Essential medications in palliative care: an updated survey of Australian and New Zealand medical practitioners practising palliative care

Abstract

Background

Palliative care patients experience complex symptoms requiring varied pharmacological management. Medication choice depends on the underlying aetiology of symptoms, guideline recommendations, clinician preference and medication availability.

Aim

The aim of this study was to canvass the views of practising Australian and New Zealand medical practitioners practising palliative care regarding the medications they perceive to be essential for managing prevalent end‐of‐life symptoms, in an update to a previous clinician survey conducted in 2000. This study also canvassed respondents’ views on whether patients experienced difficulties accessing these medications.

Methods

A cross‐sectional survey was conducted in 2022 among current members of the Australian and New Zealand Society of Palliative Medicine (ANZSPM). The questionnaire explored ideal pharmacological management for 19 different symptoms experienced at the end of life. A descriptive analysis of results was calculated in SPSS version 28.0.

Results

Survey reponses were received from 13.5% of ANZSPM members. For most symptoms (n = 70), a large number of medications were listed as essential. The most concordant first‐ranked medications included hyoscine butylbromide for noisy breathing (48.8%), benzodiazepines for breathlessness (nominated by respondents as a class, 41.2%), dexamethasone for cachexia (36.4%) and morphine for cough (35.5%). Respondents most frequently indicated there was no essential medication for oral ulceration, dry mouth, delirium and anorexia. Lidocaine for oral ulceration was ranked the most difficult for patients to access in the community (85.7%).

Conclusions

There remains various opinions on essential medications for palliative care. This suggests the need for ongoing research to reach an evidence‐based consensus for clinical practice.

Introduction

Palliative care patients frequently experience complex symptoms requiring a variety of pharmacological interventions. ¹ Effective symptom management requires understanding of the underlying aetiology and can differ compared to non‐palliative patient populations due to altered pharmacokinetics and pharmacodynamics. ²

The World Health Organisation's (WHO) ‘WHO Model Lists of Essential Medicines’ address priority medications for a range of patient populations in low‐, middle‐ and high‐income countries. The WHO acknowledges that the choice to categorise medications as ‘essential’ is dependent not only on generalised metrics of each medication itself, such as safety and efficacy, but also on epidemiology, local health needs and affordability, all of which vary between and within countries. ³

To that end, a clinician survey was conducted in 2000 to canvass the views of medical practitioners working in palliative care on the essential medications required for palliative care. ⁴ In Australia, the study contributed to the establishment of the ‘Palliative Care’ section of the Pharmaceutical Benefits Scheme (PBS), which subsidises the cost of prescription medications for patients. ⁵ Of note, that study also found significant discrepancies between the perceived and actual level of evidence to inform prescribing. Since then, changes in the evidence base, best practice and medication availability have prompted an update to reflect the current views of Australian and New Zealand (ANZ) palliative care medical practitioners practising palliative care. ⁴

The aim of this study was to canvass the views of ANZ medical practitioners practising palliative care in 2022 regarding medications and classes of medications they considered essential for managing prevalent symptoms for palliative care patients and whether their patients experienced difficulties accessing these medications in hospital or community settings.

Materials and methods

Study design

A cross‐sectional survey was conducted via anonymous online questionnaire administered on the Qualtrics platform from April to July 2022. The participation information sheet provided information about the purpose of the study and use of data. Participants were informed that completion of the questionnaire implied consent.

Study participants

Eligible participants were current members of the ANZ Society of Palliative Medicine (ANZSPM) working in Australia and New Zealand. Members of ANZSPM are palliative medicine specialists or doctors working in hospices or with a special interest in palliative care without a specialist qualification. An initial email invitation was sent to ANZSPM members in early May 2022; a further reminder was published in an ANZSPM newsletter and a second email invitation sent to members in July 2022.

Questionnaire design

The questionnaire was designed to explore ideal pharmacological management of 19 symptoms, selected from the 22 in the original 2000 survey to enable comparison. ⁴ The selected symptoms were modified from the original set by consensus among the authors to reflect the most prevalent symptoms and reduce survey length. From the original survey, ‘terminal restlessness’ was included under ‘delirium in the final hours and days of life’ and ‘pain using adjuvant analgesics’ was incorporated under ‘pain using non‐opioid analgesics’. The questionnaire (Supplementary Material 1) defined ‘essential medication’ as ‘a drug that is indispensable with no reasonable alternative’ for each symptom. Respondents were invited to list up to five individual medications which they considered ‘essential’ for treating each symptom, ranked from highest to lowest priority. The same medication could be listed for more than one indication. Respondents could also indicate that there were no essential medications for a particular indication. Where respondents listed a whole class of medication, this was assumed to indicate that medications in the class were considered interchangeable for the indication in question. Answers other than a medication or class were excluded from analysis. Results are reported according to the checklist for reporting results of internet e‐surveys (CHERRIES) checklist for online surveys. ⁶

Data analysis

Descriptive analysis was calculated in SPSS version 28.0.0.0 (IBM, Armonk, NY, USA).

Ethics proposal

Ethics approval was obtained from the University of Technology Sydney Human Research Ethics Committee (ETH21‐6482).

Results

The survey was sent to 522 ANZSPM members, of whom 70 (13.5%) responded (Table 1).

Table 1.

Characteristics of ANZSPM members responding to survey (N = 70)

Characteristics	Response (%)
Country
Australia	81.4%
New Zealand	17.1%
Other – retired	1.4%
Age (years)
Median (IQR)	45 (40–57)
Time since medical school (years)
Median (IQR)	21 (15–31)
Time in palliative care (years)
Median (IQR)	11 (7–16)
Gender
Female	67.1%
Male	34.1%
Chose not to answer	1.4%
Primary area of clinical practice†
Palliative medicine	80.0%
Palliative medicine in training	7.1%
General practice	2.9%
Geriatric medicine	1.4%
Medical oncology	1.4%
Nephrology	1.4%
Paediatric palliative oncology	1.4%
Respiratory medicine	1.4%
Retired	1.4%
Respondent's patient demographics
Adult	90.0%
Adult and paediatric	5.7%
Paediatric	2.9%
Chose not to answer	1.4%
Participated in 2000 survey
No	72.9%
Yes	7.1%
Unsure	15.7%

^†Defined as clinical area in which responder contributes > 70% of their clinical time.

Questionnaire findings

The responses for each symptom are summarised in Table 2, with the number of complete responses for that question indicated in each case. For most symptoms, there were a large number of medications identified as ‘essential’, including for pain (22 different non‐opioid analgesics), cough (24 agents) and pruritus and dry mouth (37 agents each). Indications for which respondents most frequently indicated there was no essential medication were oral ulceration (75.0%), dry mouth (46.2%), delirium (40.7%), anorexia (40.0%) and pruritus (16.1%).

Table 2.

Views of ANZSPM members on essential medications for a range of symptoms

Palliative care symptom (no. complete responses)	No. responses for ‘there is no essential medication for this symptom’ n (%)	Number of different medications identified	1st‐ranked medication (%)	2nd‐ranked medication (%)	3rd‐ranked medication (%)
Pain (opioid analgesics) (N = 38)	0	7	Morphine (21.2%)	Hydromorphone (20.1%)	Fentanyl (18.4%) Oxycodone (18.4%)
Pain (non‐opioid analgesics) (N = 36)	0	22	Gabapentinoid (22.9%)	Paracetamol (16.6%)	Dexamethasone (12.1%)
Breathlessness (N = 38)	0	11	Benzodiazepines (41.2%)	Morphine (37.8%)	Fentanyl (6.7%)
Cough (N = 37)	1 (2.7%)	24	Morphine (35.3%)	Methadone (10.5%)	Codeine (9.5%)
Noisy breathing (N = 33)	5 (15.1%)	8	Hyoscine butylbromide (48.8%)	Glycopyrolate (19.1%)	No essential medication (15.2%)
Delirium (in the last hours to days of life) (N = 38)	0	11	Midazolam (25.0%)	Haloperidol (23.5%)	Levomepromazine (19.1%)
Delirium (N = 27)	11 (40.7%)	14	No essential medication (40.7%)	Haloperidol (19.5%)	Olanzapine (9.0%)
Anorexia (N = 25)	15 (60.0%)	9	No essential medication (60.0%)	Dexamethasone (20.0%)	Mirtazapine (7.6%)
Cachexia (N = 29)	6 (20.7%)	6	Dexamethasone (37.9%)	No essential medication (20.7%)	Anamorelin (8.3%) Ibuprofen (8.3%) Megestrol (8.3%) Prednisolone (8.3%) Psychostimulants (8.3%)
Nausea/vomiting (N = 38)	0	13	Haloperidol (22.5%)	Lorazepam (21.9%)	Cyclizine (17.2%)
Hiccup (N = 32)	5 (15.6%)	14	Haloperidol (20.0%) Metoclopramide (20.0%)	N/A	No essential medication (15.6%)
Dyspepsia (N = 35)	2 (5.7%)	13	Proton pump inhibitor (44.4%)	Ranitidine (11.9%)	Gaviscon (6.5%) Metoclopramide (6.5%)
Constipation (N = 37)	0	21	Docusate and senna (23.4%) Macrogol (23.4%)	N/A	Bisacodyl (9.7%)
Oral ulceration (N = 32)	24 (75%)	29	No essential medications (75.0%)	Lidocaine (6.6%)	Kenalog (2.0%)
Dry mouth (N = 26)	12 (46.1%)	37	No essential medications (46.2%)	Biotene Mouth Gel (8.8%)	Artificial saliva (8.7%)
Depression (N = 36)	1 (2.8%)	23	Mirtazapine (28.5%)	Sertraline (13.4%)	Escitalopram (10.7%)
Insomnia (N = 35)	2 (5.7%)	16	Temazepam (22.4%)	Melatonin (19.3%)	Mirtazapine (13.2%)
Anxiety (N = 37)	0	22	Lorazepam (28.8%)	Mirtazapine (17.1%)	Clonazepam (8.1%)
Pruritus (N = 31)	5 (16.1%)	37	No essential medications (16.1%)	Pregabalin (12.1%)	Sertraline (7.3%)

Abbreviation: N/A = not applicable.

For each symptom, the most frequent first‐ranked medication or class was identified. Symptoms with the most concordant first‐ranked medication or class use included hyoscine butylbromide for noisy breathing (48.8%), proton pump inhibitors (PPIs) for dyspepsia (44.4%), benzodiazepines for breathlessness (41.2%), dexamethasone for cachexia (36.4%) and morphine for cough (35.5%).

The first‐ranked medications perceived by the largest proportions of respondents to be difficult for patients to access in the community setting were lidocaine for oral ulceration (85.7%), hyoscine butylbromide for noisy breathing (59.1%) and midazolam for terminal delirium (44.1%) (Table 3).

Table 3.

Accessibility of first‐ranked essential medicines for a range of common symptoms, as reported by medical practitioners practising palliative care from Australia and New Zealand

First‐ranked medication	Symptoms the medication was ranked first for	Percentage of respondents who find patients have difficulty in accessing the first‐ranked medication in hospital or community settings
		Hospital	Community
Morphine	Pain, cough	0%	2.7%
Gabapentinoid	Non‐opioid analgesia	0%	5.6%
Benzodiazepines	Breathlessness	10.2%	38.8%
Hyoscine butylbromide	Noisy breathing	0%	59.1%
Midazolam	Delirium (in last hours to days of life)	0%	44.1%
Haloperidol	Delirium, nausea/vomiting, hiccup	0%	2.6%
Dexamethasone	Anorexia, cachexia	0%	0%
Metoclopramide†	Hiccup	0%	0%
Proton Pump Inhibitor	Dyspepsia	0%	2.4%
Macrogol	Constipation	0%	0%
Docusate and senna‡	Constipation	0%	3.1%
Lidocaine	Oral ulceration	15%	85%
Sodium Bicarbonate Mouthwash	Dry mouth	20%	0%
Mirtazapine	Depression	0%	0%
Temazepam	Insomnia	0%	9.5%
Lorazepam	Anxiety	3.2%	25.8%
Pregabalin	Pruritus	7.1%	21.4%

Note: Sodium Bicarbonate Mouthwash combines responses of ‘Sodibic’, ‘Sodibic Mouth Wash’, ‘Sodibic Mouthwash’, ‘Sodium Bicarbonate Preparations’, ‘Sodium Bicarbonate’, and ‘Sodium Bicarbonate Mouthwash’.

^†Equal first‐ranked with haloperidol for hiccup.

^‡Equal first ranked with macrogol for constipation.

Discussion

This survey reports findings from a broad range of doctors with a special interest in palliative care, including not only palliative care specialists but also oncologists and general practitioners (GPs). The results suggest a change in views over the intervening years since the survey in 2000. Overall, the updated survey indicates a greater variability between doctors in their selection of pharmacological agents for all canvassed end‐of‐life symptoms. This could be due to the inclusion of ‘there is no essential medications for this indication’ as a response in the updated survey, a greater understanding of underlying symptom mechanisms and pharmacological targets, a wider range of available medications and/or a shift toward non‐pharmacological interventions where appropriate.

Morphine remained the most commonly nominated essential opioid analgesic for pain, as nominated by one in five (21%) of respondents, closely followed by hydromorphone (20%) and fentanyl (18%). This is in contrast to being nominated as the most essential opioid for pain by almost all (98%) respondents in 2000. This result is in keeping with a 2017 Cochrane review which found both oral morphine and transdermal fentanyl effective at reducing cancer pain to ‘no’ or ‘mild’ pain within 14 days of starting treatment and reflects international clinical guidelines recommending oral morphine as the first‐line opioid for cancer pain (NICE UK). ⁷ , ⁸ The Australian Therapeutic Guidelines for Palliative Care list both morphine and oxycodone as first‐line oral options for opioid initiation. ⁹ The paucity of evidence for opioid efficacy for non‐malignant pain, alongside major changes to opioid formulation availability and subsidy on the PBS in the years following the original survey, could also account for the variability in opioid prescribing seen in our updated survey results. ¹⁰ , ¹¹

In contrast to the 2000 survey, the most favoured non‐opioid analgesia was the gabapentinoid class, rather than paracetamol. This could reflect the updated survey's combination of ‘pain using adjuvant analgesics’ into ‘pain using nonopioid analgesics’ categories, and increased availability of gabapentinoids since the initial survey. There have been almost 27% and 17% annual increases in gabapentinoid consumption between 2008 and 2018 in Australia and New Zealand respectively. ⁴ , ¹² This may be due to the PBS listing of pregabalin in 2013 and initiatives to curtail inappropriate opioid prescribing by the Australian government. Gabapentinoids are evidence‐based as analgesia for post‐herpetic neuralgia and peripheral diabetic neuropathy. ¹³ However, there is limited evidence for gabapentinoid use outside of these specific pain conditions, and research suggests that 58% of palliative care patients discontinue gabapentin after 21 days due to adverse effects or lack of analgesic efficacy. ¹³ , ¹⁴ , ¹⁵ Given the more widespread use and accessibility of gabapentinoids since 2000, their specific efficacy in palliative care patients warrants further research. Paracetamol also lacks high‐quality evidence in palliative care patients, with a 2017 Cochrane review unable to support or refute its prescribing alongside strong opioids for pain relief in patients with cancer, despite its inclusion on the WHO's ‘Analgesic Ladder’. ¹⁶

In 2000, the single most nominated essential medication for breathlessness was morphine (94% of respondents), compared to benzodiazepines (41%) and morphine (38%) in our updated survey. Studies prior to the first survey were conducted primarily in laboratory settings and showed promise for morphine for breathlessness. ¹⁷ This benefit was not seen in high‐quality, real‐world, randomised trials conducted after 2000. That said, and while current guidelines now advise against prescribing opioids routinely for chronic breathlessness, morphine remains the preferred agent in specific cases where it may be indicated. ¹⁷ , ¹⁸ , ¹⁹ , ²⁰ However, a 2016 Cochrane review found no statistically significant effect of benzodiazepines on either episodic or general breathlessness in advanced malignant and non‐malignant disease. ¹⁸ Our findings may therefore highlight a gap in respondent knowledge of recent evidence. Despite the lack of consensus for first‐line pharmacotherapy for generalised anxiety in palliative care, benzodiazepines were used consistently between surveys, with diazepam favoured (52%) in the 2000 survey and lorazepam (29%) in 2022. ²¹ Nearly 40% of survey respondents reported patient difficulty accessing benzodiazepines in the community, though midazolam injection and clonazepam drops are both available through the PBS in Australia. ⁵

Both surveys demonstrate that prescribers utilise hyoscine butylbromide to treat noisy breathing at the end of life, though variability of opinion has increased, with only 49% of respondents regarding it as the most essential medicine compared to 86% in 2000 and 15% of clinicians reporting there to be no essential medication for this symptom. Between surveys, a 2008 Cochrane review found no evidence for pharmacological intervention over placebo in the treatment of noisy breathing and reported adverse effects such as agitation or dry mouth. ²² Emerging evidence showing benefit for prophylactic hyoscine butylbromide is perhaps yet to change clinical practice and not reflected in current clinical guidelines. ²³ Despite being publicly funded in Australia and New Zealand, 59.1% of respondents reported difficulty with community access to this medication, the reasons for which may warrant further investigation.

Interestingly, while the 2000 survey showed 69% of respondents felt dexamethasone was the most essential drug for treating anorexia, this had reduced to 20% in the 2022 survey, with 60% of respondents indicating there was no essential medicine for this indication. This could reflect a weighting of recent evidence against benefit for megestrol or dexamethasone and concerns that dexamethasone for cachexia accelerates muscle loss and is not clinically indicated. ²⁴

Metoclopramide dropped from being the most widely endorsed medication (86%) for nausea in 2000, falling behind haloperidol, lorazepam and cyclizine on the updated survey. One explanation could be that contemporary clinicians choose antiemetics by nausea aetiology, patient‐specific factors or adverse effect profiles leading to a range of first‐line therapies. ²⁵ , ²⁶ Further surveys are required to elucidate clinician rationale.

Combined docusate and senna remained the most commonly first‐ranked medications for constipation in palliative care, though this had dropped from 58% of respondents in 2000 to equal first with macrogol at 23% in 2022. Both medications were PBS listed in the interval between the two surveys and are first‐line medications for undifferentiated constipation in the Australian guidelines. ⁹ Similarly to anti‐emetics, laxative choice in palliative care patients is often driven by patient‐specific factors due to insufficient evidence on the efficacy of individual laxatives, particularly regarding rectally administered laxatives. ²⁷

Forty‐one per cent of 2022 survey respondents reported no essential medication for the treatment of delirium, with the exception being delirium in the final hours and days of life, where midazolam was ranked as most essential (25%). In comparison, 84% of respondents reported haloperidol as the most essential medication for delirium in 2000. The use of antipsychotics for delirium changed significantly in the period between surveys. Where delirium is refractory to non‐pharmacological treatment and associated with agitation, distress or safety concerns, haloperidol is recommended by expert consensus and is easily accessible through the PBS. ²⁸ Despite this ease of access, recent research does not suggest a routine role for antipsychotics in delirium. ²⁹ , ³⁰ A 2022 national survey suggests this discordance may be driven by the desire to minimise the distress and risks of harm to the patient, family and other stakeholders at the end of life. ³¹

The treatment of depression in the palliative setting has also changed between the two surveys, with 40% of respondents in 2000 indicating sertraline as the most essential medication, but only 13% of respondents reporting the same in 2022. Mirtazapine was instead ranked as most essential in 2022, possibly because its side effects of sedation and weight gain provide secondary benefits for many palliative care patients. ³²

While almost all essential medications were reported as being readily available in hospitals when required, several medications were perceived to be difficult for patients to access in the community. These included lidocaine for oral ulceration and hyoscine butylbromide for noisy breathing. The causes of difficult access were not explored within the scope of this survey, though they could relate to pharmacy supply or difficulty with subcutaneous administration without community nursing support.

Implications for clinical research, policy and practice

Our survey results reveal that in 2022 some symptoms had high variability in medications identified as essential for their management, including cough, dry mouth, oral ulceration and pruritus. This variability has many potential causes: local variations in practice, medication choice being influenced by symptom aetiology, paucity of research in the palliative care setting and/or conflicting evidence. Further, for symptoms with evidence favouring non‐pharmacological management as first line, discordance in responses for most essential medications may reflect avoidance of pharmacological management in the first instance (e.g. haloperidol use in refractory delirium). ²⁸

For some symptoms, responses aligned closely with recent evidence. For other symptoms, such as anorexia‐cachexia, breathlessness or noisy breathing, the evidence was less consistent with responses. This may reflect a lack of awareness of existing evidence, lack of faith in the quality of evidence, or a reluctance to withhold pharmacotherapy when confronted with symptoms.

We also note that for some indications (notably breathlessness and reflux), respondents ranked classes rather than individual medications as essential. One explanation is that, for some indications, respondent perception of the evidence suggests benefit is due to a class effect rather than a specific medication. This includes PPIs for reflux and benzodiazepines for breathlessness. Evidence supports the former, but, as noted above, a 2016 Cochrane review did not support benzodiazepines for breathlessness. ¹⁸ , ³³

Our survey results also reveal issues regarding changes in medication availability. Of note, hydromorphone is not available in New Zealand. There remains an ongoing need for high‐quality research to justify and guide funding of palliative care medications in Australia and New Zealand.

Limitations

The study had a low response rate (13.5%), limiting the extrapolation of results to the wider group of medical practitioners practising palliative care in Australia and New Zealand. This could have been due to survey length, which is likely to have contributed to missing data. In an effort to reduce survey length, we did not ask about the perceived level of evidence as in the previous study, which could have informed future follow‐up studies. The survey did not define an aetiology, important for some symptoms, or assess the formulation, route or dosing regimen of medicines. The survey also did not explore the specific reasons for difficulty accessing medications. Finally, some respondents ranked entire classes of medications, rather than individual medications, as requested in the survey. While we elected to not exclude these responses, respondents may have intended to assign only one or some medications as essential rather than all the medications in the class.

Conclusions

Among ANZ medical practitioners practising palliative care, there remains variable opinion of the truly essential medications for palliation of certain symptoms and considerable discordance between expert opinion and evidence bases. Reaching an evidence‐based consensus of truly essential palliative care medications would help to ensure their ongoing access and affordability for patients. This will require ongoing high‐quality clinical research at a level that can inform licensing and funding decisions by regulatory bodies.

Supporting information

Data S1. Supporting Information.

Data availability statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.