Palliative care for people with substance use disorder and multiple problems: a qualitative study on experiences of patients and proxies

A qualitative design was used, consisting of semi-structured interviews with patients with SUD+ and their proxies. Data-collection lasted from August 2017 till January 2018. For this publication, the Additional file 1: COREQ checklist (COnsolidated criteria for REporting Qualitative research) is used [42]. Previously, an extensive description of the study-design has been published [18].

To reach saturation within the qualitative data, we aimed to include ten to fifteen patients and five to ten proxies. If insufficient, more respondents would be recruited. The patients were recruited by HCPs who were specialized in SUD and had, at least, generic knowledge of PC. Proxies were approached via the patient. The main researcher (AE) included the patients based on the inclusion criteria (Table 1). AE and the patients did not have a prior relationship with each other.

Together with experts-by-experience (people with lived experience with SUD), the project-group developed an interview guide suitable for this patient group. Simple terminology and short questions were included. With regard to potential communication difficulties, the interviewer also brought so-called association cards to the interviews which could be used if patients would struggle to find words that covered their experiences. Furthermore, keeping into account patients’ vulnerability, the time was limited to one hour and the interviewer frequently asked for patients’ well-being [18].

Before the interview started, proxies and patients were asked for a few demographics. Semi-structured face-to-face, audio-taped interviews with patients and proxies were held by AE (experienced interviewer) at the respondents’ residents. This method enabled the interviewer to discover the respondent’ experiences. Respondents in turn, had room to explicate themselves [43, 44]. Respondents could bring somebody to the interview, but this did not occur.

Topics of the interview-guide were based on the main research question (What are the problems and needs of patients with SUD and multiple problems, and their proxies, in a PC phase?), literature and the project-group’s expertise: researchers, nurses, a manager and experts-by-experience. Topics concerned patients’ 1) needs and experiences in professional healthcare and; 2) well-being and needs on all life-domains. Proxies answered a few extra questions. The impact of SUD and the attention of HCPs to the needs of patients and proxies were related to every topic (Table 2). Proxies answered the questions from topic 1 till 4 about the patient's situation. The guides were tested within the project-group. AE made field notes during and after the interviews. Respondents were sent the verbatim interview transcripts.

Data-analysis consisted of several steps. Grounded theory elements for inductive analysis were used by the researchers (AE and MG). To grasp informants’ meanings and experiences, first, AE and MG applied in-vivo codes to original interview text segments. They stayed close to the data to prevent over-interpretation. Since intercoder agreement between AE and MG on the first interview was high, the second and third interview were coded by AE alone [45]. MG listened to one more interview to further familiarize herself with the data and to agree on the codes attached by AE. Second, AE clustered these in-vivo together into sub-codes, which were headed under a main-code, belonging to a theme. Eventually, five themes were identified. Connections between sub- and main-codes were described and pictured in a network (axial coding). Afterwards, sub-codes and themes were supplied with a description. In this way, a code tree was generated which served to code and analyse the remaining interviews (selective coding). AE discussed the codebook with MG. When new codes or information emerged, descriptions were updated (constant comparative method) [46, 47].

After coding six interviews with the codebook, data saturation was reached; no new codes came up and adaptations to descriptions were minimal [48]. The remaining interviews were used to confirm data saturation and supply examples. Data-analysis was supported by the research software ATLAS.ti 8.0.34. To minimize researcher bias and to maximize validity, AE and MG frequently joined for discussions about, e.g., ambiguous text fragments. No member-check was executed due to time-restraints.

During recruitment, thirty patients were found potentially eligible for inclusion, of whom nine were eventually included. Others were excluded, because they: a) denied disease (n = 7); b) were too ill to be interviewed or died before the interview took place (n = 7); c) were not willing to talk about the subject (n = 3); d) cancelled or denied participation for unknown reasons (n = 2) or; e) did not trust what would happen to their data (n = 2). Table 3 shows respondents’ characteristics in a narrative and anonymous matter (no gender and individual age). Patients (eight male, one female) were 61 years on average and often finished lower education and were single. None of the patients stayed in their own homes; seven stayed in a nursing home and two had a room in either a social care service (hostel) or an assisted living home where medical care was provided. One respondent was suffering from progressive, severe physical deterioration (cirrhosis of the liver) as a result of active addictive behaviour. The others had a serious non-reversible, life-threatening disease (COPD) or both COPD and cirrhosis of the liver. Only three proxies were interviewed, because either patients did not have close proxies or AE, based on the interview story, found asking for proxy participation inappropriate.

Five themes were derived from interviewing: 1) healthcare delivery; 2) end-of-life (EOL) preferences; 3) multidimensional problems; 4) coping and; 5) closed communication. The fourth and fifth theme do not answer the main question directly, but were considered important by the authors as these could impact communication during PC. Theme 4 is based on the interviews’ content. Theme 5 is based on the experiences of AE during the interviews, preparatory conversations with HCPs and proxy interviews.

Some patients had strong critical opinions about how they desired to be treated and which care they would like to receive. Other patients, however, were much less critical and sometimes also more grateful for the care they received. One patient said he should “have a look in the mirror first” before he could criticize care or HCPs. Also in light of great work pressure of HCPs, patients took upon a humble attitude. Furthermore, in contrast to bad experiences before, these less demanding patients stayed in places where they felt quite comfortable and taken seriously. Despite that respondents varied in their experiences and appreciations of care, some behavioural issues of HCPs and values were important in healthcare to most patients and their proxies. Quotes are presented in Table 4.

During interviewing, the involvement and willingness to talk about EOL-preferences and EOL-decisions differed greatly between patients. Also despite that the interviewer asked about the future and even afterlife, most patients spoke about desires concerning the short-term and the last days only. The period in between remained largely unspoken as patients appeared to focus on the ‘here and now’. Consequentially, EOL-preferences can be divided into ‘the current’ and ‘dying’. The extent of HCPs’ attention for these preferences and, consequentially, written or verbal agreements for future-plans varied.

Previous long-term and/or present persisting substance use influenced the extent and diversity of the problems of this study’s patient-group. Patients had different experiences in HCPs’ attention for their well-being on all four dimensions of PC: physical, psychological, social and existential.

Patients’ existential issues varied greatly. By being confronted with death and dying, patients sometimes gained insights into life or themselves and stated to be changed (positively) as a person. Furthermore, existential loneliness was implicitly expressed by some patients (Q16), whereas proxies explicitly stated so about patients.

Patients handled their multiple problems, history with substance use, their incurable disease and approaching death, differently. Various coping styles came to surface during the interviews, which can be divided in active and passive coping styles. These changed over time and differed per situation.

Going deeply into the interview topics appeared hard and sometimes was hindered by several issues. The first is expression. Patients’ answers, often, were short and closed. Some patients seemed to prefer simple questions and terminology. Others were quite avoiding in their answers or stated that they “are not feeling up” to, e.g., the interview or talking about EOL-issues. Sometimes, answers were incomprehensible, contradictory or incomplete (Q26–27). Second, during interviews, there was quite much restlessness, often caused by craving to, for example, smoke. Third, whereas patients during parts of the interviews seemed aware of their disease or their PC phase, they did not seem to be so at other moments during the interview or they stated that they did not know much about it (Q28). Proxies stated that, as a consequence of patients’ closedness in communication, they themselves were allocated bigger responsibility and a greater role within care. Also both proxies and HCPs stated that observations were important to allow for patient-centered care.

This study has several practice and educational implications. First, the high number of patients who were excluded because they were too ill, might indicate non-timely identification of PC phases within this patient-group. This assumption is confirmed by other studies, in which late identification of PC needs was, among other factors, attributed to a lack of knowledge about PC in this patient-group on the one hand and patients’ care-avoidance on the other [39, 40]. Consequentially, patients might miss opportunities for prolonged survival, higher QoL and decreased symptom burden [49‐52]. Therefore education about early identification of PC is needed. In light of patients’ limited communication, attention for observing, e.g., body language or changes in day-to-day functioning indicative of deteriorating health, is important [53, 54].

Second, patients’ EOL-preferences and ways of handling their disease seem somewhat different compared to the more general PC population. Findings show comparisons, for example, with regard to anxiety about suffering from pain, but also differences. Whereas other populations define ‘a good death’ by, for example, being prepared, saying goodbyes, and having family support [55, 56], this patient-group does not appear to do so necessarily or at least, not explicitly. Also the way respondents took up health and disease seems different. In general, being ill imposes patients with a certain responsibility and involvement in improving their situation, in exchange for a temporary removal of, e.g., employment [57, 58]. Not all respondents from this study, however, did fulfil this generally accepted ‘sick role’. Some even remained active users and some were not very willing and avoiding to talk about death or dying. Since patients with SUD+ thus seem to differ in their EOL-preferences, healthcare behaviour and coping, it is not unimaginable that HCPs who are unfamiliar with this patient-group might experience difficulties in understanding and discomfort in caring for this patient-group. A comparable study indicates the same practice implication and points at the importance of a non-judgmental attitude [59]. Stigma on SUD is a major issue in caring for this patient-group and education or bed-side teaching could be helpful [60].

Third, to provide person-centered care, HCPs need to know what is and will be of importance to a patient. This though is hindered by patients’ focus on the ‘here-and-the-now’ and their closed communication. This short-term focus was found in another study about homeless people in PC too [61]. Furthermore, for patients to be aware of and express what is important in their last phase of life, they need the ability to rationalize and understand their disease and death [33, 61‐64]. Respondents of this study sometimes temporarily lacked such realisation. If advance care planning is to be adopted for this patient-group, HCPs thus need to create trust and have alert and enduring attitudes. Repeating as well as clear conversations about EOL preferences are important [59]. A dedicated nurse supervisor may be an option. Getting insight into life stories via proxies, furthermore, might be helpful to retrieve patients’ desires. These proxies, though, need to be supported, as was suggested in other studies too [39, 65].

Future research into the diversity of the patient-group should be undertaken as it is likely that we interviewed a subgroup. The bigger patient-group might include patients that are cognitively or intellectually more disabled or not admitted to healthcare services. Also more as well as bereaved family caregivers could be interviewed to complement their experiences. Finally, observations of patients during entire last phases of life, may provide insight into what they find important in-between the ‘here-and-the-now’ and the final days.

This study adds much insight to the existing literature into the experiences of patients with SUD+ in a PC phase, and their proxies. Despite the well-adapted design of the interview guide and the respectful attitude of the interviewer, avoidance coping and possibly shame might have impacted the openness and duration of the interviews. Also only three proxies were included. Finally, the current sample includes only patients from in-patient setting. However, whether this patient-group is at all able to live at home because of their complex problems and lack of informal caregivers, is questionable.