [can’t] get any bigger, buddy. It is not going to be good. Then what are we going to do?

Nurse: Even your dexamethasone [causes swelling]. You are on a low dose.

Patient: I know.

Nurse: We’ve got some people that are taking a really high dose. Yours is low. It’s not like you’re overdoing it, and I think it’s still helps the swelling in your head, and I think it still gives you so much of the benefits that to stop it I think you’ll be more uncomfortable. Even though I know that it is an unpleasant side effect.

Patient: Umm hmm.

As seen in this interaction, often the discussion of risks and benefits was related to education of side effects of drugs or symptoms of the continued disease progression.

Element 4: Patient/Family Preferences

Patient preferences and values were heard in 60% (39) of the nursing encounters. Preferences and values involved three main themes: logistic preferences, for example how to obtain medicine, were found in 28% (18) of all visits; treatment preferences, for example which medications to take, were found in 51% (33) of the nurse visits; and finally, patient values, for example how much pain to tolerate, were identified in 43% (28) of the discussions. Below is an example coded as a logistic preference as the nurse and patient discussed her father becoming her caregiver.

Patient: I just never had my dad here to help take care of me.

Nurse: Will that be kind of weird or do you think it will feel nice?

Patient: I think it will be nice for a change.

Nurse: Yep

Values and preferences, especially logistical ones, were impacted by the ability of loved ones to actually carry through with the options most preferred.

Element 5: Patient/Family Self-efficacy

Self-efficacy involves communication by the patient and/or family that they believe they can carry out the selected option. This belief is important because if the option cannot be carried out it is not a viable plan. Communication regarding self-efficacy was identified in 22% (14) of the 65 visits. One example of this was a conversation between a family caregiver and the nurse as the caregiver demonstrated her understanding of filling the medication box.

Nurse: Oh look! Good job! This is perfect.

Caregiver: And this [box] is for stragglers during the day. So we just fill that up every morning.

Nurse: Yeah.

Caregiver: And then that way we know [by the boxes if he has taken extra] at 11:00, 3:00,5:00 [because] he has additional meds that don’t quite fit into this.

Nurse: This is wonderful! This is perfect!

Other examples also included acknowledgement for successful completion of duties by caregivers and caregiver concerns regarding their ability to handle various situations.

Element 6: Nurse Makes a Recommendation

A statement was coded as a recommendation when it was communication that provided a suggestion by the hospice nurse. Recommendations by the hospice nurse were seen in 69% (45) of the encounters. These recommendations varied depending on the problem. An example of a nurse recommendation is when the hospice nurse recommended action with the patient’s pain medicine:

Nurse: This way you are not playing catch up because if you do not give him anything, his pain might be up to here and then we have to move the medicine up to here so that we can catch up with that pain. He gets on this roller coaster. What we want to do is give him a little bit of medicine, get him feeling good, and make sure that he’s not going to have that pain up to here. And that way the pain [medicine dose] is just a little bump, which is a different thing. That is my recommendation.

Nurse recommendations were common related to symptom management and potential changes in medications.

Element 7: Assess Patient/Family understanding

The next element of the shared decision making process involved the nurse assessing the extent to which patients and caregivers fully understood the presenting problem and the options available to address it. This was only found in only 8% (5) of the visits. One example of this element was when the nurse had the patient and caregiver listen to the patient’s lungs to understand the fluid she was hearing.

Caregiver to the Nurse: So you hear all that gurgling moving around?

Nurse to the Caregiver: Do you want to listen to it?

Caregiver: yeah

Patient: The big change for me is that when I move forward, a bunch of liquid comes out.

Caregiver: Oh wow.

Nurse to Caregiver: Now let’s listen to my lungs.

Caregiver: The only thing I hear is air moving like the wind

Nurse: yeah

Caregiver to the Patient: Do you want to hear yourself?

Patient: Sure. Ok

Caregiver to Patient: Place that on your back

Nurse to Patient: Now let’s listen on the other side because this side is very amazing.

Caregiver to Patient: can you hear it?

Nurse to Patient: can you hear it?

Patient: Yeah

This element is not solely patient or family education, but included efforts by the nurse to determine if comprehension had taken place. This element was found when a discussion had the purpose of checking and clarifying the patient or caregiver understanding about a specific care issue and proposed options.

Element 8: Make a decision

Items were coded as a decision when communication indicated a choice had been made regarding a specific way to address an identified problem. Decisions were made by patients or caregivers in 31% (20) of the encounters. Decisions were varied but one example of a patient confirming a nurses decision follows:

Nurse: Alright. Well, I hate to see you uncomfortable. So that’s why we’re kind of problem solving. So just try a couple of drops of the haloperidol and see if that helps you, without putting you to sleep.

Patient: Couple of drops.

Nurse: Yes. Ok? This is an experiment.

Patient: Ok.

Decisions were by necessity also at times deferred as the nurse had to seek a physician order for final confirmation of the decision.

Element 9: Follow up

The final element of shared decision making is reflected in communication that includes a plan for returning to a specified problem in order to make sure the issue is adequately resolved (e.g., a return visit, phone call, etc). We found evidence of plans for follow up in 49% (32) of the encounters. An example is when a nurse told a patient the following:

Nurse: If you need me, you call me. If it is 2 o’clock Sunday morning, you call with that on call number, and I will be here with smiles and bells on my feet within a half hour or so, depending on what’s happening, but I want you to be comfortable calling if you need to.

Follow up often involved the nurse arranging for another visit or a follow up phone call with the family.

Discussion

These data show evidence of many shared decisions and an active shared decision making process during hospice nurse visits. The number of problems identified in these visits demonstrates the significant complexity of nursing visits in hospice. With an average of more than 4 unique problems per nursing visit and an average of 38 minutes per visit, hospice nurses must be efficient problem solvers, and the shared decision making process must be efficient to address identified problems. The two visits that used all components of the process were longer than the average visit of 38 minutes (64 and 49 minutes) however, there were other visits that were longer and did not use all of the elements. These data show hospice nurses need to be well versed in solutions for many different issues when in the home of a dying patient.

While the majority of the literature discusses decision making between patients and their physicians (2, 13, 14), once a patient enrolls in hospice their condition often prevents them from visiting the clinic and seeing their physician. The hospice nurse becomes the link between the physician and the patient and family and witnesses the results of all decisions made in the home setting. The nature of the problems found in this data and the options presented by nurses provide an insight into the complexity of the home environment.

Makoul’s comprehensive elements are behaviorally based and easily observable or interpretable. As shown in this analysis the elements are also applicable in the hospice setting and provide a comprehensive model for decision making, even in the home environment. However, the data illustrates potential challenges with shared decision making given the number of problems being addressed. For instance, identifying options and discussing risks and benefits for so many different problems in one visit may be overwhelming for patients and families in distress. These data also illustrate the tremendous autonomy for hospice nurses as they interact with patients and families, making split moment decisions on how to teach, educate, recommend and follow-up with defined problems on a regular basis.

This study data are interesting to compare with the team decision making found by Washington and colleagues. (6) The barriers in shared decision making were comparable between the team meeting process and the individual nurse visit. There were time constraints in the team meeting setting as teams had many patients to discuss in a limited timeframe. Likewise, hospice nurses have productivity expectations that prevent them from extending their visit time. Additionally, the decision making techniques used or not used in both settings reflected the communication styles of nurses, patients, and families. Finally, the assessment of patient/family understanding and self-efficacy were infrequent in both settings. The lack of use of these two elements indicates a potential intervention that would introduce the “teach back” method to assess understanding and self-efficacy which might be a helpful tool for hospice professionals. (15) The teach back approach would have caregivers or patients show their understanding by demonstrating or describing an intervention, option, or decision to the hospice nurse. For example, a caregiver demonstrating the understanding of the medication box by filling it under the nurse’s observation rather than the nurse filling the box for them.

There are limitations to the analysis. First, these data reflect only one visit to each patient and therefore do not provide the full context to understand an already established relationship between the nurse and the patient. Within these established relationships it is possible that the nurse had previously identified problems or has a history of understanding the family system’s preferences. Likewise, solutions to some problems may be straightforward and thus not require a shared decision making process. Finally, especially when under stress, some individuals just wanted someone else to make a decision. An interesting example was a caregiver who stated, “I need you to tell me what to do.”

Conclusion

While shared decision making has become common practice within oncology, the research specific to hospice is limited. Hospice is grounded in a philosophy supportive of shared decisions, yet given the limited interaction of patients and families with physicians, it must be studied in the context of relationships with nurses, a unique twist to shared decision making research. Home hospice nurse visits involve addressing and managing a number and variety of problems. While shared decisions occur in the home, there are opportunities to improve the shared decision making process, especially as it pertains to ensuring that patients and families understand the plan of care and feel comfortable implementing that plans. Hospice staff can benefit from a clear understanding of a more structured shared decision making process and helpful tools such as teach back to assure understanding and self-efficacy. The challenge with Makoul’s model is to use all the elements in a timely manner given the length of clinical encounters. This is an important issue as we translate the model into clinical practice and one which needs further research. This study also demonstrates the similarity between decisions in both hospice team and hospice visit settings. Additional research is warranted to assess the application of Makoul’s elements in settings outside the clinic and home hospice. These elements may provide an important structure in any health care setting where patient and family input are valued.