In January 2013 we asked the Portuguese Pneumology Society (SPP) to provide us with the mailing list of all affiliated Respiratory Medicine specialists and fellows with email contact. We obtained a list of 464 members. We only asked physicians who had had clinical activity in the previous year to answer the survey.

A self-administrated questionnaire was developed specifically for this study. The authors generated questions based on a review of the literature. Questions were drawn up to address four main themes: (1) palliation of symptoms; (2) frequency and topics of EoL communication; (3) timing and interveners of EoL communication; (4) obstacles to EoL communication.

The following data were collected to characterize participants: age, gender, religious beliefs, years of practice, type of institution, region of the country and number of patients with COPD treated per year.

At the beginning of the questionnaire, participants were asked to answer questions with reference to the patients they had treated in the previous year, who had two or more of the following characteristics: forced expiratory volume in first second (FEV1) under 30%, oxygen dependence, one or more hospital admissions in the previous year for an acute exacerbation of COPD, left heart failure or other comorbidities, weight loss or caquexia, decreased functional status, increasing dependence on others and were over 70 years old. We adopted these criterion from Curtis, whose purpose was to identify patients in whom a discussion on preferences about EoL care should be initiated.5

We pretested the questionnaire on a convenience sample of 6 pulmonologists to ensure clarity and comprehensiveness. A web platform was then designed to facilitate survey completion and data collection. Participants received an invitation to answer the questionnaire via an e-mail sent by the SPP, containing a link to the online platform. One month was allowed for questionnaire completion. During the one month period two e-mails of reminder were sent to each applicant, after which time the website was closed. All complete questionnaires were considered eligible so that we are characterizing the reality of our practice.

Demographic data and the several variables related to palliation and communication were analyzed using descriptive statistics. To verify a possible influence of the participant characteristics on their ability to undertake EoL communication, a univariate analysis was performed between the main characteristics – years of practice, type of institution (university hospital vs other), number of COPD patients treated – and two questions evaluated by a semi-quantitative scale: (a) “How often have you started a EoL discussion on your initiative?”; (b) “How difficult is it for you to discuss EoL issues with your patients?”. Statistical significance values were set at p≤0.05.

Although the majority of physicians reported that they treated symptoms at least occasionally, our findings suggest that symptom management may not be optimized. Only a few reported using opioids routinely in the treatment of dyspnoea and pain. Since opioids are the only drug proven to be effective in controlling dyspnoea in these patients, these results are clearly unsatisfactory.7 Previous studies have shown evidence of poor palliation in COPD.8–10 The study SUPPORT, for instance, revealed that dyspnoea occurs more often in these patients than in lung cancer patients. In the same study, 21% of patients with COPD reported suffering from severe pain.10 There is a strong possibility that analgesics are withheld in COPD because of concerns about adverse effects, in particular respiratory depression. This fear is unfounded, since there is no evidence to date that the doses of opioids used to treat breathlessness causes clinically detectable respiratory depression or increases mortality.7

It has been suggested that anxiety/depression affects as many as 90% of advanced COPD patients, but previous studies have reported that only 4–31% were receiving pharmacological treatment.11,12 We found a better performance, which may reveal a rising awareness of the beneficial effects of antidepressants in improving mood and decreasing dyspnoea in COPD patients.13 Treating depression is essential since it may modify patient decisions concerning EoL care.14

Another important outcome of this study was to realize that more than half of Portuguese Pulmonologists never or rarely introduce EoL discussions with their patients. Previous reports, both from physician and patient point of view, have also shown insufficient EoL communication.15–17 One of the topics most neglected by physicians was spiritual/religious beliefs, although there is evidence that the majority of patients want to discuss these issues.18

Although advance care planning seems to increase patient satisfaction and reduce emotional distress, our results show that physicians have used little initiative in introducing it.18 The lack of certainty on prognosis strengthens the need for advance directives in COPD, but the evidence is they are rarely used. A Survey carried out in the European Respiratory Intermediate care units showed that only 30% of patients admitted to having had a formal discussion about EoL decisions and/or signed a living will.19 As a result, COPD patients are more likely to die with aggressive, technological care directed at preservation of life and less likely to receive palliative care, despite having preferences favouring palliation.10 In fact, the majority of COPD patients die in the hospital, mostly in intensive care units, although at least half of clinically stable advanced COPD patients would prefer to die in their home environment.8,20

Hospice care in end-stage COPD is another emerging issue. The recent revised version of GOLD recommends that clinicians should identify COPD patients who could benefit from palliative care services and help find available resources within their community.4 In our study, about half the respondents indicated they occasionally/frequently refer patients to hospice care. Nonetheless, patients dying of end-stage COPD are in general less likely to be offered hospice services compared to patients with lung cancer.5 Uncertainty about prognosis as well as doubts as to whether a traditional hospice model is suitable for patients with COPD, may explain this difference, although it should not impede access to hospice care.2

Recent consensus recommends that EoL care should be available early during the course of the disease, from when the patient becomes symptomatic.3 However, contrary to this, we found that most physicians discuss it during/after a major exacerbation. Another study also reported that discussions occur mainly in advanced stages of COPD, that is when dyspnoea became severe.21

There are more growing recommendations on the use of interdisciplinary health-care teams to ensure quality EoL care.2 Nonetheless, we found poor interdisciplinary work in our study. Results were similar in the previously mentioned European survey, which found that nurses were involved in 55.9%, but only in a small minority of cases were opinions sought from other professionals. In the same study, patient's relatives were involved in discussion in only about half the cases.19 We in turn, found a better performance in terms of engaging the family, although often without patient participation. Interestingly, our participants rarely reported modifying the information provided in order to influence decisions, a phenomenon that was reported to commonly occur in previous studies.21,22

Most participants described the discussion of EoL issues as being difficult or very difficult, which makes it essential to understand what the barriers are that physicians face, especially the ones that can be modified.

As in a previous study, the feeling that patients may be unprepared to discuss these matters and the fear of taking away their hope, were pointed out as major obstacles to communication.17 Patient failure to understand COPD as a life-threatening disease and their will to concentrate on staying alive play an important role in this process.17 Cultural differences possibly have an influence as well. In a European survey evaluating public preferences for self-involvement in EoL decision-making, about 60% of the Portuguese participants wanted to be involved, which was less than the European mean (74%).23 These barriers may be overcome with education about illness progression and encouragement to communicate.

Other obstacles mentioned by our respondents have to do with organizational aspects, namely lack of time, lack of training and absence of protocols. The lack of formal protocols to initiate EoL was also pointed out by Partridge et al.24 Additionally, educational practices are inadequate, as revealed in a survey performed in the United States.25 These aspects represent important targets for intervention. Each institution should develop policies to guide health staff in initiating the kind of EoL care that is feasible in their context.2 Scientific societies, in their turn, are responsible for ensuring that there is appropriate education on EoL care in medical schools and for promoting continuous education among attendant health-care givers. There is evidence that educational programmes and even smaller interventions result in measurable improvements in communication skills.26–28 Moreover, our study demonstrates that Pulmonologists are interested in further education on this subject.

Limited resources are a significant impediment to effective EoL care. In Portugal, the need for fiscal consolidation has led to major cuts in the financing of the health sector in recent years.29 Despite the National Program for Palliative Care which was created in 2010, numbers show that resources (regarding number of hospice care beds and the availability of domiciliary services) are still far behind recommendations by the European Association for Palliative care.29 These restrictions are recognized by Portuguese citizens, who call for more hospital palliative care units and more palliative care access.30 Evidence-based service development using existing resources may be part of the solution.6

According to our results, the particular characteristics of a physician do not seem to influence their ability to communicate on EoL issues, which suggests that the barriers identified are transversal to different practice settings.

The present study has a number of potential limitations.

The Curtis criterion to identify candidates for EoL discussion may be subject to criticism. It is a presumptive criterion, since it has not been proven that patients who are included die sooner or have more symptoms. Other proposed criteria, namely the CART and BODE scores, both able to predict mortality in patients with COPD, would be equally valid.31–33 Nonetheless, we believe the former is more suitable for the purposes of this type of questionnaire, since it includes a list of clinical characteristics, which are easier to answer with a yes/no response, when mentally profiling the type of patients. Whereas CART is a tree-algorithm and BODE requires a numeric grading of each variable, both of which cases would make this process more difficult.

Concerning the use of opioids in the treatment of dyspnoea, we retrospectively realized that the question should have been put in a more objective way. Since current recommendations to treat dyspnoea include a multimodal approach, we recognize that the way the question was phrased may not reveal the true scale of opioids use for this purpose.

We recognize that our participation rate was low. However, our universe included an unknown number of retired physicians, who were asked not to answer the questionnaire, and thus underestimated the overall response rate. Also, the choosing to use a web platform to perform the survey, may have influenced the response rate and the profile of the respondents.

Another obvious limitation is that the self-reported information may not reflect what Pulmonologists actually do in practice. Future studies should find ways of quantifying EoL communication more objectively, while recognizing that research in this area is hampered by important ethical and methodological challenges.34

In conclusion, our study demonstrated that our palliative care and EoL communication are a long way from being from what they should be. Awareness, training in this field and the creation of formal protocols have been identified as major areas for improvement, and should be encouraged. Future studies should focus on developing and testing the effectiveness of practical strategies to improve EoL communication and palliative care in COPD.

The authors declare that no experiments were performed on humans or animals for this study.

The authors declare that they have followed the protocols of their work centre on the publication of patient data and that all the patients included in the study received sufficient information and gave their written informed consent to participate in the study.

The authors declare that no patient data appear in this article.